id
stringlengths 9
16
| submitter
stringlengths 3
64
⌀ | authors
stringlengths 5
6.63k
| title
stringlengths 7
245
| comments
stringlengths 1
482
⌀ | journal-ref
stringlengths 4
382
⌀ | doi
stringlengths 9
151
⌀ | report-no
stringclasses 984
values | categories
stringlengths 5
108
| license
stringclasses 9
values | abstract
stringlengths 83
3.41k
| versions
listlengths 1
20
| update_date
timestamp[s]date 2007-05-23 00:00:00
2025-04-11 00:00:00
| authors_parsed
sequencelengths 1
427
| prompt
stringlengths 166
3.49k
| label
stringclasses 2
values | prob
float64 0.5
0.98
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1409.7935 | Lior Shamir | Evan Kuminski, Joe George, John Wallin, Lior Shamir | Combining human and machine learning for morphological analysis of
galaxy images | PASP, accepted | null | 10.1086/678977 | null | astro-ph.IM astro-ph.GA cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The increasing importance of digital sky surveys collecting many millions of
galaxy images has reinforced the need for robust methods that can perform
morphological analysis of large galaxy image databases. Citizen science
initiatives such as Galaxy Zoo showed that large datasets of galaxy images can
be analyzed effectively by non-scientist volunteers, but since databases
generated by robotic telescopes grow much faster than the processing power of
any group of citizen scientists, it is clear that computer analysis is
required. Here we propose to use citizen science data for training machine
learning systems, and show experimental results demonstrating that machine
learning systems can be trained with citizen science data. Our findings show
that the performance of machine learning depends on the quality of the data,
which can be improved by using samples that have a high degree of agreement
between the citizen scientists. The source code of the method is publicly
available.
| [
{
"version": "v1",
"created": "Sun, 28 Sep 2014 17:47:35 GMT"
}
] | 2015-06-23T00:00:00 | [
[
"Kuminski",
"Evan",
""
],
[
"George",
"Joe",
""
],
[
"Wallin",
"John",
""
],
[
"Shamir",
"Lior",
""
]
] | TITLE: Combining human and machine learning for morphological analysis of
galaxy images
ABSTRACT: The increasing importance of digital sky surveys collecting many millions of
galaxy images has reinforced the need for robust methods that can perform
morphological analysis of large galaxy image databases. Citizen science
initiatives such as Galaxy Zoo showed that large datasets of galaxy images can
be analyzed effectively by non-scientist volunteers, but since databases
generated by robotic telescopes grow much faster than the processing power of
any group of citizen scientists, it is clear that computer analysis is
required. Here we propose to use citizen science data for training machine
learning systems, and show experimental results demonstrating that machine
learning systems can be trained with citizen science data. Our findings show
that the performance of machine learning depends on the quality of the data,
which can be improved by using samples that have a high degree of agreement
between the citizen scientists. The source code of the method is publicly
available.
| no_new_dataset | 0.947914 |
1410.1257 | Abhronil Sengupta | Abhronil Sengupta, Sri Harsha Choday, Yusung Kim, and Kaushik Roy | Spin Orbit Torque Based Electronic Neuron | null | null | 10.1063/1.4917011 | null | cs.ET | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A device based on current-induced spin-orbit torque (SOT) that functions as
an electronic neuron is proposed in this work. The SOT device implements an
artificial neuron's thresholding (transfer) function. In the first step of a
two-step switching scheme, a charge current places the magnetization of a
nano-magnet along the hard-axis i.e. an unstable point for the magnet. In the
second step, the SOT device (neuron) receives a current (from the synapses)
which moves the magnetization from the unstable point to one of the two stable
states. The polarity of the synaptic current encodes the excitatory and
inhibitory nature of the neuron input, and determines the final orientation of
the magnetization. A resistive crossbar array, functioning as synapses,
generates a bipolar current that is a weighted sum of the inputs. The
simulation of a two layer feed-forward Artificial Neural Network (ANN) based on
the SOT electronic neuron shows that it consumes ~3X lower power than a 45nm
digital CMOS implementation, while reaching ~80% accuracy in the classification
of one hundred images of handwritten digits from the MNIST dataset.
| [
{
"version": "v1",
"created": "Mon, 6 Oct 2014 05:36:19 GMT"
}
] | 2015-06-23T00:00:00 | [
[
"Sengupta",
"Abhronil",
""
],
[
"Choday",
"Sri Harsha",
""
],
[
"Kim",
"Yusung",
""
],
[
"Roy",
"Kaushik",
""
]
] | TITLE: Spin Orbit Torque Based Electronic Neuron
ABSTRACT: A device based on current-induced spin-orbit torque (SOT) that functions as
an electronic neuron is proposed in this work. The SOT device implements an
artificial neuron's thresholding (transfer) function. In the first step of a
two-step switching scheme, a charge current places the magnetization of a
nano-magnet along the hard-axis i.e. an unstable point for the magnet. In the
second step, the SOT device (neuron) receives a current (from the synapses)
which moves the magnetization from the unstable point to one of the two stable
states. The polarity of the synaptic current encodes the excitatory and
inhibitory nature of the neuron input, and determines the final orientation of
the magnetization. A resistive crossbar array, functioning as synapses,
generates a bipolar current that is a weighted sum of the inputs. The
simulation of a two layer feed-forward Artificial Neural Network (ANN) based on
the SOT electronic neuron shows that it consumes ~3X lower power than a 45nm
digital CMOS implementation, while reaching ~80% accuracy in the classification
of one hundred images of handwritten digits from the MNIST dataset.
| no_new_dataset | 0.958538 |
1411.1343 | Giuseppe Cataldo | Giuseppe Cataldo, Edward J. Wollack, Emily M. Barrentine, Ari D.
Brown, Samuel H. Moseley, and Kongpop U-Yen | Analysis and calibration techniques for superconducting resonators | 12 pages, 4 figures | null | 10.1063/1.4904972 | null | astro-ph.IM physics.ins-det | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A method is proposed and experimentally explored for in-situ calibration of
complex transmission data for superconducting microwave resonators. This
cryogenic calibration method accounts for the instrumental transmission
response between the vector network analyzer reference plane and the device
calibration plane. Once calibrated, the observed resonator response is analyzed
in detail by two approaches. The first, a phenomenological model based on
physically realizable rational functions, enables the extraction of multiple
resonance frequencies and widths for coupled resonators without explicit
specification of the circuit network. In the second, an ABCD-matrix
representation for the distributed transmission line circuit is used to model
the observed response from the characteristic impedance and propagation
constant. When used in conjunction with electromagnetic simulations, the
kinetic inductance fraction can be determined with this method with an accuracy
of 2%. Datasets for superconducting microstrip and coplanar-waveguide resonator
devices were investigated and a recovery within 1% of the observed complex
transmission amplitude was achieved with both analysis approaches. The
experimental configuration used in microwave characterization of the devices
and self-consistent constraints for the electromagnetic constitutive relations
for parameter extraction are also presented.
| [
{
"version": "v1",
"created": "Wed, 5 Nov 2014 17:54:42 GMT"
},
{
"version": "v2",
"created": "Mon, 17 Nov 2014 21:40:39 GMT"
},
{
"version": "v3",
"created": "Thu, 4 Dec 2014 20:00:08 GMT"
}
] | 2015-06-23T00:00:00 | [
[
"Cataldo",
"Giuseppe",
""
],
[
"Wollack",
"Edward J.",
""
],
[
"Barrentine",
"Emily M.",
""
],
[
"Brown",
"Ari D.",
""
],
[
"Moseley",
"Samuel H.",
""
],
[
"U-Yen",
"Kongpop",
""
]
] | TITLE: Analysis and calibration techniques for superconducting resonators
ABSTRACT: A method is proposed and experimentally explored for in-situ calibration of
complex transmission data for superconducting microwave resonators. This
cryogenic calibration method accounts for the instrumental transmission
response between the vector network analyzer reference plane and the device
calibration plane. Once calibrated, the observed resonator response is analyzed
in detail by two approaches. The first, a phenomenological model based on
physically realizable rational functions, enables the extraction of multiple
resonance frequencies and widths for coupled resonators without explicit
specification of the circuit network. In the second, an ABCD-matrix
representation for the distributed transmission line circuit is used to model
the observed response from the characteristic impedance and propagation
constant. When used in conjunction with electromagnetic simulations, the
kinetic inductance fraction can be determined with this method with an accuracy
of 2%. Datasets for superconducting microstrip and coplanar-waveguide resonator
devices were investigated and a recovery within 1% of the observed complex
transmission amplitude was achieved with both analysis approaches. The
experimental configuration used in microwave characterization of the devices
and self-consistent constraints for the electromagnetic constitutive relations
for parameter extraction are also presented.
| no_new_dataset | 0.953057 |
1412.5027 | Ali Borji | Ali Borji | What is a salient object? A dataset and a baseline model for salient
object detection | IEEE Transactions on Image Processing, 2014 | null | 10.1109/TIP.2014.2383320 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Salient object detection or salient region detection models, diverging from
fixation prediction models, have traditionally been dealing with locating and
segmenting the most salient object or region in a scene. While the notion of
most salient object is sensible when multiple objects exist in a scene, current
datasets for evaluation of saliency detection approaches often have scenes with
only one single object. We introduce three main contributions in this paper:
First, we take an indepth look at the problem of salient object detection by
studying the relationship between where people look in scenes and what they
choose as the most salient object when they are explicitly asked. Based on the
agreement between fixations and saliency judgments, we then suggest that the
most salient object is the one that attracts the highest fraction of fixations.
Second, we provide two new less biased benchmark datasets containing scenes
with multiple objects that challenge existing saliency models. Indeed, we
observed a severe drop in performance of 8 state-of-the-art models on our
datasets (40% to 70%). Third, we propose a very simple yet powerful model based
on superpixels to be used as a baseline for model evaluation and comparison.
While on par with the best models on MSRA-5K dataset, our model wins over other
models on our data highlighting a serious drawback of existing models, which is
convoluting the processes of locating the most salient object and its
segmentation. We also provide a review and statistical analysis of some labeled
scene datasets that can be used for evaluating salient object detection models.
We believe that our work can greatly help remedy the over-fitting of models to
existing biased datasets and opens new venues for future research in this
fast-evolving field.
| [
{
"version": "v1",
"created": "Mon, 8 Dec 2014 23:51:50 GMT"
}
] | 2015-06-23T00:00:00 | [
[
"Borji",
"Ali",
""
]
] | TITLE: What is a salient object? A dataset and a baseline model for salient
object detection
ABSTRACT: Salient object detection or salient region detection models, diverging from
fixation prediction models, have traditionally been dealing with locating and
segmenting the most salient object or region in a scene. While the notion of
most salient object is sensible when multiple objects exist in a scene, current
datasets for evaluation of saliency detection approaches often have scenes with
only one single object. We introduce three main contributions in this paper:
First, we take an indepth look at the problem of salient object detection by
studying the relationship between where people look in scenes and what they
choose as the most salient object when they are explicitly asked. Based on the
agreement between fixations and saliency judgments, we then suggest that the
most salient object is the one that attracts the highest fraction of fixations.
Second, we provide two new less biased benchmark datasets containing scenes
with multiple objects that challenge existing saliency models. Indeed, we
observed a severe drop in performance of 8 state-of-the-art models on our
datasets (40% to 70%). Third, we propose a very simple yet powerful model based
on superpixels to be used as a baseline for model evaluation and comparison.
While on par with the best models on MSRA-5K dataset, our model wins over other
models on our data highlighting a serious drawback of existing models, which is
convoluting the processes of locating the most salient object and its
segmentation. We also provide a review and statistical analysis of some labeled
scene datasets that can be used for evaluating salient object detection models.
We believe that our work can greatly help remedy the over-fitting of models to
existing biased datasets and opens new venues for future research in this
fast-evolving field.
| new_dataset | 0.926503 |
1412.7156 | Ludovic Denoyer | Gabriella Contardo and Ludovic Denoyer and Thierry Artieres | Representation Learning for cold-start recommendation | Accepted as workshop contribution at ICLR 2015 | null | null | null | cs.IR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A standard approach to Collaborative Filtering (CF), i.e. prediction of user
ratings on items, relies on Matrix Factorization techniques. Representations
for both users and items are computed from the observed ratings and used for
prediction. Unfortunatly, these transductive approaches cannot handle the case
of new users arriving in the system, with no known rating, a problem known as
user cold-start. A common approach in this context is to ask these incoming
users for a few initialization ratings. This paper presents a model to tackle
this twofold problem of (i) finding good questions to ask, (ii) building
efficient representations from this small amount of information. The model can
also be used in a more standard (warm) context. Our approach is evaluated on
the classical CF problem and on the cold-start problem on four different
datasets showing its ability to improve baseline performance in both cases.
| [
{
"version": "v1",
"created": "Mon, 22 Dec 2014 21:58:06 GMT"
},
{
"version": "v2",
"created": "Fri, 27 Feb 2015 18:56:23 GMT"
},
{
"version": "v3",
"created": "Fri, 27 Mar 2015 09:59:25 GMT"
},
{
"version": "v4",
"created": "Wed, 8 Apr 2015 15:37:19 GMT"
},
{
"version": "v5",
"created": "Mon, 22 Jun 2015 14:01:33 GMT"
}
] | 2015-06-23T00:00:00 | [
[
"Contardo",
"Gabriella",
""
],
[
"Denoyer",
"Ludovic",
""
],
[
"Artieres",
"Thierry",
""
]
] | TITLE: Representation Learning for cold-start recommendation
ABSTRACT: A standard approach to Collaborative Filtering (CF), i.e. prediction of user
ratings on items, relies on Matrix Factorization techniques. Representations
for both users and items are computed from the observed ratings and used for
prediction. Unfortunatly, these transductive approaches cannot handle the case
of new users arriving in the system, with no known rating, a problem known as
user cold-start. A common approach in this context is to ask these incoming
users for a few initialization ratings. This paper presents a model to tackle
this twofold problem of (i) finding good questions to ask, (ii) building
efficient representations from this small amount of information. The model can
also be used in a more standard (warm) context. Our approach is evaluated on
the classical CF problem and on the cold-start problem on four different
datasets showing its ability to improve baseline performance in both cases.
| no_new_dataset | 0.947721 |
1501.03252 | Ali Ajmi | Ali Ajmi, S. Uma Sankar | Muonless Events in ICAL at INO | 21 pages, 6 figures | null | 10.1088/1748-0221/10/04/P04006 | null | physics.ins-det hep-ex | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The primary physics signal events in the ICAL at INO are the ${\nu}_{\mu}$
charged current (CC) interactions with a well defined muon track. Apart from
these events, ICAL can also detect other types of neutrino interactions, i.e.
the electron neutrino charged current interactions and the neutral current
events. It is possible to have a dataset containing mostly ${\nu}_e$CC events,
by imposing appropriate selection cuts on the events. The ${\nu}_{\mu}$ CC and
the neutral current events form the background to these events. This study uses
the Monte Carlo generated neutrino events, to design the necessary selection
cuts to obtain a ${\nu}_e$ CC rich dataset. An optimized set of constraints are
developed which balance the need for improving the purity of the sample and
having a large enough event sample. Depending on the constraints used, one can
obtain a neutrino data sample, with the purity of ${\nu}_e$ events varying
between 55% to 70%.
| [
{
"version": "v1",
"created": "Wed, 14 Jan 2015 05:34:11 GMT"
}
] | 2015-06-23T00:00:00 | [
[
"Ajmi",
"Ali",
""
],
[
"Sankar",
"S. Uma",
""
]
] | TITLE: Muonless Events in ICAL at INO
ABSTRACT: The primary physics signal events in the ICAL at INO are the ${\nu}_{\mu}$
charged current (CC) interactions with a well defined muon track. Apart from
these events, ICAL can also detect other types of neutrino interactions, i.e.
the electron neutrino charged current interactions and the neutral current
events. It is possible to have a dataset containing mostly ${\nu}_e$CC events,
by imposing appropriate selection cuts on the events. The ${\nu}_{\mu}$ CC and
the neutral current events form the background to these events. This study uses
the Monte Carlo generated neutrino events, to design the necessary selection
cuts to obtain a ${\nu}_e$ CC rich dataset. An optimized set of constraints are
developed which balance the need for improving the purity of the sample and
having a large enough event sample. Depending on the constraints used, one can
obtain a neutrino data sample, with the purity of ${\nu}_e$ events varying
between 55% to 70%.
| no_new_dataset | 0.90389 |
1501.06952 | Brendon Brewer | Brendon J. Brewer, Courtney P. Donovan | Fast Bayesian Inference for Exoplanet Discovery in Radial Velocity Data | Accepted for publication in MNRAS. 9 pages, 12 figures. Code at
http://www.github.com/eggplantbren/Exoplanet | null | 10.1093/mnras/stv199 | null | astro-ph.IM astro-ph.EP physics.data-an stat.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Inferring the number of planets $N$ in an exoplanetary system from radial
velocity (RV) data is a challenging task. Recently, it has become clear that RV
data can contain periodic signals due to stellar activity, which can be
difficult to distinguish from planetary signals. However, even doing the
inference under a given set of simplifying assumptions (e.g. no stellar
activity) can be difficult. It is common for the posterior distribution for the
planet parameters, such as orbital periods, to be multimodal and to have other
awkward features. In addition, when $N$ is unknown, the marginal likelihood (or
evidence) as a function of $N$ is required. Rather than doing separate runs
with different trial values of $N$, we propose an alternative approach using a
trans-dimensional Markov Chain Monte Carlo method within Nested Sampling. The
posterior distribution for $N$ can be obtained with a single run. We apply the
method to $\nu$ Oph and Gliese 581, finding moderate evidence for additional
signals in $\nu$ Oph with periods of 36.11 $\pm$ 0.034 days, 75.58 $\pm$ 0.80
days, and 1709 $\pm$ 183 days; the posterior probability that at least one of
these exists is 85%. The results also suggest Gliese 581 hosts many (7-15)
"planets" (or other causes of other periodic signals), but only 4-6 have well
determined periods. The analysis of both of these datasets shows phase
transitions exist which are difficult to negotiate without Nested Sampling.
| [
{
"version": "v1",
"created": "Tue, 27 Jan 2015 22:54:14 GMT"
}
] | 2015-06-23T00:00:00 | [
[
"Brewer",
"Brendon J.",
""
],
[
"Donovan",
"Courtney P.",
""
]
] | TITLE: Fast Bayesian Inference for Exoplanet Discovery in Radial Velocity Data
ABSTRACT: Inferring the number of planets $N$ in an exoplanetary system from radial
velocity (RV) data is a challenging task. Recently, it has become clear that RV
data can contain periodic signals due to stellar activity, which can be
difficult to distinguish from planetary signals. However, even doing the
inference under a given set of simplifying assumptions (e.g. no stellar
activity) can be difficult. It is common for the posterior distribution for the
planet parameters, such as orbital periods, to be multimodal and to have other
awkward features. In addition, when $N$ is unknown, the marginal likelihood (or
evidence) as a function of $N$ is required. Rather than doing separate runs
with different trial values of $N$, we propose an alternative approach using a
trans-dimensional Markov Chain Monte Carlo method within Nested Sampling. The
posterior distribution for $N$ can be obtained with a single run. We apply the
method to $\nu$ Oph and Gliese 581, finding moderate evidence for additional
signals in $\nu$ Oph with periods of 36.11 $\pm$ 0.034 days, 75.58 $\pm$ 0.80
days, and 1709 $\pm$ 183 days; the posterior probability that at least one of
these exists is 85%. The results also suggest Gliese 581 hosts many (7-15)
"planets" (or other causes of other periodic signals), but only 4-6 have well
determined periods. The analysis of both of these datasets shows phase
transitions exist which are difficult to negotiate without Nested Sampling.
| no_new_dataset | 0.942718 |
1504.06044 | Radoslaw Klimek | Radoslaw Klimek and Leszek Kotulski | Towards a better understanding and behavior recognition of inhabitants
in smart cities. A public transport case | Proceedings of 14th International Conference on Arificial Inteligence
and Soft Computing (ICAISC 2015), 14-18 June, 2015, Zakopane, Poland; Lecture
Notes in Computer Science, vol. 9120, pp.237-246. Springer Verlag 2015 | null | 10.1007/978-3-319-19369-4_22 | null | cs.CY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The idea of modern urban systems and smart cities requires monitoring and
careful analysis of different signals. Such signals can originate from
different sources and one of the most promising is the BTS, i.e. base
transceiver station, an element of mobile carrier networks. This paper presents
the fundamental problems of elicitation, classification and understanding of
such signals so as to develop context-aware and pro-active systems in urban
areas. These systems are characterized by the omnipresence of computing which
is strongly focused on providing on-line support to users/inhabitants of smart
cities. A method of analyzing selected elements of mobile phone datasets
through understanding inhabitants' behavioral fingerprints to obtain smart
scenarios for public transport is proposed. Some scenarios are outlined. A
multi-agent system is proposed. A formalism based on graphs that allows
reasoning about inhabitant behaviors is also proposed.
| [
{
"version": "v1",
"created": "Thu, 23 Apr 2015 04:57:50 GMT"
},
{
"version": "v2",
"created": "Sat, 20 Jun 2015 12:59:53 GMT"
}
] | 2015-06-23T00:00:00 | [
[
"Klimek",
"Radoslaw",
""
],
[
"Kotulski",
"Leszek",
""
]
] | TITLE: Towards a better understanding and behavior recognition of inhabitants
in smart cities. A public transport case
ABSTRACT: The idea of modern urban systems and smart cities requires monitoring and
careful analysis of different signals. Such signals can originate from
different sources and one of the most promising is the BTS, i.e. base
transceiver station, an element of mobile carrier networks. This paper presents
the fundamental problems of elicitation, classification and understanding of
such signals so as to develop context-aware and pro-active systems in urban
areas. These systems are characterized by the omnipresence of computing which
is strongly focused on providing on-line support to users/inhabitants of smart
cities. A method of analyzing selected elements of mobile phone datasets
through understanding inhabitants' behavioral fingerprints to obtain smart
scenarios for public transport is proposed. Some scenarios are outlined. A
multi-agent system is proposed. A formalism based on graphs that allows
reasoning about inhabitant behaviors is also proposed.
| no_new_dataset | 0.945551 |
1505.00359 | Harm de Vries | Harm de Vries, Jason Yosinski | Can deep learning help you find the perfect match? | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Is he/she my type or not? The answer to this question depends on the personal
preferences of the one asking it. The individual process of obtaining a full
answer may generally be difficult and time consuming, but often an approximate
answer can be obtained simply by looking at a photo of the potential match.
Such approximate answers based on visual cues can be produced in a fraction of
a second, a phenomenon that has led to a series of recently successful dating
apps in which users rate others positively or negatively using primarily a
single photo. In this paper we explore using convolutional networks to create a
model of an individual's personal preferences based on rated photos. This
introduced task is difficult due to the large number of variations in profile
pictures and the noise in attractiveness labels. Toward this task we collect a
dataset comprised of $9364$ pictures and binary labels for each. We compare
performance of convolutional models trained in three ways: first directly on
the collected dataset, second with features transferred from a network trained
to predict gender, and third with features transferred from a network trained
on ImageNet. Our findings show that ImageNet features transfer best, producing
a model that attains $68.1\%$ accuracy on the test set and is moderately
successful at predicting matches.
| [
{
"version": "v1",
"created": "Sat, 2 May 2015 17:20:23 GMT"
},
{
"version": "v2",
"created": "Sat, 20 Jun 2015 15:41:45 GMT"
}
] | 2015-06-23T00:00:00 | [
[
"de Vries",
"Harm",
""
],
[
"Yosinski",
"Jason",
""
]
] | TITLE: Can deep learning help you find the perfect match?
ABSTRACT: Is he/she my type or not? The answer to this question depends on the personal
preferences of the one asking it. The individual process of obtaining a full
answer may generally be difficult and time consuming, but often an approximate
answer can be obtained simply by looking at a photo of the potential match.
Such approximate answers based on visual cues can be produced in a fraction of
a second, a phenomenon that has led to a series of recently successful dating
apps in which users rate others positively or negatively using primarily a
single photo. In this paper we explore using convolutional networks to create a
model of an individual's personal preferences based on rated photos. This
introduced task is difficult due to the large number of variations in profile
pictures and the noise in attractiveness labels. Toward this task we collect a
dataset comprised of $9364$ pictures and binary labels for each. We compare
performance of convolutional models trained in three ways: first directly on
the collected dataset, second with features transferred from a network trained
to predict gender, and third with features transferred from a network trained
on ImageNet. Our findings show that ImageNet features transfer best, producing
a model that attains $68.1\%$ accuracy on the test set and is moderately
successful at predicting matches.
| new_dataset | 0.960361 |
1506.06272 | Fei Sha | Junqi Jin, Kun Fu, Runpeng Cui, Fei Sha and Changshui Zhang | Aligning where to see and what to tell: image caption with region-based
attention and scene factorization | null | null | null | null | cs.CV cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent progress on automatic generation of image captions has shown that it
is possible to describe the most salient information conveyed by images with
accurate and meaningful sentences. In this paper, we propose an image caption
system that exploits the parallel structures between images and sentences. In
our model, the process of generating the next word, given the previously
generated ones, is aligned with the visual perception experience where the
attention shifting among the visual regions imposes a thread of visual
ordering. This alignment characterizes the flow of "abstract meaning", encoding
what is semantically shared by both the visual scene and the text description.
Our system also makes another novel modeling contribution by introducing
scene-specific contexts that capture higher-level semantic information encoded
in an image. The contexts adapt language models for word generation to specific
scene types. We benchmark our system and contrast to published results on
several popular datasets. We show that using either region-based attention or
scene-specific contexts improves systems without those components. Furthermore,
combining these two modeling ingredients attains the state-of-the-art
performance.
| [
{
"version": "v1",
"created": "Sat, 20 Jun 2015 17:25:38 GMT"
}
] | 2015-06-23T00:00:00 | [
[
"Jin",
"Junqi",
""
],
[
"Fu",
"Kun",
""
],
[
"Cui",
"Runpeng",
""
],
[
"Sha",
"Fei",
""
],
[
"Zhang",
"Changshui",
""
]
] | TITLE: Aligning where to see and what to tell: image caption with region-based
attention and scene factorization
ABSTRACT: Recent progress on automatic generation of image captions has shown that it
is possible to describe the most salient information conveyed by images with
accurate and meaningful sentences. In this paper, we propose an image caption
system that exploits the parallel structures between images and sentences. In
our model, the process of generating the next word, given the previously
generated ones, is aligned with the visual perception experience where the
attention shifting among the visual regions imposes a thread of visual
ordering. This alignment characterizes the flow of "abstract meaning", encoding
what is semantically shared by both the visual scene and the text description.
Our system also makes another novel modeling contribution by introducing
scene-specific contexts that capture higher-level semantic information encoded
in an image. The contexts adapt language models for word generation to specific
scene types. We benchmark our system and contrast to published results on
several popular datasets. We show that using either region-based attention or
scene-specific contexts improves systems without those components. Furthermore,
combining these two modeling ingredients attains the state-of-the-art
performance.
| no_new_dataset | 0.948298 |
1506.06418 | Raphael Hoffmann | Raphael Hoffmann, Luke Zettlemoyer, Daniel S. Weld | Extreme Extraction: Only One Hour per Relation | null | null | null | null | cs.CL cs.AI cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Information Extraction (IE) aims to automatically generate a large knowledge
base from natural language text, but progress remains slow. Supervised learning
requires copious human annotation, while unsupervised and weakly supervised
approaches do not deliver competitive accuracy. As a result, most fielded
applications of IE, as well as the leading TAC-KBP systems, rely on significant
amounts of manual engineering. Even "Extreme" methods, such as those reported
in Freedman et al. 2011, require about 10 hours of expert labor per relation.
This paper shows how to reduce that effort by an order of magnitude. We
present a novel system, InstaRead, that streamlines authoring with an ensemble
of methods: 1) encoding extraction rules in an expressive and compositional
representation, 2) guiding the user to promising rules based on corpus
statistics and mined resources, and 3) introducing a new interactive
development cycle that provides immediate feedback --- even on large datasets.
Experiments show that experts can create quality extractors in under an hour
and even NLP novices can author good extractors. These extractors equal or
outperform ones obtained by comparably supervised and state-of-the-art
distantly supervised approaches.
| [
{
"version": "v1",
"created": "Sun, 21 Jun 2015 22:04:39 GMT"
}
] | 2015-06-23T00:00:00 | [
[
"Hoffmann",
"Raphael",
""
],
[
"Zettlemoyer",
"Luke",
""
],
[
"Weld",
"Daniel S.",
""
]
] | TITLE: Extreme Extraction: Only One Hour per Relation
ABSTRACT: Information Extraction (IE) aims to automatically generate a large knowledge
base from natural language text, but progress remains slow. Supervised learning
requires copious human annotation, while unsupervised and weakly supervised
approaches do not deliver competitive accuracy. As a result, most fielded
applications of IE, as well as the leading TAC-KBP systems, rely on significant
amounts of manual engineering. Even "Extreme" methods, such as those reported
in Freedman et al. 2011, require about 10 hours of expert labor per relation.
This paper shows how to reduce that effort by an order of magnitude. We
present a novel system, InstaRead, that streamlines authoring with an ensemble
of methods: 1) encoding extraction rules in an expressive and compositional
representation, 2) guiding the user to promising rules based on corpus
statistics and mined resources, and 3) introducing a new interactive
development cycle that provides immediate feedback --- even on large datasets.
Experiments show that experts can create quality extractors in under an hour
and even NLP novices can author good extractors. These extractors equal or
outperform ones obtained by comparably supervised and state-of-the-art
distantly supervised approaches.
| no_new_dataset | 0.941223 |
1506.06490 | Baotian Hu | Xiaoqiang Zhou, Baotian Hu, Qingcai Chen, Buzhou Tang, Xiaolong Wang | Answer Sequence Learning with Neural Networks for Answer Selection in
Community Question Answering | 6 pages | null | null | null | cs.CL cs.IR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, the answer selection problem in community question answering
(CQA) is regarded as an answer sequence labeling task, and a novel approach is
proposed based on the recurrent architecture for this problem. Our approach
applies convolution neural networks (CNNs) to learning the joint representation
of question-answer pair firstly, and then uses the joint representation as
input of the long short-term memory (LSTM) to learn the answer sequence of a
question for labeling the matching quality of each answer. Experiments
conducted on the SemEval 2015 CQA dataset shows the effectiveness of our
approach.
| [
{
"version": "v1",
"created": "Mon, 22 Jun 2015 07:26:51 GMT"
}
] | 2015-06-23T00:00:00 | [
[
"Zhou",
"Xiaoqiang",
""
],
[
"Hu",
"Baotian",
""
],
[
"Chen",
"Qingcai",
""
],
[
"Tang",
"Buzhou",
""
],
[
"Wang",
"Xiaolong",
""
]
] | TITLE: Answer Sequence Learning with Neural Networks for Answer Selection in
Community Question Answering
ABSTRACT: In this paper, the answer selection problem in community question answering
(CQA) is regarded as an answer sequence labeling task, and a novel approach is
proposed based on the recurrent architecture for this problem. Our approach
applies convolution neural networks (CNNs) to learning the joint representation
of question-answer pair firstly, and then uses the joint representation as
input of the long short-term memory (LSTM) to learn the answer sequence of a
question for labeling the matching quality of each answer. Experiments
conducted on the SemEval 2015 CQA dataset shows the effectiveness of our
approach.
| no_new_dataset | 0.944689 |
1506.06724 | Yukun Zhu | Yukun Zhu, Ryan Kiros, Richard Zemel, Ruslan Salakhutdinov, Raquel
Urtasun, Antonio Torralba, Sanja Fidler | Aligning Books and Movies: Towards Story-like Visual Explanations by
Watching Movies and Reading Books | null | null | null | null | cs.CV cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Books are a rich source of both fine-grained information, how a character, an
object or a scene looks like, as well as high-level semantics, what someone is
thinking, feeling and how these states evolve through a story. This paper aims
to align books to their movie releases in order to provide rich descriptive
explanations for visual content that go semantically far beyond the captions
available in current datasets. To align movies and books we exploit a neural
sentence embedding that is trained in an unsupervised way from a large corpus
of books, as well as a video-text neural embedding for computing similarities
between movie clips and sentences in the book. We propose a context-aware CNN
to combine information from multiple sources. We demonstrate good quantitative
performance for movie/book alignment and show several qualitative examples that
showcase the diversity of tasks our model can be used for.
| [
{
"version": "v1",
"created": "Mon, 22 Jun 2015 19:26:56 GMT"
}
] | 2015-06-23T00:00:00 | [
[
"Zhu",
"Yukun",
""
],
[
"Kiros",
"Ryan",
""
],
[
"Zemel",
"Richard",
""
],
[
"Salakhutdinov",
"Ruslan",
""
],
[
"Urtasun",
"Raquel",
""
],
[
"Torralba",
"Antonio",
""
],
[
"Fidler",
"Sanja",
""
]
] | TITLE: Aligning Books and Movies: Towards Story-like Visual Explanations by
Watching Movies and Reading Books
ABSTRACT: Books are a rich source of both fine-grained information, how a character, an
object or a scene looks like, as well as high-level semantics, what someone is
thinking, feeling and how these states evolve through a story. This paper aims
to align books to their movie releases in order to provide rich descriptive
explanations for visual content that go semantically far beyond the captions
available in current datasets. To align movies and books we exploit a neural
sentence embedding that is trained in an unsupervised way from a large corpus
of books, as well as a video-text neural embedding for computing similarities
between movie clips and sentences in the book. We propose a context-aware CNN
to combine information from multiple sources. We demonstrate good quantitative
performance for movie/book alignment and show several qualitative examples that
showcase the diversity of tasks our model can be used for.
| no_new_dataset | 0.948106 |
1506.06726 | Ryan Kiros | Ryan Kiros, Yukun Zhu, Ruslan Salakhutdinov, Richard S. Zemel, Antonio
Torralba, Raquel Urtasun, Sanja Fidler | Skip-Thought Vectors | 11 pages | null | null | null | cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We describe an approach for unsupervised learning of a generic, distributed
sentence encoder. Using the continuity of text from books, we train an
encoder-decoder model that tries to reconstruct the surrounding sentences of an
encoded passage. Sentences that share semantic and syntactic properties are
thus mapped to similar vector representations. We next introduce a simple
vocabulary expansion method to encode words that were not seen as part of
training, allowing us to expand our vocabulary to a million words. After
training our model, we extract and evaluate our vectors with linear models on 8
tasks: semantic relatedness, paraphrase detection, image-sentence ranking,
question-type classification and 4 benchmark sentiment and subjectivity
datasets. The end result is an off-the-shelf encoder that can produce highly
generic sentence representations that are robust and perform well in practice.
