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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1606.01583 | Augustus Odena | Augustus Odena | Semi-Supervised Learning with Generative Adversarial Networks | Appearing in the Data Efficient Machine Learning workshop at ICML
2016 | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We extend Generative Adversarial Networks (GANs) to the semi-supervised
context by forcing the discriminator network to output class labels. We train a
generative model G and a discriminator D on a dataset with inputs belonging to
one of N classes. At training time, D is made to predict which of N+1 classes
the input belongs to, where an extra class is added to correspond to the
outputs of G. We show that this method can be used to create a more
data-efficient classifier and that it allows for generating higher quality
samples than a regular GAN.
| [
{
"version": "v1",
"created": "Sun, 5 Jun 2016 23:42:19 GMT"
},
{
"version": "v2",
"created": "Sat, 22 Oct 2016 01:07:38 GMT"
}
] | 2016-10-25T00:00:00 | [
[
"Odena",
"Augustus",
""
]
] | TITLE: Semi-Supervised Learning with Generative Adversarial Networks
ABSTRACT: We extend Generative Adversarial Networks (GANs) to the semi-supervised
context by forcing the discriminator network to output class labels. We train a
generative model G and a discriminator D on a dataset with inputs belonging to
one of N classes. At training time, D is made to predict which of N+1 classes
the input belongs to, where an extra class is added to correspond to the
outputs of G. We show that this method can be used to create a more
data-efficient classifier and that it allows for generating higher quality
samples than a regular GAN.
| no_new_dataset | 0.949342 |
1606.03073 | Ya\u{g}mur G\"u\c{c}l\"ut\"urk | Ya\u{g}mur G\"u\c{c}l\"ut\"urk, Umut G\"u\c{c}l\"u, Rob van Lier,
Marcel A. J. van Gerven | Convolutional Sketch Inversion | null | null | 10.1007/978-3-319-46604-0_56 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we use deep neural networks for inverting face sketches to
synthesize photorealistic face images. We first construct a semi-simulated
dataset containing a very large number of computer-generated face sketches with
different styles and corresponding face images by expanding existing
unconstrained face data sets. We then train models achieving state-of-the-art
results on both computer-generated sketches and hand-drawn sketches by
leveraging recent advances in deep learning such as batch normalization, deep
residual learning, perceptual losses and stochastic optimization in combination
with our new dataset. We finally demonstrate potential applications of our
models in fine arts and forensic arts. In contrast to existing patch-based
approaches, our deep-neural-network-based approach can be used for synthesizing
photorealistic face images by inverting face sketches in the wild.
| [
{
"version": "v1",
"created": "Thu, 9 Jun 2016 19:27:41 GMT"
}
] | 2016-10-25T00:00:00 | [
[
"Güçlütürk",
"Yağmur",
""
],
[
"Güçlü",
"Umut",
""
],
[
"van Lier",
"Rob",
""
],
[
"van Gerven",
"Marcel A. J.",
""
]
] | TITLE: Convolutional Sketch Inversion
ABSTRACT: In this paper, we use deep neural networks for inverting face sketches to
synthesize photorealistic face images. We first construct a semi-simulated
dataset containing a very large number of computer-generated face sketches with
different styles and corresponding face images by expanding existing
unconstrained face data sets. We then train models achieving state-of-the-art
results on both computer-generated sketches and hand-drawn sketches by
leveraging recent advances in deep learning such as batch normalization, deep
residual learning, perceptual losses and stochastic optimization in combination
with our new dataset. We finally demonstrate potential applications of our
models in fine arts and forensic arts. In contrast to existing patch-based
approaches, our deep-neural-network-based approach can be used for synthesizing
photorealistic face images by inverting face sketches in the wild.
| new_dataset | 0.949995 |
1607.03085 | Kamil Rocki | Kamil Rocki | Recurrent Memory Array Structures | Minor changes | null | null | null | cs.LG cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The following report introduces ideas augmenting standard Long Short Term
Memory (LSTM) architecture with multiple memory cells per hidden unit in order
to improve its generalization capabilities. It considers both deterministic and
stochastic variants of memory operation. It is shown that the nondeterministic
Array-LSTM approach improves state-of-the-art performance on character level
text prediction achieving 1.402 BPC on enwik8 dataset. Furthermore, this report
estabilishes baseline neural-based results of 1.12 BPC and 1.19 BPC for enwik9
and enwik10 datasets respectively.
| [
{
"version": "v1",
"created": "Mon, 11 Jul 2016 19:29:44 GMT"
},
{
"version": "v2",
"created": "Wed, 13 Jul 2016 16:46:33 GMT"
},
{
"version": "v3",
"created": "Sun, 23 Oct 2016 02:01:55 GMT"
}
] | 2016-10-25T00:00:00 | [
[
"Rocki",
"Kamil",
""
]
] | TITLE: Recurrent Memory Array Structures
ABSTRACT: The following report introduces ideas augmenting standard Long Short Term
Memory (LSTM) architecture with multiple memory cells per hidden unit in order
to improve its generalization capabilities. It considers both deterministic and
stochastic variants of memory operation. It is shown that the nondeterministic
Array-LSTM approach improves state-of-the-art performance on character level
text prediction achieving 1.402 BPC on enwik8 dataset. Furthermore, this report
estabilishes baseline neural-based results of 1.12 BPC and 1.19 BPC for enwik9
and enwik10 datasets respectively.
| no_new_dataset | 0.945147 |
1607.05418 | Soheil Hashemi | Hokchhay Tann, Soheil Hashemi, R. Iris Bahar, Sherief Reda | Runtime Configurable Deep Neural Networks for Energy-Accuracy Trade-off | null | null | 10.1145/2968456.2968458 | null | cs.NE cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a novel dynamic configuration technique for deep neural networks
that permits step-wise energy-accuracy trade-offs during runtime. Our
configuration technique adjusts the number of channels in the network
dynamically depending on response time, power, and accuracy targets. To enable
this dynamic configuration technique, we co-design a new training algorithm,
where the network is incrementally trained such that the weights in channels
trained in earlier steps are fixed. Our technique provides the flexibility of
multiple networks while storing and utilizing one set of weights. We evaluate
our techniques using both an ASIC-based hardware accelerator as well as a
low-power embedded GPGPU and show that our approach leads to only a small or
negligible loss in the final network accuracy. We analyze the performance of
our proposed methodology using three well-known networks for MNIST, CIFAR-10,
and SVHN datasets, and we show that we are able to achieve up to 95% energy
reduction with less than 1% accuracy loss across the three benchmarks. In
addition, compared to prior work on dynamic network reconfiguration, we show
that our approach leads to approximately 50% savings in storage requirements,
while achieving similar accuracy.
| [
{
"version": "v1",
"created": "Tue, 19 Jul 2016 06:27:05 GMT"
},
{
"version": "v2",
"created": "Wed, 20 Jul 2016 20:42:51 GMT"
}
] | 2016-10-25T00:00:00 | [
[
"Tann",
"Hokchhay",
""
],
[
"Hashemi",
"Soheil",
""
],
[
"Bahar",
"R. Iris",
""
],
[
"Reda",
"Sherief",
""
]
] | TITLE: Runtime Configurable Deep Neural Networks for Energy-Accuracy Trade-off
ABSTRACT: We present a novel dynamic configuration technique for deep neural networks
that permits step-wise energy-accuracy trade-offs during runtime. Our
configuration technique adjusts the number of channels in the network
dynamically depending on response time, power, and accuracy targets. To enable
this dynamic configuration technique, we co-design a new training algorithm,
where the network is incrementally trained such that the weights in channels
trained in earlier steps are fixed. Our technique provides the flexibility of
multiple networks while storing and utilizing one set of weights. We evaluate
our techniques using both an ASIC-based hardware accelerator as well as a
low-power embedded GPGPU and show that our approach leads to only a small or
negligible loss in the final network accuracy. We analyze the performance of
our proposed methodology using three well-known networks for MNIST, CIFAR-10,
and SVHN datasets, and we show that we are able to achieve up to 95% energy
reduction with less than 1% accuracy loss across the three benchmarks. In
addition, compared to prior work on dynamic network reconfiguration, we show
that our approach leads to approximately 50% savings in storage requirements,
while achieving similar accuracy.
| no_new_dataset | 0.948298 |
1610.04631 | Shuai Zheng | Shuai Zheng, Feiping Nie, Chris Ding, Heng Huang | A Harmonic Mean Linear Discriminant Analysis for Robust Image
Classification | IEEE 28th International Conference on Tools with Artificial
Intelligence, ICTAI 2016 | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Linear Discriminant Analysis (LDA) is a widely-used supervised dimensionality
reduction method in computer vision and pattern recognition. In null space
based LDA (NLDA), a well-known LDA extension, between-class distance is
maximized in the null space of the within-class scatter matrix. However, there
are some limitations in NLDA. Firstly, for many data sets, null space of
within-class scatter matrix does not exist, thus NLDA is not applicable to
those datasets. Secondly, NLDA uses arithmetic mean of between-class distances
and gives equal consideration to all between-class distances, which makes
larger between-class distances can dominate the result and thus limits the
performance of NLDA. In this paper, we propose a harmonic mean based Linear
Discriminant Analysis, Multi-Class Discriminant Analysis (MCDA), for image
classification, which minimizes the reciprocal of weighted harmonic mean of
pairwise between-class distance. More importantly, MCDA gives higher priority
to maximize small between-class distances. MCDA can be extended to multi-label
dimension reduction. Results on 7 single-label data sets and 4 multi-label data
sets show that MCDA has consistently better performance than 10 other
single-label approaches and 4 other multi-label approaches in terms of
classification accuracy, macro and micro average F1 score.
| [
{
"version": "v1",
"created": "Fri, 14 Oct 2016 20:36:57 GMT"
},
{
"version": "v2",
"created": "Mon, 24 Oct 2016 16:38:29 GMT"
}
] | 2016-10-25T00:00:00 | [
[
"Zheng",
"Shuai",
""
],
[
"Nie",
"Feiping",
""
],
[
"Ding",
"Chris",
""
],
[
"Huang",
"Heng",
""
]
] | TITLE: A Harmonic Mean Linear Discriminant Analysis for Robust Image
Classification
ABSTRACT: Linear Discriminant Analysis (LDA) is a widely-used supervised dimensionality
reduction method in computer vision and pattern recognition. In null space
based LDA (NLDA), a well-known LDA extension, between-class distance is
maximized in the null space of the within-class scatter matrix. However, there
are some limitations in NLDA. Firstly, for many data sets, null space of
within-class scatter matrix does not exist, thus NLDA is not applicable to
those datasets. Secondly, NLDA uses arithmetic mean of between-class distances
and gives equal consideration to all between-class distances, which makes
larger between-class distances can dominate the result and thus limits the
performance of NLDA. In this paper, we propose a harmonic mean based Linear
Discriminant Analysis, Multi-Class Discriminant Analysis (MCDA), for image
classification, which minimizes the reciprocal of weighted harmonic mean of
pairwise between-class distance. More importantly, MCDA gives higher priority
to maximize small between-class distances. MCDA can be extended to multi-label
dimension reduction. Results on 7 single-label data sets and 4 multi-label data
sets show that MCDA has consistently better performance than 10 other
single-label approaches and 4 other multi-label approaches in terms of
classification accuracy, macro and micro average F1 score.
| no_new_dataset | 0.945701 |
1610.07061 | Tanmoy Chakraborty | Dinesh Pradhan and Partha Sarathi Paul and Umesh Maheswari and Subrata
Nandi and Tanmoy Chakraborty | $C^3$-index: A PageRank based multi-faceted metric for authors'
performance measurement | 24 pages, 6 figures, 2 tables, Scientrometrics, 2016. arXiv admin
note: text overlap with arXiv:1102.1760 by other authors | null | null | null | cs.DL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Ranking scientific authors is an important but challenging task, mostly due
to the dynamic nature of the evolving scientific publications. The basic
indicators of an author's productivity and impact are still the number of
publications and the citation count (leading to the popular metrics such as
h-index, g-index etc.). H-index and its popular variants are mostly effective
in ranking highly-cited authors, thus fail to resolve ties while ranking
medium-cited and low-cited authors who are majority in number. Therefore, these
metrics are inefficient to predict the ability of promising young researchers
at the beginning of their career. In this paper, we propose $C^3$-index that
combines the effect of citations and collaborations of an author in a
systematic way using a weighted multi-layered network to rank authors. We
conduct our experiments on a massive publication dataset of Computer Science
and show that - (i) $C^3$-index is consistent over time, which is one of the
fundamental characteristics of a ranking metric, (ii) $C^3$-index is as
efficient as h-index and its variants to rank highly-cited authors, (iii)
$C^3$-index can act as a conflict resolution metric to break ties in the
ranking of medium-cited and low-cited authors, (iv) $C^3$-index can also be
used to predict future achievers at the early stage of their career.
| [
{
"version": "v1",
"created": "Sat, 22 Oct 2016 14:56:04 GMT"
}
] | 2016-10-25T00:00:00 | [
[
"Pradhan",
"Dinesh",
""
],
[
"Paul",
"Partha Sarathi",
""
],
[
"Maheswari",
"Umesh",
""
],
[
"Nandi",
"Subrata",
""
],
[
"Chakraborty",
"Tanmoy",
""
]
] | TITLE: $C^3$-index: A PageRank based multi-faceted metric for authors'
performance measurement
ABSTRACT: Ranking scientific authors is an important but challenging task, mostly due
to the dynamic nature of the evolving scientific publications. The basic
indicators of an author's productivity and impact are still the number of
publications and the citation count (leading to the popular metrics such as
h-index, g-index etc.). H-index and its popular variants are mostly effective
in ranking highly-cited authors, thus fail to resolve ties while ranking
medium-cited and low-cited authors who are majority in number. Therefore, these
metrics are inefficient to predict the ability of promising young researchers
at the beginning of their career. In this paper, we propose $C^3$-index that
combines the effect of citations and collaborations of an author in a
systematic way using a weighted multi-layered network to rank authors. We
conduct our experiments on a massive publication dataset of Computer Science
and show that - (i) $C^3$-index is consistent over time, which is one of the
fundamental characteristics of a ranking metric, (ii) $C^3$-index is as
efficient as h-index and its variants to rank highly-cited authors, (iii)
$C^3$-index can act as a conflict resolution metric to break ties in the
ranking of medium-cited and low-cited authors, (iv) $C^3$-index can also be
used to predict future achievers at the early stage of their career.
| no_new_dataset | 0.949153 |
1610.07363 | Arkaitz Zubiaga | Arkaitz Zubiaga, Maria Liakata, Rob Procter | Learning Reporting Dynamics during Breaking News for Rumour Detection in
Social Media | null | null | null | null | cs.CL cs.IR cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Breaking news leads to situations of fast-paced reporting in social media,
producing all kinds of updates related to news stories, albeit with the caveat
that some of those early updates tend to be rumours, i.e., information with an
unverified status at the time of posting. Flagging information that is
unverified can be helpful to avoid the spread of information that may turn out
to be false. Detection of rumours can also feed a rumour tracking system that
ultimately determines their veracity. In this paper we introduce a novel
approach to rumour detection that learns from the sequential dynamics of
reporting during breaking news in social media to detect rumours in new
stories. Using Twitter datasets collected during five breaking news stories, we
experiment with Conditional Random Fields as a sequential classifier that
leverages context learnt during an event for rumour detection, which we compare
with the state-of-the-art rumour detection system as well as other baselines.
In contrast to existing work, our classifier does not need to observe tweets
querying a piece of information to deem it a rumour, but instead we detect
rumours from the tweet alone by exploiting context learnt during the event. Our
classifier achieves competitive performance, beating the state-of-the-art
classifier that relies on querying tweets with improved precision and recall,
as well as outperforming our best baseline with nearly 40% improvement in terms
of F1 score. The scale and diversity of our experiments reinforces the
generalisability of our classifier.
| [
{
"version": "v1",
"created": "Mon, 24 Oct 2016 11:25:24 GMT"
}
] | 2016-10-25T00:00:00 | [
[
"Zubiaga",
"Arkaitz",
""
],
[
"Liakata",
"Maria",
""
],
[
"Procter",
"Rob",
""
]
] | TITLE: Learning Reporting Dynamics during Breaking News for Rumour Detection in
Social Media
ABSTRACT: Breaking news leads to situations of fast-paced reporting in social media,
producing all kinds of updates related to news stories, albeit with the caveat
that some of those early updates tend to be rumours, i.e., information with an
unverified status at the time of posting. Flagging information that is
unverified can be helpful to avoid the spread of information that may turn out
to be false. Detection of rumours can also feed a rumour tracking system that
ultimately determines their veracity. In this paper we introduce a novel
approach to rumour detection that learns from the sequential dynamics of
reporting during breaking news in social media to detect rumours in new
stories. Using Twitter datasets collected during five breaking news stories, we
experiment with Conditional Random Fields as a sequential classifier that
leverages context learnt during an event for rumour detection, which we compare
with the state-of-the-art rumour detection system as well as other baselines.