We will make our encoder publicly available.
| [
{
"version": "v1",
"created": "Mon, 22 Jun 2015 19:33:40 GMT"
}
] | 2015-06-23T00:00:00 | [
[
"Kiros",
"Ryan",
""
],
[
"Zhu",
"Yukun",
""
],
[
"Salakhutdinov",
"Ruslan",
""
],
[
"Zemel",
"Richard S.",
""
],
[
"Torralba",
"Antonio",
""
],
[
"Urtasun",
"Raquel",
""
],
[
"Fidler",
"Sanja",
""
]
] | TITLE: Skip-Thought Vectors
ABSTRACT: We describe an approach for unsupervised learning of a generic, distributed
sentence encoder. Using the continuity of text from books, we train an
encoder-decoder model that tries to reconstruct the surrounding sentences of an
encoded passage. Sentences that share semantic and syntactic properties are
thus mapped to similar vector representations. We next introduce a simple
vocabulary expansion method to encode words that were not seen as part of
training, allowing us to expand our vocabulary to a million words. After
training our model, we extract and evaluate our vectors with linear models on 8
tasks: semantic relatedness, paraphrase detection, image-sentence ranking,
question-type classification and 4 benchmark sentiment and subjectivity
datasets. The end result is an off-the-shelf encoder that can produce highly
generic sentence representations that are robust and perform well in practice.
We will make our encoder publicly available.
| no_new_dataset | 0.944689 |
1406.6909 | Alexey Dosovitskiy | Alexey Dosovitskiy, Philipp Fischer, Jost Tobias Springenberg, Martin
Riedmiller and Thomas Brox | Discriminative Unsupervised Feature Learning with Exemplar Convolutional
Neural Networks | PAMI submission. Includes matching experiments as in
arXiv:1405.5769v1. Also includes new network architectures, experiments on
Caltech-256, experiment on combining Exemplar-CNN with clustering | null | null | null | cs.LG cs.CV cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep convolutional networks have proven to be very successful in learning
task specific features that allow for unprecedented performance on various
computer vision tasks. Training of such networks follows mostly the supervised
learning paradigm, where sufficiently many input-output pairs are required for
training. Acquisition of large training sets is one of the key challenges, when
approaching a new task. In this paper, we aim for generic feature learning and
present an approach for training a convolutional network using only unlabeled
data. To this end, we train the network to discriminate between a set of
surrogate classes. Each surrogate class is formed by applying a variety of
transformations to a randomly sampled 'seed' image patch. In contrast to
supervised network training, the resulting feature representation is not class
specific. It rather provides robustness to the transformations that have been
applied during training. This generic feature representation allows for
classification results that outperform the state of the art for unsupervised
learning on several popular datasets (STL-10, CIFAR-10, Caltech-101,
Caltech-256). While such generic features cannot compete with class specific
features from supervised training on a classification task, we show that they
are advantageous on geometric matching problems, where they also outperform the
SIFT descriptor.
| [
{
"version": "v1",
"created": "Thu, 26 Jun 2014 15:07:14 GMT"
},
{
"version": "v2",
"created": "Fri, 19 Jun 2015 11:43:36 GMT"
}
] | 2015-06-22T00:00:00 | [
[
"Dosovitskiy",
"Alexey",
""
],
[
"Fischer",
"Philipp",
""
],
[
"Springenberg",
"Jost Tobias",
""
],
[
"Riedmiller",
"Martin",
""
],
[
"Brox",
"Thomas",
""
]
] | TITLE: Discriminative Unsupervised Feature Learning with Exemplar Convolutional
Neural Networks
ABSTRACT: Deep convolutional networks have proven to be very successful in learning
task specific features that allow for unprecedented performance on various
computer vision tasks. Training of such networks follows mostly the supervised
learning paradigm, where sufficiently many input-output pairs are required for
training. Acquisition of large training sets is one of the key challenges, when
approaching a new task. In this paper, we aim for generic feature learning and
present an approach for training a convolutional network using only unlabeled
data. To this end, we train the network to discriminate between a set of
surrogate classes. Each surrogate class is formed by applying a variety of
transformations to a randomly sampled 'seed' image patch. In contrast to
supervised network training, the resulting feature representation is not class
specific. It rather provides robustness to the transformations that have been
applied during training. This generic feature representation allows for
classification results that outperform the state of the art for unsupervised
learning on several popular datasets (STL-10, CIFAR-10, Caltech-101,
Caltech-256). While such generic features cannot compete with class specific
features from supervised training on a classification task, we show that they
are advantageous on geometric matching problems, where they also outperform the
SIFT descriptor.
| no_new_dataset | 0.944125 |
1406.7187 | Maziar Hemati | Maziar S. Hemati, Matthew O. Williams, and Clarence W. Rowley | Dynamic Mode Decomposition for Large and Streaming Datasets | null | null | 10.1063/1.4901016 | null | physics.flu-dyn | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We formulate a low-storage method for performing dynamic mode decomposition
that can be updated inexpensively as new data become available; this
formulation allows dynamical information to be extracted from large datasets
and data streams. We present two algorithms: the first is mathematically
equivalent to a standard "batch-processed" formulation; the second introduces a
compression step that maintains computational efficiency, while enhancing the
ability to isolate pertinent dynamical information from noisy measurements.
Both algorithms reliably capture dominant fluid dynamic behaviors, as
demonstrated on cylinder wake data collected from both direct numerical
simulations and particle image velocimetry experiments
| [
{
"version": "v1",
"created": "Fri, 27 Jun 2014 14:07:11 GMT"
}
] | 2015-06-22T00:00:00 | [
[
"Hemati",
"Maziar S.",
""
],
[
"Williams",
"Matthew O.",
""
],
[
"Rowley",
"Clarence W.",
""
]
] | TITLE: Dynamic Mode Decomposition for Large and Streaming Datasets
ABSTRACT: We formulate a low-storage method for performing dynamic mode decomposition
that can be updated inexpensively as new data become available; this
formulation allows dynamical information to be extracted from large datasets
and data streams. We present two algorithms: the first is mathematically
equivalent to a standard "batch-processed" formulation; the second introduces a
compression step that maintains computational efficiency, while enhancing the
ability to isolate pertinent dynamical information from noisy measurements.
Both algorithms reliably capture dominant fluid dynamic behaviors, as
demonstrated on cylinder wake data collected from both direct numerical
simulations and particle image velocimetry experiments
| no_new_dataset | 0.953362 |
1408.0365 | Will Ball | William T. Ball, Natalie A. Krivova, Yvonne C. Unruh, Joanna D. Haigh,
Sami K. Solanki | A new SATIRE-S spectral solar irradiance reconstruction for solar cycles
21--23 and its implications for stratospheric ozone | 25 pages (18 pages in main article with 6 figures; 7 pages in
supplementary materials with 6 figures) in draft mode using the American
Meteorological Society package. Submitted to Journal of Atmospheric Sciences
for publication | null | 10.1175/JAS-D-13-0241.1 | null | physics.ao-ph astro-ph.SR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a revised and extended total and spectral solar irradiance (SSI)
reconstruction, which includes a wavelength-dependent uncertainty estimate,
spanning the last three solar cycles using the SATIRE-S model. The SSI
reconstruction covers wavelengths between 115 and 160,000 nm and all dates
between August 1974 and October 2009. This represents the first full-wavelength
SATIRE-S reconstruction to cover the last three solar cycles without data gaps
and with an uncertainty estimate. SATIRE-S is compared with the NRLSSI model
and SORCE/SOLSTICE ultraviolet (UV) observations. SATIRE-S displays similar
cycle behaviour to NRLSSI for wavelengths below 242 nm and almost twice the
variability between 242 and 310 nm. During the decline of last solar cycle,
between 2003 and 2008, SSI from SORCE/SOLSTICE version 12 and 10 typically
displays more than three times the variability of SATIRE-S between 200 and 300
nm. All three datasets are used to model changes in stratospheric ozone within
a 2D atmospheric model for a decline from high solar activity to solar minimum.
The different flux changes result in different modelled ozone trends. Using
NRLSSI leads to a decline in mesospheric ozone, while SATIRE-S and
SORCE/SOLSTICE result in an increase. Recent publications have highlighted
increases in mesospheric ozone when considering version 10 SORCE/SOLSTICE
irradiances. The recalibrated SORCE/SOLSTICE version 12 irradiances result in a
much smaller mesospheric ozone response than when using version 10 and now
similar in magnitude to SATIRE-S. This shows that current knowledge of
variations in spectral irradiance is not sufficient to warrant robust
conclusions concerning the impact of solar variability on the atmosphere and
climate.
| [
{
"version": "v1",
"created": "Sat, 2 Aug 2014 12:40:51 GMT"
}
] | 2015-06-22T00:00:00 | [
[
"Ball",
"William T.",
""
],
[
"Krivova",
"Natalie A.",
""
],
[
"Unruh",
"Yvonne C.",
""
],
[
"Haigh",
"Joanna D.",
""
],
[
"Solanki",
"Sami K.",
""
]
] | TITLE: A new SATIRE-S spectral solar irradiance reconstruction for solar cycles
21--23 and its implications for stratospheric ozone
ABSTRACT: We present a revised and extended total and spectral solar irradiance (SSI)
reconstruction, which includes a wavelength-dependent uncertainty estimate,
spanning the last three solar cycles using the SATIRE-S model. The SSI
reconstruction covers wavelengths between 115 and 160,000 nm and all dates
between August 1974 and October 2009. This represents the first full-wavelength
SATIRE-S reconstruction to cover the last three solar cycles without data gaps
and with an uncertainty estimate. SATIRE-S is compared with the NRLSSI model
and SORCE/SOLSTICE ultraviolet (UV) observations. SATIRE-S displays similar
cycle behaviour to NRLSSI for wavelengths below 242 nm and almost twice the
variability between 242 and 310 nm. During the decline of last solar cycle,
between 2003 and 2008, SSI from SORCE/SOLSTICE version 12 and 10 typically
displays more than three times the variability of SATIRE-S between 200 and 300
nm. All three datasets are used to model changes in stratospheric ozone within
a 2D atmospheric model for a decline from high solar activity to solar minimum.
The different flux changes result in different modelled ozone trends. Using
NRLSSI leads to a decline in mesospheric ozone, while SATIRE-S and
SORCE/SOLSTICE result in an increase. Recent publications have highlighted
increases in mesospheric ozone when considering version 10 SORCE/SOLSTICE
irradiances. The recalibrated SORCE/SOLSTICE version 12 irradiances result in a
much smaller mesospheric ozone response than when using version 10 and now
similar in magnitude to SATIRE-S. This shows that current knowledge of
variations in spectral irradiance is not sufficient to warrant robust
conclusions concerning the impact of solar variability on the atmosphere and
climate.
| no_new_dataset | 0.948298 |
1408.1519 | Chlo\"e Brown | Chlo\"e Brown, Neal Lathia, Anastasios Noulas, Cecilia Mascolo,
Vincent Blondel | Group colocation behavior in technological social networks | 7 pages, 8 figures. Accepted for publication in PLOS One | null | 10.1371/journal.pone.0105816 | null | cs.SI physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We analyze two large datasets from technological networks with location and
social data: user location records from an online location-based social
networking service, and anonymized telecommunications data from a European
cellphone operator, in order to investigate the differences between individual
and group behavior with respect to physical location. We discover agreements
between the two datasets: firstly, that individuals are more likely to meet
with one friend at a place they have not visited before, but tend to meet at
familiar locations when with a larger group. We also find that groups of
individuals are more likely to meet at places that their other friends have
visited, and that the type of a place strongly affects the propensity for
groups to meet there. These differences between group and solo mobility has
potential technological applications, for example, in venue recommendation in
location-based social networks.
| [
{
"version": "v1",
"created": "Thu, 7 Aug 2014 09:18:17 GMT"
},
{
"version": "v2",
"created": "Fri, 8 Aug 2014 08:01:48 GMT"
}
] | 2015-06-22T00:00:00 | [
[
"Brown",
"Chloë",
""
],
[
"Lathia",
"Neal",
""
],
[
"Noulas",
"Anastasios",
""
],
[
"Mascolo",
"Cecilia",
""
],
[
"Blondel",
"Vincent",
""
]
] | TITLE: Group colocation behavior in technological social networks
ABSTRACT: We analyze two large datasets from technological networks with location and
social data: user location records from an online location-based social
networking service, and anonymized telecommunications data from a European
cellphone operator, in order to investigate the differences between individual
and group behavior with respect to physical location. We discover agreements
between the two datasets: firstly, that individuals are more likely to meet
with one friend at a place they have not visited before, but tend to meet at
familiar locations when with a larger group. We also find that groups of
individuals are more likely to meet at places that their other friends have
visited, and that the type of a place strongly affects the propensity for
groups to meet there. These differences between group and solo mobility has
potential technological applications, for example, in venue recommendation in
location-based social networks.
| no_new_dataset | 0.940463 |
1408.5240 | Shimin Cai Dr | Lili Miao, Qian-Ming Zhang, Da-Chen Nie, Shi-Min Cai | Whether Information Network Supplements Friendship Network | 8 pages, 5 figures | Physica A 419, 301 (2015) | 10.1016/j.physa.2014.10.021 | null | physics.soc-ph cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Homophily is a significant mechanism for link prediction in complex network,
of which principle describes that people with similar profiles or experiences
tend to tie with each other. In a multi-relationship network, friendship among
people has been utilized to reinforce similarity of taste for recommendation
system whose basic idea is similar to homophily, yet how the taste inversely
affects friendship prediction is little discussed. This paper contributes to
address the issue by analyzing two benchmark datasets both including user's
behavioral information of taste and friendship based on the principle of
homophily. It can be found that the creation of friendship tightly associates
with personal taste. Especially, the behavioral information of taste involving
with popular objects is much more effective to improve the performance of
friendship prediction. However, this result seems to be contradictory to the
finding in [Q.M. Zhang, et al., PLoS ONE 8(2013)e62624] that the behavior
information of taste involving with popular objects is redundant in
recommendation system. We thus discuss this inconformity to comprehensively
understand the correlation between them.
| [
{
"version": "v1",
"created": "Fri, 22 Aug 2014 09:31:48 GMT"
}
] | 2015-06-22T00:00:00 | [
[
"Miao",
"Lili",
""
],
[
"Zhang",
"Qian-Ming",
""
],
[
"Nie",
"Da-Chen",
""
],
[
"Cai",
"Shi-Min",
""
]
] | TITLE: Whether Information Network Supplements Friendship Network
ABSTRACT: Homophily is a significant mechanism for link prediction in complex network,
of which principle describes that people with similar profiles or experiences
tend to tie with each other. In a multi-relationship network, friendship among
people has been utilized to reinforce similarity of taste for recommendation
system whose basic idea is similar to homophily, yet how the taste inversely
affects friendship prediction is little discussed. This paper contributes to
address the issue by analyzing two benchmark datasets both including user's
behavioral information of taste and friendship based on the principle of
homophily. It can be found that the creation of friendship tightly associates
with personal taste. Especially, the behavioral information of taste involving
with popular objects is much more effective to improve the performance of
friendship prediction. However, this result seems to be contradictory to the
finding in [Q.M. Zhang, et al., PLoS ONE 8(2013)e62624] that the behavior
information of taste involving with popular objects is redundant in
recommendation system. We thus discuss this inconformity to comprehensively
understand the correlation between them.
| no_new_dataset | 0.945901 |
1409.2944 | Hao Wang | Hao Wang and Naiyan Wang and Dit-Yan Yeung | Collaborative Deep Learning for Recommender Systems | null | null | null | null | cs.LG cs.CL cs.IR cs.NE stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Collaborative filtering (CF) is a successful approach commonly used by many
recommender systems. Conventional CF-based methods use the ratings given to
items by users as the sole source of information for learning to make
recommendation. However, the ratings are often very sparse in many
applications, causing CF-based methods to degrade significantly in their
recommendation performance. To address this sparsity problem, auxiliary
information such as item content information may be utilized. Collaborative
topic regression (CTR) is an appealing recent method taking this approach which
tightly couples the two components that learn from two different sources of
information. Nevertheless, the latent representation learned by CTR may not be
very effective when the auxiliary information is very sparse. To address this
problem, we generalize recent advances in deep learning from i.i.d. input to
non-i.i.d. (CF-based) input and propose in this paper a hierarchical Bayesian
model called collaborative deep learning (CDL), which jointly performs deep
representation learning for the content information and collaborative filtering
for the ratings (feedback) matrix. Extensive experiments on three real-world
datasets from different domains show that CDL can significantly advance the
state of the art.
| [
{
"version": "v1",
"created": "Wed, 10 Sep 2014 03:05:22 GMT"
},
{
"version": "v2",
"created": "Thu, 18 Jun 2015 09:23:37 GMT"
}
] | 2015-06-22T00:00:00 | [
[
"Wang",
"Hao",
""
],
[
"Wang",
"Naiyan",
""
],
[
"Yeung",
"Dit-Yan",
""
]
] | TITLE: Collaborative Deep Learning for Recommender Systems
ABSTRACT: Collaborative filtering (CF) is a successful approach commonly used by many
recommender systems. Conventional CF-based methods use the ratings given to
items by users as the sole source of information for learning to make
recommendation. However, the ratings are often very sparse in many
applications, causing CF-based methods to degrade significantly in their
recommendation performance. To address this sparsity problem, auxiliary
information such as item content information may be utilized. Collaborative
topic regression (CTR) is an appealing recent method taking this approach which
tightly couples the two components that learn from two different sources of
information. Nevertheless, the latent representation learned by CTR may not be
very effective when the auxiliary information is very sparse. To address this
problem, we generalize recent advances in deep learning from i.i.d. input to
non-i.i.d. (CF-based) input and propose in this paper a hierarchical Bayesian
model called collaborative deep learning (CDL), which jointly performs deep
representation learning for the content information and collaborative filtering
for the ratings (feedback) matrix. Extensive experiments on three real-world
datasets from different domains show that CDL can significantly advance the
state of the art.
| no_new_dataset | 0.948251 |
1506.02344 | Anastasios Kyrillidis | Megasthenis Asteris, Anastasios Kyrillidis, Alexandros G. Dimakis,
Han-Gyol Yi and, Bharath Chandrasekaran | Stay on path: PCA along graph paths | 12 pages, 5 figures, In Proceedings of International Conference on
Machine Learning (ICML) 2015 | null | null | null | stat.ML cs.IT cs.LG math.IT math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We introduce a variant of (sparse) PCA in which the set of feasible support
sets is determined by a graph. In particular, we consider the following
setting: given a directed acyclic graph $G$ on $p$ vertices corresponding to
variables, the non-zero entries of the extracted principal component must
coincide with vertices lying along a path in $G$.
From a statistical perspective, information on the underlying network may
potentially reduce the number of observations required to recover the
population principal component. We consider the canonical estimator which
optimally exploits the prior knowledge by solving a non-convex quadratic
maximization on the empirical covariance. We introduce a simple network and
analyze the estimator under the spiked covariance model. We show that side
information potentially improves the statistical complexity.
We propose two algorithms to approximate the solution of the constrained
quadratic maximization, and recover a component with the desired properties. We
empirically evaluate our schemes on synthetic and real datasets.
| [
{
"version": "v1",
"created": "Mon, 8 Jun 2015 03:37:36 GMT"
},
{
"version": "v2",
"created": "Fri, 19 Jun 2015 02:27:49 GMT"
}
] | 2015-06-22T00:00:00 | [
[
"Asteris",
"Megasthenis",
""
],
[
"Kyrillidis",
"Anastasios",
""
],
[
"Dimakis",
"Alexandros G.",
""
],
[
"and",
"Han-Gyol Yi",
""
],
[
"Chandrasekaran",
"Bharath",
""
]
] | TITLE: Stay on path: PCA along graph paths
ABSTRACT: We introduce a variant of (sparse) PCA in which the set of feasible support
sets is determined by a graph. In particular, we consider the following
setting: given a directed acyclic graph $G$ on $p$ vertices corresponding to
variables, the non-zero entries of the extracted principal component must
coincide with vertices lying along a path in $G$.
From a statistical perspective, information on the underlying network may
potentially reduce the number of observations required to recover the
population principal component. We consider the canonical estimator which
optimally exploits the prior knowledge by solving a non-convex quadratic
maximization on the empirical covariance. We introduce a simple network and
analyze the estimator under the spiked covariance model. We show that side
information potentially improves the statistical complexity.
We propose two algorithms to approximate the solution of the constrained
quadratic maximization, and recover a component with the desired properties. We
empirically evaluate our schemes on synthetic and real datasets.
| no_new_dataset | 0.941708 |
1506.05247 | Zhihai Yang | Zhihai Yang | Defending Grey Attacks by Exploiting Wavelet Analysis in Collaborative
Filtering Recommender Systems | 16 pages, 16 figures | null | null | null | cs.IR cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | "Shilling" attacks or "profile injection" attacks have always major
challenges in collaborative filtering recommender systems (CFRSs). Many efforts
have been devoted to improve collaborative filtering techniques which can
eliminate the "shilling" attacks. However, most of them focused on detecting
push attack or nuke attack which is rated with the highest score or lowest
score on the target items. Few pay attention to grey attack when a target item
is rated with a lower or higher score than the average score, which shows a
more hidden rating behavior than push or nuke attack. In this paper, we present
a novel detection method to make recommender systems resistant to such attacks.
To characterize grey ratings, we exploit rating deviation of item to
discriminate between grey attack profiles and genuine profiles. In addition, we
also employ novelty and popularity of item to construct rating series. Since it
is difficult to discriminate between the rating series of attacker and genuine
users, we incorporate into discrete wavelet transform (DWT) to amplify these
differences based on the rating series of rating deviation, novelty and
popularity, respectively. Finally, we respectively extract features from rating
series of rating deviation-based, novelty-based and popularity-based by using
amplitude domain analysis method and combine all clustered results as our
detection results. We conduct a list of experiments on both the Book-Crossing
and HetRec-2011 datasets in diverse attack models. Experimental results were
included to validate the effectiveness of our approach in comparison with the
benchmarked methods.
| [
{
"version": "v1",
"created": "Wed, 17 Jun 2015 08:54:04 GMT"
},
{
"version": "v2",
"created": "Fri, 19 Jun 2015 07:30:47 GMT"
}
] | 2015-06-22T00:00:00 | [
[
"Yang",
"Zhihai",
""
]
] | TITLE: Defending Grey Attacks by Exploiting Wavelet Analysis in Collaborative
Filtering Recommender Systems
ABSTRACT: "Shilling" attacks or "profile injection" attacks have always major
challenges in collaborative filtering recommender systems (CFRSs). Many efforts
have been devoted to improve collaborative filtering techniques which can
eliminate the "shilling" attacks. However, most of them focused on detecting
push attack or nuke attack which is rated with the highest score or lowest
score on the target items. Few pay attention to grey attack when a target item
is rated with a lower or higher score than the average score, which shows a
more hidden rating behavior than push or nuke attack. In this paper, we present
a novel detection method to make recommender systems resistant to such attacks.
To characterize grey ratings, we exploit rating deviation of item to
discriminate between grey attack profiles and genuine profiles. In addition, we
also employ novelty and popularity of item to construct rating series. Since it
is difficult to discriminate between the rating series of attacker and genuine
users, we incorporate into discrete wavelet transform (DWT) to amplify these
differences based on the rating series of rating deviation, novelty and
popularity, respectively. Finally, we respectively extract features from rating
series of rating deviation-based, novelty-based and popularity-based by using
amplitude domain analysis method and combine all clustered results as our
detection results. We conduct a list of experiments on both the Book-Crossing
and HetRec-2011 datasets in diverse attack models. Experimental results were
included to validate the effectiveness of our approach in comparison with the
benchmarked methods.
| no_new_dataset | 0.948822 |
1506.05870 | Kuan-Wen Chen | Kuan-Wen Chen, Chun-Hsin Wang, Xiao Wei, Qiao Liang, Ming-Hsuan Yang,
Chu-Song Chen, Yi-Ping Hung | To Know Where We Are: Vision-Based Positioning in Outdoor Environments | 11 pages, 14 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Augmented reality (AR) displays become more and more popular recently,
because of its high intuitiveness for humans and high-quality head-mounted
display have rapidly developed. To achieve such displays with augmented
information, highly accurate image registration or ego-positioning are
required, but little attention have been paid for out-door environments. This
paper presents a method for ego-positioning in outdoor environments with low
cost monocular cameras. To reduce the computational and memory requirements as
well as the communication overheads, we formulate the model compression
algorithm as a weighted k-cover problem for better preserving model structures.
Specifically for real-world vision-based positioning applications, we consider
the issues with large scene change and propose a model update algorithm to
tackle these problems. A long- term positioning dataset with more than one
month, 106 sessions, and 14,275 images is constructed. Based on both local and
up-to-date models constructed in our approach, extensive experimental results
show that high positioning accuracy (mean ~ 30.9cm, stdev. ~ 15.4cm) can be
achieved, which outperforms existing vision-based algorithms.
| [
{
"version": "v1",
"created": "Fri, 19 Jun 2015 03:11:33 GMT"
}
] | 2015-06-22T00:00:00 | [
[
"Chen",
"Kuan-Wen",
""
],
[
"Wang",
"Chun-Hsin",
""
],
[
"Wei",
"Xiao",
""
],
[
"Liang",
"Qiao",
""
],
[
"Yang",
"Ming-Hsuan",
""
],
[
"Chen",
"Chu-Song",
""
],
[
"Hung",
"Yi-Ping",
""
]
] | TITLE: To Know Where We Are: Vision-Based Positioning in Outdoor Environments
ABSTRACT: Augmented reality (AR) displays become more and more popular recently,
because of its high intuitiveness for humans and high-quality head-mounted
display have rapidly developed. To achieve such displays with augmented
information, highly accurate image registration or ego-positioning are
required, but little attention have been paid for out-door environments. This
paper presents a method for ego-positioning in outdoor environments with low
cost monocular cameras. To reduce the computational and memory requirements as
well as the communication overheads, we formulate the model compression
algorithm as a weighted k-cover problem for better preserving model structures.
Specifically for real-world vision-based positioning applications, we consider
the issues with large scene change and propose a model update algorithm to
tackle these problems. A long- term positioning dataset with more than one
month, 106 sessions, and 14,275 images is constructed. Based on both local and
up-to-date models constructed in our approach, extensive experimental results
show that high positioning accuracy (mean ~ 30.9cm, stdev. ~ 15.4cm) can be
achieved, which outperforms existing vision-based algorithms.
| new_dataset | 0.957991 |
1506.05908 | Chris Piech | Chris Piech, Jonathan Spencer, Jonathan Huang, Surya Ganguli, Mehran
Sahami, Leonidas Guibas, Jascha Sohl-Dickstein | Deep Knowledge Tracing | null | null | null | null | cs.AI cs.CY cs.LG | http://creativecommons.org/licenses/by-nc-sa/3.0/ | Knowledge tracing---where a machine models the knowledge of a student as they
interact with coursework---is a well established problem in computer supported
education. Though effectively modeling student knowledge would have high
educational impact, the task has many inherent challenges. In this paper we
explore the utility of using Recurrent Neural Networks (RNNs) to model student
learning. The RNN family of models have important advantages over previous
methods in that they do not require the explicit encoding of human domain
knowledge, and can capture more complex representations of student knowledge.
Using neural networks results in substantial improvements in prediction
performance on a range of knowledge tracing datasets. Moreover the learned
model can be used for intelligent curriculum design and allows straightforward
interpretation and discovery of structure in student tasks. These results
suggest a promising new line of research for knowledge tracing and an exemplary
application task for RNNs.
| [
{
"version": "v1",
"created": "Fri, 19 Jun 2015 08:29:00 GMT"
}
] | 2015-06-22T00:00:00 | [
[
"Piech",
"Chris",
""
],
[
"Spencer",
"Jonathan",
""
],
[
"Huang",
"Jonathan",
""
],
[
"Ganguli",
"Surya",
""
],
[
"Sahami",
"Mehran",
""
],
[
"Guibas",
"Leonidas",
""
],
[
"Sohl-Dickstein",
"Jascha",
""
]
] | TITLE: Deep Knowledge Tracing
ABSTRACT: Knowledge tracing---where a machine models the knowledge of a student as they
interact with coursework---is a well established problem in computer supported
education. Though effectively modeling student knowledge would have high
educational impact, the task has many inherent challenges. In this paper we
explore the utility of using Recurrent Neural Networks (RNNs) to model student
learning. The RNN family of models have important advantages over previous
methods in that they do not require the explicit encoding of human domain
knowledge, and can capture more complex representations of student knowledge.
Using neural networks results in substantial improvements in prediction
performance on a range of knowledge tracing datasets. Moreover the learned
model can be used for intelligent curriculum design and allows straightforward
interpretation and discovery of structure in student tasks. These results
suggest a promising new line of research for knowledge tracing and an exemplary
application task for RNNs.
| no_new_dataset | 0.947721 |
1506.05970 | Dominique Jault | G. Hellio, N. Gillet, C. Bouligand, D. Jault | Stochastic modelling of regional archaeomagnetic series | null | Geophysical Journal International, Oxford University Press (OUP):
Policy P - Oxford Open Option A, 2014, 199 (2), pp. 931-943 | 10.1093/gji/ggu303 | null | physics.geo-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | SUMMARY We report a new method to infer continuous time series of the
declination, inclination and intensity of the magnetic field from
archeomagnetic data. Adopting a Bayesian perspective, we need to specify a
priori knowledge about the time evolution of the magnetic field. It consists in
a time correlation function that we choose to be compatible with present
knowledge about the geomagnetic time spectra. The results are presented as
distributions of possible values for the declination, inclination or intensity.
We find that the methodology can be adapted to account for the age
uncertainties of archeological artefacts and we use Markov Chain Monte Carlo to
explore the possible dates of observations. We apply the method to intensity
datasets from Mari, Syria and to intensity and directional datasets from Paris,
France. Our reconstructions display more rapid variations than previous studies
and we find that the possible values of geomagnetic field elements are not
necessarily normally distributed. Another output of the model is better age
estimates of archeological artefacts.
| [
{
"version": "v1",
"created": "Fri, 19 Jun 2015 12:14:55 GMT"
}
] | 2015-06-22T00:00:00 | [
[
"Hellio",
"G.",
""
],
[
"Gillet",
"N.",
""
],
[
"Bouligand",
"C.",
""
],
[
"Jault",
"D.",
""
]
] | TITLE: Stochastic modelling of regional archaeomagnetic series
ABSTRACT: SUMMARY We report a new method to infer continuous time series of the
declination, inclination and intensity of the magnetic field from
archeomagnetic data. Adopting a Bayesian perspective, we need to specify a
priori knowledge about the time evolution of the magnetic field. It consists in
a time correlation function that we choose to be compatible with present
knowledge about the geomagnetic time spectra. The results are presented as
distributions of possible values for the declination, inclination or intensity.