In contrast to existing work, our classifier does not need to observe tweets
querying a piece of information to deem it a rumour, but instead we detect
rumours from the tweet alone by exploiting context learnt during the event. Our
classifier achieves competitive performance, beating the state-of-the-art
classifier that relies on querying tweets with improved precision and recall,
as well as outperforming our best baseline with nearly 40% improvement in terms
of F1 score. The scale and diversity of our experiments reinforces the
generalisability of our classifier.
| no_new_dataset | 0.946597 |
1610.07569 | Jiaqi Mu Jiaqi Mu | Jiaqi Mu, Suma Bhat, Pramod Viswanath | Geometry of Polysemy | null | null | null | null | cs.CL cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Vector representations of words have heralded a transformational approach to
classical problems in NLP; the most popular example is word2vec. However, a
single vector does not suffice to model the polysemous nature of many
(frequent) words, i.e., words with multiple meanings. In this paper, we propose
a three-fold approach for unsupervised polysemy modeling: (a) context
representations, (b) sense induction and disambiguation and (c) lexeme (as a
word and sense pair) representations. A key feature of our work is the finding
that a sentence containing a target word is well represented by a low rank
subspace, instead of a point in a vector space. We then show that the subspaces
associated with a particular sense of the target word tend to intersect over a
line (one-dimensional subspace), which we use to disambiguate senses using a
clustering algorithm that harnesses the Grassmannian geometry of the
representations. The disambiguation algorithm, which we call $K$-Grassmeans,
leads to a procedure to label the different senses of the target word in the
corpus -- yielding lexeme vector representations, all in an unsupervised manner
starting from a large (Wikipedia) corpus in English. Apart from several
prototypical target (word,sense) examples and a host of empirical studies to
intuit and justify the various geometric representations, we validate our
algorithms on standard sense induction and disambiguation datasets and present
new state-of-the-art results.
| [
{
"version": "v1",
"created": "Mon, 24 Oct 2016 19:35:29 GMT"
}
] | 2016-10-25T00:00:00 | [
[
"Mu",
"Jiaqi",
""
],
[
"Bhat",
"Suma",
""
],
[
"Viswanath",
"Pramod",
""
]
] | TITLE: Geometry of Polysemy
ABSTRACT: Vector representations of words have heralded a transformational approach to
classical problems in NLP; the most popular example is word2vec. However, a
single vector does not suffice to model the polysemous nature of many
(frequent) words, i.e., words with multiple meanings. In this paper, we propose
a three-fold approach for unsupervised polysemy modeling: (a) context
representations, (b) sense induction and disambiguation and (c) lexeme (as a
word and sense pair) representations. A key feature of our work is the finding
that a sentence containing a target word is well represented by a low rank
subspace, instead of a point in a vector space. We then show that the subspaces
associated with a particular sense of the target word tend to intersect over a
line (one-dimensional subspace), which we use to disambiguate senses using a
clustering algorithm that harnesses the Grassmannian geometry of the
representations. The disambiguation algorithm, which we call $K$-Grassmeans,
leads to a procedure to label the different senses of the target word in the
corpus -- yielding lexeme vector representations, all in an unsupervised manner
starting from a large (Wikipedia) corpus in English. Apart from several
prototypical target (word,sense) examples and a host of empirical studies to
intuit and justify the various geometric representations, we validate our
algorithms on standard sense induction and disambiguation datasets and present
new state-of-the-art results.
| no_new_dataset | 0.947088 |
1610.07570 | Christoforos Charalambous | Christoforos C. Charalambous and Anil A. Bharath | A data augmentation methodology for training machine/deep learning gait
recognition algorithms | The paper and supplementary material are available on
http://www.bmva.org/bmvc/2016/papers/paper110/index.html Dataset is available
on http://www.bicv.org/datasets/m Proceedings of the BMVC 2016 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | There are several confounding factors that can reduce the accuracy of gait
recognition systems. These factors can reduce the distinctiveness, or alter the
features used to characterise gait, they include variations in clothing,
lighting, pose and environment, such as the walking surface. Full invariance to
all confounding factors is challenging in the absence of high-quality labelled
training data. We introduce a simulation-based methodology and a
subject-specific dataset which can be used for generating synthetic video
frames and sequences for data augmentation. With this methodology, we generated
a multi-modal dataset. In addition, we supply simulation files that provide the
ability to simultaneously sample from several confounding variables. The basis
of the data is real motion capture data of subjects walking and running on a
treadmill at different speeds. Results from gait recognition experiments
suggest that information about the identity of subjects is retained within
synthetically generated examples. The dataset and methodology allow studies
into fully-invariant identity recognition spanning a far greater number of
observation conditions than would otherwise be possible.
| [
{
"version": "v1",
"created": "Mon, 24 Oct 2016 19:35:35 GMT"
}
] | 2016-10-25T00:00:00 | [
[
"Charalambous",
"Christoforos C.",
""
],
[
"Bharath",
"Anil A.",
""
]
] | TITLE: A data augmentation methodology for training machine/deep learning gait
recognition algorithms
ABSTRACT: There are several confounding factors that can reduce the accuracy of gait
recognition systems. These factors can reduce the distinctiveness, or alter the
features used to characterise gait, they include variations in clothing,
lighting, pose and environment, such as the walking surface. Full invariance to
all confounding factors is challenging in the absence of high-quality labelled
training data. We introduce a simulation-based methodology and a
subject-specific dataset which can be used for generating synthetic video
frames and sequences for data augmentation. With this methodology, we generated
a multi-modal dataset. In addition, we supply simulation files that provide the
ability to simultaneously sample from several confounding variables. The basis
of the data is real motion capture data of subjects walking and running on a
treadmill at different speeds. Results from gait recognition experiments
suggest that information about the identity of subjects is retained within
synthetically generated examples. The dataset and methodology allow studies
into fully-invariant identity recognition spanning a far greater number of
observation conditions than would otherwise be possible.
| new_dataset | 0.962603 |
1603.04571 | Harry Crane | Harry Crane and Walter Dempsey | Edge exchangeable models for network data | 35 pages; 8 figures; previously cited under title "Edge exchangeable
network models and the power law" in arXiv:1509.08185 and elsewhere | null | null | null | math.ST cs.SI physics.soc-ph stat.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Exchangeable models for countable vertex-labeled graphs cannot replicate the
large sample behaviors of sparsity and power law degree distribution observed
in many network datasets. Out of this mathematical impossibility emerges the
question of how network data can be modeled in a way that reflects known
empirical behaviors and respects basic statistical principles. We address this
question by observing that edges, not vertices, act as the statistical units in
networks constructed from interaction data, making a theory of edge-labeled
networks more natural for many applications. In this context we introduce the
concept of {\em edge exchangeability}, which unlike its vertex exchangeable
counterpart admits models for networks with sparse and/or power law structure.
Our characterization of edge exchangeable networks gives rise to a class of
nonparametric models, akin to graphon models in the vertex exchangeable
setting. Within this class, we identify a tractable family of distributions
with a clear interpretation and suitable theoretical properties, whose
significance in estimation, prediction, and testing we demonstrate.
| [
{
"version": "v1",
"created": "Tue, 15 Mar 2016 06:54:36 GMT"
},
{
"version": "v2",
"created": "Thu, 17 Mar 2016 21:20:10 GMT"
},
{
"version": "v3",
"created": "Sat, 26 Mar 2016 13:38:56 GMT"
},
{
"version": "v4",
"created": "Fri, 21 Oct 2016 16:20:48 GMT"
}
] | 2016-10-24T00:00:00 | [
[
"Crane",
"Harry",
""
],
[
"Dempsey",
"Walter",
""
]
] | TITLE: Edge exchangeable models for network data
ABSTRACT: Exchangeable models for countable vertex-labeled graphs cannot replicate the
large sample behaviors of sparsity and power law degree distribution observed
in many network datasets. Out of this mathematical impossibility emerges the
question of how network data can be modeled in a way that reflects known
empirical behaviors and respects basic statistical principles. We address this
question by observing that edges, not vertices, act as the statistical units in
networks constructed from interaction data, making a theory of edge-labeled
networks more natural for many applications. In this context we introduce the
concept of {\em edge exchangeability}, which unlike its vertex exchangeable
counterpart admits models for networks with sparse and/or power law structure.
Our characterization of edge exchangeable networks gives rise to a class of
nonparametric models, akin to graphon models in the vertex exchangeable
setting. Within this class, we identify a tractable family of distributions
with a clear interpretation and suitable theoretical properties, whose
significance in estimation, prediction, and testing we demonstrate.
| no_new_dataset | 0.95222 |
1606.06135 | Markus Rempfler | Markus Rempfler, Bjoern Andres, Bjoern H. Menze | The Minimum Cost Connected Subgraph Problem in Medical Image Analysis | accepted at MICCAI 2016 | null | 10.1007/978-3-319-46726-9_46 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Several important tasks in medical image analysis can be stated in the form
of an optimization problem whose feasible solutions are connected subgraphs.
Examples include the reconstruction of neural or vascular structures under
connectedness constraints. We discuss the minimum cost connected subgraph
(MCCS) problem and its approximations from the perspective of medical
applications. We propose a) objective-dependent constraints and b) novel
constraint generation schemes to solve this optimization problem exactly by
means of a branch-and-cut algorithm. These are shown to improve scalability and
allow us to solve instances of two medical benchmark datasets to optimality for
the first time. This enables us to perform a quantitative comparison between
exact and approximative algorithms, where we identify the geodesic tree
algorithm as an excellent alternative to exact inference on the examined
datasets.
| [
{
"version": "v1",
"created": "Mon, 20 Jun 2016 14:22:31 GMT"
}
] | 2016-10-24T00:00:00 | [
[
"Rempfler",
"Markus",
""
],
[
"Andres",
"Bjoern",
""
],
[
"Menze",
"Bjoern H.",
""
]
] | TITLE: The Minimum Cost Connected Subgraph Problem in Medical Image Analysis
ABSTRACT: Several important tasks in medical image analysis can be stated in the form
of an optimization problem whose feasible solutions are connected subgraphs.
Examples include the reconstruction of neural or vascular structures under
connectedness constraints. We discuss the minimum cost connected subgraph
(MCCS) problem and its approximations from the perspective of medical
applications. We propose a) objective-dependent constraints and b) novel
constraint generation schemes to solve this optimization problem exactly by
means of a branch-and-cut algorithm. These are shown to improve scalability and
allow us to solve instances of two medical benchmark datasets to optimality for
the first time. This enables us to perform a quantitative comparison between
exact and approximative algorithms, where we identify the geodesic tree
algorithm as an excellent alternative to exact inference on the examined
datasets.
| no_new_dataset | 0.948394 |
1610.05287 | Jun Xu | Jun Xu, Gurkan Solmaz, Rouhollah Rahmatizadeh, Damla Turgut, and
Ladislau Boloni | Internet of Things Applications: Animal Monitoring with Unmanned Aerial
Vehicle | null | null | null | null | cs.AI cs.NI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In animal monitoring applications, both animal detection and their movement
prediction are major tasks. While a variety of animal monitoring strategies
exist, most of them rely on mounting devices. However, in real world, it is
difficult to find these animals and install mounting devices. In this paper, we
propose an animal monitoring application by utilizing wireless sensor networks
(WSNs) and unmanned aerial vehicle (UAV). The objective of the application is
to detect locations of endangered species in large-scale wildlife areas and
monitor movement of animals without any attached devices. In this application,
sensors deployed throughout the observation area are responsible for gathering
animal information. The UAV flies above the observation area and collects the
information from sensors. To achieve the information efficiently, we propose a
path planning approach for the UAV based on a Markov decision process (MDP)
model. The UAV receives a certain amount of reward from an area if some animals
are detected at that location. We solve the MDP using Q-learning such that the
UAV prefers going to those areas that animals are detected before. Meanwhile,
the UAV explores other areas as well to cover the entire network and detects
changes in the animal positions. We first define the mathematical model
underlying the animal monitoring problem in terms of the value of information
(VoI) and rewards. We propose a network model including clusters of sensor
nodes and a single UAV that acts as a mobile sink and visits the clusters.
Then, one MDP-based path planning approach is designed to maximize the VoI
while reducing message delays. The effectiveness of the proposed approach is
evaluated using two real-world movement datasets of zebras and leopard.
Simulation results show that our approach outperforms greedy, random heuristics
and the path planning based on the traveling salesman problem.
| [
{
"version": "v1",
"created": "Mon, 17 Oct 2016 19:39:23 GMT"
},
{
"version": "v2",
"created": "Thu, 20 Oct 2016 20:24:58 GMT"
}
] | 2016-10-24T00:00:00 | [
[
"Xu",
"Jun",
""
],
[
"Solmaz",
"Gurkan",
""
],
[
"Rahmatizadeh",
"Rouhollah",
""
],
[
"Turgut",
"Damla",
""
],
[
"Boloni",
"Ladislau",
""
]
] | TITLE: Internet of Things Applications: Animal Monitoring with Unmanned Aerial
Vehicle
ABSTRACT: In animal monitoring applications, both animal detection and their movement
prediction are major tasks. While a variety of animal monitoring strategies
exist, most of them rely on mounting devices. However, in real world, it is
difficult to find these animals and install mounting devices. In this paper, we
propose an animal monitoring application by utilizing wireless sensor networks
(WSNs) and unmanned aerial vehicle (UAV). The objective of the application is
to detect locations of endangered species in large-scale wildlife areas and
monitor movement of animals without any attached devices. In this application,
sensors deployed throughout the observation area are responsible for gathering
animal information. The UAV flies above the observation area and collects the
information from sensors. To achieve the information efficiently, we propose a
path planning approach for the UAV based on a Markov decision process (MDP)
model. The UAV receives a certain amount of reward from an area if some animals
are detected at that location. We solve the MDP using Q-learning such that the
UAV prefers going to those areas that animals are detected before. Meanwhile,
the UAV explores other areas as well to cover the entire network and detects
changes in the animal positions. We first define the mathematical model
underlying the animal monitoring problem in terms of the value of information
(VoI) and rewards. We propose a network model including clusters of sensor
nodes and a single UAV that acts as a mobile sink and visits the clusters.
Then, one MDP-based path planning approach is designed to maximize the VoI
while reducing message delays. The effectiveness of the proposed approach is
evaluated using two real-world movement datasets of zebras and leopard.
Simulation results show that our approach outperforms greedy, random heuristics
and the path planning based on the traveling salesman problem.
| no_new_dataset | 0.947817 |
1610.06669 | Chris Mattmann | Chris Mattmann and Madhav Sharan | Scalable Pooled Time Series of Big Video Data from the Deep Web | 7 pages, 5 figures | null | null | null | cs.CV | http://creativecommons.org/publicdomain/zero/1.0/ | We contribute a scalable implementation of Ryoo et al's Pooled Time Series
algorithm from CVPR 2015. The updated algorithm has been evaluated on a large
and diverse dataset of approximately 6800 videos collected from a crawl of the
deep web related to human trafficking on DARPA's MEMEX effort. We describe the
properties of Pooled Time Series and the motivation for using it to relate
videos collected from the deep web. We highlight issues that we found while
running Pooled Time Series on larger datasets and discuss solutions for those
issues. Our solution centers are re-imagining Pooled Time Series as a
Hadoop-based algorithm in which we compute portions of the eventual solution in
parallel on large commodity clusters. We demonstrate that our new Hadoop-based
algorithm works well on the 6800 video dataset and shares all of the properties
described in the CVPR 2015 paper. We suggest avenues of future work in the
project.
| [
{
"version": "v1",
"created": "Fri, 21 Oct 2016 04:43:52 GMT"
}
] | 2016-10-24T00:00:00 | [
[
"Mattmann",
"Chris",
""
],
[
"Sharan",
"Madhav",
""
]
] | TITLE: Scalable Pooled Time Series of Big Video Data from the Deep Web
ABSTRACT: We contribute a scalable implementation of Ryoo et al's Pooled Time Series
algorithm from CVPR 2015. The updated algorithm has been evaluated on a large
and diverse dataset of approximately 6800 videos collected from a crawl of the
deep web related to human trafficking on DARPA's MEMEX effort. We describe the
properties of Pooled Time Series and the motivation for using it to relate
videos collected from the deep web. We highlight issues that we found while
running Pooled Time Series on larger datasets and discuss solutions for those
issues. Our solution centers are re-imagining Pooled Time Series as a
Hadoop-based algorithm in which we compute portions of the eventual solution in
parallel on large commodity clusters. We demonstrate that our new Hadoop-based
algorithm works well on the 6800 video dataset and shares all of the properties
described in the CVPR 2015 paper. We suggest avenues of future work in the
project.
| no_new_dataset | 0.940898 |
1610.06856 | Ethan Rudd | Khudran Alzhrani, Ethan M. Rudd, Terrance E. Boult, and C. Edward Chow | Automated Big Text Security Classification | Pre-print of Best Paper Award IEEE Intelligence and Security
Informatics (ISI) 2016 Manuscript | 2016 IEEE International Conference on Intelligence and Security
Informatics (ISI) | null | null | cs.CR cs.AI cs.CL cs.CY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In recent years, traditional cybersecurity safeguards have proven ineffective
against insider threats. Famous cases of sensitive information leaks caused by
insiders, including the WikiLeaks release of diplomatic cables and the Edward
Snowden incident, have greatly harmed the U.S. government's relationship with
other governments and with its own citizens. Data Leak Prevention (DLP) is a
solution for detecting and preventing information leaks from within an
organization's network. However, state-of-art DLP detection models are only
able to detect very limited types of sensitive information, and research in the
field has been hindered due to the lack of available sensitive texts. Many
researchers have focused on document-based detection with artificially labeled
"confidential documents" for which security labels are assigned to the entire
document, when in reality only a portion of the document is sensitive. This
type of whole-document based security labeling increases the chances of
preventing authorized users from accessing non-sensitive information within
sensitive documents. In this paper, we introduce Automated Classification
Enabled by Security Similarity (ACESS), a new and innovative detection model
that penetrates the complexity of big text security classification/detection.
To analyze the ACESS system, we constructed a novel dataset, containing
formerly classified paragraphs from diplomatic cables made public by the
WikiLeaks organization. To our knowledge this paper is the first to analyze a
dataset that contains actual formerly sensitive information annotated at
paragraph granularity.
| [
{
"version": "v1",
"created": "Fri, 21 Oct 2016 16:53:09 GMT"
}
] | 2016-10-24T00:00:00 | [
[
"Alzhrani",
"Khudran",
""
],
[
"Rudd",
"Ethan M.",
""
],
[
"Boult",
"Terrance E.",
""
],
[
"Chow",
"C. Edward",
""
]
] | TITLE: Automated Big Text Security Classification
ABSTRACT: In recent years, traditional cybersecurity safeguards have proven ineffective
against insider threats. Famous cases of sensitive information leaks caused by
insiders, including the WikiLeaks release of diplomatic cables and the Edward
Snowden incident, have greatly harmed the U.S. government's relationship with
other governments and with its own citizens. Data Leak Prevention (DLP) is a
solution for detecting and preventing information leaks from within an
organization's network. However, state-of-art DLP detection models are only
able to detect very limited types of sensitive information, and research in the
field has been hindered due to the lack of available sensitive texts. Many
researchers have focused on document-based detection with artificially labeled
"confidential documents" for which security labels are assigned to the entire
document, when in reality only a portion of the document is sensitive. This
type of whole-document based security labeling increases the chances of
preventing authorized users from accessing non-sensitive information within
sensitive documents. In this paper, we introduce Automated Classification
Enabled by Security Similarity (ACESS), a new and innovative detection model
that penetrates the complexity of big text security classification/detection.
To analyze the ACESS system, we constructed a novel dataset, containing
formerly classified paragraphs from diplomatic cables made public by the
WikiLeaks organization. To our knowledge this paper is the first to analyze a
dataset that contains actual formerly sensitive information annotated at
paragraph granularity.
| new_dataset | 0.967564 |
1610.06907 | Hyungtae Lee | Yilun Cao and Hyungtae Lee and Heesung Kwon | Enhanced Object Detection via Fusion With Prior Beliefs from Image
Classification | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we introduce a novel fusion method that can enhance object
detection performance by fusing decisions from two different types of computer
vision tasks: object detection and image classification. In the proposed work,
the class label of an image obtained from image classification is viewed as
prior knowledge about existence or non-existence of certain objects. The prior
knowledge is then fused with the decisions of object detection to improve
detection accuracy by mitigating false positives of an object detector that are
strongly contradicted with the prior knowledge. A recently introduced novel
fusion approach called dynamic belief fusion (DBF) is used to fuse the detector
output with the classification prior. Experimental results show that the
detection performance of all the detection algorithms used in the proposed work
is improved on benchmark datasets via the proposed fusion framework.
| [
{
"version": "v1",
"created": "Fri, 21 Oct 2016 19:38:45 GMT"
}
] | 2016-10-24T00:00:00 | [
[
"Cao",
"Yilun",
""
],
[
"Lee",
"Hyungtae",
""
],
[
"Kwon",
"Heesung",
""
]
] | TITLE: Enhanced Object Detection via Fusion With Prior Beliefs from Image
Classification
ABSTRACT: In this paper, we introduce a novel fusion method that can enhance object
detection performance by fusing decisions from two different types of computer
vision tasks: object detection and image classification. In the proposed work,
the class label of an image obtained from image classification is viewed as
prior knowledge about existence or non-existence of certain objects. The prior
knowledge is then fused with the decisions of object detection to improve
detection accuracy by mitigating false positives of an object detector that are
strongly contradicted with the prior knowledge. A recently introduced novel
fusion approach called dynamic belief fusion (DBF) is used to fuse the detector
output with the classification prior. Experimental results show that the
detection performance of all the detection algorithms used in the proposed work
is improved on benchmark datasets via the proposed fusion framework.
| no_new_dataset | 0.952264 |
1501.02891 | Jan Korbel | Renata Rycht\'arikov\'a, Jan Korbel, Petr Mach\'a\v{c}ek, Petr
C\'isa\v{r}, Jan Urban, Dmytro Soloviov and Dalibor \v{S}tys | Point Information Gain and Multidimensional Data Analysis | 16 pages, 6 figures | Entropy 2016, 18(10), 372 | 10.3390/e18100372 | null | physics.data-an | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We generalize the Point information gain (PIG) and derived quantities, i.e.
Point information entropy (PIE) and Point information entropy density (PIED),
for the case of R\'enyi entropy and simulate the behavior of PIG for typical
distributions. We also use these methods for the analysis of multidimensional
datasets. We demonstrate the main properties of PIE/PIED spectra for the real
data on the example of several images, and discuss possible further utilization
in other fields of data processing.
| [
{
"version": "v1",
"created": "Tue, 13 Jan 2015 07:05:23 GMT"
},
{
"version": "v2",
"created": "Wed, 14 Jan 2015 07:42:46 GMT"
},
{
"version": "v3",
"created": "Mon, 26 Jan 2015 08:50:25 GMT"
},
{
"version": "v4",
"created": "Tue, 24 Mar 2015 19:29:06 GMT"
},
{
"version": "v5",
"created": "Wed, 19 Oct 2016 05:30:19 GMT"
}
] | 2016-10-21T00:00:00 | [
[
"Rychtáriková",
"Renata",
""
],
[
"Korbel",
"Jan",
""
],
[
"Macháček",
"Petr",
""
],
[
"Císař",
"Petr",
""
],
[
"Urban",
"Jan",
""
],
[
"Soloviov",
"Dmytro",
""
],
[
"Štys",
"Dalibor",
""
]
] | TITLE: Point Information Gain and Multidimensional Data Analysis
ABSTRACT: We generalize the Point information gain (PIG) and derived quantities, i.e.
Point information entropy (PIE) and Point information entropy density (PIED),
for the case of R\'enyi entropy and simulate the behavior of PIG for typical
distributions. We also use these methods for the analysis of multidimensional
datasets. We demonstrate the main properties of PIE/PIED spectra for the real
data on the example of several images, and discuss possible further utilization
in other fields of data processing.
| no_new_dataset | 0.950227 |
1610.06210 | Rik Koncel-Kedziorski | Rik Koncel-Kedziorski, Ioannis Konstas, Luke Zettlemoyer, and Hannaneh
Hajishirzi | A Theme-Rewriting Approach for Generating Algebra Word Problems | To appear EMNLP 2016 | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Texts present coherent stories that have a particular theme or overall
setting, for example science fiction or western. In this paper, we present a
text generation method called {\it rewriting} that edits existing
human-authored narratives to change their theme without changing the underlying
story. We apply the approach to math word problems, where it might help
students stay more engaged by quickly transforming all of their homework
assignments to the theme of their favorite movie without changing the math
concepts that are being taught. Our rewriting method uses a two-stage decoding
process, which proposes new words from the target theme and scores the
resulting stories according to a number of factors defining aspects of
syntactic, semantic, and thematic coherence. Experiments demonstrate that the
final stories typically represent the new theme well while still testing the
original math concepts, outperforming a number of baselines. We also release a
new dataset of human-authored rewrites of math word problems in several themes.
| [
{
"version": "v1",
"created": "Wed, 19 Oct 2016 20:49:23 GMT"
}
] | 2016-10-21T00:00:00 | [
[
"Koncel-Kedziorski",
"Rik",
""
],
[
"Konstas",
"Ioannis",
""
],
[
"Zettlemoyer",
"Luke",
""
],
[
"Hajishirzi",
"Hannaneh",
""
]
] | TITLE: A Theme-Rewriting Approach for Generating Algebra Word Problems
ABSTRACT: Texts present coherent stories that have a particular theme or overall
setting, for example science fiction or western. In this paper, we present a
text generation method called {\it rewriting} that edits existing
human-authored narratives to change their theme without changing the underlying
story. We apply the approach to math word problems, where it might help
students stay more engaged by quickly transforming all of their homework
assignments to the theme of their favorite movie without changing the math
concepts that are being taught. Our rewriting method uses a two-stage decoding
process, which proposes new words from the target theme and scores the
resulting stories according to a number of factors defining aspects of
syntactic, semantic, and thematic coherence. Experiments demonstrate that the
final stories typically represent the new theme well while still testing the
original math concepts, outperforming a number of baselines. We also release a
new dataset of human-authored rewrites of math word problems in several themes.
| new_dataset | 0.96157 |
1610.06249 | Kien Do | Kien Do and Truyen Tran and Svetha Venkatesh | Multilevel Anomaly Detection for Mixed Data | 9 pages | null | null | null | cs.LG cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Anomalies are those deviating from the norm. Unsupervised anomaly detection
often translates to identifying low density regions. Major problems arise when
data is high-dimensional and mixed of discrete and continuous attributes. We
propose MIXMAD, which stands for MIXed data Multilevel Anomaly Detection, an
ensemble method that estimates the sparse regions across multiple levels of
abstraction of mixed data. The hypothesis is for domains where multiple data
abstractions exist, a data point may be anomalous with respect to the raw
representation or more abstract representations. To this end, our method
sequentially constructs an ensemble of Deep Belief Nets (DBNs) with varying
depths. Each DBN is an energy-based detector at a predefined abstraction level.
At the bottom level of each DBN, there is a Mixed-variate Restricted Boltzmann
Machine that models the density of mixed data. Predictions across the ensemble
are finally combined via rank aggregation. The proposed MIXMAD is evaluated on
high-dimensional realworld datasets of different characteristics. The results
demonstrate that for anomaly detection, (a) multilevel abstraction of
high-dimensional and mixed data is a sensible strategy, and (b) empirically,
MIXMAD is superior to popular unsupervised detection methods for both
homogeneous and mixed data.
| [
{
"version": "v1",
"created": "Thu, 20 Oct 2016 00:04:55 GMT"
}
] | 2016-10-21T00:00:00 | [
[
"Do",
"Kien",
""
],
[
"Tran",
"Truyen",
""
],
[
"Venkatesh",
"Svetha",
""
]
] | TITLE: Multilevel Anomaly Detection for Mixed Data
ABSTRACT: Anomalies are those deviating from the norm. Unsupervised anomaly detection
often translates to identifying low density regions. Major problems arise when
data is high-dimensional and mixed of discrete and continuous attributes. We
propose MIXMAD, which stands for MIXed data Multilevel Anomaly Detection, an
ensemble method that estimates the sparse regions across multiple levels of
abstraction of mixed data. The hypothesis is for domains where multiple data
abstractions exist, a data point may be anomalous with respect to the raw
representation or more abstract representations. To this end, our method
sequentially constructs an ensemble of Deep Belief Nets (DBNs) with varying
depths. Each DBN is an energy-based detector at a predefined abstraction level.
At the bottom level of each DBN, there is a Mixed-variate Restricted Boltzmann
Machine that models the density of mixed data. Predictions across the ensemble
are finally combined via rank aggregation. The proposed MIXMAD is evaluated on
high-dimensional realworld datasets of different characteristics. The results
demonstrate that for anomaly detection, (a) multilevel abstraction of
high-dimensional and mixed data is a sensible strategy, and (b) empirically,
MIXMAD is superior to popular unsupervised detection methods for both
homogeneous and mixed data.
| no_new_dataset | 0.945298 |
1610.06370 | Georgios Spithourakis | Georgios P. Spithourakis and Steffen E. Petersen and Sebastian Riedel | Clinical Text Prediction with Numerically Grounded Conditional Language
Models | Accepted at the 7th International Workshop on Health Text Mining and
Information Analysis (LOUHI) EMNLP 2016 | null | null | null | cs.CL cs.HC cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Assisted text input techniques can save time and effort and improve text
quality. In this paper, we investigate how grounded and conditional extensions
to standard neural language models can bring improvements in the tasks of word
prediction and completion. These extensions incorporate a structured knowledge
base and numerical values from the text into the context used to predict the
next word. Our automated evaluation on a clinical dataset shows extended models
significantly outperform standard models. Our best system uses both
conditioning and grounding, because of their orthogonal benefits. For word
prediction with a list of 5 suggestions, it improves recall from 25.03% to
71.28% and for word completion it improves keystroke savings from 34.35% to
44.81%, where theoretical bound for this dataset is 58.78%. We also perform a
qualitative investigation of how models with lower perplexity occasionally fare
better at the tasks. We found that at test time numbers have more influence on
the document level than on individual word probabilities.
| [
{
"version": "v1",
"created": "Thu, 20 Oct 2016 11:48:30 GMT"
}
] | 2016-10-21T00:00:00 | [
[
"Spithourakis",
"Georgios P.",
""
],
[
"Petersen",
"Steffen E.",
""
],
[
"Riedel",
"Sebastian",
""
]
] | TITLE: Clinical Text Prediction with Numerically Grounded Conditional Language
Models
ABSTRACT: Assisted text input techniques can save time and effort and improve text
quality. In this paper, we investigate how grounded and conditional extensions
to standard neural language models can bring improvements in the tasks of word
prediction and completion. These extensions incorporate a structured knowledge
base and numerical values from the text into the context used to predict the
next word. Our automated evaluation on a clinical dataset shows extended models
significantly outperform standard models. Our best system uses both
conditioning and grounding, because of their orthogonal benefits. For word
prediction with a list of 5 suggestions, it improves recall from 25.03% to
71.28% and for word completion it improves keystroke savings from 34.35% to
44.81%, where theoretical bound for this dataset is 58.78%. We also perform a
qualitative investigation of how models with lower perplexity occasionally fare
better at the tasks. We found that at test time numbers have more influence on
the document level than on individual word probabilities.