We find that the methodology can be adapted to account for the age
uncertainties of archeological artefacts and we use Markov Chain Monte Carlo to
explore the possible dates of observations. We apply the method to intensity
datasets from Mari, Syria and to intensity and directional datasets from Paris,
France. Our reconstructions display more rapid variations than previous studies
and we find that the possible values of geomagnetic field elements are not
necessarily normally distributed. Another output of the model is better age
estimates of archeological artefacts.
| no_new_dataset | 0.947769 |
1506.05985 | Xavier Bresson | Xavier Bresson and Thomas Laurent and James von Brecht | Enhanced Lasso Recovery on Graph | null | null | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This work aims at recovering signals that are sparse on graphs. Compressed
sensing offers techniques for signal recovery from a few linear measurements
and graph Fourier analysis provides a signal representation on graph. In this
paper, we leverage these two frameworks to introduce a new Lasso recovery
algorithm on graphs. More precisely, we present a non-convex, non-smooth
algorithm that outperforms the standard convex Lasso technique. We carry out
numerical experiments on three benchmark graph datasets.
| [
{
"version": "v1",
"created": "Fri, 19 Jun 2015 12:59:18 GMT"
}
] | 2015-06-22T00:00:00 | [
[
"Bresson",
"Xavier",
""
],
[
"Laurent",
"Thomas",
""
],
[
"von Brecht",
"James",
""
]
] | TITLE: Enhanced Lasso Recovery on Graph
ABSTRACT: This work aims at recovering signals that are sparse on graphs. Compressed
sensing offers techniques for signal recovery from a few linear measurements
and graph Fourier analysis provides a signal representation on graph. In this
paper, we leverage these two frameworks to introduce a new Lasso recovery
algorithm on graphs. More precisely, we present a non-convex, non-smooth
algorithm that outperforms the standard convex Lasso technique. We carry out
numerical experiments on three benchmark graph datasets.
| no_new_dataset | 0.950915 |
1506.06006 | Srinivas S S Kruthiventi | Srinivas S. S. Kruthiventi and R. Venkatesh Babu | Crowd Flow Segmentation in Compressed Domain using CRF | In IEEE International Conference on Image Processing (ICIP), 2015 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Crowd flow segmentation is an important step in many video surveillance
tasks. In this work, we propose an algorithm for segmenting flows in H.264
compressed videos in a completely unsupervised manner. Our algorithm works on
motion vectors which can be obtained by partially decoding the compressed video
without extracting any additional features. Our approach is based on modelling
the motion vector field as a Conditional Random Field (CRF) and obtaining
oriented motion segments by finding the optimal labelling which minimises the
global energy of CRF. These oriented motion segments are recursively merged
based on gradient across their boundaries to obtain the final flow segments.
This work in compressed domain can be easily extended to pixel domain by
substituting motion vectors with motion based features like optical flow. The
proposed algorithm is experimentally evaluated on a standard crowd flow dataset
and its superior performance in both accuracy and computational time are
demonstrated through quantitative results.
| [
{
"version": "v1",
"created": "Fri, 19 Jun 2015 14:01:24 GMT"
}
] | 2015-06-22T00:00:00 | [
[
"Kruthiventi",
"Srinivas S. S.",
""
],
[
"Babu",
"R. Venkatesh",
""
]
] | TITLE: Crowd Flow Segmentation in Compressed Domain using CRF
ABSTRACT: Crowd flow segmentation is an important step in many video surveillance
tasks. In this work, we propose an algorithm for segmenting flows in H.264
compressed videos in a completely unsupervised manner. Our algorithm works on
motion vectors which can be obtained by partially decoding the compressed video
without extracting any additional features. Our approach is based on modelling
the motion vector field as a Conditional Random Field (CRF) and obtaining
oriented motion segments by finding the optimal labelling which minimises the
global energy of CRF. These oriented motion segments are recursively merged
based on gradient across their boundaries to obtain the final flow segments.
This work in compressed domain can be easily extended to pixel domain by
substituting motion vectors with motion based features like optical flow. The
proposed algorithm is experimentally evaluated on a standard crowd flow dataset
and its superior performance in both accuracy and computational time are
demonstrated through quantitative results.
| no_new_dataset | 0.957715 |
1506.06068 | Teng Qiu | Teng Qiu, Yongjie Li | A general framework for the IT-based clustering methods | 17 pages | null | null | null | cs.CV cs.LG stat.ML | http://creativecommons.org/licenses/by-nc-sa/3.0/ | Previously, we proposed a physically inspired rule to organize the data
points in a sparse yet effective structure, called the in-tree (IT) graph,
which is able to capture a wide class of underlying cluster structures in the
datasets, especially for the density-based datasets. Although there are some
redundant edges or lines between clusters requiring to be removed by computer,
this IT graph has a big advantage compared with the k-nearest-neighborhood
(k-NN) or the minimal spanning tree (MST) graph, in that the redundant edges in
the IT graph are much more distinguishable and thus can be easily determined by
several methods previously proposed by us.
In this paper, we propose a general framework to re-construct the IT graph,
based on an initial neighborhood graph, such as the k-NN or MST, etc, and the
corresponding graph distances. For this general framework, our previous way of
constructing the IT graph turns out to be a special case of it. This general
framework 1) can make the IT graph capture a wider class of underlying cluster
structures in the datasets, especially for the manifolds, and 2) should be more
effective to cluster the sparse or graph-based datasets.
| [
{
"version": "v1",
"created": "Fri, 19 Jun 2015 16:03:31 GMT"
}
] | 2015-06-22T00:00:00 | [
[
"Qiu",
"Teng",
""
],
[
"Li",
"Yongjie",
""
]
] | TITLE: A general framework for the IT-based clustering methods
ABSTRACT: Previously, we proposed a physically inspired rule to organize the data
points in a sparse yet effective structure, called the in-tree (IT) graph,
which is able to capture a wide class of underlying cluster structures in the
datasets, especially for the density-based datasets. Although there are some
redundant edges or lines between clusters requiring to be removed by computer,
this IT graph has a big advantage compared with the k-nearest-neighborhood
(k-NN) or the minimal spanning tree (MST) graph, in that the redundant edges in
the IT graph are much more distinguishable and thus can be easily determined by
several methods previously proposed by us.
In this paper, we propose a general framework to re-construct the IT graph,
based on an initial neighborhood graph, such as the k-NN or MST, etc, and the
corresponding graph distances. For this general framework, our previous way of
constructing the IT graph turns out to be a special case of it. This general
framework 1) can make the IT graph capture a wider class of underlying cluster
structures in the datasets, especially for the manifolds, and 2) should be more
effective to cluster the sparse or graph-based datasets.
| no_new_dataset | 0.952706 |
1403.4462 | Andrzej Cichocki | A. Cichocki, D. Mandic, A-H. Phan, C. Caiafa, G. Zhou, Q. Zhao, and L.
De Lathauwer | Tensor Decompositions for Signal Processing Applications From Two-way to
Multiway Component Analysis | null | null | 10.1109/MSP.2013.2297439 | null | cs.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The widespread use of multi-sensor technology and the emergence of big
datasets has highlighted the limitations of standard flat-view matrix models
and the necessity to move towards more versatile data analysis tools. We show
that higher-order tensors (i.e., multiway arrays) enable such a fundamental
paradigm shift towards models that are essentially polynomial and whose
uniqueness, unlike the matrix methods, is guaranteed under verymild and natural
conditions. Benefiting fromthe power ofmultilinear algebra as theirmathematical
backbone, data analysis techniques using tensor decompositions are shown to
have great flexibility in the choice of constraints that match data properties,
and to find more general latent components in the data than matrix-based
methods. A comprehensive introduction to tensor decompositions is provided from
a signal processing perspective, starting from the algebraic foundations, via
basic Canonical Polyadic and Tucker models, through to advanced cause-effect
and multi-view data analysis schemes. We show that tensor decompositions enable
natural generalizations of some commonly used signal processing paradigms, such
as canonical correlation and subspace techniques, signal separation, linear
regression, feature extraction and classification. We also cover computational
aspects, and point out how ideas from compressed sensing and scientific
computing may be used for addressing the otherwise unmanageable storage and
manipulation problems associated with big datasets. The concepts are supported
by illustrative real world case studies illuminating the benefits of the tensor
framework, as efficient and promising tools for modern signal processing, data
analysis and machine learning applications; these benefits also extend to
vector/matrix data through tensorization. Keywords: ICA, NMF, CPD, Tucker
decomposition, HOSVD, tensor networks, Tensor Train.
| [
{
"version": "v1",
"created": "Mon, 17 Mar 2014 11:03:58 GMT"
}
] | 2015-06-19T00:00:00 | [
[
"Cichocki",
"A.",
""
],
[
"Mandic",
"D.",
""
],
[
"Phan",
"A-H.",
""
],
[
"Caiafa",
"C.",
""
],
[
"Zhou",
"G.",
""
],
[
"Zhao",
"Q.",
""
],
[
"De Lathauwer",
"L.",
""
]
] | TITLE: Tensor Decompositions for Signal Processing Applications From Two-way to
Multiway Component Analysis
ABSTRACT: The widespread use of multi-sensor technology and the emergence of big
datasets has highlighted the limitations of standard flat-view matrix models
and the necessity to move towards more versatile data analysis tools. We show
that higher-order tensors (i.e., multiway arrays) enable such a fundamental
paradigm shift towards models that are essentially polynomial and whose
uniqueness, unlike the matrix methods, is guaranteed under verymild and natural
conditions. Benefiting fromthe power ofmultilinear algebra as theirmathematical
backbone, data analysis techniques using tensor decompositions are shown to
have great flexibility in the choice of constraints that match data properties,
and to find more general latent components in the data than matrix-based
methods. A comprehensive introduction to tensor decompositions is provided from
a signal processing perspective, starting from the algebraic foundations, via
basic Canonical Polyadic and Tucker models, through to advanced cause-effect
and multi-view data analysis schemes. We show that tensor decompositions enable
natural generalizations of some commonly used signal processing paradigms, such
as canonical correlation and subspace techniques, signal separation, linear
regression, feature extraction and classification. We also cover computational
aspects, and point out how ideas from compressed sensing and scientific
computing may be used for addressing the otherwise unmanageable storage and
manipulation problems associated with big datasets. The concepts are supported
by illustrative real world case studies illuminating the benefits of the tensor
framework, as efficient and promising tools for modern signal processing, data
analysis and machine learning applications; these benefits also extend to
vector/matrix data through tensorization. Keywords: ICA, NMF, CPD, Tucker
decomposition, HOSVD, tensor networks, Tensor Train.
| no_new_dataset | 0.944382 |
1403.4590 | Kareem Osman | K. T. Osman, W. H. Matthaeus, J. T. Gosling, A. Greco, S. Servidio, B.
Hnat, S. C. Chapman, and T. D. Phan | Magnetic Reconnection and Intermittent Turbulence in the Solar Wind | 5 pages, 3 figures, submitted to Physical Review Letters | null | 10.1103/PhysRevLett.112.215002 | null | physics.space-ph physics.data-an physics.plasm-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A statistical relationship between magnetic reconnection, current sheets and
intermittent turbulence in the solar wind is reported for the first time using
in-situ measurements from the Wind spacecraft at 1 AU. We identify
intermittency as non-Gaussian fluctuations in increments of the magnetic field
vector, $\mathbf{B}$, that are spatially and temporally non-uniform. The
reconnection events and current sheets are found to be concentrated in
intervals of intermittent turbulence, identified using the partial variance of
increments method: within the most non-Gaussian 1% of fluctuations in
$\mathbf{B}$, we find 87%-92% of reconnection exhausts and $\sim$9% of current
sheets. Also, the likelihood that an identified current sheet will also
correspond to a reconnection exhaust increases dramatically as the least
intermittent fluctuations are removed from the dataset. Hence, the turbulent
solar wind contains a hierarchy of intermittent magnetic field structures that
are increasingly linked to current sheets, which in turn are progressively more
likely to correspond to sites of magnetic reconnection. These results could
have far reaching implications for laboratory and astrophysical plasmas where
turbulence and magnetic reconnection are ubiquitous.
| [
{
"version": "v1",
"created": "Tue, 18 Mar 2014 19:45:07 GMT"
}
] | 2015-06-19T00:00:00 | [
[
"Osman",
"K. T.",
""
],
[
"Matthaeus",
"W. H.",
""
],
[
"Gosling",
"J. T.",
""
],
[
"Greco",
"A.",
""
],
[
"Servidio",
"S.",
""
],
[
"Hnat",
"B.",
""
],
[
"Chapman",
"S. C.",
""
],
[
"Phan",
"T. D.",
""
]
] | TITLE: Magnetic Reconnection and Intermittent Turbulence in the Solar Wind
ABSTRACT: A statistical relationship between magnetic reconnection, current sheets and
intermittent turbulence in the solar wind is reported for the first time using
in-situ measurements from the Wind spacecraft at 1 AU. We identify
intermittency as non-Gaussian fluctuations in increments of the magnetic field
vector, $\mathbf{B}$, that are spatially and temporally non-uniform. The
reconnection events and current sheets are found to be concentrated in
intervals of intermittent turbulence, identified using the partial variance of
increments method: within the most non-Gaussian 1% of fluctuations in
$\mathbf{B}$, we find 87%-92% of reconnection exhausts and $\sim$9% of current
sheets. Also, the likelihood that an identified current sheet will also
correspond to a reconnection exhaust increases dramatically as the least
intermittent fluctuations are removed from the dataset. Hence, the turbulent
solar wind contains a hierarchy of intermittent magnetic field structures that
are increasingly linked to current sheets, which in turn are progressively more
likely to correspond to sites of magnetic reconnection. These results could
have far reaching implications for laboratory and astrophysical plasmas where
turbulence and magnetic reconnection are ubiquitous.
| no_new_dataset | 0.949576 |
1403.5156 | Daniele Marinazzo | Sebastiano Stramaglia, Jesus M. Cortes, Daniele Marinazzo | Synergy and redundancy in the Granger causal analysis of dynamical
networks | null | null | 10.1088/1367-2630/16/10/105003 | null | q-bio.QM cs.IT math.IT physics.data-an | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We analyze by means of Granger causality the effect of synergy and redundancy
in the inference (from time series data) of the information flow between
subsystems of a complex network. Whilst we show that fully conditioned Granger
causality is not affected by synergy, the pairwise analysis fails to put in
evidence synergetic effects.
In cases when the number of samples is low, thus making the fully conditioned
approach unfeasible, we show that partially conditioned Granger causality is an
effective approach if the set of conditioning variables is properly chosen. We
consider here two different strategies (based either on informational content
for the candidate driver or on selecting the variables with highest pairwise
influences) for partially conditioned Granger causality and show that depending
on the data structure either one or the other might be valid. On the other
hand, we observe that fully conditioned approaches do not work well in presence
of redundancy, thus suggesting the strategy of separating the pairwise links in
two subsets: those corresponding to indirect connections of the fully
conditioned Granger causality (which should thus be excluded) and links that
can be ascribed to redundancy effects and, together with the results from the
fully connected approach, provide a better description of the causality pattern
in presence of redundancy. We finally apply these methods to two different real
datasets. First, analyzing electrophysiological data from an epileptic brain,
we show that synergetic effects are dominant just before seizure occurrences.
Second, our analysis applied to gene expression time series from HeLa culture
shows that the underlying regulatory networks are characterized by both
redundancy and synergy.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2014 14:49:27 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2014 22:38:24 GMT"
}
] | 2015-06-19T00:00:00 | [
[
"Stramaglia",
"Sebastiano",
""
],
[
"Cortes",
"Jesus M.",
""
],
[
"Marinazzo",
"Daniele",
""
]
] | TITLE: Synergy and redundancy in the Granger causal analysis of dynamical
networks
ABSTRACT: We analyze by means of Granger causality the effect of synergy and redundancy
in the inference (from time series data) of the information flow between
subsystems of a complex network. Whilst we show that fully conditioned Granger
causality is not affected by synergy, the pairwise analysis fails to put in
evidence synergetic effects.
In cases when the number of samples is low, thus making the fully conditioned
approach unfeasible, we show that partially conditioned Granger causality is an
effective approach if the set of conditioning variables is properly chosen. We
consider here two different strategies (based either on informational content
for the candidate driver or on selecting the variables with highest pairwise
influences) for partially conditioned Granger causality and show that depending
on the data structure either one or the other might be valid. On the other
hand, we observe that fully conditioned approaches do not work well in presence
of redundancy, thus suggesting the strategy of separating the pairwise links in
two subsets: those corresponding to indirect connections of the fully
conditioned Granger causality (which should thus be excluded) and links that
can be ascribed to redundancy effects and, together with the results from the
fully connected approach, provide a better description of the causality pattern
in presence of redundancy. We finally apply these methods to two different real
datasets. First, analyzing electrophysiological data from an epileptic brain,
we show that synergetic effects are dominant just before seizure occurrences.
Second, our analysis applied to gene expression time series from HeLa culture
shows that the underlying regulatory networks are characterized by both
redundancy and synergy.
| no_new_dataset | 0.944177 |
1403.7595 | Zi-Ke Zhang Dr. | Da-Cheng Nie, Zi-Ke Zhang, Jun-lin Zhou, Yan Fu, Kui Zhang | Information Filtering on Coupled Social Networks | null | null | 10.1371/journal.pone.0101675 | null | cs.SI physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, based on the coupled social networks (CSN), we propose a
hybrid algorithm to nonlinearly integrate both social and behavior information
of online users. Filtering algorithm based on the coupled social networks,
which considers the effects of both social influence and personalized
preference. Experimental results on two real datasets, \emph{Epinions} and
\emph{Friendfeed}, show that hybrid pattern can not only provide more accurate
recommendations, but also can enlarge the recommendation coverage while
adopting global metric. Further empirical analyses demonstrate that the mutual
reinforcement and rich-club phenomenon can also be found in coupled social
networks where the identical individuals occupy the core position of the online
system. This work may shed some light on the in-depth understanding structure
and function of coupled social networks.
| [
{
"version": "v1",
"created": "Sat, 29 Mar 2014 06:20:25 GMT"
}
] | 2015-06-19T00:00:00 | [
[
"Nie",
"Da-Cheng",
""
],
[
"Zhang",
"Zi-Ke",
""
],
[
"Zhou",
"Jun-lin",
""
],
[
"Fu",
"Yan",
""
],
[
"Zhang",
"Kui",
""
]
] | TITLE: Information Filtering on Coupled Social Networks
ABSTRACT: In this paper, based on the coupled social networks (CSN), we propose a
hybrid algorithm to nonlinearly integrate both social and behavior information
of online users. Filtering algorithm based on the coupled social networks,
which considers the effects of both social influence and personalized
preference. Experimental results on two real datasets, \emph{Epinions} and
\emph{Friendfeed}, show that hybrid pattern can not only provide more accurate
recommendations, but also can enlarge the recommendation coverage while
adopting global metric. Further empirical analyses demonstrate that the mutual
reinforcement and rich-club phenomenon can also be found in coupled social
networks where the identical individuals occupy the core position of the online
system. This work may shed some light on the in-depth understanding structure
and function of coupled social networks.
| no_new_dataset | 0.949669 |
1404.2342 | Ko-Jen Hsiao | Ko-Jen Hsiao, Alex Kulesza, Alfred Hero | Social Collaborative Retrieval | 10 pages | null | 10.1109/JSTSP.2014.2317286 | null | cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Socially-based recommendation systems have recently attracted significant
interest, and a number of studies have shown that social information can
dramatically improve a system's predictions of user interests. Meanwhile, there
are now many potential applications that involve aspects of both recommendation
and information retrieval, and the task of collaborative retrieval---a
combination of these two traditional problems---has recently been introduced.
Successful collaborative retrieval requires overcoming severe data sparsity,
making additional sources of information, such as social graphs, particularly
valuable. In this paper we propose a new model for collaborative retrieval, and
show that our algorithm outperforms current state-of-the-art approaches by
incorporating information from social networks. We also provide empirical
analyses of the ways in which cultural interests propagate along a social graph
using a real-world music dataset.
| [
{
"version": "v1",
"created": "Wed, 9 Apr 2014 01:18:05 GMT"
}
] | 2015-06-19T00:00:00 | [
[
"Hsiao",
"Ko-Jen",
""
],
[
"Kulesza",
"Alex",
""
],
[
"Hero",
"Alfred",
""
]
] | TITLE: Social Collaborative Retrieval
ABSTRACT: Socially-based recommendation systems have recently attracted significant
interest, and a number of studies have shown that social information can
dramatically improve a system's predictions of user interests. Meanwhile, there
are now many potential applications that involve aspects of both recommendation
and information retrieval, and the task of collaborative retrieval---a
combination of these two traditional problems---has recently been introduced.
Successful collaborative retrieval requires overcoming severe data sparsity,
making additional sources of information, such as social graphs, particularly
valuable. In this paper we propose a new model for collaborative retrieval, and
show that our algorithm outperforms current state-of-the-art approaches by
incorporating information from social networks. We also provide empirical
analyses of the ways in which cultural interests propagate along a social graph
using a real-world music dataset.
| no_new_dataset | 0.941601 |
1404.4667 | Morteza Mardani | Morteza Mardani, Gonzalo Mateos, and Georgios B. Giannakis | Subspace Learning and Imputation for Streaming Big Data Matrices and
Tensors | null | null | 10.1109/TSP.2015.2417491 | null | stat.ML cs.IT cs.LG math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Extracting latent low-dimensional structure from high-dimensional data is of
paramount importance in timely inference tasks encountered with `Big Data'
analytics. However, increasingly noisy, heterogeneous, and incomplete datasets
as well as the need for {\em real-time} processing of streaming data pose major
challenges to this end. In this context, the present paper permeates benefits
from rank minimization to scalable imputation of missing data, via tracking
low-dimensional subspaces and unraveling latent (possibly multi-way) structure
from \emph{incomplete streaming} data. For low-rank matrix data, a subspace
estimator is proposed based on an exponentially-weighted least-squares
criterion regularized with the nuclear norm. After recasting the non-separable
nuclear norm into a form amenable to online optimization, real-time algorithms
with complementary strengths are developed and their convergence is established
under simplifying technical assumptions. In a stationary setting, the
asymptotic estimates obtained offer the well-documented performance guarantees
of the {\em batch} nuclear-norm regularized estimator. Under the same unifying
framework, a novel online (adaptive) algorithm is developed to obtain multi-way
decompositions of \emph{low-rank tensors} with missing entries, and perform
imputation as a byproduct. Simulated tests with both synthetic as well as real
Internet and cardiac magnetic resonance imagery (MRI) data confirm the efficacy
of the proposed algorithms, and their superior performance relative to
state-of-the-art alternatives.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2014 22:55:08 GMT"
}
] | 2015-06-19T00:00:00 | [
[
"Mardani",
"Morteza",
""
],
[
"Mateos",
"Gonzalo",
""
],
[
"Giannakis",
"Georgios B.",
""
]
] | TITLE: Subspace Learning and Imputation for Streaming Big Data Matrices and
Tensors
ABSTRACT: Extracting latent low-dimensional structure from high-dimensional data is of
paramount importance in timely inference tasks encountered with `Big Data'
analytics. However, increasingly noisy, heterogeneous, and incomplete datasets
as well as the need for {\em real-time} processing of streaming data pose major
challenges to this end. In this context, the present paper permeates benefits
from rank minimization to scalable imputation of missing data, via tracking
low-dimensional subspaces and unraveling latent (possibly multi-way) structure
from \emph{incomplete streaming} data. For low-rank matrix data, a subspace
estimator is proposed based on an exponentially-weighted least-squares
criterion regularized with the nuclear norm. After recasting the non-separable
nuclear norm into a form amenable to online optimization, real-time algorithms
with complementary strengths are developed and their convergence is established
under simplifying technical assumptions. In a stationary setting, the
asymptotic estimates obtained offer the well-documented performance guarantees
of the {\em batch} nuclear-norm regularized estimator. Under the same unifying
framework, a novel online (adaptive) algorithm is developed to obtain multi-way
decompositions of \emph{low-rank tensors} with missing entries, and perform
imputation as a byproduct. Simulated tests with both synthetic as well as real
Internet and cardiac magnetic resonance imagery (MRI) data confirm the efficacy
of the proposed algorithms, and their superior performance relative to
state-of-the-art alternatives.
| no_new_dataset | 0.945601 |
1404.4923 | Jie Shen | Jie Shen, Guangcan Liu, Jia Chen, Yuqiang Fang, Jianbin Xie, Yong Yu,
Shuicheng Yan | Unified Structured Learning for Simultaneous Human Pose Estimation and
Garment Attribute Classification | Accepted to IEEE Trans. on Image Processing | null | 10.1109/TIP.2014.2358082 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we utilize structured learning to simultaneously address two
intertwined problems: human pose estimation (HPE) and garment attribute
classification (GAC), which are valuable for a variety of computer vision and
multimedia applications. Unlike previous works that usually handle the two
problems separately, our approach aims to produce a jointly optimal estimation
for both HPE and GAC via a unified inference procedure. To this end, we adopt a
preprocessing step to detect potential human parts from each image (i.e., a set
of "candidates") that allows us to have a manageable input space. In this way,
the simultaneous inference of HPE and GAC is converted to a structured learning
problem, where the inputs are the collections of candidate ensembles, the
outputs are the joint labels of human parts and garment attributes, and the
joint feature representation involves various cues such as pose-specific
features, garment-specific features, and cross-task features that encode
correlations between human parts and garment attributes. Furthermore, we
explore the "strong edge" evidence around the potential human parts so as to
derive more powerful representations for oriented human parts. Such evidences
can be seamlessly integrated into our structured learning model as a kind of
energy function, and the learning process could be performed by standard
structured Support Vector Machines (SVM) algorithm. However, the joint
structure of the two problems is a cyclic graph, which hinders efficient
inference. To resolve this issue, we compute instead approximate optima by
using an iterative procedure, where in each iteration the variables of one
problem are fixed. In this way, satisfactory solutions can be efficiently
computed by dynamic programming. Experimental results on two benchmark datasets
show the state-of-the-art performance of our approach.
| [
{
"version": "v1",
"created": "Sat, 19 Apr 2014 04:51:06 GMT"
},
{
"version": "v2",
"created": "Tue, 16 Sep 2014 19:50:41 GMT"
},
{
"version": "v3",
"created": "Mon, 22 Sep 2014 19:09:38 GMT"
}
] | 2015-06-19T00:00:00 | [
[
"Shen",
"Jie",
""
],
[
"Liu",
"Guangcan",
""
],
[
"Chen",
"Jia",
""
],
[
"Fang",
"Yuqiang",
""
],
[
"Xie",
"Jianbin",
""
],
[
"Yu",
"Yong",
""
],
[
"Yan",
"Shuicheng",
""
]
] | TITLE: Unified Structured Learning for Simultaneous Human Pose Estimation and
Garment Attribute Classification
ABSTRACT: In this paper, we utilize structured learning to simultaneously address two
intertwined problems: human pose estimation (HPE) and garment attribute
classification (GAC), which are valuable for a variety of computer vision and
multimedia applications. Unlike previous works that usually handle the two
problems separately, our approach aims to produce a jointly optimal estimation
for both HPE and GAC via a unified inference procedure. To this end, we adopt a
preprocessing step to detect potential human parts from each image (i.e., a set
of "candidates") that allows us to have a manageable input space. In this way,
the simultaneous inference of HPE and GAC is converted to a structured learning
problem, where the inputs are the collections of candidate ensembles, the
outputs are the joint labels of human parts and garment attributes, and the
joint feature representation involves various cues such as pose-specific
features, garment-specific features, and cross-task features that encode
correlations between human parts and garment attributes. Furthermore, we
explore the "strong edge" evidence around the potential human parts so as to
derive more powerful representations for oriented human parts. Such evidences
can be seamlessly integrated into our structured learning model as a kind of
energy function, and the learning process could be performed by standard
structured Support Vector Machines (SVM) algorithm. However, the joint
structure of the two problems is a cyclic graph, which hinders efficient
inference. To resolve this issue, we compute instead approximate optima by
using an iterative procedure, where in each iteration the variables of one
problem are fixed. In this way, satisfactory solutions can be efficiently
computed by dynamic programming. Experimental results on two benchmark datasets
show the state-of-the-art performance of our approach.
| no_new_dataset | 0.942718 |
1404.7170 | Paul Expert | Giovanni Petri, Paul Expert | Temporal stability of network partitions | 15 pages, 12 figures | Phys. Rev. E 90, 022813, 2014 | 10.1103/PhysRevE.90.022813 | null | physics.soc-ph cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a method to find the best temporal partition at any time-scale and
rank the relevance of partitions found at different time-scales. This method is
based on random walkers coevolving with the network and as such constitutes a
generalization of partition stability to the case of temporal networks. We show
that, when applied to a toy model and real datasets, temporal stability
uncovers structures that are persistent over meaningful time-scales as well as
important isolated events, making it an effective tool to study both abrupt
changes and gradual evolution of a network mesoscopic structures.
| [
{
"version": "v1",
"created": "Mon, 28 Apr 2014 21:18:49 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Aug 2014 09:25:50 GMT"
}
] | 2015-06-19T00:00:00 | [
[
"Petri",
"Giovanni",
""
],
[
"Expert",
"Paul",
""
]
] | TITLE: Temporal stability of network partitions
ABSTRACT: We present a method to find the best temporal partition at any time-scale and
rank the relevance of partitions found at different time-scales. This method is
based on random walkers coevolving with the network and as such constitutes a
generalization of partition stability to the case of temporal networks. We show
that, when applied to a toy model and real datasets, temporal stability
uncovers structures that are persistent over meaningful time-scales as well as
important isolated events, making it an effective tool to study both abrupt
changes and gradual evolution of a network mesoscopic structures.
| no_new_dataset | 0.943608 |
1405.4574 | Kristjan Greenewald | Kristjan H. Greenewald and Alfred O. Hero III | Kronecker PCA Based Spatio-Temporal Modeling of Video for Dismount
Classification | 8 pages. To appear in Proceeding of SPIE DSS. arXiv admin note: text
overlap with arXiv:1402.5568 | null | 10.1117/12.2050184 | null | cs.CV stat.ME | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider the application of KronPCA spatio-temporal modeling techniques
[Greenewald et al 2013, Tsiligkaridis et al 2013] to the extraction of
spatiotemporal features for video dismount classification. KronPCA performs a
low-rank type of dimensionality reduction that is adapted to spatio-temporal
data and is characterized by the T frame multiframe mean and covariance of p
spatial features. For further regularization and improved inverse estimation,
we also use the diagonally corrected KronPCA shrinkage methods we presented in
[Greenewald et al 2013]. We apply this very general method to the modeling of
the multivariate temporal behavior of HOG features extracted from pedestrian
bounding boxes in video, with gender classification in a challenging dataset
chosen as a specific application. The learned covariances for each class are
used to extract spatiotemporal features which are then classified, achieving
competitive classification performance.
| [
{
"version": "v1",
"created": "Mon, 19 May 2014 01:22:34 GMT"
}
] | 2015-06-19T00:00:00 | [
[
"Greenewald",
"Kristjan H.",
""
],
[
"Hero",
"Alfred O.",
"III"
]
] | TITLE: Kronecker PCA Based Spatio-Temporal Modeling of Video for Dismount
Classification
ABSTRACT: We consider the application of KronPCA spatio-temporal modeling techniques
[Greenewald et al 2013, Tsiligkaridis et al 2013] to the extraction of
spatiotemporal features for video dismount classification. KronPCA performs a
low-rank type of dimensionality reduction that is adapted to spatio-temporal
data and is characterized by the T frame multiframe mean and covariance of p
spatial features. For further regularization and improved inverse estimation,
we also use the diagonally corrected KronPCA shrinkage methods we presented in
[Greenewald et al 2013]. We apply this very general method to the modeling of
the multivariate temporal behavior of HOG features extracted from pedestrian
bounding boxes in video, with gender classification in a challenging dataset
chosen as a specific application. The learned covariances for each class are
used to extract spatiotemporal features which are then classified, achieving
competitive classification performance.
| no_new_dataset | 0.947575 |
1406.3295 | Cesar Caiafa | Cesar F. Caiafa and Andrzej Cichocki | Stable, Robust and Super Fast Reconstruction of Tensors Using Multi-Way
Projections | Submitted to IEEE Transactions on Signal Processing | null | 10.1109/TSP.2014.2385040 | null | cs.IT cs.DS math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In the framework of multidimensional Compressed Sensing (CS), we introduce an
analytical reconstruction formula that allows one to recover an $N$th-order
$(I_1\times I_2\times \cdots \times I_N)$ data tensor $\underline{\mathbf{X}}$
from a reduced set of multi-way compressive measurements by exploiting its low
multilinear-rank structure. Moreover, we show that, an interesting property of
multi-way measurements allows us to build the reconstruction based on
compressive linear measurements taken only in two selected modes, independently
of the tensor order $N$. In addition, it is proved that, in the matrix case and
in a particular case with $3$rd-order tensors where the same 2D sensor operator
is applied to all mode-3 slices, the proposed reconstruction
$\underline{\mathbf{X}}_\tau$ is stable in the sense that the approximation
error is comparable to the one provided by the best low-multilinear-rank
approximation, where $\tau$ is a threshold parameter that controls the
approximation error. Through the analysis of the upper bound of the
approximation error we show that, in the 2D case, an optimal value for the
threshold parameter $\tau=\tau_0 > 0$ exists, which is confirmed by our
simulation results. On the other hand, our experiments on 3D datasets show that
very good reconstructions are obtained using $\tau=0$, which means that this
parameter does not need to be tuned. Our extensive simulation results
demonstrate the stability and robustness of the method when it is applied to
real-world 2D and 3D signals. A comparison with state-of-the-arts sparsity
based CS methods specialized for multidimensional signals is also included. A
very attractive characteristic of the proposed method is that it provides a
direct computation, i.e. it is non-iterative in contrast to all existing
sparsity based CS algorithms, thus providing super fast computations, even for
large datasets.