| no_new_dataset | 0.951051 |
1610.06494 | Ahmed Ibrahim | Ahmed Ibrahim, A. Lynn Abbott, Mohamed E. Hussein | An Image Dataset of Text Patches in Everyday Scenes | Accepted in the 12th International Symposium on Visual Computing
(ISVC'16) | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper describes a dataset containing small images of text from everyday
scenes. The purpose of the dataset is to support the development of new
automated systems that can detect and analyze text. Although much research has
been devoted to text detection and recognition in scanned documents, relatively
little attention has been given to text detection in other types of images,
such as photographs that are posted on social-media sites. This new dataset,
known as COCO-Text-Patch, contains approximately 354,000 small images that are
each labeled as "text" or "non-text". This dataset particularly addresses the
problem of text verification, which is an essential stage in the end-to-end
text detection and recognition pipeline. In order to evaluate the utility of
this dataset, it has been used to train two deep convolution neural networks to
distinguish text from non-text. One network is inspired by the GoogLeNet
architecture, and the second one is based on CaffeNet. Accuracy levels of 90.2%
and 90.9% were obtained using the two networks, respectively. All of the
images, source code, and deep-learning trained models described in this paper
will be publicly available
| [
{
"version": "v1",
"created": "Thu, 20 Oct 2016 16:38:42 GMT"
}
] | 2016-10-21T00:00:00 | [
[
"Ibrahim",
"Ahmed",
""
],
[
"Abbott",
"A. Lynn",
""
],
[
"Hussein",
"Mohamed E.",
""
]
] | TITLE: An Image Dataset of Text Patches in Everyday Scenes
ABSTRACT: This paper describes a dataset containing small images of text from everyday
scenes. The purpose of the dataset is to support the development of new
automated systems that can detect and analyze text. Although much research has
been devoted to text detection and recognition in scanned documents, relatively
little attention has been given to text detection in other types of images,
such as photographs that are posted on social-media sites. This new dataset,
known as COCO-Text-Patch, contains approximately 354,000 small images that are
each labeled as "text" or "non-text". This dataset particularly addresses the
problem of text verification, which is an essential stage in the end-to-end
text detection and recognition pipeline. In order to evaluate the utility of
this dataset, it has been used to train two deep convolution neural networks to
distinguish text from non-text. One network is inspired by the GoogLeNet
architecture, and the second one is based on CaffeNet. Accuracy levels of 90.2%
and 90.9% were obtained using the two networks, respectively. All of the
images, source code, and deep-learning trained models described in this paper
will be publicly available
| new_dataset | 0.967225 |
1610.04154 | Sergio Ram\'irez-Gallego | Sergio Ram\'irez-Gallego, H\'ector Mouri\~no-Tal\'in, David
Mart\'inez-Rego, Ver\'onica Bol\'on-Canedo, Jos\'e Manuel Ben\'itez, Amparo
Alonso-Betanzos, Francisco Herrera | An Information Theoretic Feature Selection Framework for Big Data under
Apache Spark | null | null | null | null | cs.AI cs.DC cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the advent of extremely high dimensional datasets, dimensionality
reduction techniques are becoming mandatory. Among many techniques, feature
selection has been growing in interest as an important tool to identify
relevant features on huge datasets --both in number of instances and
features--. The purpose of this work is to demonstrate that standard feature
selection methods can be parallelized in Big Data platforms like Apache Spark,
boosting both performance and accuracy. We thus propose a distributed
implementation of a generic feature selection framework which includes a wide
group of well-known Information Theoretic methods. Experimental results on a
wide set of real-world datasets show that our distributed framework is capable
of dealing with ultra-high dimensional datasets as well as those with a huge
number of samples in a short period of time, outperforming the sequential
version in all the cases studied.
| [
{
"version": "v1",
"created": "Thu, 13 Oct 2016 16:17:07 GMT"
},
{
"version": "v2",
"created": "Wed, 19 Oct 2016 16:46:28 GMT"
}
] | 2016-10-20T00:00:00 | [
[
"Ramírez-Gallego",
"Sergio",
""
],
[
"Mouriño-Talín",
"Héctor",
""
],
[
"Martínez-Rego",
"David",
""
],
[
"Bolón-Canedo",
"Verónica",
""
],
[
"Benítez",
"José Manuel",
""
],
[
"Alonso-Betanzos",
"Amparo",
""
],
[
"Herrera",
"Francisco",
""
]
] | TITLE: An Information Theoretic Feature Selection Framework for Big Data under
Apache Spark
ABSTRACT: With the advent of extremely high dimensional datasets, dimensionality
reduction techniques are becoming mandatory. Among many techniques, feature
selection has been growing in interest as an important tool to identify
relevant features on huge datasets --both in number of instances and
features--. The purpose of this work is to demonstrate that standard feature
selection methods can be parallelized in Big Data platforms like Apache Spark,
boosting both performance and accuracy. We thus propose a distributed
implementation of a generic feature selection framework which includes a wide
group of well-known Information Theoretic methods. Experimental results on a
wide set of real-world datasets show that our distributed framework is capable
of dealing with ultra-high dimensional datasets as well as those with a huge
number of samples in a short period of time, outperforming the sequential
version in all the cases studied.
| no_new_dataset | 0.942295 |
1610.05796 | Koray Mancuhan | Koray Mancuhan and Chris Clifton | Decision Tree Classification on Outsourced Data | Presented in the Data Ethics Workshop at the 20th ACM SIGKDD
Conference on Knowledge Discovery and Data Mining | null | null | null | cs.LG cs.CR cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper proposes a client-server decision tree learning method for
outsourced private data. The privacy model is anatomization/fragmentation: the
server sees data values, but the link between sensitive and identifying
information is encrypted with a key known only to clients. Clients have limited
processing and storage capability. Both sensitive and identifying information
thus are stored on the server. The approach presented also retains most
processing at the server, and client-side processing is amortized over
predictions made by the clients. Experiments on various datasets show that the
method produces decision trees approaching the accuracy of a non-private
decision tree, while substantially reducing the client's computing resource
requirements.
| [
{
"version": "v1",
"created": "Tue, 18 Oct 2016 20:49:21 GMT"
}
] | 2016-10-20T00:00:00 | [
[
"Mancuhan",
"Koray",
""
],
[
"Clifton",
"Chris",
""
]
] | TITLE: Decision Tree Classification on Outsourced Data
ABSTRACT: This paper proposes a client-server decision tree learning method for
outsourced private data. The privacy model is anatomization/fragmentation: the
server sees data values, but the link between sensitive and identifying
information is encrypted with a key known only to clients. Clients have limited
processing and storage capability. Both sensitive and identifying information
thus are stored on the server. The approach presented also retains most
processing at the server, and client-side processing is amortized over
predictions made by the clients. Experiments on various datasets show that the
method produces decision trees approaching the accuracy of a non-private
decision tree, while substantially reducing the client's computing resource
requirements.
| no_new_dataset | 0.947914 |
1610.05815 | Koray Mancuhan | Koray Mancuhan and Chris Clifton | Statistical Learning Theory Approach for Data Classification with
l-diversity | Technical Report | null | null | null | cs.LG cs.CR cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Corporations are retaining ever-larger corpuses of personal data; the
frequency or breaches and corresponding privacy impact have been rising
accordingly. One way to mitigate this risk is through use of anonymized data,
limiting the exposure of individual data to only where it is absolutely needed.
This would seem particularly appropriate for data mining, where the goal is
generalizable knowledge rather than data on specific individuals. In practice,
corporate data miners often insist on original data, for fear that they might
"miss something" with anonymized or differentially private approaches. This
paper provides a theoretical justification for the use of anonymized data.
Specifically, we show that a support vector classifier trained on anatomized
data satisfying l-diversity should be expected to do as well as on the original
data. Anatomy preserves all data values, but introduces uncertainty in the
mapping between identifying and sensitive values, thus satisfying l-diversity.
The theoretical effectiveness of the proposed approach is validated using
several publicly available datasets, showing that we outperform the state of
the art for support vector classification using training data protected by
k-anonymity, and are comparable to learning on the original data.
| [
{
"version": "v1",
"created": "Tue, 18 Oct 2016 22:14:27 GMT"
}
] | 2016-10-20T00:00:00 | [
[
"Mancuhan",
"Koray",
""
],
[
"Clifton",
"Chris",
""
]
] | TITLE: Statistical Learning Theory Approach for Data Classification with
l-diversity
ABSTRACT: Corporations are retaining ever-larger corpuses of personal data; the
frequency or breaches and corresponding privacy impact have been rising
accordingly. One way to mitigate this risk is through use of anonymized data,
limiting the exposure of individual data to only where it is absolutely needed.
This would seem particularly appropriate for data mining, where the goal is
generalizable knowledge rather than data on specific individuals. In practice,
corporate data miners often insist on original data, for fear that they might
"miss something" with anonymized or differentially private approaches. This
paper provides a theoretical justification for the use of anonymized data.
Specifically, we show that a support vector classifier trained on anatomized
data satisfying l-diversity should be expected to do as well as on the original
data. Anatomy preserves all data values, but introduces uncertainty in the
mapping between identifying and sensitive values, thus satisfying l-diversity.
The theoretical effectiveness of the proposed approach is validated using
several publicly available datasets, showing that we outperform the state of
the art for support vector classification using training data protected by
k-anonymity, and are comparable to learning on the original data.
| no_new_dataset | 0.953708 |
1610.05883 | Thanh Nguyen | Duc Thanh Nguyen, Binh-Son Hua, Lap-Fai Yu, and Sai-Kit Yeung | A Robust 3D-2D Interactive Tool for Scene Segmentation and Annotation | 14 pages | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advances of 3D acquisition devices have enabled large-scale
acquisition of 3D scene data. Such data, if completely and well annotated, can
serve as useful ingredients for a wide spectrum of computer vision and graphics
works such as data-driven modeling and scene understanding, object detection
and recognition. However, annotating a vast amount of 3D scene data remains
challenging due to the lack of an effective tool and/or the complexity of 3D
scenes (e.g. clutter, varying illumination conditions). This paper aims to
build a robust annotation tool that effectively and conveniently enables the
segmentation and annotation of massive 3D data. Our tool works by coupling 2D
and 3D information via an interactive framework, through which users can
provide high-level semantic annotation for objects. We have experimented our
tool and found that a typical indoor scene could be well segmented and
annotated in less than 30 minutes by using the tool, as opposed to a few hours
if done manually. Along with the tool, we created a dataset of over a hundred
3D scenes associated with complete annotations using our tool. The tool and
dataset are available at www.scenenn.net.
| [
{
"version": "v1",
"created": "Wed, 19 Oct 2016 06:54:02 GMT"
}
] | 2016-10-20T00:00:00 | [
[
"Nguyen",
"Duc Thanh",
""
],
[
"Hua",
"Binh-Son",
""
],
[
"Yu",
"Lap-Fai",
""
],
[
"Yeung",
"Sai-Kit",
""
]
] | TITLE: A Robust 3D-2D Interactive Tool for Scene Segmentation and Annotation
ABSTRACT: Recent advances of 3D acquisition devices have enabled large-scale
acquisition of 3D scene data. Such data, if completely and well annotated, can
serve as useful ingredients for a wide spectrum of computer vision and graphics
works such as data-driven modeling and scene understanding, object detection
and recognition. However, annotating a vast amount of 3D scene data remains
challenging due to the lack of an effective tool and/or the complexity of 3D
scenes (e.g. clutter, varying illumination conditions). This paper aims to
build a robust annotation tool that effectively and conveniently enables the
segmentation and annotation of massive 3D data. Our tool works by coupling 2D
and 3D information via an interactive framework, through which users can
provide high-level semantic annotation for objects. We have experimented our
tool and found that a typical indoor scene could be well segmented and
annotated in less than 30 minutes by using the tool, as opposed to a few hours
if done manually. Along with the tool, we created a dataset of over a hundred
3D scenes associated with complete annotations using our tool. The tool and
dataset are available at www.scenenn.net.
| new_dataset | 0.952926 |
1610.05929 | Luyan Ji | Luyan Ji, Xiurui Geng, Yongchao Zhao, Fuxiang Wang | An automatic bad band preremoval algorithm for hyperspectral imagery | 17 pages, 8 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | For most hyperspectral remote sensing applications, removing bad bands, such
as water absorption bands, is a required preprocessing step. Currently, the
commonly applied method is by visual inspection, which is very time-consuming
and it is easy to overlook some noisy bands. In this study, we find an inherent
connection between target detection algorithms and the corrupted band removal.
As an example, for the matched filter (MF), which is the most widely used
target detection method for hyperspectral data, we present an automatic
MF-based algorithm for bad band identification. The MF detector is a filter
vector, and the resulting filter output is the sum of all bands weighted by the
MF coefficients. Therefore, we can identify bad bands only by using the MF
filter vector itself, the absolute value of whose entry accounts for the
importance of each band for the target detection. For a specific target of
interest, the bands with small MF weights correspond to the noisy or bad ones.
Based on this fact, we develop an automatic bad band preremoval algorithm by
utilizing the average absolute value of MF weights for multiple targets within
a scene. Experiments with three well known hyperspectral datasets show that our
method can always identify the water absorption and other low signal-to-noise
(SNR) bands that are usually chosen as bad bands manually.
| [
{
"version": "v1",
"created": "Wed, 19 Oct 2016 09:31:31 GMT"
}
] | 2016-10-20T00:00:00 | [
[
"Ji",
"Luyan",
""
],
[
"Geng",
"Xiurui",
""
],
[
"Zhao",
"Yongchao",
""
],
[
"Wang",
"Fuxiang",
""
]
] | TITLE: An automatic bad band preremoval algorithm for hyperspectral imagery
ABSTRACT: For most hyperspectral remote sensing applications, removing bad bands, such
as water absorption bands, is a required preprocessing step. Currently, the
commonly applied method is by visual inspection, which is very time-consuming
and it is easy to overlook some noisy bands. In this study, we find an inherent
connection between target detection algorithms and the corrupted band removal.
As an example, for the matched filter (MF), which is the most widely used
target detection method for hyperspectral data, we present an automatic
MF-based algorithm for bad band identification. The MF detector is a filter
vector, and the resulting filter output is the sum of all bands weighted by the
MF coefficients. Therefore, we can identify bad bands only by using the MF
filter vector itself, the absolute value of whose entry accounts for the
importance of each band for the target detection. For a specific target of
interest, the bands with small MF weights correspond to the noisy or bad ones.
Based on this fact, we develop an automatic bad band preremoval algorithm by
utilizing the average absolute value of MF weights for multiple targets within
a scene. Experiments with three well known hyperspectral datasets show that our
method can always identify the water absorption and other low signal-to-noise
(SNR) bands that are usually chosen as bad bands manually.
| no_new_dataset | 0.947769 |
1610.05994 | Jos\'e Fuentes | Jos\'e Fuentes-Sep\'ulveda and Erick Elejalde and Leo Ferres and Diego
Seco | Parallel Construction of Wavelet Trees on Multicore Architectures | This research has received funding from the European Union's Horizon
2020 research and innovation programme under the Marie Sk{\l}odowska-Curie
Actions H2020-MSCA-RISE-2015 BIRDS GA No. 690941 | Knowl Inf Syst (2016) | 10.1007/s10115-016-1000-6 | null | cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The wavelet tree has become a very useful data structure to efficiently
represent and query large volumes of data in many different domains, from
bioinformatics to geographic information systems. One problem with wavelet
trees is their construction time. In this paper, we introduce two algorithms
that reduce the time complexity of a wavelet tree's construction by taking
advantage of nowadays ubiquitous multicore machines.
Our first algorithm constructs all the levels of the wavelet in parallel in
$O(n)$ time and $O(n\lg\sigma + \sigma\lg n)$ bits of working space, where $n$
is the size of the input sequence and $\sigma$ is the size of the alphabet. Our
second algorithm constructs the wavelet tree in a domain-decomposition fashion,
using our first algorithm in each segment, reaching $O(\lg n)$ time and
$O(n\lg\sigma + p\sigma\lg n/\lg\sigma)$ bits of extra space, where $p$ is the
number of available cores. Both algorithms are practical and report good
speedup for large real datasets.
| [
{
"version": "v1",
"created": "Wed, 19 Oct 2016 12:57:48 GMT"
}
] | 2016-10-20T00:00:00 | [
[
"Fuentes-Sepúlveda",
"José",
""
],
[
"Elejalde",
"Erick",
""
],
[
"Ferres",
"Leo",
""
],
[
"Seco",
"Diego",
""
]
] | TITLE: Parallel Construction of Wavelet Trees on Multicore Architectures
ABSTRACT: The wavelet tree has become a very useful data structure to efficiently
represent and query large volumes of data in many different domains, from
bioinformatics to geographic information systems. One problem with wavelet
trees is their construction time. In this paper, we introduce two algorithms
that reduce the time complexity of a wavelet tree's construction by taking
advantage of nowadays ubiquitous multicore machines.
Our first algorithm constructs all the levels of the wavelet in parallel in
$O(n)$ time and $O(n\lg\sigma + \sigma\lg n)$ bits of working space, where $n$
is the size of the input sequence and $\sigma$ is the size of the alphabet. Our
second algorithm constructs the wavelet tree in a domain-decomposition fashion,
using our first algorithm in each segment, reaching $O(\lg n)$ time and
$O(n\lg\sigma + p\sigma\lg n/\lg\sigma)$ bits of extra space, where $p$ is the
number of available cores. Both algorithms are practical and report good
speedup for large real datasets.
| no_new_dataset | 0.947575 |
1610.06049 | Michael Breu{\ss} | Martin B\"ahr, Michael Breu{\ss}, Yvain Qu\'eau, Ali Sharifi
Boroujerdi, Jean-Denis Durou | Fast and Accurate Surface Normal Integration on Non-Rectangular Domains | null | null | null | null | cs.NA cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The integration of surface normals for the purpose of computing the shape of
a surface in 3D space is a classic problem in computer vision. However, even
nowadays it is still a challenging task to devise a method that combines the
flexibility to work on non-trivial computational domains with high accuracy,
robustness and computational efficiency. By uniting a classic approach for
surface normal integration with modern computational techniques we construct a
solver that fulfils these requirements. Building upon the Poisson integration
model we propose to use an iterative Krylov subspace solver as a core step in
tackling the task. While such a method can be very efficient, it may only show
its full potential when combined with a suitable numerical preconditioning and
a problem-specific initialisation. We perform a thorough numerical study in
order to identify an appropriate preconditioner for our purpose. To address the
issue of a suitable initialisation we propose to compute this initial state via
a recently developed fast marching integrator. Detailed numerical experiments
illuminate the benefits of this novel combination. In addition, we show on
real-world photometric stereo datasets that the developed numerical framework
is flexible enough to tackle modern computer vision applications.
| [
{
"version": "v1",
"created": "Wed, 19 Oct 2016 15:01:09 GMT"
}
] | 2016-10-20T00:00:00 | [
[
"Bähr",
"Martin",
""
],
[
"Breuß",
"Michael",
""
],
[
"Quéau",
"Yvain",
""
],
[
"Boroujerdi",
"Ali Sharifi",
""
],
[
"Durou",
"Jean-Denis",
""
]
] | TITLE: Fast and Accurate Surface Normal Integration on Non-Rectangular Domains
ABSTRACT: The integration of surface normals for the purpose of computing the shape of
a surface in 3D space is a classic problem in computer vision. However, even
nowadays it is still a challenging task to devise a method that combines the
flexibility to work on non-trivial computational domains with high accuracy,
robustness and computational efficiency. By uniting a classic approach for
surface normal integration with modern computational techniques we construct a
solver that fulfils these requirements. Building upon the Poisson integration
model we propose to use an iterative Krylov subspace solver as a core step in
tackling the task. While such a method can be very efficient, it may only show
its full potential when combined with a suitable numerical preconditioning and
a problem-specific initialisation. We perform a thorough numerical study in
order to identify an appropriate preconditioner for our purpose. To address the
issue of a suitable initialisation we propose to compute this initial state via
a recently developed fast marching integrator. Detailed numerical experiments
illuminate the benefits of this novel combination. In addition, we show on
real-world photometric stereo datasets that the developed numerical framework
is flexible enough to tackle modern computer vision applications.
| no_new_dataset | 0.936576 |
1610.06072 | Tom Bosc | Tom Bosc | Learning to Learn Neural Networks | presented at "Reasoning, Attention, Memory" workshop, NIPS 2015 | null | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Meta-learning consists in learning learning algorithms. We use a Long Short
Term Memory (LSTM) based network to learn to compute on-line updates of the
parameters of another neural network. These parameters are stored in the cell
state of the LSTM. Our framework allows to compare learned algorithms to
hand-made algorithms within the traditional train and test methodology. In an
experiment, we learn a learning algorithm for a one-hidden layer Multi-Layer
Perceptron (MLP) on non-linearly separable datasets. The learned algorithm is
able to update parameters of both layers and generalise well on similar
datasets.
| [
{
"version": "v1",
"created": "Wed, 19 Oct 2016 15:46:30 GMT"
}
] | 2016-10-20T00:00:00 | [
[
"Bosc",
"Tom",
""
]
] | TITLE: Learning to Learn Neural Networks
ABSTRACT: Meta-learning consists in learning learning algorithms. We use a Long Short
Term Memory (LSTM) based network to learn to compute on-line updates of the
parameters of another neural network. These parameters are stored in the cell
state of the LSTM. Our framework allows to compare learned algorithms to
hand-made algorithms within the traditional train and test methodology. In an
experiment, we learn a learning algorithm for a one-hidden layer Multi-Layer
Perceptron (MLP) on non-linearly separable datasets. The learned algorithm is
able to update parameters of both layers and generalise well on similar
datasets.
| no_new_dataset | 0.945096 |
1504.07107 | Wenbo Hu | Wenbo Hu, Jun Zhu, Bo Zhang | Fast Sampling for Bayesian Max-Margin Models | null | null | null | null | stat.ML cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Bayesian max-margin models have shown superiority in various practical
applications, such as text categorization, collaborative prediction, social
network link prediction and crowdsourcing, and they conjoin the flexibility of
Bayesian modeling and predictive strengths of max-margin learning. However,
Monte Carlo sampling for these models still remains challenging, especially for
applications that involve large-scale datasets. In this paper, we present the
stochastic subgradient Hamiltonian Monte Carlo (HMC) methods, which are easy to
implement and computationally efficient. We show the approximate detailed
balance property of subgradient HMC which reveals a natural and validated
generalization of the ordinary HMC. Furthermore, we investigate the variants
that use stochastic subsampling and thermostats for better scalability and
mixing. Using stochastic subgradient Markov Chain Monte Carlo (MCMC), we
efficiently solve the posterior inference task of various Bayesian max-margin
models and extensive experimental results demonstrate the effectiveness of our
approach.
| [
{
"version": "v1",
"created": "Mon, 27 Apr 2015 14:29:40 GMT"
},
{
"version": "v2",
"created": "Wed, 29 Apr 2015 12:28:41 GMT"
},
{
"version": "v3",
"created": "Sat, 9 May 2015 07:26:02 GMT"
},
{
"version": "v4",
"created": "Sat, 20 Jun 2015 12:53:35 GMT"
},
{
"version": "v5",
"created": "Tue, 18 Oct 2016 13:44:30 GMT"
}
] | 2016-10-19T00:00:00 | [
[
"Hu",
"Wenbo",
""
],
[
"Zhu",
"Jun",
""
],
[
"Zhang",
"Bo",
""
]
] | TITLE: Fast Sampling for Bayesian Max-Margin Models
ABSTRACT: Bayesian max-margin models have shown superiority in various practical
applications, such as text categorization, collaborative prediction, social
network link prediction and crowdsourcing, and they conjoin the flexibility of
Bayesian modeling and predictive strengths of max-margin learning. However,
Monte Carlo sampling for these models still remains challenging, especially for
applications that involve large-scale datasets. In this paper, we present the
stochastic subgradient Hamiltonian Monte Carlo (HMC) methods, which are easy to
implement and computationally efficient. We show the approximate detailed
balance property of subgradient HMC which reveals a natural and validated
generalization of the ordinary HMC. Furthermore, we investigate the variants
that use stochastic subsampling and thermostats for better scalability and
mixing. Using stochastic subgradient Markov Chain Monte Carlo (MCMC), we
efficiently solve the posterior inference task of various Bayesian max-margin
models and extensive experimental results demonstrate the effectiveness of our
approach.
| no_new_dataset | 0.945801 |
1511.06789 | Jonathan Krause | Jonathan Krause, Benjamin Sapp, Andrew Howard, Howard Zhou, Alexander
Toshev, Tom Duerig, James Philbin, Li Fei-Fei | The Unreasonable Effectiveness of Noisy Data for Fine-Grained
Recognition | ECCV 2016, data is released | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Current approaches for fine-grained recognition do the following: First,
recruit experts to annotate a dataset of images, optionally also collecting
more structured data in the form of part annotations and bounding boxes.
Second, train a model utilizing this data. Toward the goal of solving
fine-grained recognition, we introduce an alternative approach, leveraging
free, noisy data from the web and simple, generic methods of recognition. This
approach has benefits in both performance and scalability. We demonstrate its
efficacy on four fine-grained datasets, greatly exceeding existing state of the
art without the manual collection of even a single label, and furthermore show
first results at scaling to more than 10,000 fine-grained categories.
Quantitatively, we achieve top-1 accuracies of 92.3% on CUB-200-2011, 85.4% on
Birdsnap, 93.4% on FGVC-Aircraft, and 80.8% on Stanford Dogs without using
their annotated training sets. We compare our approach to an active learning
approach for expanding fine-grained datasets.
| [
{
"version": "v1",
"created": "Fri, 20 Nov 2015 22:40:30 GMT"
},
{
"version": "v2",
"created": "Sat, 30 Jul 2016 08:22:52 GMT"
},
{
"version": "v3",
"created": "Tue, 18 Oct 2016 18:35:31 GMT"
}
] | 2016-10-19T00:00:00 | [
[
"Krause",
"Jonathan",
""
],
[
"Sapp",
"Benjamin",
""
],
[
"Howard",
"Andrew",
""
],
[
"Zhou",
"Howard",
""
],
[
"Toshev",
"Alexander",
""
],
[
"Duerig",
"Tom",
""
],
[
"Philbin",
"James",
""
],
[
"Fei-Fei",
"Li",
""
]
] | TITLE: The Unreasonable Effectiveness of Noisy Data for Fine-Grained
Recognition
ABSTRACT: Current approaches for fine-grained recognition do the following: First,
recruit experts to annotate a dataset of images, optionally also collecting
more structured data in the form of part annotations and bounding boxes.
Second, train a model utilizing this data. Toward the goal of solving
fine-grained recognition, we introduce an alternative approach, leveraging
free, noisy data from the web and simple, generic methods of recognition. This
approach has benefits in both performance and scalability. We demonstrate its
efficacy on four fine-grained datasets, greatly exceeding existing state of the
art without the manual collection of even a single label, and furthermore show
first results at scaling to more than 10,000 fine-grained categories.
Quantitatively, we achieve top-1 accuracies of 92.3% on CUB-200-2011, 85.4% on
Birdsnap, 93.4% on FGVC-Aircraft, and 80.8% on Stanford Dogs without using
their annotated training sets. We compare our approach to an active learning
approach for expanding fine-grained datasets.
| no_new_dataset | 0.947235 |
1605.07157 | Chelsea Finn | Chelsea Finn, Ian Goodfellow, Sergey Levine | Unsupervised Learning for Physical Interaction through Video Prediction | To appear in NIPS '16; Video results, code, and data available at:
http://www.sites.google.com/site/robotprediction | null | null | null | cs.LG cs.AI cs.CV cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A core challenge for an agent learning to interact with the world is to
predict how its actions affect objects in its environment. Many existing
methods for learning the dynamics of physical interactions require labeled
object information. However, to scale real-world interaction learning to a
variety of scenes and objects, acquiring labeled data becomes increasingly
impractical. To learn about physical object motion without labels, we develop
an action-conditioned video prediction model that explicitly models pixel
motion, by predicting a distribution over pixel motion from previous frames.
Because our model explicitly predicts motion, it is partially invariant to
object appearance, enabling it to generalize to previously unseen objects. To
explore video prediction for real-world interactive agents, we also introduce a
dataset of 59,000 robot interactions involving pushing motions, including a
test set with novel objects. In this dataset, accurate prediction of videos
conditioned on the robot's future actions amounts to learning a "visual
imagination" of different futures based on different courses of action. Our
experiments show that our proposed method produces more accurate video
predictions both quantitatively and qualitatively, when compared to prior
methods.
| [
{
"version": "v1",
"created": "Mon, 23 May 2016 19:45:55 GMT"
},
{
"version": "v2",
"created": "Tue, 24 May 2016 19:33:23 GMT"
},
{
"version": "v3",
"created": "Thu, 9 Jun 2016 00:29:37 GMT"
},
{
"version": "v4",
"created": "Mon, 17 Oct 2016 20:09:56 GMT"
}
] | 2016-10-19T00:00:00 | [
[
"Finn",
"Chelsea",
""
],
[
"Goodfellow",
"Ian",
""
],
[
"Levine",
"Sergey",
""
]
] | TITLE: Unsupervised Learning for Physical Interaction through Video Prediction
ABSTRACT: A core challenge for an agent learning to interact with the world is to
predict how its actions affect objects in its environment. Many existing
methods for learning the dynamics of physical interactions require labeled
object information. However, to scale real-world interaction learning to a
variety of scenes and objects, acquiring labeled data becomes increasingly
impractical. To learn about physical object motion without labels, we develop
an action-conditioned video prediction model that explicitly models pixel
motion, by predicting a distribution over pixel motion from previous frames.
Because our model explicitly predicts motion, it is partially invariant to
object appearance, enabling it to generalize to previously unseen objects. To
explore video prediction for real-world interactive agents, we also introduce a
dataset of 59,000 robot interactions involving pushing motions, including a
test set with novel objects. In this dataset, accurate prediction of videos
conditioned on the robot's future actions amounts to learning a "visual
imagination" of different futures based on different courses of action. Our
experiments show that our proposed method produces more accurate video
predictions both quantitatively and qualitatively, when compared to prior
methods.
| new_dataset | 0.960063 |
1606.01021 | Mario Taschwer | Mario Taschwer and Oge Marques | Automatic Separation of Compound Figures in Scientific Articles | accepted for Multimedia Tools and Applications with minor revisions | null | null | null | cs.CV cs.MM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Content-based analysis and retrieval of digital images found in scientific
articles is often hindered by images consisting of multiple subfigures
(compound figures). We address this problem by proposing a method to
automatically classify and separate compound figures, which consists of two
main steps: (i) a supervised compound figure classifier (CFC) discriminates
between compound and non-compound figures using task-specific image features;
and (ii) an image processing algorithm is applied to predicted compound images
to perform compound figure separation (CFS). Our CFC approach is shown to
achieve state-of-the-art classification performance on a published dataset. Our
CFS algorithm shows superior separation accuracy on two different datasets
compared to other known automatic approaches. Finally, we propose a method to
evaluate the effectiveness of the CFC-CFS process chain and use it to optimize
the misclassification loss of CFC for maximal effectiveness in the process
chain.
| [
{
"version": "v1",
"created": "Fri, 3 Jun 2016 09:53:01 GMT"
},
{
"version": "v2",
"created": "Tue, 18 Oct 2016 06:26:58 GMT"
}
] | 2016-10-19T00:00:00 | [
[
"Taschwer",
"Mario",
""
],
[
"Marques",
"Oge",
""
]
] | TITLE: Automatic Separation of Compound Figures in Scientific Articles
ABSTRACT: Content-based analysis and retrieval of digital images found in scientific
articles is often hindered by images consisting of multiple subfigures
(compound figures). We address this problem by proposing a method to
automatically classify and separate compound figures, which consists of two
main steps: (i) a supervised compound figure classifier (CFC) discriminates
between compound and non-compound figures using task-specific image features;
and (ii) an image processing algorithm is applied to predicted compound images
to perform compound figure separation (CFS). Our CFC approach is shown to
achieve state-of-the-art classification performance on a published dataset. Our
CFS algorithm shows superior separation accuracy on two different datasets
compared to other known automatic approaches. Finally, we propose a method to
evaluate the effectiveness of the CFC-CFS process chain and use it to optimize
the misclassification loss of CFC for maximal effectiveness in the process
chain.
| no_new_dataset | 0.950227 |
1607.05954 | Carlos Dafonte | C. Dafonte, D. Fustes, M. Manteiga, D. Garabato, M. A. Alvarez, A.