| [
{
"version": "v1",
"created": "Thu, 22 May 2014 18:35:07 GMT"
},
{
"version": "v2",
"created": "Mon, 30 Jun 2014 17:05:36 GMT"
}
] | 2015-06-19T00:00:00 | [
[
"Caiafa",
"Cesar F.",
""
],
[
"Cichocki",
"Andrzej",
""
]
] | TITLE: Stable, Robust and Super Fast Reconstruction of Tensors Using Multi-Way
Projections
ABSTRACT: In the framework of multidimensional Compressed Sensing (CS), we introduce an
analytical reconstruction formula that allows one to recover an $N$th-order
$(I_1\times I_2\times \cdots \times I_N)$ data tensor $\underline{\mathbf{X}}$
from a reduced set of multi-way compressive measurements by exploiting its low
multilinear-rank structure. Moreover, we show that, an interesting property of
multi-way measurements allows us to build the reconstruction based on
compressive linear measurements taken only in two selected modes, independently
of the tensor order $N$. In addition, it is proved that, in the matrix case and
in a particular case with $3$rd-order tensors where the same 2D sensor operator
is applied to all mode-3 slices, the proposed reconstruction
$\underline{\mathbf{X}}_\tau$ is stable in the sense that the approximation
error is comparable to the one provided by the best low-multilinear-rank
approximation, where $\tau$ is a threshold parameter that controls the
approximation error. Through the analysis of the upper bound of the
approximation error we show that, in the 2D case, an optimal value for the
threshold parameter $\tau=\tau_0 > 0$ exists, which is confirmed by our
simulation results. On the other hand, our experiments on 3D datasets show that
very good reconstructions are obtained using $\tau=0$, which means that this
parameter does not need to be tuned. Our extensive simulation results
demonstrate the stability and robustness of the method when it is applied to
real-world 2D and 3D signals. A comparison with state-of-the-arts sparsity
based CS methods specialized for multidimensional signals is also included. A
very attractive characteristic of the proposed method is that it provides a
direct computation, i.e. it is non-iterative in contrast to all existing
sparsity based CS algorithms, thus providing super fast computations, even for
large datasets.
| no_new_dataset | 0.943191 |
1503.02128 | Qingming Tang | Qingming Tang, Chao Yang, Jian Peng and Jinbo Xu | Exact Hybrid Covariance Thresholding for Joint Graphical Lasso | null | null | null | null | cs.LG cs.AI stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper considers the problem of estimating multiple related Gaussian
graphical models from a $p$-dimensional dataset consisting of different
classes. Our work is based upon the formulation of this problem as group
graphical lasso. This paper proposes a novel hybrid covariance thresholding
algorithm that can effectively identify zero entries in the precision matrices
and split a large joint graphical lasso problem into small subproblems. Our
hybrid covariance thresholding method is superior to existing uniform
thresholding methods in that our method can split the precision matrix of each
individual class using different partition schemes and thus split group
graphical lasso into much smaller subproblems, each of which can be solved very
fast. In addition, this paper establishes necessary and sufficient conditions
for our hybrid covariance thresholding algorithm. The superior performance of
our thresholding method is thoroughly analyzed and illustrated by a few
experiments on simulated data and real gene expression data.
| [
{
"version": "v1",
"created": "Sat, 7 Mar 2015 03:34:48 GMT"
},
{
"version": "v2",
"created": "Thu, 18 Jun 2015 02:52:51 GMT"
}
] | 2015-06-19T00:00:00 | [
[
"Tang",
"Qingming",
""
],
[
"Yang",
"Chao",
""
],
[
"Peng",
"Jian",
""
],
[
"Xu",
"Jinbo",
""
]
] | TITLE: Exact Hybrid Covariance Thresholding for Joint Graphical Lasso
ABSTRACT: This paper considers the problem of estimating multiple related Gaussian
graphical models from a $p$-dimensional dataset consisting of different
classes. Our work is based upon the formulation of this problem as group
graphical lasso. This paper proposes a novel hybrid covariance thresholding
algorithm that can effectively identify zero entries in the precision matrices
and split a large joint graphical lasso problem into small subproblems. Our
hybrid covariance thresholding method is superior to existing uniform
thresholding methods in that our method can split the precision matrix of each
individual class using different partition schemes and thus split group
graphical lasso into much smaller subproblems, each of which can be solved very
fast. In addition, this paper establishes necessary and sufficient conditions
for our hybrid covariance thresholding algorithm. The superior performance of
our thresholding method is thoroughly analyzed and illustrated by a few
experiments on simulated data and real gene expression data.
| no_new_dataset | 0.949248 |
1506.04729 | Shohreh Shaghaghian Ms | Shohreh Shaghaghian, Mark Coates | Optimal Forwarding in Opportunistic Delay Tolerant Networks with Meeting
Rate Estimations | null | null | null | null | cs.NI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Data transfer in opportunistic Delay Tolerant Networks (DTNs) must rely on
unscheduled sporadic meetings between nodes. The main challenge in these
networks is to develop a mechanism based on which nodes can learn to make
nearly optimal forwarding decision rules despite having no a-priori knowledge
of the network topology. The forwarding mechanism should ideally result in a
high delivery probability, low average latency and efficient usage of the
network resources. In this paper, we propose both centralized and decentralized
single-copy message forwarding algorithms that, under relatively strong
assumptions about the networks behaviour, minimize the expected latencies from
any node in the network to a particular destination. After proving the
optimality of our proposed algorithms, we develop a decentralized algorithm
that involves a recursive maximum likelihood procedure to estimate the meeting
rates. We confirm the improvement that our proposed algorithms make in the
system performance through numerical simulations on datasets from synthetic and
real-world opportunistic networks.
| [
{
"version": "v1",
"created": "Mon, 15 Jun 2015 19:49:48 GMT"
},
{
"version": "v2",
"created": "Tue, 16 Jun 2015 18:20:30 GMT"
},
{
"version": "v3",
"created": "Wed, 17 Jun 2015 14:38:37 GMT"
},
{
"version": "v4",
"created": "Thu, 18 Jun 2015 14:12:07 GMT"
}
] | 2015-06-19T00:00:00 | [
[
"Shaghaghian",
"Shohreh",
""
],
[
"Coates",
"Mark",
""
]
] | TITLE: Optimal Forwarding in Opportunistic Delay Tolerant Networks with Meeting
Rate Estimations
ABSTRACT: Data transfer in opportunistic Delay Tolerant Networks (DTNs) must rely on
unscheduled sporadic meetings between nodes. The main challenge in these
networks is to develop a mechanism based on which nodes can learn to make
nearly optimal forwarding decision rules despite having no a-priori knowledge
of the network topology. The forwarding mechanism should ideally result in a
high delivery probability, low average latency and efficient usage of the
network resources. In this paper, we propose both centralized and decentralized
single-copy message forwarding algorithms that, under relatively strong
assumptions about the networks behaviour, minimize the expected latencies from
any node in the network to a particular destination. After proving the
optimality of our proposed algorithms, we develop a decentralized algorithm
that involves a recursive maximum likelihood procedure to estimate the meeting
rates. We confirm the improvement that our proposed algorithms make in the
system performance through numerical simulations on datasets from synthetic and
real-world opportunistic networks.
| no_new_dataset | 0.950319 |
1506.05514 | Ubai Sandouk | Ubai Sandouk, Ke Chen | Learning Contextualized Semantics from Co-occurring Terms via a Siamese
Architecture | null | null | null | 2015-06-18 | cs.IR cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | One of the biggest challenges in Multimedia information retrieval and
understanding is to bridge the semantic gap by properly modeling concept
semantics in context. The presence of out of vocabulary (OOV) concepts
exacerbates this difficulty. To address the semantic gap issues, we formulate a
problem on learning contextualized semantics from descriptive terms and propose
a novel Siamese architecture to model the contextualized semantics from
descriptive terms. By means of pattern aggregation and probabilistic topic
models, our Siamese architecture captures contextualized semantics from the
co-occurring descriptive terms via unsupervised learning, which leads to a
concept embedding space of the terms in context. Furthermore, the co-occurring
OOV concepts can be easily represented in the learnt concept embedding space.
The main properties of the concept embedding space are demonstrated via
visualization. Using various settings in semantic priming, we have carried out
a thorough evaluation by comparing our approach to a number of state-of-the-art
methods on six annotation corpora in different domains, i.e., MagTag5K, CAL500
and Million Song Dataset in the music domain as well as Corel5K, LabelMe and
SUNDatabase in the image domain. Experimental results on semantic priming
suggest that our approach outperforms those state-of-the-art methods
considerably in various aspects.
| [
{
"version": "v1",
"created": "Wed, 17 Jun 2015 23:03:43 GMT"
}
] | 2015-06-19T00:00:00 | [
[
"Sandouk",
"Ubai",
""
],
[
"Chen",
"Ke",
""
]
] | TITLE: Learning Contextualized Semantics from Co-occurring Terms via a Siamese
Architecture
ABSTRACT: One of the biggest challenges in Multimedia information retrieval and
understanding is to bridge the semantic gap by properly modeling concept
semantics in context. The presence of out of vocabulary (OOV) concepts
exacerbates this difficulty. To address the semantic gap issues, we formulate a
problem on learning contextualized semantics from descriptive terms and propose
a novel Siamese architecture to model the contextualized semantics from
descriptive terms. By means of pattern aggregation and probabilistic topic
models, our Siamese architecture captures contextualized semantics from the
co-occurring descriptive terms via unsupervised learning, which leads to a
concept embedding space of the terms in context. Furthermore, the co-occurring
OOV concepts can be easily represented in the learnt concept embedding space.
The main properties of the concept embedding space are demonstrated via
visualization. Using various settings in semantic priming, we have carried out
a thorough evaluation by comparing our approach to a number of state-of-the-art
methods on six annotation corpora in different domains, i.e., MagTag5K, CAL500
and Million Song Dataset in the music domain as well as Corel5K, LabelMe and
SUNDatabase in the image domain. Experimental results on semantic priming
suggest that our approach outperforms those state-of-the-art methods
considerably in various aspects.
| no_new_dataset | 0.946399 |
1506.05541 | Yi Sun | Yi Sun, Xiaoqi Yin, Nanshu Wang, Junchen Jiang, Vyas Sekar, Yun Jin,
Bruno Sinopoli | Analyzing TCP Throughput Stability and Predictability with Implications
for Adaptive Video Streaming | null | null | null | null | cs.NI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent work suggests that TCP throughput stability and predictability within
a video viewing session can inform the design of better video bitrate
adaptation algorithms. Despite a rich tradition of Internet measurement,
however, our understanding of throughput stability and predictability is quite
limited. To bridge this gap, we present a measurement study of throughput
stability using a large-scale dataset from a video service provider. Drawing on
this analysis, we propose a simple-but-effective prediction mechanism based on
a hidden Markov model and demonstrate that it outperforms other approaches. We
also show the practical implications in improving the user experience of
adaptive video streaming.
| [
{
"version": "v1",
"created": "Thu, 18 Jun 2015 03:36:24 GMT"
}
] | 2015-06-19T00:00:00 | [
[
"Sun",
"Yi",
""
],
[
"Yin",
"Xiaoqi",
""
],
[
"Wang",
"Nanshu",
""
],
[
"Jiang",
"Junchen",
""
],
[
"Sekar",
"Vyas",
""
],
[
"Jin",
"Yun",
""
],
[
"Sinopoli",
"Bruno",
""
]
] | TITLE: Analyzing TCP Throughput Stability and Predictability with Implications
for Adaptive Video Streaming
ABSTRACT: Recent work suggests that TCP throughput stability and predictability within
a video viewing session can inform the design of better video bitrate
adaptation algorithms. Despite a rich tradition of Internet measurement,
however, our understanding of throughput stability and predictability is quite
limited. To bridge this gap, we present a measurement study of throughput
stability using a large-scale dataset from a video service provider. Drawing on
this analysis, we propose a simple-but-effective prediction mechanism based on
a hidden Markov model and demonstrate that it outperforms other approaches. We
also show the practical implications in improving the user experience of
adaptive video streaming.
| no_new_dataset | 0.933309 |
1506.05751 | Rob Fergus | Emily Denton, Soumith Chintala, Arthur Szlam, Rob Fergus | Deep Generative Image Models using a Laplacian Pyramid of Adversarial
Networks | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper we introduce a generative parametric model capable of producing
high quality samples of natural images. Our approach uses a cascade of
convolutional networks within a Laplacian pyramid framework to generate images
in a coarse-to-fine fashion. At each level of the pyramid, a separate
generative convnet model is trained using the Generative Adversarial Nets (GAN)
approach (Goodfellow et al.). Samples drawn from our model are of significantly
higher quality than alternate approaches. In a quantitative assessment by human
evaluators, our CIFAR10 samples were mistaken for real images around 40% of the
time, compared to 10% for samples drawn from a GAN baseline model. We also show
samples from models trained on the higher resolution images of the LSUN scene
dataset.
| [
{
"version": "v1",
"created": "Thu, 18 Jun 2015 17:03:54 GMT"
}
] | 2015-06-19T00:00:00 | [
[
"Denton",
"Emily",
""
],
[
"Chintala",
"Soumith",
""
],
[
"Szlam",
"Arthur",
""
],
[
"Fergus",
"Rob",
""
]
] | TITLE: Deep Generative Image Models using a Laplacian Pyramid of Adversarial
Networks
ABSTRACT: In this paper we introduce a generative parametric model capable of producing
high quality samples of natural images. Our approach uses a cascade of
convolutional networks within a Laplacian pyramid framework to generate images
in a coarse-to-fine fashion. At each level of the pyramid, a separate
generative convnet model is trained using the Generative Adversarial Nets (GAN)
approach (Goodfellow et al.). Samples drawn from our model are of significantly
higher quality than alternate approaches. In a quantitative assessment by human
evaluators, our CIFAR10 samples were mistaken for real images around 40% of the
time, compared to 10% for samples drawn from a GAN baseline model. We also show
samples from models trained on the higher resolution images of the LSUN scene
dataset.
| no_new_dataset | 0.953492 |
1312.0317 | Chunxiao Jiang | Chunxiao Jiang and Yan Chen and K. J. Ray Liu | Evolutionary Dynamics of Information Diffusion over Social Networks | arXiv admin note: substantial text overlap with arXiv:1309.2920 | null | 10.1109/JSTSP.2014.2313024 | null | cs.SI physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Current social networks are of extremely large-scale generating tremendous
information flows at every moment. How information diffuse over social networks
has attracted much attention from both industry and academics. Most of the
existing works on information diffusion analysis are based on machine learning
methods focusing on social network structure analysis and empirical data
mining. However, the dynamics of information diffusion, which are heavily
influenced by network users' decisions, actions and their socio-economic
interactions, is generally ignored by most of existing works. In this paper, we
propose an evolutionary game theoretic framework to model the dynamic
information diffusion process in social networks. Specifically, we derive the
information diffusion dynamics in complete networks, uniform degree and
non-uniform degree networks, with the highlight of two special networks,
Erd\H{o}s-R\'enyi random network and the Barab\'asi-Albert scale-free network.
We find that the dynamics of information diffusion over these three kinds of
networks are scale-free and the same with each other when the network scale is
sufficiently large. To verify our theoretical analysis, we perform simulations
for the information diffusion over synthetic networks and real-world Facebook
networks. Moreover, we also conduct experiment on Twitter hashtags dataset,
which shows that the proposed game theoretic model can well fit and predict the
information diffusion over real social networks.
| [
{
"version": "v1",
"created": "Mon, 2 Dec 2013 03:21:28 GMT"
}
] | 2015-06-18T00:00:00 | [
[
"Jiang",
"Chunxiao",
""
],
[
"Chen",
"Yan",
""
],
[
"Liu",
"K. J. Ray",
""
]
] | TITLE: Evolutionary Dynamics of Information Diffusion over Social Networks
ABSTRACT: Current social networks are of extremely large-scale generating tremendous
information flows at every moment. How information diffuse over social networks
has attracted much attention from both industry and academics. Most of the
existing works on information diffusion analysis are based on machine learning
methods focusing on social network structure analysis and empirical data
mining. However, the dynamics of information diffusion, which are heavily
influenced by network users' decisions, actions and their socio-economic
interactions, is generally ignored by most of existing works. In this paper, we
propose an evolutionary game theoretic framework to model the dynamic
information diffusion process in social networks. Specifically, we derive the
information diffusion dynamics in complete networks, uniform degree and
non-uniform degree networks, with the highlight of two special networks,
Erd\H{o}s-R\'enyi random network and the Barab\'asi-Albert scale-free network.
We find that the dynamics of information diffusion over these three kinds of
networks are scale-free and the same with each other when the network scale is
sufficiently large. To verify our theoretical analysis, we perform simulations
for the information diffusion over synthetic networks and real-world Facebook
networks. Moreover, we also conduct experiment on Twitter hashtags dataset,
which shows that the proposed game theoretic model can well fit and predict the
information diffusion over real social networks.
| no_new_dataset | 0.949012 |
1401.0887 | Dorina Thanou | Dorina Thanou, David I Shuman, Pascal Frossard | Learning parametric dictionaries for graph signals | null | null | 10.1109/TSP.2014.2332441 | null | cs.LG cs.SI stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In sparse signal representation, the choice of a dictionary often involves a
tradeoff between two desirable properties -- the ability to adapt to specific
signal data and a fast implementation of the dictionary. To sparsely represent
signals residing on weighted graphs, an additional design challenge is to
incorporate the intrinsic geometric structure of the irregular data domain into
the atoms of the dictionary. In this work, we propose a parametric dictionary
learning algorithm to design data-adapted, structured dictionaries that
sparsely represent graph signals. In particular, we model graph signals as
combinations of overlapping local patterns. We impose the constraint that each
dictionary is a concatenation of subdictionaries, with each subdictionary being
a polynomial of the graph Laplacian matrix, representing a single pattern
translated to different areas of the graph. The learning algorithm adapts the
patterns to a training set of graph signals. Experimental results on both
synthetic and real datasets demonstrate that the dictionaries learned by the
proposed algorithm are competitive with and often better than unstructured
dictionaries learned by state-of-the-art numerical learning algorithms in terms
of sparse approximation of graph signals. In contrast to the unstructured
dictionaries, however, the dictionaries learned by the proposed algorithm
feature localized atoms and can be implemented in a computationally efficient
manner in signal processing tasks such as compression, denoising, and
classification.
| [
{
"version": "v1",
"created": "Sun, 5 Jan 2014 12:17:51 GMT"
}
] | 2015-06-18T00:00:00 | [
[
"Thanou",
"Dorina",
""
],
[
"Shuman",
"David I",
""
],
[
"Frossard",
"Pascal",
""
]
] | TITLE: Learning parametric dictionaries for graph signals
ABSTRACT: In sparse signal representation, the choice of a dictionary often involves a
tradeoff between two desirable properties -- the ability to adapt to specific
signal data and a fast implementation of the dictionary. To sparsely represent
signals residing on weighted graphs, an additional design challenge is to
incorporate the intrinsic geometric structure of the irregular data domain into
the atoms of the dictionary. In this work, we propose a parametric dictionary
learning algorithm to design data-adapted, structured dictionaries that
sparsely represent graph signals. In particular, we model graph signals as
combinations of overlapping local patterns. We impose the constraint that each
dictionary is a concatenation of subdictionaries, with each subdictionary being
a polynomial of the graph Laplacian matrix, representing a single pattern
translated to different areas of the graph. The learning algorithm adapts the
patterns to a training set of graph signals. Experimental results on both
synthetic and real datasets demonstrate that the dictionaries learned by the
proposed algorithm are competitive with and often better than unstructured
dictionaries learned by state-of-the-art numerical learning algorithms in terms
of sparse approximation of graph signals. In contrast to the unstructured
dictionaries, however, the dictionaries learned by the proposed algorithm
feature localized atoms and can be implemented in a computationally efficient
manner in signal processing tasks such as compression, denoising, and
classification.
| no_new_dataset | 0.947332 |
1412.6583 | Brian Cheung | Brian Cheung, Jesse A. Livezey, Arjun K. Bansal, Bruno A. Olshausen | Discovering Hidden Factors of Variation in Deep Networks | Presented at International Conference on Learning Representations
2015 Workshop | null | null | null | cs.LG cs.CV cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep learning has enjoyed a great deal of success because of its ability to
learn useful features for tasks such as classification. But there has been less
exploration in learning the factors of variation apart from the classification
signal. By augmenting autoencoders with simple regularization terms during
training, we demonstrate that standard deep architectures can discover and
explicitly represent factors of variation beyond those relevant for
categorization. We introduce a cross-covariance penalty (XCov) as a method to
disentangle factors like handwriting style for digits and subject identity in
faces. We demonstrate this on the MNIST handwritten digit database, the Toronto
Faces Database (TFD) and the Multi-PIE dataset by generating manipulated
instances of the data. Furthermore, we demonstrate these deep networks can
extrapolate `hidden' variation in the supervised signal.
| [
{
"version": "v1",
"created": "Sat, 20 Dec 2014 02:52:03 GMT"
},
{
"version": "v2",
"created": "Fri, 27 Feb 2015 20:41:40 GMT"
},
{
"version": "v3",
"created": "Fri, 17 Apr 2015 17:15:02 GMT"
},
{
"version": "v4",
"created": "Wed, 17 Jun 2015 06:47:48 GMT"
}
] | 2015-06-18T00:00:00 | [
[
"Cheung",
"Brian",
""
],
[
"Livezey",
"Jesse A.",
""
],
[
"Bansal",
"Arjun K.",
""
],
[
"Olshausen",
"Bruno A.",
""
]
] | TITLE: Discovering Hidden Factors of Variation in Deep Networks
ABSTRACT: Deep learning has enjoyed a great deal of success because of its ability to
learn useful features for tasks such as classification. But there has been less
exploration in learning the factors of variation apart from the classification
signal. By augmenting autoencoders with simple regularization terms during
training, we demonstrate that standard deep architectures can discover and
explicitly represent factors of variation beyond those relevant for
categorization. We introduce a cross-covariance penalty (XCov) as a method to
disentangle factors like handwriting style for digits and subject identity in
faces. We demonstrate this on the MNIST handwritten digit database, the Toronto
Faces Database (TFD) and the Multi-PIE dataset by generating manipulated
instances of the data. Furthermore, we demonstrate these deep networks can
extrapolate `hidden' variation in the supervised signal.
| no_new_dataset | 0.944944 |
1504.06852 | Philipp Fischer | Philipp Fischer, Alexey Dosovitskiy, Eddy Ilg, Philip H\"ausser, Caner
Haz{\i}rba\c{s}, Vladimir Golkov, Patrick van der Smagt, Daniel Cremers,
Thomas Brox | FlowNet: Learning Optical Flow with Convolutional Networks | Added supplementary material | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Convolutional neural networks (CNNs) have recently been very successful in a
variety of computer vision tasks, especially on those linked to recognition.
Optical flow estimation has not been among the tasks where CNNs were
successful. In this paper we construct appropriate CNNs which are capable of
solving the optical flow estimation problem as a supervised learning task. We
propose and compare two architectures: a generic architecture and another one
including a layer that correlates feature vectors at different image locations.
Since existing ground truth data sets are not sufficiently large to train a
CNN, we generate a synthetic Flying Chairs dataset. We show that networks
trained on this unrealistic data still generalize very well to existing
datasets such as Sintel and KITTI, achieving competitive accuracy at frame
rates of 5 to 10 fps.
| [
{
"version": "v1",
"created": "Sun, 26 Apr 2015 17:30:32 GMT"
},
{
"version": "v2",
"created": "Mon, 4 May 2015 08:50:57 GMT"
}
] | 2015-06-18T00:00:00 | [
[
"Fischer",
"Philipp",
""
],
[
"Dosovitskiy",
"Alexey",
""
],
[
"Ilg",
"Eddy",
""
],
[
"Häusser",
"Philip",
""
],
[
"Hazırbaş",
"Caner",
""
],
[
"Golkov",
"Vladimir",
""
],
[
"van der Smagt",
"Patrick",
""
],
[
"Cremers",
"Daniel",
""
],
[
"Brox",
"Thomas",
""
]
] | TITLE: FlowNet: Learning Optical Flow with Convolutional Networks
ABSTRACT: Convolutional neural networks (CNNs) have recently been very successful in a
variety of computer vision tasks, especially on those linked to recognition.
Optical flow estimation has not been among the tasks where CNNs were
successful. In this paper we construct appropriate CNNs which are capable of
solving the optical flow estimation problem as a supervised learning task. We
propose and compare two architectures: a generic architecture and another one
including a layer that correlates feature vectors at different image locations.
Since existing ground truth data sets are not sufficiently large to train a
CNN, we generate a synthetic Flying Chairs dataset. We show that networks
trained on this unrealistic data still generalize very well to existing
datasets such as Sintel and KITTI, achieving competitive accuracy at frame
rates of 5 to 10 fps.
| new_dataset | 0.960025 |
1505.06454 | Qing Ke | Qing Ke, Emilio Ferrara, Filippo Radicchi, Alessandro Flammini | Defining and identifying Sleeping Beauties in science | 40 pages, Supporting Information included, top examples listed at
http://qke.github.io/projects/beauty/beauty.html | Proc. Natl. Acad. Sci. USA 112, 7426-7431 (2015) | 10.1073/pnas.1424329112 | null | physics.soc-ph cs.DL cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A Sleeping Beauty (SB) in science refers to a paper whose importance is not
recognized for several years after publication. Its citation history exhibits a
long hibernation period followed by a sudden spike of popularity. Previous
studies suggest a relative scarcity of SBs. The reliability of this conclusion
is, however, heavily dependent on identification methods based on arbitrary
threshold parameters for sleeping time and number of citations, applied to
small or monodisciplinary bibliographic datasets. Here we present a systematic,
large-scale, and multidisciplinary analysis of the SB phenomenon in science. We
introduce a parameter-free measure that quantifies the extent to which a
specific paper can be considered an SB. We apply our method to 22 million
scientific papers published in all disciplines of natural and social sciences
over a time span longer than a century. Our results reveal that the SB
phenomenon is not exceptional. There is a continuous spectrum of delayed
recognition where both the hibernation period and the awakening intensity are
taken into account. Although many cases of SBs can be identified by looking at
monodisciplinary bibliographic data, the SB phenomenon becomes much more
apparent with the analysis of multidisciplinary datasets, where we can observe
many examples of papers achieving delayed yet exceptional importance in
disciplines different from those where they were originally published. Our
analysis emphasizes a complex feature of citation dynamics that so far has
received little attention, and also provides empirical evidence against the use
of short-term citation metrics in the quantification of scientific impact.
| [
{
"version": "v1",
"created": "Sun, 24 May 2015 16:38:14 GMT"
}
] | 2015-06-18T00:00:00 | [
[
"Ke",
"Qing",
""
],
[
"Ferrara",
"Emilio",
""
],
[
"Radicchi",
"Filippo",
""
],
[
"Flammini",
"Alessandro",
""
]
] | TITLE: Defining and identifying Sleeping Beauties in science
ABSTRACT: A Sleeping Beauty (SB) in science refers to a paper whose importance is not
recognized for several years after publication. Its citation history exhibits a
long hibernation period followed by a sudden spike of popularity. Previous
studies suggest a relative scarcity of SBs. The reliability of this conclusion
is, however, heavily dependent on identification methods based on arbitrary
threshold parameters for sleeping time and number of citations, applied to
small or monodisciplinary bibliographic datasets. Here we present a systematic,
large-scale, and multidisciplinary analysis of the SB phenomenon in science. We
introduce a parameter-free measure that quantifies the extent to which a
specific paper can be considered an SB. We apply our method to 22 million
scientific papers published in all disciplines of natural and social sciences
over a time span longer than a century. Our results reveal that the SB
phenomenon is not exceptional. There is a continuous spectrum of delayed
recognition where both the hibernation period and the awakening intensity are
taken into account. Although many cases of SBs can be identified by looking at
monodisciplinary bibliographic data, the SB phenomenon becomes much more
apparent with the analysis of multidisciplinary datasets, where we can observe
many examples of papers achieving delayed yet exceptional importance in
disciplines different from those where they were originally published. Our
analysis emphasizes a complex feature of citation dynamics that so far has
received little attention, and also provides empirical evidence against the use
of short-term citation metrics in the quantification of scientific impact.
| no_new_dataset | 0.945349 |
1506.02085 | Min Xu | Min Xu, Rudy Setiono | Gene selection for cancer classification using a hybrid of univariate
and multivariate feature selection methods | null | Applied Genomics and Proteomics. 2003:2(2)79-91 | null | null | q-bio.QM cs.CE cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Various approaches to gene selection for cancer classification based on
microarray data can be found in the literature and they may be grouped into two
categories: univariate methods and multivariate methods. Univariate methods
look at each gene in the data in isolation from others. They measure the
contribution of a particular gene to the classification without considering the
presence of the other genes. In contrast, multivariate methods measure the
relative contribution of a gene to the classification by taking the other genes
in the data into consideration. Multivariate methods select fewer genes in
general. However, the selection process of multivariate methods may be
sensitive to the presence of irrelevant genes, noises in the expression and
outliers in the training data. At the same time, the computational cost of
multivariate methods is high. To overcome the disadvantages of the two types of
approaches, we propose a hybrid method to obtain gene sets that are small and
highly discriminative.
We devise our hybrid method from the univariate Maximum Likelihood method
(LIK) and the multivariate Recursive Feature Elimination method (RFE). We
analyze the properties of these methods and systematically test the
effectiveness of our proposed method on two cancer microarray datasets. Our
experiments on a leukemia dataset and a small, round blue cell tumors dataset
demonstrate the effectiveness of our hybrid method. It is able to discover sets
consisting of fewer genes than those reported in the literature and at the same
time achieve the same or better prediction accuracy.
| [
{
"version": "v1",
"created": "Fri, 5 Jun 2015 23:29:06 GMT"
}
] | 2015-06-18T00:00:00 | [
[
"Xu",
"Min",
""
],
[
"Setiono",
"Rudy",
""
]
] | TITLE: Gene selection for cancer classification using a hybrid of univariate
and multivariate feature selection methods
ABSTRACT: Various approaches to gene selection for cancer classification based on
microarray data can be found in the literature and they may be grouped into two
categories: univariate methods and multivariate methods. Univariate methods
look at each gene in the data in isolation from others. They measure the
contribution of a particular gene to the classification without considering the
presence of the other genes. In contrast, multivariate methods measure the
relative contribution of a gene to the classification by taking the other genes
in the data into consideration. Multivariate methods select fewer genes in
general. However, the selection process of multivariate methods may be
sensitive to the presence of irrelevant genes, noises in the expression and
outliers in the training data. At the same time, the computational cost of
multivariate methods is high. To overcome the disadvantages of the two types of
approaches, we propose a hybrid method to obtain gene sets that are small and
highly discriminative.
We devise our hybrid method from the univariate Maximum Likelihood method
(LIK) and the multivariate Recursive Feature Elimination method (RFE). We
analyze the properties of these methods and systematically test the
effectiveness of our proposed method on two cancer microarray datasets. Our
experiments on a leukemia dataset and a small, round blue cell tumors dataset
demonstrate the effectiveness of our hybrid method. It is able to discover sets
consisting of fewer genes than those reported in the literature and at the same
time achieve the same or better prediction accuracy.
| no_new_dataset | 0.946001 |
1506.02087 | Min Xu | Min Xu | Global Gene Expression Analysis Using Machine Learning Methods | Author's master thesis (National University of Singapore, May 2003).
Adviser: Rudy Setiono | null | null | null | q-bio.QM cs.CE cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Microarray is a technology to quantitatively monitor the expression of large
number of genes in parallel. It has become one of the main tools for global
gene expression analysis in molecular biology research in recent years. The
large amount of expression data generated by this technology makes the study of
certain complex biological problems possible and machine learning methods are
playing a crucial role in the analysis process. At present, many machine
learning methods have been or have the potential to be applied to major areas
of gene expression analysis. These areas include clustering, classification,
dynamic modeling and reverse engineering.