Ulla, C. Allende Prieto | On the estimation of stellar parameters with uncertainty prediction from
Generative Artificial Neural Networks: application to Gaia RVS simulated
spectra | null | A&A 594, A68 (2016) | 10.1051/0004-6361/201527045 | null | astro-ph.IM astro-ph.SR cs.AI cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Aims. We present an innovative artificial neural network (ANN) architecture,
called Generative ANN (GANN), that computes the forward model, that is it
learns the function that relates the unknown outputs (stellar atmospheric
parameters, in this case) to the given inputs (spectra). Such a model can be
integrated in a Bayesian framework to estimate the posterior distribution of
the outputs. Methods. The architecture of the GANN follows the same scheme as a
normal ANN, but with the inputs and outputs inverted. We train the network with
the set of atmospheric parameters (Teff, logg, [Fe/H] and [alpha/Fe]),
obtaining the stellar spectra for such inputs. The residuals between the
spectra in the grid and the estimated spectra are minimized using a validation
dataset to keep solutions as general as possible. Results. The performance of
both conventional ANNs and GANNs to estimate the stellar parameters as a
function of the star brightness is presented and compared for different
Galactic populations. GANNs provide significantly improved parameterizations
for early and intermediate spectral types with rich and intermediate
metallicities. The behaviour of both algorithms is very similar for our sample
of late-type stars, obtaining residuals in the derivation of [Fe/H] and
[alpha/Fe] below 0.1dex for stars with Gaia magnitude Grvs<12, which accounts
for a number in the order of four million stars to be observed by the Radial
Velocity Spectrograph of the Gaia satellite. Conclusions. Uncertainty
estimation of computed astrophysical parameters is crucial for the validation
of the parameterization itself and for the subsequent exploitation by the
astronomical community. GANNs produce not only the parameters for a given
spectrum, but a goodness-of-fit between the observed spectrum and the predicted
one for a given set of parameters. Moreover, they allow us to obtain the full
posterior distribution...
| [
{
"version": "v1",
"created": "Tue, 19 Jul 2016 15:16:56 GMT"
}
] | 2016-10-19T00:00:00 | [
[
"Dafonte",
"C.",
""
],
[
"Fustes",
"D.",
""
],
[
"Manteiga",
"M.",
""
],
[
"Garabato",
"D.",
""
],
[
"Alvarez",
"M. A.",
""
],
[
"Ulla",
"A.",
""
],
[
"Prieto",
"C. Allende",
""
]
] | TITLE: On the estimation of stellar parameters with uncertainty prediction from
Generative Artificial Neural Networks: application to Gaia RVS simulated
spectra
ABSTRACT: Aims. We present an innovative artificial neural network (ANN) architecture,
called Generative ANN (GANN), that computes the forward model, that is it
learns the function that relates the unknown outputs (stellar atmospheric
parameters, in this case) to the given inputs (spectra). Such a model can be
integrated in a Bayesian framework to estimate the posterior distribution of
the outputs. Methods. The architecture of the GANN follows the same scheme as a
normal ANN, but with the inputs and outputs inverted. We train the network with
the set of atmospheric parameters (Teff, logg, [Fe/H] and [alpha/Fe]),
obtaining the stellar spectra for such inputs. The residuals between the
spectra in the grid and the estimated spectra are minimized using a validation
dataset to keep solutions as general as possible. Results. The performance of
both conventional ANNs and GANNs to estimate the stellar parameters as a
function of the star brightness is presented and compared for different
Galactic populations. GANNs provide significantly improved parameterizations
for early and intermediate spectral types with rich and intermediate
metallicities. The behaviour of both algorithms is very similar for our sample
of late-type stars, obtaining residuals in the derivation of [Fe/H] and
[alpha/Fe] below 0.1dex for stars with Gaia magnitude Grvs<12, which accounts
for a number in the order of four million stars to be observed by the Radial
Velocity Spectrograph of the Gaia satellite. Conclusions. Uncertainty
estimation of computed astrophysical parameters is crucial for the validation
of the parameterization itself and for the subsequent exploitation by the
astronomical community. GANNs produce not only the parameters for a given
spectrum, but a goodness-of-fit between the observed spectrum and the predicted
one for a given set of parameters. Moreover, they allow us to obtain the full
posterior distribution...
| no_new_dataset | 0.946349 |
1609.02806 | Kai Wang | Kai Wang, Zhan Bin Chen, Ran Si, Per J\"onsson, J\"orgen Ekman, Xue
Lin Guo, Shuang Li, Fei Yun Long, Wei Dang, Xiao Hui Zhao, Roger Hutton,
Chong Yang Chen, Jun Yan, and Xu Yang | Extended relativistic configuration interaction and many-body
perturbation calculations of spectroscopic data for the $n \leq 6$
configurationsin ne-like ions between Cr XV and Kr XXVII | null | null | 10.3847/0067-0049/226/2/14 | null | physics.atom-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Level energies, wavelengths, electric dipole, magnetic dipole, electric
quadrupole, and magnetic quadrupole transition rates, oscillator strengths, and
line strengths from combined relativistic configuration interaction and
many-body perturbation calculations are reported for the 201 fine-structure
states of the $2s^2 2p^6$, $2s^2 2p^5 3l$, $2s 2p^6 3l$, $2s^2 2p^5 4l$, $2s
2p^6 4l$, $2s^2 2p^5 5l$, and $2s^2 2p^5 6l$ configurations in all Ne-like ions
between Cr XV and Kr XXVII. Calculated level energies and transition data are
compared with experiments from the NIST and CHIANTI databases, and other recent
benchmark calculations. The mean energy difference with the NIST experiments is
only 0.05%. The present calculations significantly increase the amount of
accurate spectroscopic data for the $n >3$ states in a number of Ne-like ions
of astrophysics interest. A complete dataset should be helpful in analyzing new
observations from the solar and other astrophysical sources, and is also likely
to be useful for modeling and diagnosing a variety of plasmas including
astronomical and fusion plasma.
| [
{
"version": "v1",
"created": "Fri, 9 Sep 2016 14:17:00 GMT"
},
{
"version": "v2",
"created": "Mon, 12 Sep 2016 00:48:48 GMT"
}
] | 2016-10-19T00:00:00 | [
[
"Wang",
"Kai",
""
],
[
"Chen",
"Zhan Bin",
""
],
[
"Si",
"Ran",
""
],
[
"Jönsson",
"Per",
""
],
[
"Ekman",
"Jörgen",
""
],
[
"Guo",
"Xue Lin",
""
],
[
"Li",
"Shuang",
""
],
[
"Long",
"Fei Yun",
""
],
[
"Dang",
"Wei",
""
],
[
"Zhao",
"Xiao Hui",
""
],
[
"Hutton",
"Roger",
""
],
[
"Chen",
"Chong Yang",
""
],
[
"Yan",
"Jun",
""
],
[
"Yang",
"Xu",
""
]
] | TITLE: Extended relativistic configuration interaction and many-body
perturbation calculations of spectroscopic data for the $n \leq 6$
configurationsin ne-like ions between Cr XV and Kr XXVII
ABSTRACT: Level energies, wavelengths, electric dipole, magnetic dipole, electric
quadrupole, and magnetic quadrupole transition rates, oscillator strengths, and
line strengths from combined relativistic configuration interaction and
many-body perturbation calculations are reported for the 201 fine-structure
states of the $2s^2 2p^6$, $2s^2 2p^5 3l$, $2s 2p^6 3l$, $2s^2 2p^5 4l$, $2s
2p^6 4l$, $2s^2 2p^5 5l$, and $2s^2 2p^5 6l$ configurations in all Ne-like ions
between Cr XV and Kr XXVII. Calculated level energies and transition data are
compared with experiments from the NIST and CHIANTI databases, and other recent
benchmark calculations. The mean energy difference with the NIST experiments is
only 0.05%. The present calculations significantly increase the amount of
accurate spectroscopic data for the $n >3$ states in a number of Ne-like ions
of astrophysics interest. A complete dataset should be helpful in analyzing new
observations from the solar and other astrophysical sources, and is also likely
to be useful for modeling and diagnosing a variety of plasmas including
astronomical and fusion plasma.
| no_new_dataset | 0.949295 |
1609.08864 | Mrutyunjaya Panda | Mrutyunjaya Panda (Utkal University, Vani Vihar, Bhubaneswar, India) | Towards the effectiveness of Deep Convolutional Neural Network based
Fast Random Forest Classifier | 11 pages, 9 figures, 1 table | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep Learning is considered to be a quite young in the area of machine
learning research, found its effectiveness in dealing complex yet high
dimensional dataset that includes but limited to images, text and speech etc.
with multiple levels of representation and abstraction. As there are a plethora
of research on these datasets by various researchers , a win over them needs
lots of attention. Careful setting of Deep learning parameters is of paramount
importance in order to avoid the overfitting unlike conventional methods with
limited parameter settings. Deep Convolutional neural network (DCNN) with
multiple layers of compositions and appropriate settings might be is an
efficient machine learning method that can outperform the conventional methods
in a great way. However, due to its slow adoption in learning, there are also
always a chance of overfitting during feature selection process, which can be
addressed by employing a regularization method called dropout. Fast Random
Forest (FRF) is a powerful ensemble classifier especially when the datasets are
noisy and when the number of attributes is large in comparison to the number of
instances, as is the case of Bioinformatics datasets. Several publicly
available Bioinformatics dataset, Handwritten digits recognition and Image
segmentation dataset are considered for evaluation of the proposed approach.
The excellent performance obtained by the proposed DCNN based feature selection
with FRF classifier on high dimensional datasets makes it a fast and accurate
classifier in comparison the state-of-the-art.
| [
{
"version": "v1",
"created": "Wed, 28 Sep 2016 11:35:17 GMT"
}
] | 2016-10-19T00:00:00 | [
[
"Panda",
"Mrutyunjaya",
"",
"Utkal University, Vani Vihar, Bhubaneswar, India"
]
] | TITLE: Towards the effectiveness of Deep Convolutional Neural Network based
Fast Random Forest Classifier
ABSTRACT: Deep Learning is considered to be a quite young in the area of machine
learning research, found its effectiveness in dealing complex yet high
dimensional dataset that includes but limited to images, text and speech etc.
with multiple levels of representation and abstraction. As there are a plethora
of research on these datasets by various researchers , a win over them needs
lots of attention. Careful setting of Deep learning parameters is of paramount
importance in order to avoid the overfitting unlike conventional methods with
limited parameter settings. Deep Convolutional neural network (DCNN) with
multiple layers of compositions and appropriate settings might be is an
efficient machine learning method that can outperform the conventional methods
in a great way. However, due to its slow adoption in learning, there are also
always a chance of overfitting during feature selection process, which can be
addressed by employing a regularization method called dropout. Fast Random
Forest (FRF) is a powerful ensemble classifier especially when the datasets are
noisy and when the number of attributes is large in comparison to the number of
instances, as is the case of Bioinformatics datasets. Several publicly
available Bioinformatics dataset, Handwritten digits recognition and Image
segmentation dataset are considered for evaluation of the proposed approach.
The excellent performance obtained by the proposed DCNN based feature selection
with FRF classifier on high dimensional datasets makes it a fast and accurate
classifier in comparison the state-of-the-art.
| no_new_dataset | 0.947866 |
1610.03108 | Eamon Duede | Yadu N. Babuji, Kyle Chard, Aaron Gerow, and Eamon Duede | Cloud Kotta: Enabling Secure and Scalable Data Analytics in the Cloud | A version of this paper is forthcoming at BigData 2016 | null | null | null | cs.DC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Distributed communities of researchers rely increasingly on valuable,
proprietary, or sensitive datasets. Given the growth of such data, especially
in fields new to data-driven, computationally intensive research like the
social sciences and humanities, coupled with what are often strict and complex
data-use agreements, many research communities now require methods that allow
secure, scalable and cost-effective storage and analysis. Here we present CLOUD
KOTTA: a cloud-based data management and analytics framework. CLOUD KOTTA
delivers an end-to-end solution for coordinating secure access to large
datasets, and an execution model that provides both automated infrastructure
scaling and support for executing analytics near to the data. CLOUD KOTTA
implements a fine-grained security model ensuring that only authorized users
may access, analyze, and download protected data. It also implements automated
methods for acquiring and configuring low-cost storage and compute resources as
they are needed. We present the architecture and implementation of CLOUD KOTTA
and demonstrate the advantages it provides in terms of increased performance
and flexibility. We show that CLOUD KOTTA's elastic provisioning model can
reduce costs by up to 16x when compared with statically provisioned models.
| [
{
"version": "v1",
"created": "Mon, 10 Oct 2016 21:58:09 GMT"
},
{
"version": "v2",
"created": "Tue, 18 Oct 2016 19:07:46 GMT"
}
] | 2016-10-19T00:00:00 | [
[
"Babuji",
"Yadu N.",
""
],
[
"Chard",
"Kyle",
""
],
[
"Gerow",
"Aaron",
""
],
[
"Duede",
"Eamon",
""
]
] | TITLE: Cloud Kotta: Enabling Secure and Scalable Data Analytics in the Cloud
ABSTRACT: Distributed communities of researchers rely increasingly on valuable,
proprietary, or sensitive datasets. Given the growth of such data, especially
in fields new to data-driven, computationally intensive research like the
social sciences and humanities, coupled with what are often strict and complex
data-use agreements, many research communities now require methods that allow
secure, scalable and cost-effective storage and analysis. Here we present CLOUD
KOTTA: a cloud-based data management and analytics framework. CLOUD KOTTA
delivers an end-to-end solution for coordinating secure access to large
datasets, and an execution model that provides both automated infrastructure
scaling and support for executing analytics near to the data. CLOUD KOTTA
implements a fine-grained security model ensuring that only authorized users
may access, analyze, and download protected data. It also implements automated
methods for acquiring and configuring low-cost storage and compute resources as
they are needed. We present the architecture and implementation of CLOUD KOTTA
and demonstrate the advantages it provides in terms of increased performance
and flexibility. We show that CLOUD KOTTA's elastic provisioning model can
reduce costs by up to 16x when compared with statically provisioned models.
| no_new_dataset | 0.943867 |
1610.04662 | Noel Codella | Noel Codella, Quoc-Bao Nguyen, Sharath Pankanti, David Gutman, Brian
Helba, Allan Halpern, John R. Smith | Deep Learning Ensembles for Melanoma Recognition in Dermoscopy Images | URL for the IBM Journal of Research and Development:
http://www.research.ibm.com/journal/ | IBM Journal of Research and Development, vol. 61, no. 4/5, 2017 | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Melanoma is the deadliest form of skin cancer. While curable with early
detection, only highly trained specialists are capable of accurately
recognizing the disease. As expertise is in limited supply, automated systems
capable of identifying disease could save lives, reduce unnecessary biopsies,
and reduce costs. Toward this goal, we propose a system that combines recent
developments in deep learning with established machine learning approaches,
creating ensembles of methods that are capable of segmenting skin lesions, as
well as analyzing the detected area and surrounding tissue for melanoma
detection. The system is evaluated using the largest publicly available
benchmark dataset of dermoscopic images, containing 900 training and 379
testing images. New state-of-the-art performance levels are demonstrated,
leading to an improvement in the area under receiver operating characteristic
curve of 7.5% (0.843 vs. 0.783), in average precision of 4% (0.649 vs. 0.624),
and in specificity measured at the clinically relevant 95% sensitivity
operating point 2.9 times higher than the previous state-of-the-art (36.8%
specificity compared to 12.5%). Compared to the average of 8 expert
dermatologists on a subset of 100 test images, the proposed system produces a
higher accuracy (76% vs. 70.5%), and specificity (62% vs. 59%) evaluated at an
equivalent sensitivity (82%).
| [
{
"version": "v1",
"created": "Fri, 14 Oct 2016 22:31:34 GMT"
},
{
"version": "v2",
"created": "Tue, 18 Oct 2016 00:25:35 GMT"
}
] | 2016-10-19T00:00:00 | [
[
"Codella",
"Noel",
""
],
[
"Nguyen",
"Quoc-Bao",
""
],
[
"Pankanti",
"Sharath",
""
],
[
"Gutman",
"David",
""
],
[
"Helba",
"Brian",
""
],
[
"Halpern",
"Allan",
""
],
[
"Smith",
"John R.",
""
]
] | TITLE: Deep Learning Ensembles for Melanoma Recognition in Dermoscopy Images
ABSTRACT: Melanoma is the deadliest form of skin cancer. While curable with early
detection, only highly trained specialists are capable of accurately
recognizing the disease. As expertise is in limited supply, automated systems
capable of identifying disease could save lives, reduce unnecessary biopsies,
and reduce costs. Toward this goal, we propose a system that combines recent
developments in deep learning with established machine learning approaches,
creating ensembles of methods that are capable of segmenting skin lesions, as
well as analyzing the detected area and surrounding tissue for melanoma
detection. The system is evaluated using the largest publicly available
benchmark dataset of dermoscopic images, containing 900 training and 379
testing images. New state-of-the-art performance levels are demonstrated,
leading to an improvement in the area under receiver operating characteristic
curve of 7.5% (0.843 vs. 0.783), in average precision of 4% (0.649 vs. 0.624),
and in specificity measured at the clinically relevant 95% sensitivity
operating point 2.9 times higher than the previous state-of-the-art (36.8%
specificity compared to 12.5%). Compared to the average of 8 expert
dermatologists on a subset of 100 test images, the proposed system produces a
higher accuracy (76% vs. 70.5%), and specificity (62% vs. 59%) evaluated at an
equivalent sensitivity (82%).
| no_new_dataset | 0.749408 |
1610.05394 | Venkatesh Saligrama | Manjesh Hanawal and Csaba Szepesvari and Venkatesh Saligrama | Sequential Learning without Feedback | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In many security and healthcare systems a sequence of features/sensors/tests
are used for detection and diagnosis. Each test outputs a prediction of the
latent state, and carries with it inherent costs. Our objective is to {\it
learn} strategies for selecting tests to optimize accuracy \& costs.
Unfortunately it is often impossible to acquire in-situ ground truth
annotations and we are left with the problem of unsupervised sensor selection
(USS). We pose USS as a version of stochastic partial monitoring problem with
an {\it unusual} reward structure (even noisy annotations are unavailable).
Unsurprisingly no learner can achieve sublinear regret without further
assumptions. To this end we propose the notion of weak-dominance. This is a
condition on the joint probability distribution of test outputs and latent
state and says that whenever a test is accurate on an example, a later test in
the sequence is likely to be accurate as well. We empirically verify that weak
dominance holds on real datasets and prove that it is a maximal condition for
achieving sublinear regret. We reduce USS to a special case of multi-armed
bandit problem with side information and develop polynomial time algorithms
that achieve sublinear regret.
| [
{
"version": "v1",
"created": "Tue, 18 Oct 2016 01:15:57 GMT"
}
] | 2016-10-19T00:00:00 | [
[
"Hanawal",
"Manjesh",
""
],
[
"Szepesvari",
"Csaba",
""
],
[
"Saligrama",
"Venkatesh",
""
]
] | TITLE: Sequential Learning without Feedback
ABSTRACT: In many security and healthcare systems a sequence of features/sensors/tests
are used for detection and diagnosis. Each test outputs a prediction of the
latent state, and carries with it inherent costs. Our objective is to {\it
learn} strategies for selecting tests to optimize accuracy \& costs.
Unfortunately it is often impossible to acquire in-situ ground truth
annotations and we are left with the problem of unsupervised sensor selection
(USS). We pose USS as a version of stochastic partial monitoring problem with
an {\it unusual} reward structure (even noisy annotations are unavailable).
Unsurprisingly no learner can achieve sublinear regret without further
assumptions. To this end we propose the notion of weak-dominance. This is a
condition on the joint probability distribution of test outputs and latent
state and says that whenever a test is accurate on an example, a later test in
the sequence is likely to be accurate as well. We empirically verify that weak
dominance holds on real datasets and prove that it is a maximal condition for
achieving sublinear regret. We reduce USS to a special case of multi-armed
bandit problem with side information and develop polynomial time algorithms
that achieve sublinear regret.
| no_new_dataset | 0.95018 |
1610.05455 | Steve Chang | Adam Wang, Steve Chang, John Wilson | Predict Moves | null | null | null | null | cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Mobile applications and on-body devices are becoming increasingly ubiquitous
tools for physical activity tracking. We propose utilizing a self-tracker's
habits to support continuous prediction of whether they will reach their daily
step goal, thus enabling a variety of potential persuasive interventions. Our
aim is to improve the prediction by leveraging historical data and other
qualitative (motivation for using the systems, location, gender) and,
quantitative (age) features. We have collected datasets from two activity
tracking platforms (Moves and Fitbit) and aim to check if the model we derive
from one is generalizable over the other. In the following paper we establish a
pipeline for extracting the data and formatting it for modeling. We discuss the
approach we took and our findings while selecting the features and
classification models for the dataset. We further discuss the notion of
generalizability of the model across different types of dataset and the
probable inclusion of non standard features to further improve the model's
accuracy.
| [
{
"version": "v1",
"created": "Tue, 18 Oct 2016 07:01:57 GMT"
}
] | 2016-10-19T00:00:00 | [
[
"Wang",
"Adam",
""
],
[
"Chang",
"Steve",
""
],
[
"Wilson",
"John",
""
]
] | TITLE: Predict Moves
ABSTRACT: Mobile applications and on-body devices are becoming increasingly ubiquitous
tools for physical activity tracking. We propose utilizing a self-tracker's
habits to support continuous prediction of whether they will reach their daily
step goal, thus enabling a variety of potential persuasive interventions. Our
aim is to improve the prediction by leveraging historical data and other
qualitative (motivation for using the systems, location, gender) and,
quantitative (age) features. We have collected datasets from two activity
tracking platforms (Moves and Fitbit) and aim to check if the model we derive
from one is generalizable over the other. In the following paper we establish a
pipeline for extracting the data and formatting it for modeling. We discuss the
approach we took and our findings while selecting the features and
classification models for the dataset. We further discuss the notion of
generalizability of the model across different types of dataset and the
probable inclusion of non standard features to further improve the model's
accuracy.
| no_new_dataset | 0.95275 |
1610.05463 | Steve Chang | Teng Lee, James Johnson, Steve Cheng | An Interactive Machine Learning Framework | null | null | null | null | cs.HC cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Machine learning (ML) is believed to be an effective and efficient tool to
build reliable prediction model or extract useful structure from an avalanche
of data. However, ML is also criticized by its difficulty in interpretation and
complicated parameter tuning. In contrast, visualization is able to well
organize and visually encode the entangled information in data and guild
audiences to simpler perceptual inferences and analytic thinking. But large
scale and high dimensional data will usually lead to the failure of many
visualization methods. In this paper, we close a loop between ML and
visualization via interaction between ML algorithm and users, so machine
intelligence and human intelligence can cooperate and improve each other in a
mutually rewarding way. In particular, we propose "transparent boosting tree
(TBT)", which visualizes both the model structure and prediction statistics of
each step in the learning process of gradient boosting tree to user, and
involves user's feedback operations to trees into the learning process. In TBT,
ML is in charge of updating weights in learning model and filtering information
shown to user from the big data, while visualization is in charge of providing
a visual understanding of ML model to facilitate user exploration. It combines
the advantages of both ML in big data statistics and human in decision making
based on domain knowledge. We develop a user friendly interface for this novel
learning method, and apply it to two datasets collected from real applications.
Our study shows that making ML transparent by using interactive visualization
can significantly improve the exploration of ML algorithms, give rise to novel
insights of ML models, and integrates both machine and human intelligence.
| [
{
"version": "v1",
"created": "Tue, 18 Oct 2016 07:46:11 GMT"
}
] | 2016-10-19T00:00:00 | [
[
"Lee",
"Teng",
""
],
[
"Johnson",
"James",
""
],
[
"Cheng",
"Steve",
""
]
] | TITLE: An Interactive Machine Learning Framework
ABSTRACT: Machine learning (ML) is believed to be an effective and efficient tool to
build reliable prediction model or extract useful structure from an avalanche
of data. However, ML is also criticized by its difficulty in interpretation and
complicated parameter tuning. In contrast, visualization is able to well
organize and visually encode the entangled information in data and guild
audiences to simpler perceptual inferences and analytic thinking. But large
scale and high dimensional data will usually lead to the failure of many
visualization methods. In this paper, we close a loop between ML and
visualization via interaction between ML algorithm and users, so machine
intelligence and human intelligence can cooperate and improve each other in a
mutually rewarding way. In particular, we propose "transparent boosting tree
(TBT)", which visualizes both the model structure and prediction statistics of
each step in the learning process of gradient boosting tree to user, and
involves user's feedback operations to trees into the learning process. In TBT,
ML is in charge of updating weights in learning model and filtering information
shown to user from the big data, while visualization is in charge of providing
a visual understanding of ML model to facilitate user exploration. It combines
the advantages of both ML in big data statistics and human in decision making
based on domain knowledge. We develop a user friendly interface for this novel
learning method, and apply it to two datasets collected from real applications.
Our study shows that making ML transparent by using interactive visualization
can significantly improve the exploration of ML algorithms, give rise to novel
insights of ML models, and integrates both machine and human intelligence.
| no_new_dataset | 0.949201 |
1610.05465 | Gwenole Quellec | Katia Charri\`ere, Gwenol\'e Quellec, Mathieu Lamard, David Martiano,
Guy Cazuguel, Gouenou Coatrieux, B\'eatrice Cochener | Real-time analysis of cataract surgery videos using statistical models | This is an extended version of a paper presented at the CBMI 2016
conference | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The automatic analysis of the surgical process, from videos recorded during
surgeries, could be very useful to surgeons, both for training and for
acquiring new techniques. The training process could be optimized by
automatically providing some targeted recommendations or warnings, similar to
the expert surgeon's guidance. In this paper, we propose to reuse videos
recorded and stored during cataract surgeries to perform the analysis. The
proposed system allows to automatically recognize, in real time, what the
surgeon is doing: what surgical phase or, more precisely, what surgical step he
or she is performing. This recognition relies on the inference of a multilevel
statistical model which uses 1) the conditional relations between levels of
description (steps and phases) and 2) the temporal relations among steps and
among phases. The model accepts two types of inputs: 1) the presence of
surgical tools, manually provided by the surgeons, or 2) motion in videos,
automatically analyzed through the Content Based Video retrieval (CBVR)
paradigm. Different data-driven statistical models are evaluated in this paper.
For this project, a dataset of 30 cataract surgery videos was collected at
Brest University hospital. The system was evaluated in terms of area under the
ROC curve. Promising results were obtained using either the presence of
surgical tools ($A_z$ = 0.983) or motion analysis ($A_z$ = 0.759). The
generality of the method allows to adapt it to any kinds of surgeries. The
proposed solution could be used in a computer assisted surgery tool to support
surgeons during the surgery.
| [
{
"version": "v1",
"created": "Tue, 18 Oct 2016 07:55:48 GMT"
}
] | 2016-10-19T00:00:00 | [
[
"Charrière",
"Katia",
""
],
[
"Quellec",
"Gwenolé",
""
],
[
"Lamard",
"Mathieu",
""
],
[
"Martiano",
"David",
""
],
[
"Cazuguel",
"Guy",
""
],
[
"Coatrieux",
"Gouenou",
""
],
[
"Cochener",
"Béatrice",
""
]
] | TITLE: Real-time analysis of cataract surgery videos using statistical models
ABSTRACT: The automatic analysis of the surgical process, from videos recorded during
surgeries, could be very useful to surgeons, both for training and for
acquiring new techniques. The training process could be optimized by
automatically providing some targeted recommendations or warnings, similar to
the expert surgeon's guidance. In this paper, we propose to reuse videos
recorded and stored during cataract surgeries to perform the analysis. The
proposed system allows to automatically recognize, in real time, what the
surgeon is doing: what surgical phase or, more precisely, what surgical step he
or she is performing. This recognition relies on the inference of a multilevel
statistical model which uses 1) the conditional relations between levels of
description (steps and phases) and 2) the temporal relations among steps and
among phases. The model accepts two types of inputs: 1) the presence of
surgical tools, manually provided by the surgeons, or 2) motion in videos,
automatically analyzed through the Content Based Video retrieval (CBVR)
paradigm. Different data-driven statistical models are evaluated in this paper.
For this project, a dataset of 30 cataract surgery videos was collected at
Brest University hospital. The system was evaluated in terms of area under the
ROC curve. Promising results were obtained using either the presence of
surgical tools ($A_z$ = 0.983) or motion analysis ($A_z$ = 0.759). The
generality of the method allows to adapt it to any kinds of surgeries. The
proposed solution could be used in a computer assisted surgery tool to support
surgeons during the surgery.
| new_dataset | 0.794146 |
1610.05518 | Gianni D'Angelo | Gianni D'Angelo, Salvatore Rampone | Shape-based defect classification for Non Destructive Testing | 5 pages, IEEE International Workshop | IEEE International Workshop on Metrology for Aerospace, Benevento,
Italy, June 4-5, 2015 | 10.1109/MetroAeroSpace.2015.7180691 | null | cs.CV cs.CE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The aim of this work is to classify the aerospace structure defects detected
by eddy current non-destructive testing. The proposed method is based on the
assumption that the defect is bound to the reaction of the probe coil impedance
during the test. Impedance plane analysis is used to extract a feature vector
from the shape of the coil impedance in the complex plane, through the use of
some geometric parameters. Shape recognition is tested with three different
machine-learning based classifiers: decision trees, neural networks and Naive
Bayes. The performance of the proposed detection system are measured in terms
of accuracy, sensitivity, specificity, precision and Matthews correlation
coefficient. Several experiments are performed on dataset of eddy current
signal samples for aircraft structures. The obtained results demonstrate the
usefulness of our approach and the competiveness against existing descriptors.
| [
{
"version": "v1",
"created": "Tue, 18 Oct 2016 10:03:25 GMT"
}
] | 2016-10-19T00:00:00 | [
[
"D'Angelo",
"Gianni",
""
],
[
"Rampone",
"Salvatore",
""
]
] | TITLE: Shape-based defect classification for Non Destructive Testing
ABSTRACT: The aim of this work is to classify the aerospace structure defects detected
by eddy current non-destructive testing. The proposed method is based on the
assumption that the defect is bound to the reaction of the probe coil impedance
during the test. Impedance plane analysis is used to extract a feature vector
from the shape of the coil impedance in the complex plane, through the use of
some geometric parameters. Shape recognition is tested with three different
machine-learning based classifiers: decision trees, neural networks and Naive
Bayes. The performance of the proposed detection system are measured in terms
of accuracy, sensitivity, specificity, precision and Matthews correlation
coefficient. Several experiments are performed on dataset of eddy current
signal samples for aircraft structures. The obtained results demonstrate the
usefulness of our approach and the competiveness against existing descriptors.
| no_new_dataset | 0.955775 |
1610.05522 | Giovanni Da San Martino | Giovanni Da San Martino, Alberto Barr\'on-Cede\~no, Salvatore Romeo,
Alessandro Moschitti, Shafiq Joty, Fahad A. Al Obaidli, Kateryna Tymoshenko,
Antonio Uva | Addressing Community Question Answering in English and Arabic | presented at Second WebQA workshop, SIGIR2016
(http://plg2.cs.uwaterloo.ca/~avtyurin/WebQA2016/) | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper studies the impact of different types of features applied to
learning to re-rank questions in community Question Answering. We tested our
models on two datasets released in SemEval-2016 Task 3 on "Community Question
Answering". Task 3 targeted real-life Web fora both in English and Arabic. Our
models include bag-of-words features (BoW), syntactic tree kernels (TKs), rank
features, embeddings, and machine translation evaluation features. To the best
of our knowledge, structural kernels have barely been applied to the question
reranking task, where they have to model paraphrase relations. In the case of
the English question re-ranking task, we compare our learning to rank (L2R)
algorithms against a strong baseline given by the Google-generated ranking
(GR). The results show that i) the shallow structures used in our TKs are
robust enough to noisy data and ii) improving GR is possible, but effective BoW
features and TKs along with an accurate model of GR features in the used L2R
algorithm are required. In the case of the Arabic question re-ranking task, for
the first time we applied tree kernels on syntactic trees of Arabic sentences.