In this thesis, we focus our work on using machine learning methods to solve
the classification problems arising from microarray data. We first identify the
major types of the classification problems; then apply several machine learning
methods to solve the problems and perform systematic tests on real and
artificial datasets. We propose improvement to existing methods. Specifically,
we develop a multivariate and a hybrid feature selection method to obtain high
classification performance for high dimension classification problems. Using
the hybrid feature selection method, we are able to identify small sets of
features that give predictive accuracy that is as good as that from other
methods which require many more features.
| [
{
"version": "v1",
"created": "Fri, 5 Jun 2015 23:37:20 GMT"
}
] | 2015-06-18T00:00:00 | [
[
"Xu",
"Min",
""
]
] | TITLE: Global Gene Expression Analysis Using Machine Learning Methods
ABSTRACT: Microarray is a technology to quantitatively monitor the expression of large
number of genes in parallel. It has become one of the main tools for global
gene expression analysis in molecular biology research in recent years. The
large amount of expression data generated by this technology makes the study of
certain complex biological problems possible and machine learning methods are
playing a crucial role in the analysis process. At present, many machine
learning methods have been or have the potential to be applied to major areas
of gene expression analysis. These areas include clustering, classification,
dynamic modeling and reverse engineering.
In this thesis, we focus our work on using machine learning methods to solve
the classification problems arising from microarray data. We first identify the
major types of the classification problems; then apply several machine learning
methods to solve the problems and perform systematic tests on real and
artificial datasets. We propose improvement to existing methods. Specifically,
we develop a multivariate and a hybrid feature selection method to obtain high
classification performance for high dimension classification problems. Using
the hybrid feature selection method, we are able to identify small sets of
features that give predictive accuracy that is as good as that from other
methods which require many more features.
| no_new_dataset | 0.950503 |
1506.04924 | Seunghoon Hong | Seunghoon Hong, Hyeonwoo Noh, Bohyung Han | Decoupled Deep Neural Network for Semi-supervised Semantic Segmentation | Added a link to the project page for more comprehensive illustration
of results | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a novel deep neural network architecture for semi-supervised
semantic segmentation using heterogeneous annotations. Contrary to existing
approaches posing semantic segmentation as a single task of region-based
classification, our algorithm decouples classification and segmentation, and
learns a separate network for each task. In this architecture, labels
associated with an image are identified by classification network, and binary
segmentation is subsequently performed for each identified label in
segmentation network. The decoupled architecture enables us to learn
classification and segmentation networks separately based on the training data
with image-level and pixel-wise class labels, respectively. It facilitates to
reduce search space for segmentation effectively by exploiting class-specific
activation maps obtained from bridging layers. Our algorithm shows outstanding
performance compared to other semi-supervised approaches even with much less
training images with strong annotations in PASCAL VOC dataset.
| [
{
"version": "v1",
"created": "Tue, 16 Jun 2015 11:20:04 GMT"
},
{
"version": "v2",
"created": "Wed, 17 Jun 2015 08:38:32 GMT"
}
] | 2015-06-18T00:00:00 | [
[
"Hong",
"Seunghoon",
""
],
[
"Noh",
"Hyeonwoo",
""
],
[
"Han",
"Bohyung",
""
]
] | TITLE: Decoupled Deep Neural Network for Semi-supervised Semantic Segmentation
ABSTRACT: We propose a novel deep neural network architecture for semi-supervised
semantic segmentation using heterogeneous annotations. Contrary to existing
approaches posing semantic segmentation as a single task of region-based
classification, our algorithm decouples classification and segmentation, and
learns a separate network for each task. In this architecture, labels
associated with an image are identified by classification network, and binary
segmentation is subsequently performed for each identified label in
segmentation network. The decoupled architecture enables us to learn
classification and segmentation networks separately based on the training data
with image-level and pixel-wise class labels, respectively. It facilitates to
reduce search space for segmentation effectively by exploiting class-specific
activation maps obtained from bridging layers. Our algorithm shows outstanding
performance compared to other semi-supervised approaches even with much less
training images with strong annotations in PASCAL VOC dataset.
| no_new_dataset | 0.953405 |
1506.05158 | Taylor Arnold | Taylor Arnold | An Entropy Maximizing Geohash for Distributed Spatiotemporal Database
Indexing | 12 pages, 4 figures | null | null | null | cs.DB cs.DC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a modification of the standard geohash algorithm based on maximum
entropy encoding in which the data volume is approximately constant for a given
hash prefix length. Distributed spatiotemporal databases, which typically
require interleaving spatial and temporal elements into a single key, reap
large benefits from a balanced geohash by creating a consistent ratio between
spatial and temporal precision even across areas of varying data density. This
property is also useful for indexing purely spatial datasets, where the load
distribution of large range scans is an important aspect of query performance.
We apply our algorithm to data generated proportional to population as given by
census block population counts provided from the US Census Bureau.
| [
{
"version": "v1",
"created": "Tue, 16 Jun 2015 21:54:12 GMT"
}
] | 2015-06-18T00:00:00 | [
[
"Arnold",
"Taylor",
""
]
] | TITLE: An Entropy Maximizing Geohash for Distributed Spatiotemporal Database
Indexing
ABSTRACT: We present a modification of the standard geohash algorithm based on maximum
entropy encoding in which the data volume is approximately constant for a given
hash prefix length. Distributed spatiotemporal databases, which typically
require interleaving spatial and temporal elements into a single key, reap
large benefits from a balanced geohash by creating a consistent ratio between
spatial and temporal precision even across areas of varying data density. This
property is also useful for indexing purely spatial datasets, where the load
distribution of large range scans is an important aspect of query performance.
We apply our algorithm to data generated proportional to population as given by
census block population counts provided from the US Census Bureau.
| no_new_dataset | 0.948155 |
1506.05257 | Daniel J Mankowitz | Daniel J. Mankowitz and Ehud Rivlin | CFORB: Circular FREAK-ORB Visual Odometry | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a novel Visual Odometry algorithm entitled Circular FREAK-ORB
(CFORB). This algorithm detects features using the well-known ORB algorithm
[12] and computes feature descriptors using the FREAK algorithm [14]. CFORB is
invariant to both rotation and scale changes, and is suitable for use in
environments with uneven terrain. Two visual geometric constraints have been
utilized in order to remove invalid feature descriptor matches. These
constraints have not previously been utilized in a Visual Odometry algorithm. A
variation to circular matching [16] has also been implemented. This allows
features to be matched between images without having to be dependent upon the
epipolar constraint. This algorithm has been run on the KITTI benchmark dataset
and achieves a competitive average translational error of $3.73 \%$ and average
rotational error of $0.0107 deg/m$. CFORB has also been run in an indoor
environment and achieved an average translational error of $3.70 \%$. After
running CFORB in a highly textured environment with an approximately uniform
feature spread across the images, the algorithm achieves an average
translational error of $2.4 \%$ and an average rotational error of $0.009
deg/m$.
| [
{
"version": "v1",
"created": "Wed, 17 Jun 2015 09:44:42 GMT"
}
] | 2015-06-18T00:00:00 | [
[
"Mankowitz",
"Daniel J.",
""
],
[
"Rivlin",
"Ehud",
""
]
] | TITLE: CFORB: Circular FREAK-ORB Visual Odometry
ABSTRACT: We present a novel Visual Odometry algorithm entitled Circular FREAK-ORB
(CFORB). This algorithm detects features using the well-known ORB algorithm
[12] and computes feature descriptors using the FREAK algorithm [14]. CFORB is
invariant to both rotation and scale changes, and is suitable for use in
environments with uneven terrain. Two visual geometric constraints have been
utilized in order to remove invalid feature descriptor matches. These
constraints have not previously been utilized in a Visual Odometry algorithm. A
variation to circular matching [16] has also been implemented. This allows
features to be matched between images without having to be dependent upon the
epipolar constraint. This algorithm has been run on the KITTI benchmark dataset
and achieves a competitive average translational error of $3.73 \%$ and average
rotational error of $0.0107 deg/m$. CFORB has also been run in an indoor
environment and achieved an average translational error of $3.70 \%$. After
running CFORB in a highly textured environment with an approximately uniform
feature spread across the images, the algorithm achieves an average
translational error of $2.4 \%$ and an average rotational error of $0.009
deg/m$.
| no_new_dataset | 0.948822 |
1309.0691 | Zi-Ke Zhang Dr. | Chu-Xu Zhang, Zi-Ke Zhang, Lu Yu, Chuang Liu, Hao Liu, Xiao-Yong Yan | Information Filtering via Collaborative User Clustering Modeling | null | null | 10.1016/j.physa.2013.11.024 | null | cs.IR cs.SI physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The past few years have witnessed the great success of recommender systems,
which can significantly help users find out personalized items for them from
the information era. One of the most widely applied recommendation methods is
the Matrix Factorization (MF). However, most of researches on this topic have
focused on mining the direct relationships between users and items. In this
paper, we optimize the standard MF by integrating the user clustering
regularization term. Our model considers not only the user-item rating
information, but also takes into account the user interest. We compared the
proposed model with three typical other methods: User-Mean (UM), Item-Mean (IM)
and standard MF. Experimental results on a real-world dataset,
\emph{MovieLens}, show that our method performs much better than other three
methods in the accuracy of recommendation.
| [
{
"version": "v1",
"created": "Tue, 3 Sep 2013 14:20:00 GMT"
},
{
"version": "v2",
"created": "Wed, 4 Sep 2013 09:20:30 GMT"
},
{
"version": "v3",
"created": "Thu, 7 Nov 2013 16:29:26 GMT"
},
{
"version": "v4",
"created": "Mon, 10 Feb 2014 08:40:21 GMT"
}
] | 2015-06-17T00:00:00 | [
[
"Zhang",
"Chu-Xu",
""
],
[
"Zhang",
"Zi-Ke",
""
],
[
"Yu",
"Lu",
""
],
[
"Liu",
"Chuang",
""
],
[
"Liu",
"Hao",
""
],
[
"Yan",
"Xiao-Yong",
""
]
] | TITLE: Information Filtering via Collaborative User Clustering Modeling
ABSTRACT: The past few years have witnessed the great success of recommender systems,
which can significantly help users find out personalized items for them from
the information era. One of the most widely applied recommendation methods is
the Matrix Factorization (MF). However, most of researches on this topic have
focused on mining the direct relationships between users and items. In this
paper, we optimize the standard MF by integrating the user clustering
regularization term. Our model considers not only the user-item rating
information, but also takes into account the user interest. We compared the
proposed model with three typical other methods: User-Mean (UM), Item-Mean (IM)
and standard MF. Experimental results on a real-world dataset,
\emph{MovieLens}, show that our method performs much better than other three
methods in the accuracy of recommendation.
| no_new_dataset | 0.949482 |
1309.2920 | Chunxiao Jiang | Chunxiao Jiang and Yan Chen and K. J. Ray Liu | Evolutionary Information Diffusion over Social Networks | null | null | 10.1109/TSP.2014.2339799 | null | cs.GT cs.SI physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Social networks have become ubiquitous in our daily life, as such it has
attracted great research interests recently. A key challenge is that it is of
extremely large-scale with tremendous information flow, creating the phenomenon
of "Big Data". Under such a circumstance, understanding information diffusion
over social networks has become an important research issue. Most of the
existing works on information diffusion analysis are based on either network
structure modeling or empirical approach with dataset mining. However, the
information diffusion is also heavily influenced by network users' decisions,
actions and their socio-economic connections, which is generally ignored in
existing works. In this paper, we propose an evolutionary game theoretic
framework to model the dynamic information diffusion process in social
networks. Specifically, we analyze the framework in uniform degree and
non-uniform degree networks and derive the closed-form expressions of the
evolutionary stable network states. Moreover, the information diffusion over
two special networks, Erd\H{o}s-R\'enyi random network and the
Barab\'asi-Albert scale-free network, are also highlighted. To verify our
theoretical analysis, we conduct experiments by using both synthetic networks
and real-world Facebook network, as well as real-world information spreading
dataset of Twitter and Memetracker. Experiments shows that the proposed game
theoretic framework is effective and practical in modeling the social network
users' information forwarding behaviors.
| [
{
"version": "v1",
"created": "Wed, 11 Sep 2013 19:22:33 GMT"
}
] | 2015-06-17T00:00:00 | [
[
"Jiang",
"Chunxiao",
""
],
[
"Chen",
"Yan",
""
],
[
"Liu",
"K. J. Ray",
""
]
] | TITLE: Evolutionary Information Diffusion over Social Networks
ABSTRACT: Social networks have become ubiquitous in our daily life, as such it has
attracted great research interests recently. A key challenge is that it is of
extremely large-scale with tremendous information flow, creating the phenomenon
of "Big Data". Under such a circumstance, understanding information diffusion
over social networks has become an important research issue. Most of the
existing works on information diffusion analysis are based on either network
structure modeling or empirical approach with dataset mining. However, the
information diffusion is also heavily influenced by network users' decisions,
actions and their socio-economic connections, which is generally ignored in
existing works. In this paper, we propose an evolutionary game theoretic
framework to model the dynamic information diffusion process in social
networks. Specifically, we analyze the framework in uniform degree and
non-uniform degree networks and derive the closed-form expressions of the
evolutionary stable network states. Moreover, the information diffusion over
two special networks, Erd\H{o}s-R\'enyi random network and the
Barab\'asi-Albert scale-free network, are also highlighted. To verify our
theoretical analysis, we conduct experiments by using both synthetic networks
and real-world Facebook network, as well as real-world information spreading
dataset of Twitter and Memetracker. Experiments shows that the proposed game
theoretic framework is effective and practical in modeling the social network
users' information forwarding behaviors.
| no_new_dataset | 0.952442 |
1309.3330 | Aditya Vempaty | Aditya Vempaty, Lav R. Varshney and Pramod K. Varshney | Reliable Crowdsourcing for Multi-Class Labeling using Coding Theory | 20 pages, 11 figures, under revision, IEEE Journal of Selected Topics
in Signal Processing | null | 10.1109/JSTSP.2014.2316116 | null | cs.IT cs.SI math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Crowdsourcing systems often have crowd workers that perform unreliable work
on the task they are assigned. In this paper, we propose the use of
error-control codes and decoding algorithms to design crowdsourcing systems for
reliable classification despite unreliable crowd workers. Coding-theory based
techniques also allow us to pose easy-to-answer binary questions to the crowd
workers. We consider three different crowdsourcing models: systems with
independent crowd workers, systems with peer-dependent reward schemes, and
systems where workers have common sources of information. For each of these
models, we analyze classification performance with the proposed coding-based
scheme. We develop an ordering principle for the quality of crowds and describe
how system performance changes with the quality of the crowd. We also show that
pairing among workers and diversification of the questions help in improving
system performance. We demonstrate the effectiveness of the proposed
coding-based scheme using both simulated data and real datasets from Amazon
Mechanical Turk, a crowdsourcing microtask platform. Results suggest that use
of good codes may improve the performance of the crowdsourcing task over
typical majority-voting approaches.
| [
{
"version": "v1",
"created": "Thu, 12 Sep 2013 23:10:32 GMT"
},
{
"version": "v2",
"created": "Wed, 22 Jan 2014 21:23:43 GMT"
}
] | 2015-06-17T00:00:00 | [
[
"Vempaty",
"Aditya",
""
],
[
"Varshney",
"Lav R.",
""
],
[
"Varshney",
"Pramod K.",
""
]
] | TITLE: Reliable Crowdsourcing for Multi-Class Labeling using Coding Theory
ABSTRACT: Crowdsourcing systems often have crowd workers that perform unreliable work
on the task they are assigned. In this paper, we propose the use of
error-control codes and decoding algorithms to design crowdsourcing systems for
reliable classification despite unreliable crowd workers. Coding-theory based
techniques also allow us to pose easy-to-answer binary questions to the crowd
workers. We consider three different crowdsourcing models: systems with
independent crowd workers, systems with peer-dependent reward schemes, and
systems where workers have common sources of information. For each of these
models, we analyze classification performance with the proposed coding-based
scheme. We develop an ordering principle for the quality of crowds and describe
how system performance changes with the quality of the crowd. We also show that
pairing among workers and diversification of the questions help in improving
system performance. We demonstrate the effectiveness of the proposed
coding-based scheme using both simulated data and real datasets from Amazon
Mechanical Turk, a crowdsourcing microtask platform. Results suggest that use
of good codes may improve the performance of the crowdsourcing task over
typical majority-voting approaches.
| no_new_dataset | 0.953449 |
1309.4411 | Ginestra Bianconi | Arda Halu, Satyam Mukherjee and Ginestra Bianconi | Emergence of overlap in ensembles of spatial multiplexes and statistical
mechanics of spatial interacting networks ensembles | (12 pages, 4 figures) for downloading data see URL
http://sites.google.com/site/satyammukherjee/pubs | Phys. Rev. E 89, 012806 (2014) | 10.1103/PhysRevE.89.012806 | null | physics.soc-ph cond-mat.dis-nn cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Spatial networks range from the brain networks, to transportation networks
and infrastructures. Recently interacting and multiplex networks are attracting
great attention because their dynamics and robustness cannot be understood
without treating at the same time several networks. Here we present maximal
entropy ensembles of spatial multiplex and spatial interacting networks that
can be used in order to model spatial multilayer network structures and to
build null models of real datasets. We show that spatial multiplex naturally
develop a significant overlap of the links, a noticeable property of many
multiplexes that can affect significantly the dynamics taking place on them.
Additionally, we characterize ensembles of spatial interacting networks and we
analyse the structure of interacting airport and railway networks in India,
showing the effect of space in determining the link probability.
| [
{
"version": "v1",
"created": "Tue, 17 Sep 2013 18:05:29 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Dec 2013 17:59:58 GMT"
},
{
"version": "v3",
"created": "Wed, 29 Apr 2015 19:49:40 GMT"
}
] | 2015-06-17T00:00:00 | [
[
"Halu",
"Arda",
""
],
[
"Mukherjee",
"Satyam",
""
],
[
"Bianconi",
"Ginestra",
""
]
] | TITLE: Emergence of overlap in ensembles of spatial multiplexes and statistical
mechanics of spatial interacting networks ensembles
ABSTRACT: Spatial networks range from the brain networks, to transportation networks
and infrastructures. Recently interacting and multiplex networks are attracting
great attention because their dynamics and robustness cannot be understood
without treating at the same time several networks. Here we present maximal
entropy ensembles of spatial multiplex and spatial interacting networks that
can be used in order to model spatial multilayer network structures and to
build null models of real datasets. We show that spatial multiplex naturally
develop a significant overlap of the links, a noticeable property of many
multiplexes that can affect significantly the dynamics taking place on them.
Additionally, we characterize ensembles of spatial interacting networks and we
analyse the structure of interacting airport and railway networks in India,
showing the effect of space in determining the link probability.
| no_new_dataset | 0.949248 |
1309.7031 | Nicola Perra | Suyu Liu, Nicola Perra, Marton Karsai, Alessandro Vespignani | Controlling Contagion Processes in Time-Varying Networks | null | null | 10.1103/PhysRevLett.112.118702 | null | physics.soc-ph cs.SI q-bio.PE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The vast majority of strategies aimed at controlling contagion processes on
networks considers the connectivity pattern of the system as either quenched or
annealed. However, in the real world many networks are highly dynamical and
evolve in time concurrently to the contagion process. Here, we derive an
analytical framework for the study of control strategies specifically devised
for time-varying networks. We consider the removal/immunization of individual
nodes according the their activity in the network and develop a block variable
mean-field approach that allows the derivation of the equations describing the
evolution of the contagion process concurrently to the network dynamic. We
derive the critical immunization threshold and assess the effectiveness of the
control strategies. Finally, we validate the theoretical picture by simulating
numerically the information spreading process and control strategies in both
synthetic networks and a large-scale, real-world mobile telephone call dataset
| [
{
"version": "v1",
"created": "Thu, 26 Sep 2013 19:50:15 GMT"
}
] | 2015-06-17T00:00:00 | [
[
"Liu",
"Suyu",
""
],
[
"Perra",
"Nicola",
""
],
[
"Karsai",
"Marton",
""
],
[
"Vespignani",
"Alessandro",
""
]
] | TITLE: Controlling Contagion Processes in Time-Varying Networks
ABSTRACT: The vast majority of strategies aimed at controlling contagion processes on
networks considers the connectivity pattern of the system as either quenched or
annealed. However, in the real world many networks are highly dynamical and
evolve in time concurrently to the contagion process. Here, we derive an
analytical framework for the study of control strategies specifically devised
for time-varying networks. We consider the removal/immunization of individual
nodes according the their activity in the network and develop a block variable
mean-field approach that allows the derivation of the equations describing the
evolution of the contagion process concurrently to the network dynamic. We
derive the critical immunization threshold and assess the effectiveness of the
control strategies. Finally, we validate the theoretical picture by simulating
numerically the information spreading process and control strategies in both
synthetic networks and a large-scale, real-world mobile telephone call dataset
| no_new_dataset | 0.946547 |
1310.2632 | Philip Schniter | Jason T. Parker, Philip Schniter, and Volkan Cevher | Bilinear Generalized Approximate Message Passing | null | null | 10.1109/TSP.2014.2357776 | null | cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We extend the generalized approximate message passing (G-AMP) approach,
originally proposed for high-dimensional generalized-linear regression in the
context of compressive sensing, to the generalized-bilinear case, which enables
its application to matrix completion, robust PCA, dictionary learning, and
related matrix-factorization problems. In the first part of the paper, we
derive our Bilinear G-AMP (BiG-AMP) algorithm as an approximation of the
sum-product belief propagation algorithm in the high-dimensional limit, where
central-limit theorem arguments and Taylor-series approximations apply, and
under the assumption of statistically independent matrix entries with known
priors. In addition, we propose an adaptive damping mechanism that aids
convergence under finite problem sizes, an expectation-maximization (EM)-based
method to automatically tune the parameters of the assumed priors, and two
rank-selection strategies. In the second part of the paper, we discuss the
specializations of EM-BiG-AMP to the problems of matrix completion, robust PCA,
and dictionary learning, and present the results of an extensive empirical
study comparing EM-BiG-AMP to state-of-the-art algorithms on each problem. Our
numerical results, using both synthetic and real-world datasets, demonstrate
that EM-BiG-AMP yields excellent reconstruction accuracy (often best in class)
while maintaining competitive runtimes and avoiding the need to tune
algorithmic parameters.
| [
{
"version": "v1",
"created": "Wed, 9 Oct 2013 21:08:40 GMT"
},
{
"version": "v2",
"created": "Fri, 1 Nov 2013 16:45:09 GMT"
},
{
"version": "v3",
"created": "Thu, 5 Jun 2014 14:32:06 GMT"
}
] | 2015-06-17T00:00:00 | [
[
"Parker",
"Jason T.",
""
],
[
"Schniter",
"Philip",
""
],
[
"Cevher",
"Volkan",
""
]
] | TITLE: Bilinear Generalized Approximate Message Passing
ABSTRACT: We extend the generalized approximate message passing (G-AMP) approach,
originally proposed for high-dimensional generalized-linear regression in the
context of compressive sensing, to the generalized-bilinear case, which enables
its application to matrix completion, robust PCA, dictionary learning, and
related matrix-factorization problems. In the first part of the paper, we
derive our Bilinear G-AMP (BiG-AMP) algorithm as an approximation of the
sum-product belief propagation algorithm in the high-dimensional limit, where
central-limit theorem arguments and Taylor-series approximations apply, and
under the assumption of statistically independent matrix entries with known
priors. In addition, we propose an adaptive damping mechanism that aids
convergence under finite problem sizes, an expectation-maximization (EM)-based
method to automatically tune the parameters of the assumed priors, and two
rank-selection strategies. In the second part of the paper, we discuss the
specializations of EM-BiG-AMP to the problems of matrix completion, robust PCA,
and dictionary learning, and present the results of an extensive empirical
study comparing EM-BiG-AMP to state-of-the-art algorithms on each problem. Our
numerical results, using both synthetic and real-world datasets, demonstrate
that EM-BiG-AMP yields excellent reconstruction accuracy (often best in class)
while maintaining competitive runtimes and avoiding the need to tune
algorithmic parameters.
| no_new_dataset | 0.947332 |
1311.1753 | Rolf Andreassen | R. Andreassen, B. T. Meadows, M. de Silva, M. D. Sokoloff, and K.
Tomko | GooFit: A library for massively parallelising maximum-likelihood fits | Presented at the CHEP 2013 conference | null | 10.1088/1742-6596/513/5/052003 | null | cs.DC cs.MS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Fitting complicated models to large datasets is a bottleneck of many
analyses. We present GooFit, a library and tool for constructing
arbitrarily-complex probability density functions (PDFs) to be evaluated on
nVidia GPUs or on multicore CPUs using OpenMP. The massive parallelisation of
dividing up event calculations between hundreds of processors can achieve
speedups of factors 200-300 in real-world problems.
| [
{
"version": "v1",
"created": "Thu, 7 Nov 2013 17:18:42 GMT"
}
] | 2015-06-17T00:00:00 | [
[
"Andreassen",
"R.",
""
],
[
"Meadows",
"B. T.",
""
],
[
"de Silva",
"M.",
""
],
[
"Sokoloff",
"M. D.",
""
],
[
"Tomko",
"K.",
""
]
] | TITLE: GooFit: A library for massively parallelising maximum-likelihood fits
ABSTRACT: Fitting complicated models to large datasets is a bottleneck of many
analyses. We present GooFit, a library and tool for constructing
arbitrarily-complex probability density functions (PDFs) to be evaluated on
nVidia GPUs or on multicore CPUs using OpenMP. The massive parallelisation of
dividing up event calculations between hundreds of processors can achieve
speedups of factors 200-300 in real-world problems.
| no_new_dataset | 0.944228 |
1311.2911 | Kevin Kung | Kevin S. Kung, Kael Greco, Stanislav Sobolevsky, and Carlo Ratti | Exploring universal patterns in human home-work commuting from mobile
phone data | null | Kung KS, Greco K, Sobolevsky S, Ratti C (2014) Exploring Universal
Patterns in Human Home-Work Commuting from Mobile Phone Data. PLoS ONE 9(6):
e96180 | 10.1371/journal.pone.0096180 | null | cs.SI cs.CY physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Home-work commuting has always attracted significant research attention
because of its impact on human mobility. One of the key assumptions in this
domain of study is the universal uniformity of commute times. However, a true
comparison of commute patterns has often been hindered by the intrinsic
differences in data collection methods, which make observation from different
countries potentially biased and unreliable. In the present work, we approach
this problem through the use of mobile phone call detail records (CDRs), which
offers a consistent method for investigating mobility patterns in wholly
different parts of the world. We apply our analysis to a broad range of
datasets, at both the country and city scale. Additionally, we compare these
results with those obtained from vehicle GPS traces in Milan. While different
regions have some unique commute time characteristics, we show that the
home-work time distributions and average values within a single region are
indeed largely independent of commute distance or country (Portugal, Ivory
Coast, and Boston)--despite substantial spatial and infrastructural
differences. Furthermore, a comparative analysis demonstrates that such
distance-independence holds true only if we consider multimodal commute
behaviors--as consistent with previous studies. In car-only (Milan GPS traces)
and car-heavy (Saudi Arabia) commute datasets, we see that commute time is
indeed influenced by commute distance.
| [
{
"version": "v1",
"created": "Tue, 12 Nov 2013 20:29:14 GMT"
},
{
"version": "v2",
"created": "Wed, 24 Sep 2014 20:46:56 GMT"
}
] | 2015-06-17T00:00:00 | [
[
"Kung",
"Kevin S.",
""
],
[
"Greco",
"Kael",
""
],
[
"Sobolevsky",
"Stanislav",
""
],
[
"Ratti",
"Carlo",
""
]
] | TITLE: Exploring universal patterns in human home-work commuting from mobile
phone data
ABSTRACT: Home-work commuting has always attracted significant research attention
because of its impact on human mobility. One of the key assumptions in this
domain of study is the universal uniformity of commute times. However, a true
comparison of commute patterns has often been hindered by the intrinsic
differences in data collection methods, which make observation from different
countries potentially biased and unreliable. In the present work, we approach
this problem through the use of mobile phone call detail records (CDRs), which
offers a consistent method for investigating mobility patterns in wholly
different parts of the world. We apply our analysis to a broad range of
datasets, at both the country and city scale. Additionally, we compare these
results with those obtained from vehicle GPS traces in Milan. While different
regions have some unique commute time characteristics, we show that the
home-work time distributions and average values within a single region are
indeed largely independent of commute distance or country (Portugal, Ivory
Coast, and Boston)--despite substantial spatial and infrastructural
differences. Furthermore, a comparative analysis demonstrates that such
distance-independence holds true only if we consider multimodal commute
behaviors--as consistent with previous studies. In car-only (Milan GPS traces)
and car-heavy (Saudi Arabia) commute datasets, we see that commute time is
indeed influenced by commute distance.
| no_new_dataset | 0.921922 |
1407.7390 | Jos\'e Ram\'on Padilla-L\'opez | Jos\'e Ram\'on Padilla-L\'opez and Alexandros Andr\'e Chaaraoui and
Francisco Fl\'orez-Revuelta | A discussion on the validation tests employed to compare human action
recognition methods using the MSR Action3D dataset | 16 pages and 7 tables | null | null | hdl:10045/39889 | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper aims to determine which is the best human action recognition
method based on features extracted from RGB-D devices, such as the Microsoft
Kinect. A review of all the papers that make reference to MSR Action3D, the
most used dataset that includes depth information acquired from a RGB-D device,
has been performed. We found that the validation method used by each work
differs from the others. So, a direct comparison among works cannot be made.
However, almost all the works present their results comparing them without
taking into account this issue. Therefore, we present different rankings
according to the methodology used for the validation in orden to clarify the
existing confusion.
| [
{
"version": "v1",
"created": "Mon, 28 Jul 2014 11:59:30 GMT"
},
{
"version": "v2",
"created": "Mon, 12 Jan 2015 11:30:40 GMT"
},
{
"version": "v3",
"created": "Tue, 16 Jun 2015 19:57:45 GMT"
}
] | 2015-06-17T00:00:00 | [
[
"Padilla-López",
"José Ramón",
""
],
[
"Chaaraoui",
"Alexandros André",
""
],
[
"Flórez-Revuelta",
"Francisco",
""
]
] | TITLE: A discussion on the validation tests employed to compare human action
recognition methods using the MSR Action3D dataset
ABSTRACT: This paper aims to determine which is the best human action recognition
method based on features extracted from RGB-D devices, such as the Microsoft
Kinect. A review of all the papers that make reference to MSR Action3D, the
most used dataset that includes depth information acquired from a RGB-D device,
has been performed. We found that the validation method used by each work
differs from the others. So, a direct comparison among works cannot be made.
However, almost all the works present their results comparing them without
taking into account this issue. Therefore, we present different rankings
according to the methodology used for the validation in orden to clarify the
existing confusion.
| no_new_dataset | 0.867598 |
1506.01071 | Aleksey Buzmakov | Aleksey Buzmakov and Sergei O. Kuznetsov and Amedeo Napoli | Fast Generation of Best Interval Patterns for Nonmonotonic Constraints | 18 pages; 2 figures; 2 tables; 1 algorithm; PKDD 2015 Conference
Scientific Track | null | null | null | cs.AI cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In pattern mining, the main challenge is the exponential explosion of the set
of patterns. Typically, to solve this problem, a constraint for pattern
selection is introduced. One of the first constraints proposed in pattern
mining is support (frequency) of a pattern in a dataset. Frequency is an
anti-monotonic function, i.e., given an infrequent pattern, all its
superpatterns are not frequent. However, many other constraints for pattern
selection are not (anti-)monotonic, which makes it difficult to generate
patterns satisfying these constraints. In this paper we introduce the notion of
projection-antimonotonicity and $\theta$-$\Sigma\o\phi\iota\alpha$ algorithm
that allows efficient generation of the best patterns for some nonmonotonic
constraints. In this paper we consider stability and $\Delta$-measure, which
are nonmonotonic constraints, and apply them to interval tuple datasets. In the
experiments, we compute best interval tuple patterns w.r.t. these measures and
show the advantage of our approach over postfiltering approaches.
KEYWORDS: Pattern mining, nonmonotonic constraints, interval tuple data
| [
{
"version": "v1",
"created": "Tue, 2 Jun 2015 21:32:14 GMT"
},
{
"version": "v2",
"created": "Tue, 16 Jun 2015 15:31:19 GMT"
}
] | 2015-06-17T00:00:00 | [
[
"Buzmakov",
"Aleksey",
""
],
[
"Kuznetsov",
"Sergei O.",
""
],
[
"Napoli",
"Amedeo",
""
]
] | TITLE: Fast Generation of Best Interval Patterns for Nonmonotonic Constraints
ABSTRACT: In pattern mining, the main challenge is the exponential explosion of the set
of patterns. Typically, to solve this problem, a constraint for pattern
selection is introduced. One of the first constraints proposed in pattern
mining is support (frequency) of a pattern in a dataset. Frequency is an
anti-monotonic function, i.e., given an infrequent pattern, all its
superpatterns are not frequent. However, many other constraints for pattern
selection are not (anti-)monotonic, which makes it difficult to generate
patterns satisfying these constraints. In this paper we introduce the notion of
projection-antimonotonicity and $\theta$-$\Sigma\o\phi\iota\alpha$ algorithm
that allows efficient generation of the best patterns for some nonmonotonic
constraints. In this paper we consider stability and $\Delta$-measure, which
are nonmonotonic constraints, and apply them to interval tuple datasets. In the
experiments, we compute best interval tuple patterns w.r.t. these measures and
show the advantage of our approach over postfiltering approaches.
KEYWORDS: Pattern mining, nonmonotonic constraints, interval tuple data
| no_new_dataset | 0.953232 |
1506.04757 | Julian McAuley | Julian McAuley, Christopher Targett, Qinfeng Shi, Anton van den Hengel | Image-based Recommendations on Styles and Substitutes | 11 pages, 10 figures, SIGIR 2015 | null | null | null | cs.CV cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Humans inevitably develop a sense of the relationships between objects, some
of which are based on their appearance. Some pairs of objects might be seen as
being alternatives to each other (such as two pairs of jeans), while others may
be seen as being complementary (such as a pair of jeans and a matching shirt).
This information guides many of the choices that people make, from buying
clothes to their interactions with each other. We seek here to model this human
sense of the relationships between objects based on their appearance. Our
approach is not based on fine-grained modeling of user annotations but rather
on capturing the largest dataset possible and developing a scalable method for
uncovering human notions of the visual relationships within. We cast this as a
network inference problem defined on graphs of related images, and provide a
large-scale dataset for the training and evaluation of the same. The system we
develop is capable of recommending which clothes and accessories will go well
together (and which will not), amongst a host of other applications.
| [
{
"version": "v1",
"created": "Mon, 15 Jun 2015 20:01:49 GMT"
}
] | 2015-06-17T00:00:00 | [
[
"McAuley",
"Julian",
""
],
[
"Targett",
"Christopher",
""
],
[
"Shi",
"Qinfeng",
""
],
[
"Hengel",
"Anton van den",
""
]
] | TITLE: Image-based Recommendations on Styles and Substitutes
ABSTRACT: Humans inevitably develop a sense of the relationships between objects, some
of which are based on their appearance. Some pairs of objects might be seen as
being alternatives to each other (such as two pairs of jeans), while others may
be seen as being complementary (such as a pair of jeans and a matching shirt).
This information guides many of the choices that people make, from buying
clothes to their interactions with each other. We seek here to model this human
sense of the relationships between objects based on their appearance. Our
approach is not based on fine-grained modeling of user annotations but rather
on capturing the largest dataset possible and developing a scalable method for
uncovering human notions of the visual relationships within. We cast this as a
network inference problem defined on graphs of related images, and provide a
large-scale dataset for the training and evaluation of the same. The system we
develop is capable of recommending which clothes and accessories will go well
together (and which will not), amongst a host of other applications.
| no_new_dataset | 0.939582 |
1506.04776 | Jeff Heaton | Jeff Heaton | Encog: Library of Interchangeable Machine Learning Models for Java and
C# | null | null | null | null | cs.MS cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper introduces the Encog library for Java and C#, a scalable,
adaptable, multiplatform machine learning framework that was 1st released in
2008. Encog allows a variety of machine learning models to be applied to
datasets using regression, classification, and clustering. Various supported
machine learning models can be used interchangeably with minimal recoding.
Encog uses efficient multithreaded code to reduce training time by exploiting
modern multicore processors. The current version of Encog can be downloaded
from http://www.encog.org.
| [
{
"version": "v1",
"created": "Mon, 15 Jun 2015 21:20:06 GMT"
}
] | 2015-06-17T00:00:00 | [
[
"Heaton",
"Jeff",
""
]
] | TITLE: Encog: Library of Interchangeable Machine Learning Models for Java and
C#
ABSTRACT: This paper introduces the Encog library for Java and C#, a scalable,
adaptable, multiplatform machine learning framework that was 1st released in
2008. Encog allows a variety of machine learning models to be applied to
datasets using regression, classification, and clustering. Various supported
machine learning models can be used interchangeably with minimal recoding.
Encog uses efficient multithreaded code to reduce training time by exploiting
modern multicore processors. The current version of Encog can be downloaded
from http://www.encog.org.
| no_new_dataset | 0.947137 |
1506.04803 | Afshin Rahimi | Afshin Rahimi, Duy Vu, Trevor Cohn, and Timothy Baldwin | Exploiting Text and Network Context for Geolocation of Social Media
Users | null | null | null | null | cs.CL cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Research on automatically geolocating social media users has conventionally
been based on the text content of posts from a given user or the social network
of the user, with very little crossover between the two, and no bench-marking
of the two approaches over compara- ble datasets. We bring the two threads of
research together in first proposing a text-based method based on adaptive
grids, followed by a hybrid network- and text-based method. Evaluating over
three Twitter datasets, we show that the empirical difference between text- and
network-based methods is not great, and that hybridisation of the two is
superior to the component methods, especially in contexts where the user graph
is not well connected. We achieve state-of-the-art results on all three
datasets.
| [
{
"version": "v1",
"created": "Tue, 16 Jun 2015 00:32:33 GMT"
}
] | 2015-06-17T00:00:00 | [
[
"Rahimi",
"Afshin",
""
],
[
"Vu",
"Duy",
""
],
[
"Cohn",
"Trevor",
""
],
[
"Baldwin",
"Timothy",
""
]
] | TITLE: Exploiting Text and Network Context for Geolocation of Social Media
Users
ABSTRACT: Research on automatically geolocating social media users has conventionally
been based on the text content of posts from a given user or the social network
of the user, with very little crossover between the two, and no bench-marking
of the two approaches over compara- ble datasets. We bring the two threads of
research together in first proposing a text-based method based on adaptive
grids, followed by a hybrid network- and text-based method. Evaluating over
three Twitter datasets, we show that the empirical difference between text- and
network-based methods is not great, and that hybridisation of the two is
superior to the component methods, especially in contexts where the user graph
is not well connected. We achieve state-of-the-art results on all three
datasets.
| no_new_dataset | 0.948155 |
1506.04815 | Amit Chavan | Amit Chavan, Silu Huang, Amol Deshpande, Aaron Elmore, Samuel Madden
and Aditya Parameswaran | Towards a unified query language for provenance and versioning | Theory and Practice of Provenance, 2015 | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Organizations and teams collect and acquire data from various sources, such
as social interactions, financial transactions, sensor data, and genome
sequencers. Different teams in an organization as well as different data
scientists within a team are interested in extracting a variety of insights
which require combining and collaboratively analyzing datasets in diverse ways.
DataHub is a system that aims to provide robust version control and provenance
management for such a scenario. To be truly useful for collaborative data
science, one also needs the ability to specify queries and analysis tasks over
the versioning and the provenance information in a unified manner. In this
paper, we present an initial design of our query language, called VQuel, that
aims to support such unified querying over both types of information, as well
as the intermediate and final results of analyses. We also discuss some of the
key language design and implementation challenges moving forward.
| [
{
"version": "v1",
"created": "Tue, 16 Jun 2015 01:32:51 GMT"
}
] | 2015-06-17T00:00:00 | [
[
"Chavan",
"Amit",
""
],
[
"Huang",
"Silu",
""
],
[
"Deshpande",
"Amol",
""
],
[
"Elmore",
"Aaron",
""
],
[
"Madden",
"Samuel",
""
],
[
"Parameswaran",
"Aditya",
""
]
] | TITLE: Towards a unified query language for provenance and versioning
ABSTRACT: Organizations and teams collect and acquire data from various sources, such
as social interactions, financial transactions, sensor data, and genome
sequencers. Different teams in an organization as well as different data
scientists within a team are interested in extracting a variety of insights
which require combining and collaboratively analyzing datasets in diverse ways.
DataHub is a system that aims to provide robust version control and provenance
management for such a scenario. To be truly useful for collaborative data
science, one also needs the ability to specify queries and analysis tasks over
the versioning and the provenance information in a unified manner. In this
paper, we present an initial design of our query language, called VQuel, that
aims to support such unified querying over both types of information, as well
as the intermediate and final results of analyses. We also discuss some of the
key language design and implementation challenges moving forward.
| no_new_dataset | 0.928926 |
1506.05101 | Dhruba Bhattacharyya | Hirak Kashyap, Hasin Afzal Ahmed, Nazrul Hoque, Swarup Roy and Dhruba
Kumar Bhattacharyya | Big Data Analytics in Bioinformatics: A Machine Learning Perspective | 20 pages survey paper on Big data analytics in Bioinformatics | null | null | null | cs.CE cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Bioinformatics research is characterized by voluminous and incremental
datasets and complex data analytics methods. The machine learning methods used
in bioinformatics are iterative and parallel. These methods can be scaled to
handle big data using the distributed and parallel computing technologies.
Usually big data tools perform computation in batch-mode and are not
optimized for iterative processing and high data dependency among operations.
In the recent years, parallel, incremental, and multi-view machine learning
algorithms have been proposed. Similarly, graph-based architectures and
in-memory big data tools have been developed to minimize I/O cost and optimize
iterative processing.
However, there lack standard big data architectures and tools for many
important bioinformatics problems, such as fast construction of co-expression
and regulatory networks and salient module identification, detection of
complexes over growing protein-protein interaction data, fast analysis of
massive DNA, RNA, and protein sequence data, and fast querying on incremental
and heterogeneous disease networks. This paper addresses the issues and
challenges posed by several big data problems in bioinformatics, and gives an
overview of the state of the art and the future research opportunities.
| [
{
"version": "v1",
"created": "Mon, 15 Jun 2015 11:32:00 GMT"
}
] | 2015-06-17T00:00:00 | [
[
"Kashyap",
"Hirak",
""
],
[
"Ahmed",
"Hasin Afzal",
""
],
[
"Hoque",
"Nazrul",
""
],
[
"Roy",
"Swarup",
""
],
[
"Bhattacharyya",
"Dhruba Kumar",
""
]
] | TITLE: Big Data Analytics in Bioinformatics: A Machine Learning Perspective
ABSTRACT: Bioinformatics research is characterized by voluminous and incremental
datasets and complex data analytics methods. The machine learning methods used
in bioinformatics are iterative and parallel. These methods can be scaled to
handle big data using the distributed and parallel computing technologies.
Usually big data tools perform computation in batch-mode and are not
optimized for iterative processing and high data dependency among operations.
In the recent years, parallel, incremental, and multi-view machine learning
algorithms have been proposed. Similarly, graph-based architectures and
in-memory big data tools have been developed to minimize I/O cost and optimize
iterative processing.
However, there lack standard big data architectures and tools for many
important bioinformatics problems, such as fast construction of co-expression
and regulatory networks and salient module identification, detection of
complexes over growing protein-protein interaction data, fast analysis of
massive DNA, RNA, and protein sequence data, and fast querying on incremental
and heterogeneous disease networks. This paper addresses the issues and
challenges posed by several big data problems in bioinformatics, and gives an
overview of the state of the art and the future research opportunities.
| no_new_dataset | 0.944689 |
1306.6455 | Sergio Servidio | S. Servidio, K.T. Osman, F. Valentini, D. Perrone, F. Califano, S.
Chapman, W. H. Matthaeus, and P. Veltri | Proton Kinetic Effects in Vlasov and Solar Wind Turbulence | 12 pages, 3 figures | null | 10.1088/2041-8205/781/2/L27 | null | physics.space-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Kinetic plasma processes have been investigated in the framework of solar
wind turbulence, employing Hybrid Vlasov-Maxwell (HVM) simulations. The
dependency of proton temperature anisotropy T_{\perp}/T_{\parallel} on the
parallel plasma beta \beta_{\parallel}, commonly observed in spacecraft data,
has been recovered using an ensemble of HVM simulations. By varying plasma
parameters, such as plasma beta and fluctuation level, the simulations explore
distinct regions of the parameter space given by T_{\perp}/T_{\parallel} and
\beta_{\parallel}, similar to solar wind sub-datasets. Moreover, both
simulation and solar wind data suggest that temperature anisotropy is not only
associated with magnetic intermittent events, but also with gradient-type
structures in the flow and in the density. This connection between
non-Maxwellian kinetic effects and various types of intermittency may be a key
point for understanding the complex nature of plasma turbulence.
| [
{
"version": "v1",
"created": "Thu, 27 Jun 2013 10:21:26 GMT"
}
] | 2015-06-16T00:00:00 | [
[
"Servidio",
"S.",
""
],
[
"Osman",
"K. T.",
""
],
[
"Valentini",
"F.",
""
],
[
"Perrone",
"D.",
""
],
[
"Califano",
"F.",
""
],
[
"Chapman",
"S.",
""
],
[
"Matthaeus",
"W. H.",
""
],
[
"Veltri",
"P.",
""
]
] | TITLE: Proton Kinetic Effects in Vlasov and Solar Wind Turbulence
ABSTRACT: Kinetic plasma processes have been investigated in the framework of solar
wind turbulence, employing Hybrid Vlasov-Maxwell (HVM) simulations. The
dependency of proton temperature anisotropy T_{\perp}/T_{\parallel} on the
parallel plasma beta \beta_{\parallel}, commonly observed in spacecraft data,
has been recovered using an ensemble of HVM simulations. By varying plasma
parameters, such as plasma beta and fluctuation level, the simulations explore
distinct regions of the parameter space given by T_{\perp}/T_{\parallel} and
\beta_{\parallel}, similar to solar wind sub-datasets. Moreover, both
simulation and solar wind data suggest that temperature anisotropy is not only
associated with magnetic intermittent events, but also with gradient-type
structures in the flow and in the density. This connection between
non-Maxwellian kinetic effects and various types of intermittency may be a key
point for understanding the complex nature of plasma turbulence.
| no_new_dataset | 0.956227 |
1307.3756 | Jean Golay | J. Golay, M. Kanevski, C. Vega Orozco, M. Leuenberger | The Multipoint Morisita Index for the Analysis of Spatial Patterns | null | null | 10.1016/j.physa.2014.03.063 | null | physics.data-an | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In many fields, the spatial clustering of sampled data points has many
consequences. Therefore, several indices have been proposed to assess the level
of clustering affecting datasets (e.g. the Morisita index, Ripley's K-function
and R\'enyi's generalized entropy). The classical Morisita index measures how
many times it is more likely to select two measurement points from the same
quadrats (the data set is covered by a regular grid of changing size) than it
would be in the case of a random distribution generated from a Poisson process.
The multipoint version (k-Morisita) takes into account k points with k greater
than or equal to 2. The present research deals with a new development of the
k-Morisita index for (1) monitoring network characterization and for (2) the
detection of patterns in monitored phenomena. From a theoretical perspective, a
connection between the k-Morisita index and multifractality has also been found
and highlighted on a mathematical multifractal set.
| [
{
"version": "v1",
"created": "Sun, 14 Jul 2013 17:17:24 GMT"
},
{
"version": "v2",
"created": "Thu, 5 Dec 2013 18:13:33 GMT"
},
{
"version": "v3",
"created": "Mon, 13 Jan 2014 16:20:35 GMT"
}
] | 2015-06-16T00:00:00 | [
[
"Golay",
"J.",
""
],
[
"Kanevski",
"M.",
""
],
[
"Orozco",
"C. Vega",
""
],
[
"Leuenberger",
"M.",
""
]
] | TITLE: The Multipoint Morisita Index for the Analysis of Spatial Patterns
ABSTRACT: In many fields, the spatial clustering of sampled data points has many
consequences. Therefore, several indices have been proposed to assess the level
of clustering affecting datasets (e.g. the Morisita index, Ripley's K-function
and R\'enyi's generalized entropy). The classical Morisita index measures how
many times it is more likely to select two measurement points from the same
quadrats (the data set is covered by a regular grid of changing size) than it
would be in the case of a random distribution generated from a Poisson process.
The multipoint version (k-Morisita) takes into account k points with k greater
than or equal to 2. The present research deals with a new development of the
k-Morisita index for (1) monitoring network characterization and for (2) the
detection of patterns in monitored phenomena. From a theoretical perspective, a
connection between the k-Morisita index and multifractality has also been found
and highlighted on a mathematical multifractal set.
| no_new_dataset | 0.950134 |
1506.04257 | Matthew Malloy | Matthew L. Malloy, Scott Alfeld, Paul Barford | Contamination Estimation via Convex Relaxations | To appear, ISIT 2015 | null | null | null | cs.IT cs.LG math.IT math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Identifying anomalies and contamination in datasets is important in a wide
variety of settings. In this paper, we describe a new technique for estimating
contamination in large, discrete valued datasets. Our approach considers the
normal condition of the data to be specified by a model consisting of a set of
distributions. Our key contribution is in our approach to contamination
estimation. Specifically, we develop a technique that identifies the minimum
number of data points that must be discarded (i.e., the level of contamination)
from an empirical data set in order to match the model to within a specified
goodness-of-fit, controlled by a p-value. Appealing to results from large
deviations theory, we show a lower bound on the level of contamination is
obtained by solving a series of convex programs. Theoretical results guarantee
the bound converges at a rate of $O(\sqrt{\log(p)/p})$, where p is the size of
the empirical data set.
| [
{
"version": "v1",
"created": "Sat, 13 Jun 2015 11:51:52 GMT"
}
] | 2015-06-16T00:00:00 | [
[
"Malloy",
"Matthew L.",
""
],
[
"Alfeld",
"Scott",
""
],
[
"Barford",
"Paul",
""
]
] | TITLE: Contamination Estimation via Convex Relaxations
ABSTRACT: Identifying anomalies and contamination in datasets is important in a wide
variety of settings. In this paper, we describe a new technique for estimating
contamination in large, discrete valued datasets. Our approach considers the
normal condition of the data to be specified by a model consisting of a set of
distributions. Our key contribution is in our approach to contamination
estimation. Specifically, we develop a technique that identifies the minimum
number of data points that must be discarded (i.e., the level of contamination)
from an empirical data set in order to match the model to within a specified
goodness-of-fit, controlled by a p-value. Appealing to results from large
deviations theory, we show a lower bound on the level of contamination is
obtained by solving a series of convex programs. Theoretical results guarantee
the bound converges at a rate of $O(\sqrt{\log(p)/p})$, where p is the size of
the empirical data set.
| no_new_dataset | 0.947624 |
1506.04359 | Yunwen Lei | Yunwen Lei and \"Ur\"un Dogan and Alexander Binder and Marius Kloft | Multi-class SVMs: From Tighter Data-Dependent Generalization Bounds to
Novel Algorithms | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper studies the generalization performance of multi-class
classification algorithms, for which we obtain, for the first time, a
data-dependent generalization error bound with a logarithmic dependence on the
class size, substantially improving the state-of-the-art linear dependence in
the existing data-dependent generalization analysis. The theoretical analysis
motivates us to introduce a new multi-class classification machine based on
$\ell_p$-norm regularization, where the parameter $p$ controls the complexity
of the corresponding bounds. We derive an efficient optimization algorithm
based on Fenchel duality theory. Benchmarks on several real-world datasets show
that the proposed algorithm can achieve significant accuracy gains over the
state of the art.
| [
{
"version": "v1",
"created": "Sun, 14 Jun 2015 08:07:23 GMT"
}
] | 2015-06-16T00:00:00 | [
[
"Lei",
"Yunwen",
""
],
[
"Dogan",
"Ürün",
""
],
[
"Binder",
"Alexander",
""
],
[
"Kloft",
"Marius",
""
]
] | TITLE: Multi-class SVMs: From Tighter Data-Dependent Generalization Bounds to
Novel Algorithms
ABSTRACT: This paper studies the generalization performance of multi-class
classification algorithms, for which we obtain, for the first time, a
data-dependent generalization error bound with a logarithmic dependence on the
class size, substantially improving the state-of-the-art linear dependence in
the existing data-dependent generalization analysis. The theoretical analysis
motivates us to introduce a new multi-class classification machine based on
$\ell_p$-norm regularization, where the parameter $p$ controls the complexity
of the corresponding bounds. We derive an efficient optimization algorithm
based on Fenchel duality theory. Benchmarks on several real-world datasets show
that the proposed algorithm can achieve significant accuracy gains over the
state of the art.
| no_new_dataset | 0.946892 |
1506.04608 | Daja Abdul | Javairia Nazir, Mehreen Sirshar | Flow Segmentation in Dense Crowds | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A framework is proposed in this paper that is used to segment flow of dense
crowds. The flow field that is generated by the movement in the crowd is
treated just like an aperiodic dynamic system. On this flow field a grid of
particles is put over for particle advection by the use of a numerical
integration scheme. Then flow maps are generated which associates the initial
position of the particles with final position. The gradient of the flow maps
gives the amount of divergence of the neighboring particles. For forward
integration and analysis forward Finite time Lyapunov Exponent is calculated
and backward Finite time Lyapunov Exponent is also calculated it gives the
Lagrangian coherent structures of the flow in crowd. Lagrangian Coherent
Structures basically divides the flow in crowd into regions and these regions
have different dynamics. These regions are then used to get the boundary in the
different flow segments by using water shed algorithm. The experiment is
conducted on the crowd dataset of UCF (University of central Florida).
| [
{
"version": "v1",
"created": "Mon, 15 Jun 2015 14:14:20 GMT"
}
] | 2015-06-16T00:00:00 | [
[
"Nazir",
"Javairia",
""
],
[
"Sirshar",
"Mehreen",
""
]
] | TITLE: Flow Segmentation in Dense Crowds
ABSTRACT: A framework is proposed in this paper that is used to segment flow of dense
crowds. The flow field that is generated by the movement in the crowd is
treated just like an aperiodic dynamic system. On this flow field a grid of
particles is put over for particle advection by the use of a numerical
integration scheme. Then flow maps are generated which associates the initial
position of the particles with final position. The gradient of the flow maps
gives the amount of divergence of the neighboring particles. For forward
integration and analysis forward Finite time Lyapunov Exponent is calculated
and backward Finite time Lyapunov Exponent is also calculated it gives the
Lagrangian coherent structures of the flow in crowd. Lagrangian Coherent
Structures basically divides the flow in crowd into regions and these regions
have different dynamics. These regions are then used to get the boundary in the
different flow segments by using water shed algorithm. The experiment is
conducted on the crowd dataset of UCF (University of central Florida).
| no_new_dataset | 0.954351 |
1506.04693 | Vincent Labatut | G\"unce Orman (BIT Lab), Vincent Labatut (LIA), Marc Plantevit
(LIRIS), Jean-Fran\c{c}ois Boulicaut (LIRIS) | Interpreting communities based on the evolution of a dynamic attributed
network | null | Social Network Analysis and Mining Journal (SNAM), 2015, 5, pp.20.
\<http://link.springer.com/article/10.1007%2Fs13278-015-0262-4\>.
\<10.1007/s13278-015-0262-4\> | 10.1007/s13278-015-0262-4 | null | cs.SI physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Many methods have been proposed to detect communities, not only in plain, but
also in attributed, directed or even dynamic complex networks. From the
modeling point of view, to be of some utility, the community structure must be
characterized relatively to the properties of the studied system. However, most
of the existing works focus on the detection of communities, and only very few
try to tackle this interpretation problem. Moreover, the existing approaches
are limited either by the type of data they handle, or by the nature of the
results they output. In this work, we see the interpretation of communities as
a problem independent from the detection process, consisting in identifying the
most characteristic features of communities. We give a formal definition of
this problem and propose a method to solve it. To this aim, we first define a
sequence-based representation of networks, combining temporal information,
community structure, topological measures, and nodal attributes. We then
describe how to identify the most emerging sequential patterns of this dataset,
and use them to characterize the communities. We study the performance of our
method on artificially generated dynamic attributed networks. We also
empirically validate our framework on real-world systems: a DBLP network of
scientific collaborations, and a LastFM network of social and musical
interactions.
| [
{
"version": "v1",
"created": "Mon, 15 Jun 2015 18:22:38 GMT"
}
] | 2015-06-16T00:00:00 | [
[
"Orman",
"Günce",
"",
"BIT Lab"
],
[
"Labatut",
"Vincent",
"",
"LIA"
],
[
"Plantevit",
"Marc",
"",
"LIRIS"
],
[
"Boulicaut",
"Jean-François",
"",
"LIRIS"
]
] | TITLE: Interpreting communities based on the evolution of a dynamic attributed
network
ABSTRACT: Many methods have been proposed to detect communities, not only in plain, but
also in attributed, directed or even dynamic complex networks. From the
modeling point of view, to be of some utility, the community structure must be
characterized relatively to the properties of the studied system. However, most
of the existing works focus on the detection of communities, and only very few
try to tackle this interpretation problem. Moreover, the existing approaches
are limited either by the type of data they handle, or by the nature of the
results they output. In this work, we see the interpretation of communities as
a problem independent from the detection process, consisting in identifying the
most characteristic features of communities. We give a formal definition of
this problem and propose a method to solve it. To this aim, we first define a
sequence-based representation of networks, combining temporal information,
community structure, topological measures, and nodal attributes. We then
describe how to identify the most emerging sequential patterns of this dataset,
and use them to characterize the communities. We study the performance of our
method on artificially generated dynamic attributed networks. We also
empirically validate our framework on real-world systems: a DBLP network of
scientific collaborations, and a LastFM network of social and musical
interactions.
| no_new_dataset | 0.939192 |
1506.04720 | Siqi Nie | Siqi Nie, Qiang Ji | Latent Regression Bayesian Network for Data Representation | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep directed generative models have attracted much attention recently due to
their expressive representation power and the ability of ancestral sampling.
One major difficulty of learning directed models with many latent variables is
the intractable inference. To address this problem, most existing algorithms
make assumptions to render the latent variables independent of each other,
either by designing specific priors, or by approximating the true posterior
using a factorized distribution. We believe the correlations among latent
variables are crucial for faithful data representation. Driven by this idea, we
propose an inference method based on the conditional pseudo-likelihood that
preserves the dependencies among the latent variables. For learning, we propose
to employ the hard Expectation Maximization (EM) algorithm, which avoids the
intractability of the traditional EM by max-out instead of sum-out to compute
the data likelihood. Qualitative and quantitative evaluations of our model
against state of the art deep models on benchmark datasets demonstrate the
effectiveness of the proposed algorithm in data representation and
reconstruction.
| [
{
"version": "v1",
"created": "Mon, 15 Jun 2015 19:34:59 GMT"
}
] | 2015-06-16T00:00:00 | [
[
"Nie",
"Siqi",
""
],
[
"Ji",
"Qiang",
""
]
] | TITLE: Latent Regression Bayesian Network for Data Representation
ABSTRACT: Deep directed generative models have attracted much attention recently due to
their expressive representation power and the ability of ancestral sampling.
One major difficulty of learning directed models with many latent variables is
the intractable inference. To address this problem, most existing algorithms
make assumptions to render the latent variables independent of each other,
either by designing specific priors, or by approximating the true posterior
using a factorized distribution. We believe the correlations among latent
variables are crucial for faithful data representation. Driven by this idea, we
propose an inference method based on the conditional pseudo-likelihood that
preserves the dependencies among the latent variables. For learning, we propose
to employ the hard Expectation Maximization (EM) algorithm, which avoids the
intractability of the traditional EM by max-out instead of sum-out to compute
the data likelihood. Qualitative and quantitative evaluations of our model
against state of the art deep models on benchmark datasets demonstrate the
effectiveness of the proposed algorithm in data representation and
reconstruction.
| no_new_dataset | 0.946843 |
1303.5577 | Ilaria Ermolli | I. Ermolli, K. Matthes, T. Dudok de Wit, N. A. Krivova, K. Tourpali,
M. Weber, Y. C. Unruh, L. Gray, U. Langematz, P. Pilewskie, E. Rozanov, W.
Schmutz, A. Shapiro, S. K. Solanki, and T. N. Woods | Recent variability of the solar spectral irradiance and its impact on
climate modelling | 34 pages, 12 figures, accepted for publication in ACP | null | 10.5194/acp-13-3945-2013 | null | astro-ph.SR physics.ao-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The lack of long and reliable time series of solar spectral irradiance (SSI)
measurements makes an accurate quantification of solar contributions to recent
climate change difficult. Whereas earlier SSI observations and models provided
a qualitatively consistent picture of the SSI variability, recent measurements
by the SORCE satellite suggest a significantly stronger variability in the
ultraviolet (UV) spectral range and changes in the visible and near-infrared
(NIR) bands in anti-phase with the solar cycle. A number of recent
chemistry-climate model (CCM) simulations have shown that this might have
significant implications on the Earth's atmosphere. Motivated by these results,
we summarize here our current knowledge of SSI variability and its impact on
Earth's climate. We present a detailed overview of existing SSI measurements
and provide thorough comparison of models available to date. SSI changes
influence the Earth's atmosphere, both directly, through changes in shortwave
(SW) heating and therefore, temperature and ozone distributions in the
stratosphere, and indirectly, through dynamical feedbacks. We investigate these
direct and indirect effects using several state-of-the art CCM simulations
forced with measured and modeled SSI changes. A unique asset of this study is
the use of a common comprehensive approach for an issue that is usually
addressed separately by different communities. Omissis. Finally, we discuss the
reliability of the available data and we propose additional coordinated work,
first to build composite SSI datasets out of scattered observations and to
refine current SSI models, and second, to run coordinated CCM experiments.
| [
{
"version": "v1",
"created": "Fri, 22 Mar 2013 10:51:01 GMT"
}
] | 2015-06-15T00:00:00 | [
[
"Ermolli",
"I.",
""
],
[
"Matthes",
"K.",
""
],
[
"de Wit",
"T. Dudok",
""
],
[
"Krivova",
"N. A.",
""
],
[
"Tourpali",
"K.",
""
],
[
"Weber",
"M.",
""
],
[
"Unruh",
"Y. C.",
""
],
[
"Gray",
"L.",
""
],
[
"Langematz",
"U.",
""
],
[
"Pilewskie",
"P.",
""
],
[
"Rozanov",
"E.",
""
],
[
"Schmutz",
"W.",
""
],
[
"Shapiro",
"A.",
""
],
[
"Solanki",
"S. K.",
""
],
[
"Woods",
"T. N.",
""
]
] | TITLE: Recent variability of the solar spectral irradiance and its impact on
climate modelling
ABSTRACT: The lack of long and reliable time series of solar spectral irradiance (SSI)
measurements makes an accurate quantification of solar contributions to recent
climate change difficult. Whereas earlier SSI observations and models provided
a qualitatively consistent picture of the SSI variability, recent measurements
by the SORCE satellite suggest a significantly stronger variability in the
ultraviolet (UV) spectral range and changes in the visible and near-infrared
(NIR) bands in anti-phase with the solar cycle. A number of recent
chemistry-climate model (CCM) simulations have shown that this might have
significant implications on the Earth's atmosphere. Motivated by these results,
we summarize here our current knowledge of SSI variability and its impact on
Earth's climate. We present a detailed overview of existing SSI measurements
and provide thorough comparison of models available to date. SSI changes
influence the Earth's atmosphere, both directly, through changes in shortwave
(SW) heating and therefore, temperature and ozone distributions in the
stratosphere, and indirectly, through dynamical feedbacks. We investigate these
direct and indirect effects using several state-of-the art CCM simulations
forced with measured and modeled SSI changes. A unique asset of this study is
the use of a common comprehensive approach for an issue that is usually
addressed separately by different communities. Omissis. Finally, we discuss the
reliability of the available data and we propose additional coordinated work,
first to build composite SSI datasets out of scattered observations and to
refine current SSI models, and second, to run coordinated CCM experiments.
| no_new_dataset | 0.947866 |
1303.6170 | Brandon Jones | Brandon Jones, Mark Campbell, Lang Tong | Maximum Likelihood Fusion of Stochastic Maps | 10 pages, 8 figures, submitted to IEEE Transactions on Signal
Processing on 24-March-2013 | null | 10.1109/TSP.2014.2304435 | null | stat.AP cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The fusion of independently obtained stochastic maps by collaborating mobile
agents is considered. The proposed approach includes two parts: matching of
stochastic maps and maximum likelihood alignment. In particular, an affine
invariant hypergraph is constructed for each stochastic map, and a bipartite
matching via a linear program is used to establish landmark correspondence
between stochastic maps. A maximum likelihood alignment procedure is proposed
to determine rotation and translation between common landmarks in order to
construct a global map within a common frame of reference. A main feature of
the proposed approach is its scalability with respect to the number of
landmarks: the matching step has polynomial complexity and the maximum
likelihood alignment is obtained in closed form. Experimental validation of the
proposed fusion approach is performed using the Victoria Park benchmark
dataset.
| [
{
"version": "v1",
"created": "Mon, 25 Mar 2013 15:34:26 GMT"
}
] | 2015-06-15T00:00:00 | [
[
"Jones",
"Brandon",
""
],
[
"Campbell",
"Mark",
""
],
[
"Tong",
"Lang",
""
]
] | TITLE: Maximum Likelihood Fusion of Stochastic Maps
ABSTRACT: The fusion of independently obtained stochastic maps by collaborating mobile
agents is considered. The proposed approach includes two parts: matching of
stochastic maps and maximum likelihood alignment. In particular, an affine
invariant hypergraph is constructed for each stochastic map, and a bipartite
matching via a linear program is used to establish landmark correspondence
between stochastic maps. A maximum likelihood alignment procedure is proposed
to determine rotation and translation between common landmarks in order to
construct a global map within a common frame of reference. A main feature of
the proposed approach is its scalability with respect to the number of
landmarks: the matching step has polynomial complexity and the maximum
likelihood alignment is obtained in closed form. Experimental validation of the
proposed fusion approach is performed using the Victoria Park benchmark
dataset.
| no_new_dataset | 0.95222 |
1304.5302 | Satoshi Eguchi | Satoshi Eguchi | "Superluminal" FITS File Processing on Multiprocessors: Zero Time Endian
Conversion Technique | 25 pages, 9 figures, 12 tables, accepted for publication in PASP | null | 10.1086/671105 | null | astro-ph.IM cs.PF | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The FITS is the standard file format in astronomy, and it has been extended
to agree with astronomical needs of the day. However, astronomical datasets
have been inflating year by year. In case of ALMA telescope, a ~ TB scale
4-dimensional data cube may be produced for one target. Considering that
typical Internet bandwidth is a few 10 MB/s at most, the original data cubes in
FITS format are hosted on a VO server, and the region which a user is
interested in should be cut out and transferred to the user (Eguchi et al.,
2012). The system will equip a very high-speed disk array to process a TB scale
data cube in a few 10 seconds, and disk I/O speed, endian conversion and data
processing one will be comparable. Hence to reduce the endian conversion time
is one of issues to realize our system. In this paper, I introduce a technique
named "just-in-time endian conversion", which delays the endian conversion for
each pixel just before it is really needed, to sweep out the endian conversion
time; by applying this method, the FITS processing speed increases 20% for
single threading, and 40% for multi-threading compared to CFITSIO. The speed-up
by the method tightly relates to modern CPU architecture to improve the
efficiency of instruction pipelines due to break of "causality", a programmed
instruction code sequence.
| [
{
"version": "v1",
"created": "Fri, 19 Apr 2013 03:29:36 GMT"
}
] | 2015-06-15T00:00:00 | [
[
"Eguchi",
"Satoshi",
""
]
] | TITLE: "Superluminal" FITS File Processing on Multiprocessors: Zero Time Endian
Conversion Technique
ABSTRACT: The FITS is the standard file format in astronomy, and it has been extended
to agree with astronomical needs of the day. However, astronomical datasets
have been inflating year by year. In case of ALMA telescope, a ~ TB scale
4-dimensional data cube may be produced for one target. Considering that
typical Internet bandwidth is a few 10 MB/s at most, the original data cubes in
FITS format are hosted on a VO server, and the region which a user is
interested in should be cut out and transferred to the user (Eguchi et al.,
2012). The system will equip a very high-speed disk array to process a TB scale
data cube in a few 10 seconds, and disk I/O speed, endian conversion and data
processing one will be comparable. Hence to reduce the endian conversion time
is one of issues to realize our system. In this paper, I introduce a technique
named "just-in-time endian conversion", which delays the endian conversion for
each pixel just before it is really needed, to sweep out the endian conversion
time; by applying this method, the FITS processing speed increases 20% for
single threading, and 40% for multi-threading compared to CFITSIO. The speed-up
by the method tightly relates to modern CPU architecture to improve the
efficiency of instruction pipelines due to break of "causality", a programmed
instruction code sequence.
| no_new_dataset | 0.949809 |
1305.3532 | Alain Barrat | Alain Barrat, Ciro Cattuto | Temporal networks of face-to-face human interactions | Chapter of the book "Temporal Networks", Springer, 2013. Series:
Understanding Complex Systems. Holme, Petter; Saram\"aki, Jari (Eds.) | null | 10.1007/978-3-642-36461-7_10 | null | physics.soc-ph cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The ever increasing adoption of mobile technologies and ubiquitous services
allows to sense human behavior at unprecedented levels of details and scale.
Wearable sensors are opening up a new window on human mobility and proximity at
the finest resolution of face-to-face proximity. As a consequence, empirical
data describing social and behavioral networks are acquiring a longitudinal
dimension that brings forth new challenges for analysis and modeling. Here we
review recent work on the representation and analysis of temporal networks of
face-to-face human proximity, based on large-scale datasets collected in the
context of the SocioPatterns collaboration. We show that the raw behavioral
data can be studied at various levels of coarse-graining, which turn out to be
complementary to one another, with each level exposing different features of
the underlying system. We briefly review a generative model of temporal contact
networks that reproduces some statistical observables. Then, we shift our focus
from surface statistical features to dynamical processes on empirical temporal
networks. We discuss how simple dynamical processes can be used as probes to
expose important features of the interaction patterns, such as burstiness and
causal constraints. We show that simulating dynamical processes on empirical
temporal networks can unveil differences between datasets that would otherwise
look statistically similar. Moreover, we argue that, due to the temporal
heterogeneity of human dynamics, in order to investigate the temporal
properties of spreading processes it may be necessary to abandon the notion of
wall-clock time in favour of an intrinsic notion of time for each individual
node, defined in terms of its activity level. We conclude highlighting several
open research questions raised by the nature of the data at hand.
| [
{
"version": "v1",
"created": "Wed, 15 May 2013 16:18:24 GMT"
}
] | 2015-06-15T00:00:00 | [
[
"Barrat",
"Alain",
""
],
[
"Cattuto",
"Ciro",
""
]
] | TITLE: Temporal networks of face-to-face human interactions
ABSTRACT: The ever increasing adoption of mobile technologies and ubiquitous services
allows to sense human behavior at unprecedented levels of details and scale.
Wearable sensors are opening up a new window on human mobility and proximity at
the finest resolution of face-to-face proximity. As a consequence, empirical
data describing social and behavioral networks are acquiring a longitudinal
dimension that brings forth new challenges for analysis and modeling. Here we
review recent work on the representation and analysis of temporal networks of
face-to-face human proximity, based on large-scale datasets collected in the
context of the SocioPatterns collaboration. We show that the raw behavioral
data can be studied at various levels of coarse-graining, which turn out to be
complementary to one another, with each level exposing different features of
the underlying system. We briefly review a generative model of temporal contact
networks that reproduces some statistical observables. Then, we shift our focus
from surface statistical features to dynamical processes on empirical temporal
networks. We discuss how simple dynamical processes can be used as probes to
expose important features of the interaction patterns, such as burstiness and
causal constraints. We show that simulating dynamical processes on empirical
temporal networks can unveil differences between datasets that would otherwise
look statistically similar. Moreover, we argue that, due to the temporal
heterogeneity of human dynamics, in order to investigate the temporal
properties of spreading processes it may be necessary to abandon the notion of
wall-clock time in favour of an intrinsic notion of time for each individual
node, defined in terms of its activity level. We conclude highlighting several
open research questions raised by the nature of the data at hand.
| no_new_dataset | 0.946001 |
1412.3421 | Juan Eugenio Iglesias | Juan Eugenio Iglesias and Mert Rory Sabuncu | Multi-Atlas Segmentation of Biomedical Images: A Survey | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multi-atlas segmentation (MAS), first introduced and popularized by the
pioneering work of Rohlfing, Brandt, Menzel and Maurer Jr (2004), Klein, Mensh,
Ghosh, Tourville and Hirsch (2005), and Heckemann, Hajnal, Aljabar, Rueckert
and Hammers (2006), is becoming one of the most widely-used and successful
image segmentation techniques in biomedical applications. By manipulating and
utilizing the entire dataset of "atlases" (training images that have been
previously labeled, e.g., manually by an expert), rather than some model-based
average representation, MAS has the flexibility to better capture anatomical
variation, thus offering superior segmentation accuracy. This benefit, however,
typically comes at a high computational cost. Recent advancements in computer
hardware and image processing software have been instrumental in addressing
this challenge and facilitated the wide adoption of MAS. Today, MAS has come a
long way and the approach includes a wide array of sophisticated algorithms
that employ ideas from machine learning, probabilistic modeling, optimization,
and computer vision, among other fields. This paper presents a survey of
published MAS algorithms and studies that have applied these methods to various
biomedical problems. In writing this survey, we have three distinct aims. Our
primary goal is to document how MAS was originally conceived, later evolved,
and now relates to alternative methods. Second, this paper is intended to be a
detailed reference of past research activity in MAS, which now spans over a
decade (2003 - 2014) and entails novel methodological developments and
application-specific solutions. Finally, our goal is to also present a
perspective on the future of MAS, which, we believe, will be one of the
dominant approaches in biomedical image segmentation.
| [
{
"version": "v1",
"created": "Wed, 10 Dec 2014 19:28:09 GMT"
},
{
"version": "v2",
"created": "Fri, 12 Jun 2015 14:35:30 GMT"
}
] | 2015-06-15T00:00:00 | [
[
"Iglesias",
"Juan Eugenio",
""
],
[
"Sabuncu",
"Mert Rory",
""
]
] | TITLE: Multi-Atlas Segmentation of Biomedical Images: A Survey
ABSTRACT: Multi-atlas segmentation (MAS), first introduced and popularized by the
pioneering work of Rohlfing, Brandt, Menzel and Maurer Jr (2004), Klein, Mensh,
Ghosh, Tourville and Hirsch (2005), and Heckemann, Hajnal, Aljabar, Rueckert
and Hammers (2006), is becoming one of the most widely-used and successful
image segmentation techniques in biomedical applications. By manipulating and
utilizing the entire dataset of "atlases" (training images that have been
previously labeled, e.g., manually by an expert), rather than some model-based
average representation, MAS has the flexibility to better capture anatomical
variation, thus offering superior segmentation accuracy. This benefit, however,
typically comes at a high computational cost. Recent advancements in computer
hardware and image processing software have been instrumental in addressing
this challenge and facilitated the wide adoption of MAS. Today, MAS has come a
long way and the approach includes a wide array of sophisticated algorithms
that employ ideas from machine learning, probabilistic modeling, optimization,
and computer vision, among other fields. This paper presents a survey of
published MAS algorithms and studies that have applied these methods to various
biomedical problems. In writing this survey, we have three distinct aims. Our
primary goal is to document how MAS was originally conceived, later evolved,
and now relates to alternative methods. Second, this paper is intended to be a
detailed reference of past research activity in MAS, which now spans over a
decade (2003 - 2014) and entails novel methodological developments and
application-specific solutions. Finally, our goal is to also present a
perspective on the future of MAS, which, we believe, will be one of the
dominant approaches in biomedical image segmentation.
| no_new_dataset | 0.945801 |
1506.00815 | Yuhuang Hu | Yuhuang Hu, M.S. Ishwarya, Chu Kiong Loo | Classify Images with Conceptor Network | This paper has been withdrawn by the author due to a crucial sign
error in experiments | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This article demonstrates a new conceptor network based classifier in
classifying images. Mathematical descriptions and analysis are presented.
Various tests are experimented using three benchmark datasets: MNIST, CIFAR-10
and CIFAR-100. The experiments displayed that conceptor network can offer
superior results and flexible configurations than conventional classifiers such
as Softmax Regression and Support Vector Machine (SVM).
| [
{
"version": "v1",
"created": "Tue, 2 Jun 2015 09:49:45 GMT"
},
{
"version": "v2",
"created": "Wed, 3 Jun 2015 13:57:14 GMT"
},
{
"version": "v3",
"created": "Sat, 6 Jun 2015 16:58:41 GMT"
},
{
"version": "v4",
"created": "Fri, 12 Jun 2015 01:13:06 GMT"
}
] | 2015-06-15T00:00:00 | [
[
"Hu",
"Yuhuang",
""
],
[
"Ishwarya",
"M. S.",
""
],
[
"Loo",
"Chu Kiong",
""
]
] | TITLE: Classify Images with Conceptor Network
ABSTRACT: This article demonstrates a new conceptor network based classifier in
classifying images. Mathematical descriptions and analysis are presented.
Various tests are experimented using three benchmark datasets: MNIST, CIFAR-10
and CIFAR-100. The experiments displayed that conceptor network can offer
superior results and flexible configurations than conventional classifiers such
as Softmax Regression and Support Vector Machine (SVM).
| no_new_dataset | 0.952175 |
1506.03837 | Weinan Zhang | Weinan Zhang, Jun Wang | Statistical Arbitrage Mining for Display Advertising | In the proceedings of the 21st ACM SIGKDD international conference on
Knowledge discovery and data mining (KDD 2015) | null | 10.1145/2783258.2783269 | null | cs.GT cs.DB | http://creativecommons.org/licenses/publicdomain/ | We study and formulate arbitrage in display advertising. Real-Time Bidding
(RTB) mimics stock spot exchanges and utilises computers to algorithmically buy
display ads per impression via a real-time auction. Despite the new automation,
the ad markets are still informationally inefficient due to the heavily
fragmented marketplaces. Two display impressions with similar or identical
effectiveness (e.g., measured by conversion or click-through rates for a
targeted audience) may sell for quite different prices at different market
segments or pricing schemes. In this paper, we propose a novel data mining
paradigm called Statistical Arbitrage Mining (SAM) focusing on mining and
exploiting price discrepancies between two pricing schemes. In essence, our
SAMer is a meta-bidder that hedges advertisers' risk between CPA (cost per
action)-based campaigns and CPM (cost per mille impressions)-based ad
inventories; it statistically assesses the potential profit and cost for an
incoming CPM bid request against a portfolio of CPA campaigns based on the
estimated conversion rate, bid landscape and other statistics learned from
historical data. In SAM, (i) functional optimisation is utilised to seek for
optimal bidding to maximise the expected arbitrage net profit, and (ii) a
portfolio-based risk management solution is leveraged to reallocate bid volume
and budget across the set of campaigns to make a risk and return trade-off. We
propose to jointly optimise both components in an EM fashion with high
efficiency to help the meta-bidder successfully catch the transient statistical
arbitrage opportunities in RTB. Both the offline experiments on a real-world
large-scale dataset and online A/B tests on a commercial platform demonstrate
the effectiveness of our proposed solution in exploiting arbitrage in various
model settings and market environments.
| [
{
"version": "v1",
"created": "Thu, 11 Jun 2015 21:05:26 GMT"
}
] | 2015-06-15T00:00:00 | [
[
"Zhang",
"Weinan",
""
],
[
"Wang",
"Jun",
""
]
] | TITLE: Statistical Arbitrage Mining for Display Advertising
ABSTRACT: We study and formulate arbitrage in display advertising. Real-Time Bidding
(RTB) mimics stock spot exchanges and utilises computers to algorithmically buy
display ads per impression via a real-time auction. Despite the new automation,
the ad markets are still informationally inefficient due to the heavily
fragmented marketplaces. Two display impressions with similar or identical
effectiveness (e.g., measured by conversion or click-through rates for a
targeted audience) may sell for quite different prices at different market
segments or pricing schemes. In this paper, we propose a novel data mining
paradigm called Statistical Arbitrage Mining (SAM) focusing on mining and
exploiting price discrepancies between two pricing schemes. In essence, our
SAMer is a meta-bidder that hedges advertisers' risk between CPA (cost per
action)-based campaigns and CPM (cost per mille impressions)-based ad
inventories; it statistically assesses the potential profit and cost for an
incoming CPM bid request against a portfolio of CPA campaigns based on the
estimated conversion rate, bid landscape and other statistics learned from
historical data. In SAM, (i) functional optimisation is utilised to seek for
optimal bidding to maximise the expected arbitrage net profit, and (ii) a
portfolio-based risk management solution is leveraged to reallocate bid volume
and budget across the set of campaigns to make a risk and return trade-off. We
propose to jointly optimise both components in an EM fashion with high
efficiency to help the meta-bidder successfully catch the transient statistical
arbitrage opportunities in RTB. Both the offline experiments on a real-world
large-scale dataset and online A/B tests on a commercial platform demonstrate
the effectiveness of our proposed solution in exploiting arbitrage in various
model settings and market environments.
| no_new_dataset | 0.943608 |
1506.04046 | Jason Byrne | Jason P. Byrne | Investigating the Kinematics of Coronal Mass Ejections with the
Automated CORIMP Catalog | 23 pages, 11 figures, 1 table | null | null | null | astro-ph.SR astro-ph.EP physics.data-an physics.space-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Studying coronal mass ejections (CMEs) in coronagraph data can be challenging
due to their diffuse structure and transient nature, compounded by the
variations in their dynamics, morphology, and frequency of occurrence. The
large amounts of data available from missions like the Solar and Heliospheric
Observatory (SOHO) make manual cataloging of CMEs tedious and prone to human
error, and so a robust method of detection and analysis is required and often
preferred. A new coronal image processing catalog called CORIMP has been
developed in an effort to achieve this, through the implementation of a dynamic
background separation technique and multiscale edge detection. These algorithms
together isolate and characterise CME structure in the field-of-view of the
Large Angle Spectrometric Coronagraph (LASCO) onboard SOHO. CORIMP also applies
a Savitzky-Golay filter, along with quadratic and linear fits, to the
height-time measurements for better revealing the true CME speed and
acceleration profiles across the plane-of-sky. Here we present a sample of new
results from the CORIMP CME catalog, and directly compare them with the other
automated catalogs of Computer Aided CME Tracking (CACTus) and Solar Eruptive
Events Detection System (SEEDS), as well as the manual CME catalog at the
Coordinated Data Analysis Workshop (CDAW) Data Center and a previously
published study of the sample events. We further investigate a form of
unsupervised machine learning by using a k-means clustering algorithm to
distinguish detections of multiple CMEs that occur close together in space and
time. While challenges still exist, this investigation and comparison of
results demonstrates the reliability and robustness of the CORIMP catalog,
proving its effectiveness at detecting and tracking CMEs throughout the LASCO
dataset.
| [
{
"version": "v1",
"created": "Fri, 12 Jun 2015 15:39:27 GMT"
}
] | 2015-06-15T00:00:00 | [
[
"Byrne",
"Jason P.",
""
]
] | TITLE: Investigating the Kinematics of Coronal Mass Ejections with the
Automated CORIMP Catalog
ABSTRACT: Studying coronal mass ejections (CMEs) in coronagraph data can be challenging
due to their diffuse structure and transient nature, compounded by the
variations in their dynamics, morphology, and frequency of occurrence. The
large amounts of data available from missions like the Solar and Heliospheric
Observatory (SOHO) make manual cataloging of CMEs tedious and prone to human
error, and so a robust method of detection and analysis is required and often
preferred. A new coronal image processing catalog called CORIMP has been
developed in an effort to achieve this, through the implementation of a dynamic
background separation technique and multiscale edge detection. These algorithms
together isolate and characterise CME structure in the field-of-view of the
Large Angle Spectrometric Coronagraph (LASCO) onboard SOHO. CORIMP also applies
a Savitzky-Golay filter, along with quadratic and linear fits, to the
height-time measurements for better revealing the true CME speed and
acceleration profiles across the plane-of-sky. Here we present a sample of new
results from the CORIMP CME catalog, and directly compare them with the other
automated catalogs of Computer Aided CME Tracking (CACTus) and Solar Eruptive
Events Detection System (SEEDS), as well as the manual CME catalog at the
Coordinated Data Analysis Workshop (CDAW) Data Center and a previously
published study of the sample events. We further investigate a form of
unsupervised machine learning by using a k-means clustering algorithm to
distinguish detections of multiple CMEs that occur close together in space and
time. While challenges still exist, this investigation and comparison of
results demonstrates the reliability and robustness of the CORIMP catalog,
proving its effectiveness at detecting and tracking CMEs throughout the LASCO
dataset.
| no_new_dataset | 0.949856 |
1506.04051 | Lucia Maddalena | Lucia Maddalena and Alfredo Petrosino | Towards Benchmarking Scene Background Initialization | 6 pages, SBI dataset, SBMI2015 Workshop | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/3.0/ | Given a set of images of a scene taken at different times, the availability
of an initial background model that describes the scene without foreground
objects is the prerequisite for a wide range of applications, ranging from
video surveillance to computational photography. Even though several methods
have been proposed for scene background initialization, the lack of a common
groundtruthed dataset and of a common set of metrics makes it difficult to
compare their performance. To move first steps towards an easy and fair
comparison of these methods, we assembled a dataset of sequences frequently
adopted for background initialization, selected or created ground truths for
quantitative evaluation through a selected suite of metrics, and compared
results obtained by some existing methods, making all the material publicly
available.
| [
{
"version": "v1",
"created": "Fri, 12 Jun 2015 15:52:46 GMT"
}
] | 2015-06-15T00:00:00 | [
[
"Maddalena",
"Lucia",
""
],
[
"Petrosino",
"Alfredo",
""
]
] | TITLE: Towards Benchmarking Scene Background Initialization
ABSTRACT: Given a set of images of a scene taken at different times, the availability
of an initial background model that describes the scene without foreground
objects is the prerequisite for a wide range of applications, ranging from
video surveillance to computational photography. Even though several methods
have been proposed for scene background initialization, the lack of a common
groundtruthed dataset and of a common set of metrics makes it difficult to
compare their performance. To move first steps towards an easy and fair
comparison of these methods, we assembled a dataset of sequences frequently
adopted for background initialization, selected or created ground truths for
quantitative evaluation through a selected suite of metrics, and compared
results obtained by some existing methods, making all the material publicly
available.
| new_dataset | 0.955527 |
1211.1073 | Venkat Chandrasekaran | Venkat Chandrasekaran and Michael I. Jordan | Computational and Statistical Tradeoffs via Convex Relaxation | null | null | 10.1073/pnas.1302293110 | null | math.ST cs.IT math.IT math.OC stat.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In modern data analysis, one is frequently faced with statistical inference
problems involving massive datasets. Processing such large datasets is usually
viewed as a substantial computational challenge. However, if data are a
statistician's main resource then access to more data should be viewed as an
asset rather than as a burden. In this paper we describe a computational
framework based on convex relaxation to reduce the computational complexity of
an inference procedure when one has access to increasingly larger datasets.
Convex relaxation techniques have been widely used in theoretical computer
science as they give tractable approximation algorithms to many computationally
intractable tasks. We demonstrate the efficacy of this methodology in
statistical estimation in providing concrete time-data tradeoffs in a class of
denoising problems. Thus, convex relaxation offers a principled approach to
exploit the statistical gains from larger datasets to reduce the runtime of
inference algorithms.
| [
{
"version": "v1",
"created": "Mon, 5 Nov 2012 23:28:44 GMT"
},
{
"version": "v2",
"created": "Mon, 26 Nov 2012 22:02:27 GMT"
}
] | 2015-06-12T00:00:00 | [
[
"Chandrasekaran",
"Venkat",
""
],
[
"Jordan",
"Michael I.",
""
]
] | TITLE: Computational and Statistical Tradeoffs via Convex Relaxation
ABSTRACT: In modern data analysis, one is frequently faced with statistical inference
problems involving massive datasets. Processing such large datasets is usually
viewed as a substantial computational challenge. However, if data are a
statistician's main resource then access to more data should be viewed as an
asset rather than as a burden. In this paper we describe a computational
framework based on convex relaxation to reduce the computational complexity of
an inference procedure when one has access to increasingly larger datasets.
Convex relaxation techniques have been widely used in theoretical computer
science as they give tractable approximation algorithms to many computationally
intractable tasks. We demonstrate the efficacy of this methodology in
statistical estimation in providing concrete time-data tradeoffs in a class of
denoising problems. Thus, convex relaxation offers a principled approach to
exploit the statistical gains from larger datasets to reduce the runtime of
inference algorithms.
| no_new_dataset | 0.946448 |
1211.6688 | Jaroslav Hlinka | Jaroslav Hlinka, David Hartman, Martin Vejmelka, Dagmar Novotn\'a,
Milan Palu\v{s} | Non-linear dependence and teleconnections in climate data: sources,
relevance, nonstationarity | null | null | 10.1007/s00382-013-1780-2 | null | stat.ME physics.ao-ph physics.data-an | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Quantification of relations between measured variables of interest by
statistical measures of dependence is a common step in analysis of climate
data. The term "connectivity" is used in the network context including the
study of complex coupled dynamical systems. The choice of dependence measure is
key for the results of the subsequent analysis and interpretation. The use of
linear Pearson's correlation coefficient is widespread and convenient. On the
other side, as the climate is widely acknowledged to be a nonlinear system,
nonlinear connectivity quantification methods, such as those based on
information-theoretical concepts, are increasingly used for this purpose.
In this paper we outline an approach that enables well informed choice of
connectivity method for a given type of data, improving the subsequent
interpretation of the results. The presented multi-step approach includes
statistical testing, quantification of the specific non-linear contribution to
the interaction information, localization of nodes with strongest nonlinear
contribution and assessment of the role of specific temporal patterns,
including signal nonstationarities. In detail we study the consequences of the
choice of a general nonlinear connectivity measure, namely mutual information,
focusing on its relevance and potential alterations in the discovered
dependence structure.
We document the method by applying it on monthly mean temperature data from
the NCEP/NCAR reanalysis dataset as well as the ERA dataset. We have been able
to identify main sources of observed non-linearity in inter-node couplings.
Detailed analysis suggested an important role of several sources of
nonstationarity within the climate data. The quantitative role of genuine
nonlinear coupling at this scale has proven to be almost negligible, providing
quantitative support for the use of linear methods for this type of data.
| [
{
"version": "v1",
"created": "Wed, 28 Nov 2012 18:06:06 GMT"
}
] | 2015-06-12T00:00:00 | [
[
"Hlinka",
"Jaroslav",
""
],
[
"Hartman",
"David",
""
],
[
"Vejmelka",
"Martin",
""
],
[
"Novotná",
"Dagmar",
""
],
[
"Paluš",
"Milan",
""
]
] | TITLE: Non-linear dependence and teleconnections in climate data: sources,
relevance, nonstationarity
ABSTRACT: Quantification of relations between measured variables of interest by
statistical measures of dependence is a common step in analysis of climate
data. The term "connectivity" is used in the network context including the
study of complex coupled dynamical systems. The choice of dependence measure is
key for the results of the subsequent analysis and interpretation. The use of
linear Pearson's correlation coefficient is widespread and convenient. On the
other side, as the climate is widely acknowledged to be a nonlinear system,
nonlinear connectivity quantification methods, such as those based on
information-theoretical concepts, are increasingly used for this purpose.
In this paper we outline an approach that enables well informed choice of
connectivity method for a given type of data, improving the subsequent
interpretation of the results. The presented multi-step approach includes
statistical testing, quantification of the specific non-linear contribution to
the interaction information, localization of nodes with strongest nonlinear
contribution and assessment of the role of specific temporal patterns,
including signal nonstationarities. In detail we study the consequences of the
choice of a general nonlinear connectivity measure, namely mutual information,
focusing on its relevance and potential alterations in the discovered
dependence structure.
We document the method by applying it on monthly mean temperature data from
the NCEP/NCAR reanalysis dataset as well as the ERA dataset. We have been able
to identify main sources of observed non-linearity in inter-node couplings.
Detailed analysis suggested an important role of several sources of
nonstationarity within the climate data. The quantitative role of genuine
nonlinear coupling at this scale has proven to be almost negligible, providing
quantitative support for the use of linear methods for this type of data.
| no_new_dataset | 0.948632 |
1212.3333 | Ralf Kaehler | Ralf Kaehler, Tom Abel | Single-Pass GPU-Raycasting for Structured Adaptive Mesh Refinement Data | 12 pages, 7 figures. submitted to Visualization and Data Analysis
2013 | null | 10.1117/12.2008552 | null | astro-ph.IM cs.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Structured Adaptive Mesh Refinement (SAMR) is a popular numerical technique
to study processes with high spatial and temporal dynamic range. It reduces
computational requirements by adapting the lattice on which the underlying
differential equations are solved to most efficiently represent the solution.
Particularly in astrophysics and cosmology such simulations now can capture
spatial scales ten orders of magnitude apart and more. The irregular locations
and extensions of the refined regions in the SAMR scheme and the fact that
different resolution levels partially overlap, poses a challenge for GPU-based
direct volume rendering methods. kD-trees have proven to be advantageous to
subdivide the data domain into non-overlapping blocks of equally sized cells,
optimal for the texture units of current graphics hardware, but previous
GPU-supported raycasting approaches for SAMR data using this data structure
required a separate rendering pass for each node, preventing the application of
many advanced lighting schemes that require simultaneous access to more than
one block of cells. In this paper we present a single-pass GPU-raycasting
algorithm for SAMR data that is based on a kD-tree. The tree is efficiently
encoded by a set of 3D-textures, which allows to adaptively sample complete
rays entirely on the GPU without any CPU interaction. We discuss two different
data storage strategies to access the grid data on the GPU and apply them to
several datasets to prove the benefits of the proposed method.
| [
{
"version": "v1",
"created": "Thu, 13 Dec 2012 21:00:02 GMT"
}
] | 2015-06-12T00:00:00 | [
[
"Kaehler",
"Ralf",
""
],
[
"Abel",
"Tom",
""
]
] | TITLE: Single-Pass GPU-Raycasting for Structured Adaptive Mesh Refinement Data
ABSTRACT: Structured Adaptive Mesh Refinement (SAMR) is a popular numerical technique
to study processes with high spatial and temporal dynamic range. It reduces
computational requirements by adapting the lattice on which the underlying
differential equations are solved to most efficiently represent the solution.
Particularly in astrophysics and cosmology such simulations now can capture
spatial scales ten orders of magnitude apart and more. The irregular locations
and extensions of the refined regions in the SAMR scheme and the fact that
different resolution levels partially overlap, poses a challenge for GPU-based
direct volume rendering methods. kD-trees have proven to be advantageous to
subdivide the data domain into non-overlapping blocks of equally sized cells,
optimal for the texture units of current graphics hardware, but previous
GPU-supported raycasting approaches for SAMR data using this data structure
required a separate rendering pass for each node, preventing the application of
many advanced lighting schemes that require simultaneous access to more than
one block of cells. In this paper we present a single-pass GPU-raycasting
algorithm for SAMR data that is based on a kD-tree. The tree is efficiently
encoded by a set of 3D-textures, which allows to adaptively sample complete
rays entirely on the GPU without any CPU interaction. We discuss two different
data storage strategies to access the grid data on the GPU and apply them to
several datasets to prove the benefits of the proposed method.
| no_new_dataset | 0.949482 |
1301.4546 | Akira Kageyama | Akira Kageyama and Tomoki Yamada | An Approach to Exascale Visualization: Interactive Viewing of In-Situ
Visualization | Will appear in Comput. Phys. Comm | null | 10.1016/j.cpc.2013.08.017 | null | physics.comp-ph cs.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In the coming era of exascale supercomputing, in-situ visualization will be a
crucial approach for reducing the output data size. A problem of in-situ
visualization is that it loses interactivity if a steering method is not
adopted. In this paper, we propose a new method for the interactive analysis of
in-situ visualization images produced by a batch simulation job. A key idea is
to apply numerous (thousands to millions) in-situ visualizations
simultaneously. The viewer then analyzes the image database interactively
during postprocessing. If each movie can be compressed to 100 MB, one million
movies will only require 100 TB, which is smaller than the size of the raw
numerical data in exascale supercomputing. We performed a feasibility study
using the proposed method. Multiple movie files were produced by a simulation
and they were analyzed using a specially designed movie player. The user could
change the viewing angle, the visualization method, and the parameters
interactively by retrieving an appropriate sequence of images from the movie
dataset.
| [
{
"version": "v1",
"created": "Sat, 19 Jan 2013 08:39:58 GMT"
},
{
"version": "v2",
"created": "Sat, 24 Aug 2013 00:21:14 GMT"
},
{
"version": "v3",
"created": "Fri, 13 Sep 2013 08:55:28 GMT"
}
] | 2015-06-12T00:00:00 | [
[
"Kageyama",
"Akira",
""
],
[
"Yamada",
"Tomoki",
""
]
] | TITLE: An Approach to Exascale Visualization: Interactive Viewing of In-Situ
Visualization
ABSTRACT: In the coming era of exascale supercomputing, in-situ visualization will be a
crucial approach for reducing the output data size. A problem of in-situ
visualization is that it loses interactivity if a steering method is not
adopted. In this paper, we propose a new method for the interactive analysis of
in-situ visualization images produced by a batch simulation job. A key idea is
to apply numerous (thousands to millions) in-situ visualizations
simultaneously. The viewer then analyzes the image database interactively
during postprocessing. If each movie can be compressed to 100 MB, one million
movies will only require 100 TB, which is smaller than the size of the raw
numerical data in exascale supercomputing. We performed a feasibility study
using the proposed method. Multiple movie files were produced by a simulation
and they were analyzed using a specially designed movie player. The user could
change the viewing angle, the visualization method, and the parameters
interactively by retrieving an appropriate sequence of images from the movie
dataset.
| no_new_dataset | 0.943867 |
1409.5400 | Tobias Weyand | Tobias Weyand and Bastian Leibe | Visual Landmark Recognition from Internet Photo Collections: A
Large-Scale Evaluation | null | null | 10.1016/j.cviu.2015.02.002 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The task of a visual landmark recognition system is to identify photographed
buildings or objects in query photos and to provide the user with relevant
information on them. With their increasing coverage of the world's landmark
buildings and objects, Internet photo collections are now being used as a
source for building such systems in a fully automatic fashion. This process
typically consists of three steps: clustering large amounts of images by the
objects they depict; determining object names from user-provided tags; and
building a robust, compact, and efficient recognition index. To this date,
however, there is little empirical information on how well current approaches
for those steps perform in a large-scale open-set mining and recognition task.
Furthermore, there is little empirical information on how recognition
performance varies for different types of landmark objects and where there is
still potential for improvement. With this paper, we intend to fill these gaps.
Using a dataset of 500k images from Paris, we analyze each component of the
landmark recognition pipeline in order to answer the following questions: How
many and what kinds of objects can be discovered automatically? How can we best
use the resulting image clusters to recognize the object in a query? How can
the object be efficiently represented in memory for recognition? How reliably
can semantic information be extracted? And finally: What are the limiting
factors in the resulting pipeline from query to semantics? We evaluate how
different choices of methods and parameters for the individual pipeline steps
affect overall system performance and examine their effects for different query
categories such as buildings, paintings or sculptures.
| [
{
"version": "v1",
"created": "Thu, 18 Sep 2014 18:28:20 GMT"
}
] | 2015-06-12T00:00:00 | [
[
"Weyand",
"Tobias",
""
],
[
"Leibe",
"Bastian",
""
]
] | TITLE: Visual Landmark Recognition from Internet Photo Collections: A
Large-Scale Evaluation
ABSTRACT: The task of a visual landmark recognition system is to identify photographed
buildings or objects in query photos and to provide the user with relevant
information on them. With their increasing coverage of the world's landmark
buildings and objects, Internet photo collections are now being used as a
source for building such systems in a fully automatic fashion. This process
typically consists of three steps: clustering large amounts of images by the
objects they depict; determining object names from user-provided tags; and
building a robust, compact, and efficient recognition index. To this date,
however, there is little empirical information on how well current approaches
for those steps perform in a large-scale open-set mining and recognition task.
Furthermore, there is little empirical information on how recognition
performance varies for different types of landmark objects and where there is
still potential for improvement. With this paper, we intend to fill these gaps.
Using a dataset of 500k images from Paris, we analyze each component of the
landmark recognition pipeline in order to answer the following questions: How
many and what kinds of objects can be discovered automatically? How can we best
use the resulting image clusters to recognize the object in a query? How can
the object be efficiently represented in memory for recognition? How reliably
can semantic information be extracted? And finally: What are the limiting
factors in the resulting pipeline from query to semantics? We evaluate how
different choices of methods and parameters for the individual pipeline steps
affect overall system performance and examine their effects for different query
categories such as buildings, paintings or sculptures.
| no_new_dataset | 0.929951 |
1412.6632 | Junhua Mao | Junhua Mao, Wei Xu, Yi Yang, Jiang Wang, Zhiheng Huang, Alan Yuille | Deep Captioning with Multimodal Recurrent Neural Networks (m-RNN) | Add a simple strategy to boost the performance of image captioning
task significantly. More details are shown in Section 8 of the paper. The
code and related data are available at https://github.com/mjhucla/mRNN-CR ;.
arXiv admin note: substantial text overlap with arXiv:1410.1090 | ICLR 2015 | null | null | cs.CV cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we present a multimodal Recurrent Neural Network (m-RNN) model
for generating novel image captions. It directly models the probability
distribution of generating a word given previous words and an image. Image
captions are generated by sampling from this distribution. The model consists
of two sub-networks: a deep recurrent neural network for sentences and a deep
convolutional network for images. These two sub-networks interact with each
other in a multimodal layer to form the whole m-RNN model. The effectiveness of
our model is validated on four benchmark datasets (IAPR TC-12, Flickr 8K,
Flickr 30K and MS COCO). Our model outperforms the state-of-the-art methods. In
addition, we apply the m-RNN model to retrieval tasks for retrieving images or
sentences, and achieves significant performance improvement over the
state-of-the-art methods which directly optimize the ranking objective function
for retrieval. The project page of this work is:
www.stat.ucla.edu/~junhua.mao/m-RNN.html .
| [
{
"version": "v1",
"created": "Sat, 20 Dec 2014 08:10:04 GMT"
},
{
"version": "v2",
"created": "Fri, 26 Dec 2014 08:24:11 GMT"
},
{
"version": "v3",
"created": "Tue, 10 Mar 2015 04:17:48 GMT"
},
{
"version": "v4",
"created": "Fri, 10 Apr 2015 21:03:35 GMT"
},
{
"version": "v5",
"created": "Thu, 11 Jun 2015 15:26:58 GMT"
}
] | 2015-06-12T00:00:00 | [
[
"Mao",
"Junhua",
""
],
[
"Xu",
"Wei",
""
],
[
"Yang",
"Yi",
""
],
[
"Wang",
"Jiang",
""
],
[
"Huang",
"Zhiheng",
""
],
[
"Yuille",
"Alan",
""
]
] | TITLE: Deep Captioning with Multimodal Recurrent Neural Networks (m-RNN)
ABSTRACT: In this paper, we present a multimodal Recurrent Neural Network (m-RNN) model
for generating novel image captions. It directly models the probability
distribution of generating a word given previous words and an image. Image
captions are generated by sampling from this distribution. The model consists
of two sub-networks: a deep recurrent neural network for sentences and a deep
convolutional network for images. These two sub-networks interact with each
other in a multimodal layer to form the whole m-RNN model. The effectiveness of
our model is validated on four benchmark datasets (IAPR TC-12, Flickr 8K,
Flickr 30K and MS COCO). Our model outperforms the state-of-the-art methods. In
addition, we apply the m-RNN model to retrieval tasks for retrieving images or
sentences, and achieves significant performance improvement over the
state-of-the-art methods which directly optimize the ranking objective function
for retrieval. The project page of this work is:
www.stat.ucla.edu/~junhua.mao/m-RNN.html .
| no_new_dataset | 0.950503 |
1506.03623 | Han Xiao Bookman | Han Xiao, Xiaoyan Zhu | Max-Entropy Feed-Forward Clustering Neural Network | This paper has been published in ICANN 2015 | ICANN 2015: International Conference on Artificial Neural
Networks, Amsterdam, The Netherlands, (May 14-15, 2015) | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The outputs of non-linear feed-forward neural network are positive, which
could be treated as probability when they are normalized to one. If we take
Entropy-Based Principle into consideration, the outputs for each sample could
be represented as the distribution of this sample for different clusters.
Entropy-Based Principle is the principle with which we could estimate the
unknown distribution under some limited conditions. As this paper defines two
processes in Feed-Forward Neural Network, our limited condition is the
abstracted features of samples which are worked out in the abstraction process.
And the final outputs are the probability distribution for different clusters
in the clustering process. As Entropy-Based Principle is considered into the
feed-forward neural network, a clustering method is born. We have conducted
some experiments on six open UCI datasets, comparing with a few baselines and
applied purity as the measurement . The results illustrate that our method
outperforms all the other baselines that are most popular clustering methods.
| [
{
"version": "v1",
"created": "Thu, 11 Jun 2015 11:01:40 GMT"
}
] | 2015-06-12T00:00:00 | [
[
"Xiao",
"Han",
""
],
[
"Zhu",
"Xiaoyan",
""
]
] | TITLE: Max-Entropy Feed-Forward Clustering Neural Network
ABSTRACT: The outputs of non-linear feed-forward neural network are positive, which
could be treated as probability when they are normalized to one. If we take
Entropy-Based Principle into consideration, the outputs for each sample could
be represented as the distribution of this sample for different clusters.
Entropy-Based Principle is the principle with which we could estimate the
unknown distribution under some limited conditions. As this paper defines two
processes in Feed-Forward Neural Network, our limited condition is the
abstracted features of samples which are worked out in the abstraction process.
And the final outputs are the probability distribution for different clusters
in the clustering process. As Entropy-Based Principle is considered into the
feed-forward neural network, a clustering method is born. We have conducted
some experiments on six open UCI datasets, comparing with a few baselines and
applied purity as the measurement . The results illustrate that our method
outperforms all the other baselines that are most popular clustering methods.
| no_new_dataset | 0.948106 |
1506.03626 | Han Xiao Bookman | Han Xiao, Xiaoyan Zhu | Margin-Based Feed-Forward Neural Network Classifiers | This paper has been published in ICANN 2015: International Conference
on Artificial Neural Networks, Amsterdam, The Netherlands, (May 14-15, 2015) | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Margin-Based Principle has been proposed for a long time, it has been proved
that this principle could reduce the structural risk and improve the
performance in both theoretical and practical aspects. Meanwhile, feed-forward
neural network is a traditional classifier, which is very hot at present with a
deeper architecture. However, the training algorithm of feed-forward neural
network is developed and generated from Widrow-Hoff Principle that means to
minimize the squared error. In this paper, we propose a new training algorithm
for feed-forward neural networks based on Margin-Based Principle, which could
effectively promote the accuracy and generalization ability of neural network
classifiers with less labelled samples and flexible network. We have conducted
experiments on four UCI open datasets and achieved good results as expected. In
conclusion, our model could handle more sparse labelled and more high-dimension
dataset in a high accuracy while modification from old ANN method to our method
is easy and almost free of work.
| [
{
"version": "v1",
"created": "Thu, 11 Jun 2015 11:10:25 GMT"
}
] | 2015-06-12T00:00:00 | [
[
"Xiao",
"Han",
""
],
[
"Zhu",
"Xiaoyan",
""
]
] | TITLE: Margin-Based Feed-Forward Neural Network Classifiers
ABSTRACT: Margin-Based Principle has been proposed for a long time, it has been proved
that this principle could reduce the structural risk and improve the
performance in both theoretical and practical aspects. Meanwhile, feed-forward
neural network is a traditional classifier, which is very hot at present with a
deeper architecture. However, the training algorithm of feed-forward neural
network is developed and generated from Widrow-Hoff Principle that means to
minimize the squared error. In this paper, we propose a new training algorithm
for feed-forward neural networks based on Margin-Based Principle, which could
effectively promote the accuracy and generalization ability of neural network
classifiers with less labelled samples and flexible network. We have conducted
experiments on four UCI open datasets and achieved good results as expected. In
conclusion, our model could handle more sparse labelled and more high-dimension
dataset in a high accuracy while modification from old ANN method to our method
is easy and almost free of work.
| no_new_dataset | 0.949529 |
1506.03668 | Zolzaya Dashdorj | Zolzaya Dashdorj and Stanislav Sobolevsky | Impact of the spatial context on human communication activity | 12 pages, 11 figures | null | null | null | cs.SI physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Technology development produces terabytes of data generated by hu- man
activity in space and time. This enormous amount of data often called big data
becomes crucial for delivering new insights to decision makers. It contains
behavioral information on different types of human activity influenced by many
external factors such as geographic infor- mation and weather forecast. Early
recognition and prediction of those human behaviors are of great importance in
many societal applications like health-care, risk management and urban
planning, etc. In this pa- per, we investigate relevant geographical areas
based on their categories of human activities (i.e., working and shopping)
which identified from ge- ographic information (i.e., Openstreetmap). We use
spectral clustering followed by k-means clustering algorithm based on TF/IDF
cosine simi- larity metric. We evaluate the quality of those observed clusters
with the use of silhouette coefficients which are estimated based on the
similari- ties of the mobile communication activity temporal patterns. The area
clusters are further used to explain typical or exceptional communication
activities. We demonstrate the study using a real dataset containing 1 million
Call Detailed Records. This type of analysis and its application are important
for analyzing the dependency of human behaviors from the external factors and
hidden relationships and unknown correlations and other useful information that
can support decision-making.
| [
{
"version": "v1",
"created": "Thu, 11 Jun 2015 13:46:16 GMT"
}
] | 2015-06-12T00:00:00 | [
[
"Dashdorj",
"Zolzaya",
""
],
[
"Sobolevsky",
"Stanislav",
""
]
] | TITLE: Impact of the spatial context on human communication activity
ABSTRACT: Technology development produces terabytes of data generated by hu- man
activity in space and time. This enormous amount of data often called big data
becomes crucial for delivering new insights to decision makers. It contains
behavioral information on different types of human activity influenced by many
external factors such as geographic infor- mation and weather forecast. Early
recognition and prediction of those human behaviors are of great importance in
many societal applications like health-care, risk management and urban
planning, etc. In this pa- per, we investigate relevant geographical areas
based on their categories of human activities (i.e., working and shopping)
which identified from ge- ographic information (i.e., Openstreetmap). We use
spectral clustering followed by k-means clustering algorithm based on TF/IDF
cosine simi- larity metric. We evaluate the quality of those observed clusters
with the use of silhouette coefficients which are estimated based on the
similari- ties of the mobile communication activity temporal patterns. The area
clusters are further used to explain typical or exceptional communication
activities. We demonstrate the study using a real dataset containing 1 million
Call Detailed Records. This type of analysis and its application are important
for analyzing the dependency of human behaviors from the external factors and
hidden relationships and unknown correlations and other useful information that
can support decision-making.
| no_new_dataset | 0.922273 |
1208.3953 | Vasyl Palchykov | Vasyl Palchykov, J\'anos Kert\'esz, Robin I. M. Dunbar, Kimmo Kaski | Close relationships: A study of mobile communication records | 11 pages, 7 figures | J. Stat. Phys. 151 (2013) 735-744 | 10.1007/s10955-013-0705-0 | null | physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Mobile phone communication as digital service generates ever-increasing
datasets of human communication actions, which in turn allow us to investigate
the structure and evolution of social interactions and their networks. These
datasets can be used to study the structuring of such egocentric networks with
respect to the strength of the relationships by assuming direct dependence of
the communication intensity on the strength of the social tie. Recently we have
discovered that there are significant differences between the first and further
"best friends" from the point of view of age and gender preferences. Here we
introduce a control parameter $p_{\rm max}$ based on the statistics of
communication with the first and second "best friend" and use it to filter the
data. We find that when $p_{\rm max}$ is decreased the identification of the
"best friend" becomes less ambiguous and the earlier observed effects get
stronger, thus corroborating them.
| [
{
"version": "v1",
"created": "Mon, 20 Aug 2012 09:18:55 GMT"
},
{
"version": "v2",
"created": "Thu, 24 Jan 2013 16:52:20 GMT"
}
] | 2015-06-11T00:00:00 | [
[
"Palchykov",
"Vasyl",
""
],
[
"Kertész",
"János",
""
],
[
"Dunbar",
"Robin I. M.",
""
],
[
"Kaski",
"Kimmo",
""
]
] | TITLE: Close relationships: A study of mobile communication records
ABSTRACT: Mobile phone communication as digital service generates ever-increasing
datasets of human communication actions, which in turn allow us to investigate
the structure and evolution of social interactions and their networks. These
datasets can be used to study the structuring of such egocentric networks with
respect to the strength of the relationships by assuming direct dependence of
the communication intensity on the strength of the social tie. Recently we have
discovered that there are significant differences between the first and further
"best friends" from the point of view of age and gender preferences. Here we
introduce a control parameter $p_{\rm max}$ based on the statistics of
communication with the first and second "best friend" and use it to filter the
data. We find that when $p_{\rm max}$ is decreased the identification of the
"best friend" becomes less ambiguous and the earlier observed effects get
stronger, thus corroborating them.
| no_new_dataset | 0.845369 |
1208.4122 | Stephen Bailey | Stephen Bailey | Principal Component Analysis with Noisy and/or Missing Data | Accepted for publication in PASP; v2 with minor updates, mostly to
bibliography | null | 10.1086/668105 | null | astro-ph.IM physics.data-an | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a method for performing Principal Component Analysis (PCA) on
noisy datasets with missing values. Estimates of the measurement error are used
to weight the input data such that compared to classic PCA, the resulting
eigenvectors are more sensitive to the true underlying signal variations rather
than being pulled by heteroskedastic measurement noise. Missing data is simply
the limiting case of weight=0. The underlying algorithm is a noise weighted
Expectation Maximization (EM) PCA, which has additional benefits of
implementation speed and flexibility for smoothing eigenvectors to reduce the
noise contribution. We present applications of this method on simulated data
and QSO spectra from the Sloan Digital Sky Survey.
| [
{
"version": "v1",
"created": "Mon, 20 Aug 2012 20:59:10 GMT"
},
{
"version": "v2",
"created": "Fri, 14 Sep 2012 18:27:56 GMT"
}
] | 2015-06-11T00:00:00 | [
[
"Bailey",
"Stephen",
""
]
] | TITLE: Principal Component Analysis with Noisy and/or Missing Data
ABSTRACT: We present a method for performing Principal Component Analysis (PCA) on
noisy datasets with missing values. Estimates of the measurement error are used
to weight the input data such that compared to classic PCA, the resulting
eigenvectors are more sensitive to the true underlying signal variations rather
than being pulled by heteroskedastic measurement noise. Missing data is simply
the limiting case of weight=0. The underlying algorithm is a noise weighted
Expectation Maximization (EM) PCA, which has additional benefits of
implementation speed and flexibility for smoothing eigenvectors to reduce the
noise contribution. We present applications of this method on simulated data
and QSO spectra from the Sloan Digital Sky Survey.
| no_new_dataset | 0.954308 |
1208.5582 | Davide Faranda | Davide Faranda, Jorge Milhazes Freitas, Valerio Lucarini, Giorgio
Turchetti and Sandro Vaienti | Extreme value statistics for dynamical systems with noise | 34 pages, 8 figures | null | 10.1088/0951-7715/26/9/2597 | null | math.DS physics.data-an | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study the distribution of maxima (Extreme Value Statistics) for sequences
of observables computed along orbits generated by random transformations. The
underlying, deterministic, dynamical system can be regular or chaotic. In the
former case, we will show that by perturbing rational or irrational rotations
with additive noise, an extreme value law appears, regardless of the intensity
of the noise, while unperturbed rotations do not admit such limiting
distributions. In the case of deterministic chaotic dynamics, we will consider
observables specially designed to study the recurrence properties in the
neighbourhood of periodic points. Hence, the exponential limiting law for the
distribution of maxima is modified by the presence of the extremal index, a
positive parameter not larger than one, whose inverse gives the average size of
the clusters of extreme events. The theory predicts that such a parameter is
unitary when the system is perturbed randomly. We perform sophisticated
numerical tests to assess how strong is the impact of noise level, when finite
time series are considered. We find agreement with the asymptotic theoretical
results but also non-trivial behaviour in the finite range. In particular our
results suggest that in many applications where finite datasets can be produced
or analysed one must be careful in assuming that the smoothing nature of noise
prevails over the underlying deterministic dynamics.
| [
{
"version": "v1",
"created": "Tue, 28 Aug 2012 08:03:07 GMT"
},
{
"version": "v2",
"created": "Mon, 25 Mar 2013 11:21:05 GMT"
}
] | 2015-06-11T00:00:00 | [
[
"Faranda",
"Davide",
""
],
[
"Freitas",
"Jorge Milhazes",
""
],
[
"Lucarini",
"Valerio",
""
],
[
"Turchetti",
"Giorgio",
""
],
[
"Vaienti",
"Sandro",
""
]
] | TITLE: Extreme value statistics for dynamical systems with noise
ABSTRACT: We study the distribution of maxima (Extreme Value Statistics) for sequences
of observables computed along orbits generated by random transformations. The
underlying, deterministic, dynamical system can be regular or chaotic. In the
former case, we will show that by perturbing rational or irrational rotations
with additive noise, an extreme value law appears, regardless of the intensity
of the noise, while unperturbed rotations do not admit such limiting
distributions. In the case of deterministic chaotic dynamics, we will consider
observables specially designed to study the recurrence properties in the
neighbourhood of periodic points. Hence, the exponential limiting law for the
distribution of maxima is modified by the presence of the extremal index, a
positive parameter not larger than one, whose inverse gives the average size of
the clusters of extreme events. The theory predicts that such a parameter is
unitary when the system is perturbed randomly. We perform sophisticated
numerical tests to assess how strong is the impact of noise level, when finite
time series are considered. We find agreement with the asymptotic theoretical
results but also non-trivial behaviour in the finite range. In particular our
results suggest that in many applications where finite datasets can be produced
or analysed one must be careful in assuming that the smoothing nature of noise
prevails over the underlying deterministic dynamics.
| no_new_dataset | 0.943919 |
1209.4826 | Dr. Anirudh Pradhan | Anirudh Pradhan | Accelerating dark energy models with anisotropic fluid in Bianchi
type-$VI_{0}$ space-time | 22 pages, 8 figures. arXiv admin note: substantial text overlap with
arXiv:1010.1121, arXiv:1108.2133, arXiv:1010.2362 | Res. Astron. Astrophys., Vol. 13, No. 2, (2013), 139-158 | 10.1088/1674-4527/13/2/002 | null | physics.gen-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Motivated by the increasing evidence for the need of a geometry that
resembles Bianchi morphology to explain the observed anisotropy in the WMAP
data, we have discussed some features of the Bianchi type-$VI_{0}$ universes in
the presence of a fluid that wields an anisotropic equation of state (EoS)
parameter in general relativity. We present two accelerating dark energy (DE)
models with an anisotropic fluid in Bianchi type-$VI_{0}$ space-time. To
prevail the deterministic solution we choose the scale factor $a(t) =
\sqrt{t^{n}e^{t}}$, which yields a time-dependent deceleration parameter (DP),
representing a class of models which generate a transition of the universe from
the early decelerating phase to the recent accelerating phase. Under the
suitable condition, the anisotropic models approach to isotropic scenario. The
EoS for dark energy $\omega$ is found to be time-dependent and its existing
range for derived models is in good agreement with the recent observations of
SNe Ia data (Knop et al. 2003), SNe Ia data with CMBR anisotropy and galaxy
clustering statistics (Tegmark et al. 2004) and latest combination of
cosmological datasets coming from CMB anisotropies, luminosity distances of
high redshift type Ia supernovae and galaxy clustering (Hinshaw et al. 2009;
Komatsu et al. 2009). For different values of $n$, we can generate a class of
physically viable DE models.The cosmological constant $\Lambda$ is found to be
a positive decreasing function of time and it approaches to a small positive
value at late time (i.e. the present epoch) which is corroborated by results
from recent type Ia supernovae observations. We also observe that our solutions
are stable. The physical and geometric aspects of both the models are also
discussed in detail.
| [
{
"version": "v1",
"created": "Mon, 17 Sep 2012 04:55:54 GMT"
}
] | 2015-06-11T00:00:00 | [
[
"Pradhan",
"Anirudh",
""
]
] | TITLE: Accelerating dark energy models with anisotropic fluid in Bianchi
type-$VI_{0}$ space-time
ABSTRACT: Motivated by the increasing evidence for the need of a geometry that
resembles Bianchi morphology to explain the observed anisotropy in the WMAP
data, we have discussed some features of the Bianchi type-$VI_{0}$ universes in
the presence of a fluid that wields an anisotropic equation of state (EoS)
parameter in general relativity. We present two accelerating dark energy (DE)
models with an anisotropic fluid in Bianchi type-$VI_{0}$ space-time. To
prevail the deterministic solution we choose the scale factor $a(t) =
\sqrt{t^{n}e^{t}}$, which yields a time-dependent deceleration parameter (DP),
representing a class of models which generate a transition of the universe from
the early decelerating phase to the recent accelerating phase. Under the
suitable condition, the anisotropic models approach to isotropic scenario. The
EoS for dark energy $\omega$ is found to be time-dependent and its existing
range for derived models is in good agreement with the recent observations of
SNe Ia data (Knop et al. 2003), SNe Ia data with CMBR anisotropy and galaxy
clustering statistics (Tegmark et al. 2004) and latest combination of
cosmological datasets coming from CMB anisotropies, luminosity distances of
high redshift type Ia supernovae and galaxy clustering (Hinshaw et al. 2009;
Komatsu et al. 2009). For different values of $n$, we can generate a class of
physically viable DE models.The cosmological constant $\Lambda$ is found to be
a positive decreasing function of time and it approaches to a small positive
value at late time (i.e. the present epoch) which is corroborated by results
from recent type Ia supernovae observations. We also observe that our solutions
are stable. The physical and geometric aspects of both the models are also
discussed in detail.
| no_new_dataset | 0.952574 |
1210.1095 | Francesco Vezzi | Francesco Vezzi, Giuseppe Narzisi and Bud Mishra | Reevaluating Assembly Evaluations with Feature Response Curves: GAGE and
Assemblathons | Submitted to PLoS One. Supplementary material available at
http://www.nada.kth.se/~vezzi/publications/supplementary.pdf and
http://cs.nyu.edu/mishra/PUBLICATIONS/12.supplementaryFRC.pdf | null | 10.1371/journal.pone.0052210 | null | q-bio.GN cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In just the last decade, a multitude of bio-technologies and software
pipelines have emerged to revolutionize genomics. To further their central
goal, they aim to accelerate and improve the quality of de novo whole-genome
assembly starting from short DNA reads. However, the performance of each of
these tools is contingent on the length and quality of the sequencing data, the
structure and complexity of the genome sequence, and the resolution and quality
of long-range information. Furthermore, in the absence of any metric that
captures the most fundamental "features" of a high-quality assembly, there is
no obvious recipe for users to select the most desirable assembler/assembly.
International competitions such as Assemblathons or GAGE tried to identify the
best assembler(s) and their features. Some what circuitously, the only
available approach to gauge de novo assemblies and assemblers relies solely on
the availability of a high-quality fully assembled reference genome sequence.
Still worse, reference-guided evaluations are often both difficult to analyze,
leading to conclusions that are difficult to interpret. In this paper, we
circumvent many of these issues by relying upon a tool, dubbed FRCbam, which is
capable of evaluating de novo assemblies from the read-layouts even when no
reference exists. We extend the FRCurve approach to cases where lay-out
information may have been obscured, as is true in many deBruijn-graph-based
algorithms. As a by-product, FRCurve now expands its applicability to a much
wider class of assemblers -- thus, identifying higher-quality members of this
group, their inter-relations as well as sensitivity to carefully selected
features, with or without the support of a reference sequence or layout for the
reads. The paper concludes by reevaluating several recently conducted assembly
competitions and the datasets that have resulted from them.
| [
{
"version": "v1",
"created": "Wed, 3 Oct 2012 13:02:30 GMT"
}
] | 2015-06-11T00:00:00 | [
[
"Vezzi",
"Francesco",
""
],
[
"Narzisi",
"Giuseppe",
""
],
[
"Mishra",
"Bud",
""
]
] | TITLE: Reevaluating Assembly Evaluations with Feature Response Curves: GAGE and
Assemblathons
ABSTRACT: In just the last decade, a multitude of bio-technologies and software
pipelines have emerged to revolutionize genomics. To further their central
goal, they aim to accelerate and improve the quality of de novo whole-genome
assembly starting from short DNA reads. However, the performance of each of
these tools is contingent on the length and quality of the sequencing data, the
structure and complexity of the genome sequence, and the resolution and quality
of long-range information. Furthermore, in the absence of any metric that
captures the most fundamental "features" of a high-quality assembly, there is
no obvious recipe for users to select the most desirable assembler/assembly.
International competitions such as Assemblathons or GAGE tried to identify the
best assembler(s) and their features. Some what circuitously, the only
available approach to gauge de novo assemblies and assemblers relies solely on
the availability of a high-quality fully assembled reference genome sequence.
Still worse, reference-guided evaluations are often both difficult to analyze,
leading to conclusions that are difficult to interpret. In this paper, we
circumvent many of these issues by relying upon a tool, dubbed FRCbam, which is
capable of evaluating de novo assemblies from the read-layouts even when no
reference exists. We extend the FRCurve approach to cases where lay-out
information may have been obscured, as is true in many deBruijn-graph-based
algorithms. As a by-product, FRCurve now expands its applicability to a much
wider class of assemblers -- thus, identifying higher-quality members of this
group, their inter-relations as well as sensitivity to carefully selected
features, with or without the support of a reference sequence or layout for the
reads. The paper concludes by reevaluating several recently conducted assembly
competitions and the datasets that have resulted from them.
| no_new_dataset | 0.939858 |
1404.0466 | Da Kuang | Da Kuang, Alex Gittens, Raffay Hamid | piCholesky: Polynomial Interpolation of Multiple Cholesky Factors for
Efficient Approximate Cross-Validation | null | null | null | null | cs.LG cs.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The dominant cost in solving least-square problems using Newton's method is
often that of factorizing the Hessian matrix over multiple values of the
regularization parameter ($\lambda$). We propose an efficient way to
interpolate the Cholesky factors of the Hessian matrix computed over a small
set of $\lambda$ values. This approximation enables us to optimally minimize
the hold-out error while incurring only a fraction of the cost compared to
exact cross-validation. We provide a formal error bound for our approximation
scheme and present solutions to a set of key implementation challenges that
allow our approach to maximally exploit the compute power of modern
architectures. We present a thorough empirical analysis over multiple datasets
to show the effectiveness of our approach.
| [
{
"version": "v1",
"created": "Wed, 2 Apr 2014 05:33:41 GMT"
},
{
"version": "v2",
"created": "Wed, 10 Jun 2015 18:20:16 GMT"
}
] | 2015-06-11T00:00:00 | [
[
"Kuang",
"Da",
""
],
[
"Gittens",
"Alex",
""
],
[
"Hamid",
"Raffay",
""
]
] | TITLE: piCholesky: Polynomial Interpolation of Multiple Cholesky Factors for
Efficient Approximate Cross-Validation
ABSTRACT: The dominant cost in solving least-square problems using Newton's method is
often that of factorizing the Hessian matrix over multiple values of the
regularization parameter ($\lambda$). We propose an efficient way to
interpolate the Cholesky factors of the Hessian matrix computed over a small
set of $\lambda$ values. This approximation enables us to optimally minimize
the hold-out error while incurring only a fraction of the cost compared to
exact cross-validation. We provide a formal error bound for our approximation
scheme and present solutions to a set of key implementation challenges that
allow our approach to maximally exploit the compute power of modern
architectures. We present a thorough empirical analysis over multiple datasets
to show the effectiveness of our approach.
| no_new_dataset | 0.942135 |
1408.2617 | Ludwig Ritschl | Ludwig Ritschl, Jan Kuntz and Marc Kachelrie{\ss} | The rotate-plus-shift C-arm trajectory: Complete CT data with less than
180{\deg} rotation | null | null | 10.1117/12.2081925 | null | physics.med-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In the last decade C-arm-based cone-beam CT became a widely used modality for
intraoperative imaging. Typically a C-arm scan is performed using a circle-like
trajectory around a region of interest. Therefor an angular range of at least
180{\deg} plus fan-angle must be covered to ensure a completely sampled data
set. This fact defines some constraints on the geometry and technical
specifications of a C-arm system, for example a larger C radius or a smaller C
opening respectively. These technical modifications are usually not benificial
in terms of handling and usability of the C-arm during classical 2D
applications like fluoroscopy. The method proposed in this paper relaxes the
constraint of 180{\deg} plus fan-angle rotation to acquire a complete data set.
The proposed C-arm trajectory requires a motorization of the orbital axis of
the C and of ideally two orthogonal axis in the C plane. The trajectory
consists of three parts: A rotation of the C around a defined iso-center and
two translational movements parallel to the detector plane at the begin and at
the end of the rotation. Combining these three parts to one trajectory enables
for the acquisition of a completely sampled dataset using only 180{\deg} minus
fan-angle of rotation. To evaluate the method we show animal scans acquired
with a mobile C-arm prototype. We expect that the transition of this method
into clinical routine will lead to a much broader use of intraoperative 3D
imaging in a wide field of clinical applications.
| [
{
"version": "v1",
"created": "Tue, 12 Aug 2014 05:01:18 GMT"
}
] | 2015-06-11T00:00:00 | [
[
"Ritschl",
"Ludwig",
""
],
[
"Kuntz",
"Jan",
""
],
[
"Kachelrieß",
"Marc",
""
]
] | TITLE: The rotate-plus-shift C-arm trajectory: Complete CT data with less than
180{\deg} rotation
ABSTRACT: In the last decade C-arm-based cone-beam CT became a widely used modality for
intraoperative imaging. Typically a C-arm scan is performed using a circle-like
trajectory around a region of interest. Therefor an angular range of at least
180{\deg} plus fan-angle must be covered to ensure a completely sampled data
set. This fact defines some constraints on the geometry and technical
specifications of a C-arm system, for example a larger C radius or a smaller C
opening respectively. These technical modifications are usually not benificial
in terms of handling and usability of the C-arm during classical 2D
applications like fluoroscopy. The method proposed in this paper relaxes the
constraint of 180{\deg} plus fan-angle rotation to acquire a complete data set.
The proposed C-arm trajectory requires a motorization of the orbital axis of
the C and of ideally two orthogonal axis in the C plane. The trajectory
consists of three parts: A rotation of the C around a defined iso-center and
two translational movements parallel to the detector plane at the begin and at
the end of the rotation. Combining these three parts to one trajectory enables
for the acquisition of a completely sampled dataset using only 180{\deg} minus
fan-angle of rotation. To evaluate the method we show animal scans acquired
with a mobile C-arm prototype. We expect that the transition of this method
into clinical routine will lead to a much broader use of intraoperative 3D
imaging in a wide field of clinical applications.
| no_new_dataset | 0.95018 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.