Our approaches to both tasks obtained the second best results on SemEval-2016
subtasks B on English and D on Arabic.
| [
{
"version": "v1",
"created": "Tue, 18 Oct 2016 10:22:46 GMT"
}
] | 2016-10-19T00:00:00 | [
[
"Martino",
"Giovanni Da San",
""
],
[
"Barrón-Cedeño",
"Alberto",
""
],
[
"Romeo",
"Salvatore",
""
],
[
"Moschitti",
"Alessandro",
""
],
[
"Joty",
"Shafiq",
""
],
[
"Obaidli",
"Fahad A. Al",
""
],
[
"Tymoshenko",
"Kateryna",
""
],
[
"Uva",
"Antonio",
""
]
] | TITLE: Addressing Community Question Answering in English and Arabic
ABSTRACT: This paper studies the impact of different types of features applied to
learning to re-rank questions in community Question Answering. We tested our
models on two datasets released in SemEval-2016 Task 3 on "Community Question
Answering". Task 3 targeted real-life Web fora both in English and Arabic. Our
models include bag-of-words features (BoW), syntactic tree kernels (TKs), rank
features, embeddings, and machine translation evaluation features. To the best
of our knowledge, structural kernels have barely been applied to the question
reranking task, where they have to model paraphrase relations. In the case of
the English question re-ranking task, we compare our learning to rank (L2R)
algorithms against a strong baseline given by the Google-generated ranking
(GR). The results show that i) the shallow structures used in our TKs are
robust enough to noisy data and ii) improving GR is possible, but effective BoW
features and TKs along with an accurate model of GR features in the used L2R
algorithm are required. In the case of the Arabic question re-ranking task, for
the first time we applied tree kernels on syntactic trees of Arabic sentences.
Our approaches to both tasks obtained the second best results on SemEval-2016
subtasks B on English and D on Arabic.
| no_new_dataset | 0.953622 |
1610.05555 | Decebal Constantin Mocanu | Decebal Constantin Mocanu and Maria Torres Vega and Eric Eaton and
Peter Stone and Antonio Liotta | Online Contrastive Divergence with Generative Replay: Experience Replay
without Storing Data | null | null | null | null | cs.LG cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Conceived in the early 1990s, Experience Replay (ER) has been shown to be a
successful mechanism to allow online learning algorithms to reuse past
experiences. Traditionally, ER can be applied to all machine learning paradigms
(i.e., unsupervised, supervised, and reinforcement learning). Recently, ER has
contributed to improving the performance of deep reinforcement learning. Yet,
its application to many practical settings is still limited by the memory
requirements of ER, necessary to explicitly store previous observations. To
remedy this issue, we explore a novel approach, Online Contrastive Divergence
with Generative Replay (OCD_GR), which uses the generative capability of
Restricted Boltzmann Machines (RBMs) instead of recorded past experiences. The
RBM is trained online, and does not require the system to store any of the
observed data points. We compare OCD_GR to ER on 9 real-world datasets,
considering a worst-case scenario (data points arriving in sorted order) as
well as a more realistic one (sequential random-order data points). Our results
show that in 64.28% of the cases OCD_GR outperforms ER and in the remaining
35.72% it has an almost equal performance, while having a considerably reduced
space complexity (i.e., memory usage) at a comparable time complexity.
| [
{
"version": "v1",
"created": "Tue, 18 Oct 2016 12:06:14 GMT"
}
] | 2016-10-19T00:00:00 | [
[
"Mocanu",
"Decebal Constantin",
""
],
[
"Vega",
"Maria Torres",
""
],
[
"Eaton",
"Eric",
""
],
[
"Stone",
"Peter",
""
],
[
"Liotta",
"Antonio",
""
]
] | TITLE: Online Contrastive Divergence with Generative Replay: Experience Replay
without Storing Data
ABSTRACT: Conceived in the early 1990s, Experience Replay (ER) has been shown to be a
successful mechanism to allow online learning algorithms to reuse past
experiences. Traditionally, ER can be applied to all machine learning paradigms
(i.e., unsupervised, supervised, and reinforcement learning). Recently, ER has
contributed to improving the performance of deep reinforcement learning. Yet,
its application to many practical settings is still limited by the memory
requirements of ER, necessary to explicitly store previous observations. To
remedy this issue, we explore a novel approach, Online Contrastive Divergence
with Generative Replay (OCD_GR), which uses the generative capability of
Restricted Boltzmann Machines (RBMs) instead of recorded past experiences. The
RBM is trained online, and does not require the system to store any of the
observed data points. We compare OCD_GR to ER on 9 real-world datasets,
considering a worst-case scenario (data points arriving in sorted order) as
well as a more realistic one (sequential random-order data points). Our results
show that in 64.28% of the cases OCD_GR outperforms ER and in the remaining
35.72% it has an almost equal performance, while having a considerably reduced
space complexity (i.e., memory usage) at a comparable time complexity.
| no_new_dataset | 0.947088 |
1610.05567 | R\'emi Cad\`ene | R\'emi Cad\`ene, Nicolas Thome, Matthieu Cord | Master's Thesis : Deep Learning for Visual Recognition | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The goal of our research is to develop methods advancing automatic visual
recognition. In order to predict the unique or multiple labels associated to an
image, we study different kind of Deep Neural Networks architectures and
methods for supervised features learning. We first draw up a state-of-the-art
review of the Convolutional Neural Networks aiming to understand the history
behind this family of statistical models, the limit of modern architectures and
the novel techniques currently used to train deep CNNs. The originality of our
work lies in our approach focusing on tasks with a low amount of data. We
introduce different models and techniques to achieve the best accuracy on
several kind of datasets, such as a medium dataset of food recipes (100k
images) for building a web API, or a small dataset of satellite images (6,000)
for the DSG online challenge that we've won. We also draw up the
state-of-the-art in Weakly Supervised Learning, introducing different kind of
CNNs able to localize regions of interest. Our last contribution is a
framework, build on top of Torch7, for training and testing deep models on any
visual recognition tasks and on datasets of any scale.
| [
{
"version": "v1",
"created": "Tue, 18 Oct 2016 12:26:49 GMT"
}
] | 2016-10-19T00:00:00 | [
[
"Cadène",
"Rémi",
""
],
[
"Thome",
"Nicolas",
""
],
[
"Cord",
"Matthieu",
""
]
] | TITLE: Master's Thesis : Deep Learning for Visual Recognition
ABSTRACT: The goal of our research is to develop methods advancing automatic visual
recognition. In order to predict the unique or multiple labels associated to an
image, we study different kind of Deep Neural Networks architectures and
methods for supervised features learning. We first draw up a state-of-the-art
review of the Convolutional Neural Networks aiming to understand the history
behind this family of statistical models, the limit of modern architectures and
the novel techniques currently used to train deep CNNs. The originality of our
work lies in our approach focusing on tasks with a low amount of data. We
introduce different models and techniques to achieve the best accuracy on
several kind of datasets, such as a medium dataset of food recipes (100k
images) for building a web API, or a small dataset of satellite images (6,000)
for the DSG online challenge that we've won. We also draw up the
state-of-the-art in Weakly Supervised Learning, introducing different kind of
CNNs able to localize regions of interest. Our last contribution is a
framework, build on top of Torch7, for training and testing deep models on any
visual recognition tasks and on datasets of any scale.
| new_dataset | 0.865793 |
1610.05613 | Aditya Singh | Aditya Singh, Saurabh Saini, Rajvi Shah, and P J Narayanan | From Traditional to Modern : Domain Adaptation for Action Classification
in Short Social Video Clips | 9 pages, GCPR, 2016 | Pattern Recognition,38th German Conference, GCPR 2016, Hannover,
Germany, September 12-15, 2016, Proceedings,pp 245-257 | 10.1007/978-3-319-45886-1_20 | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Short internet video clips like vines present a significantly wild
distribution compared to traditional video datasets. In this paper, we focus on
the problem of unsupervised action classification in wild vines using
traditional labeled datasets. To this end, we use a data augmentation based
simple domain adaptation strategy. We utilise semantic word2vec space as a
common subspace to embed video features from both, labeled source domain and
unlablled target domain. Our method incrementally augments the labeled source
with target samples and iteratively modifies the embedding function to bring
the source and target distributions together. Additionally, we utilise a
multi-modal representation that incorporates noisy semantic information
available in form of hash-tags. We show the effectiveness of this simple
adaptation technique on a test set of vines and achieve notable improvements in
performance.
| [
{
"version": "v1",
"created": "Tue, 18 Oct 2016 13:45:32 GMT"
}
] | 2016-10-19T00:00:00 | [
[
"Singh",
"Aditya",
""
],
[
"Saini",
"Saurabh",
""
],
[
"Shah",
"Rajvi",
""
],
[
"Narayanan",
"P J",
""
]
] | TITLE: From Traditional to Modern : Domain Adaptation for Action Classification
in Short Social Video Clips
ABSTRACT: Short internet video clips like vines present a significantly wild
distribution compared to traditional video datasets. In this paper, we focus on
the problem of unsupervised action classification in wild vines using
traditional labeled datasets. To this end, we use a data augmentation based
simple domain adaptation strategy. We utilise semantic word2vec space as a
common subspace to embed video features from both, labeled source domain and
unlablled target domain. Our method incrementally augments the labeled source
with target samples and iteratively modifies the embedding function to bring
the source and target distributions together. Additionally, we utilise a
multi-modal representation that incorporates noisy semantic information
available in form of hash-tags. We show the effectiveness of this simple
adaptation technique on a test set of vines and achieve notable improvements in
performance.
| no_new_dataset | 0.952086 |
1508.04999 | Juhan Nam | Juhan Nam, Jorge Herrera, Kyogu Lee | A Deep Bag-of-Features Model for Music Auto-Tagging | We resubmit a new version to revive the paper and record it as a
technical report. We did not add any incremental work to the previous work
but removed out some sections (criticized by a review process) and polished
sentences accordingly | null | null | null | cs.LG cs.SD stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Feature learning and deep learning have drawn great attention in recent years
as a way of transforming input data into more effective representations using
learning algorithms. Such interest has grown in the area of music information
retrieval (MIR) as well, particularly in music audio classification tasks such
as auto-tagging. In this paper, we present a two-stage learning model to
effectively predict multiple labels from music audio. The first stage learns to
project local spectral patterns of an audio track onto a high-dimensional
sparse space in an unsupervised manner and summarizes the audio track as a
bag-of-features. The second stage successively performs the unsupervised
learning on the bag-of-features in a layer-by-layer manner to initialize a deep
neural network and finally fine-tunes it with the tag labels. Through the
experiment, we rigorously examine training choices and tuning parameters, and
show that the model achieves high performance on Magnatagatune, a popularly
used dataset in music auto-tagging.
| [
{
"version": "v1",
"created": "Thu, 20 Aug 2015 14:38:56 GMT"
},
{
"version": "v2",
"created": "Sat, 18 Jun 2016 02:45:04 GMT"
},
{
"version": "v3",
"created": "Sun, 16 Oct 2016 13:03:20 GMT"
}
] | 2016-10-18T00:00:00 | [
[
"Nam",
"Juhan",
""
],
[
"Herrera",
"Jorge",
""
],
[
"Lee",
"Kyogu",
""
]
] | TITLE: A Deep Bag-of-Features Model for Music Auto-Tagging
ABSTRACT: Feature learning and deep learning have drawn great attention in recent years
as a way of transforming input data into more effective representations using
learning algorithms. Such interest has grown in the area of music information
retrieval (MIR) as well, particularly in music audio classification tasks such
as auto-tagging. In this paper, we present a two-stage learning model to
effectively predict multiple labels from music audio. The first stage learns to
project local spectral patterns of an audio track onto a high-dimensional
sparse space in an unsupervised manner and summarizes the audio track as a
bag-of-features. The second stage successively performs the unsupervised
learning on the bag-of-features in a layer-by-layer manner to initialize a deep
neural network and finally fine-tunes it with the tag labels. Through the
experiment, we rigorously examine training choices and tuning parameters, and
show that the model achieves high performance on Magnatagatune, a popularly
used dataset in music auto-tagging.
| no_new_dataset | 0.941708 |
1603.00772 | Azad Naik | Azad Naik, Huzefa Rangwala | Filter based Taxonomy Modification for Improving Hierarchical
Classification | The conference version of the paper is submitted for publication | null | null | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Hierarchical Classification (HC) is a supervised learning problem where
unlabeled instances are classified into a taxonomy of classes. Several methods
that utilize the hierarchical structure have been developed to improve the HC
performance. However, in most cases apriori defined hierarchical structure by
domain experts is inconsistent; as a consequence performance improvement is not
noticeable in comparison to flat classification methods. We propose a scalable
data-driven filter based rewiring approach to modify an expert-defined
hierarchy. Experimental comparisons of top-down HC with our modified hierarchy,
on a wide range of datasets shows classification performance improvement over
the baseline hierarchy (i:e:, defined by expert), clustered hierarchy and
flattening based hierarchy modification approaches. In comparison to existing
rewiring approaches, our developed method (rewHier) is computationally
efficient, enabling it to scale to datasets with large numbers of classes,
instances and features. We also show that our modified hierarchy leads to
improved classification performance for classes with few training samples in
comparison to flat and state-of-the-art HC approaches.
| [
{
"version": "v1",
"created": "Wed, 2 Mar 2016 16:14:49 GMT"
},
{
"version": "v2",
"created": "Thu, 9 Jun 2016 06:41:42 GMT"
},
{
"version": "v3",
"created": "Sat, 15 Oct 2016 06:21:54 GMT"
}
] | 2016-10-18T00:00:00 | [
[
"Naik",
"Azad",
""
],
[
"Rangwala",
"Huzefa",
""
]
] | TITLE: Filter based Taxonomy Modification for Improving Hierarchical
Classification
ABSTRACT: Hierarchical Classification (HC) is a supervised learning problem where
unlabeled instances are classified into a taxonomy of classes. Several methods
that utilize the hierarchical structure have been developed to improve the HC
performance. However, in most cases apriori defined hierarchical structure by
domain experts is inconsistent; as a consequence performance improvement is not
noticeable in comparison to flat classification methods. We propose a scalable
data-driven filter based rewiring approach to modify an expert-defined
hierarchy. Experimental comparisons of top-down HC with our modified hierarchy,
on a wide range of datasets shows classification performance improvement over
the baseline hierarchy (i:e:, defined by expert), clustered hierarchy and
flattening based hierarchy modification approaches. In comparison to existing
rewiring approaches, our developed method (rewHier) is computationally
efficient, enabling it to scale to datasets with large numbers of classes,
instances and features. We also show that our modified hierarchy leads to
improved classification performance for classes with few training samples in
comparison to flat and state-of-the-art HC approaches.
| no_new_dataset | 0.954942 |
1603.04146 | Shuhan Chen | Shuhan Chen, Jindong Li, Xuelong Hu, Ping Zhou | Saliency Detection for Improving Object Proposals | IEEE DSP 2016 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Object proposals greatly benefit object detection task in recent
state-of-the-art works. However, the existing object proposals usually have low
localization accuracy at high intersection over union threshold. To address it,
we apply saliency detection to each bounding box to improve their quality in
this paper. We first present a geodesic saliency detection method in contour,
which is designed to find closed contours. Then, we apply it to each candidate
box with multi-sizes, and refined boxes can be easily produced in the obtained
saliency maps which are further used to calculate saliency scores for proposal
ranking. Experiments on PASCAL VOC 2007 test dataset demonstrate the proposed
refinement approach can greatly improve existing models.
| [
{
"version": "v1",
"created": "Mon, 14 Mar 2016 06:44:43 GMT"
},
{
"version": "v2",
"created": "Tue, 15 Mar 2016 02:01:08 GMT"
},
{
"version": "v3",
"created": "Mon, 17 Oct 2016 06:30:08 GMT"
}
] | 2016-10-18T00:00:00 | [
[
"Chen",
"Shuhan",
""
],
[
"Li",
"Jindong",
""
],
[
"Hu",
"Xuelong",
""
],
[
"Zhou",
"Ping",
""
]
] | TITLE: Saliency Detection for Improving Object Proposals
ABSTRACT: Object proposals greatly benefit object detection task in recent
state-of-the-art works. However, the existing object proposals usually have low
localization accuracy at high intersection over union threshold. To address it,
we apply saliency detection to each bounding box to improve their quality in
this paper. We first present a geodesic saliency detection method in contour,
which is designed to find closed contours. Then, we apply it to each candidate
box with multi-sizes, and refined boxes can be easily produced in the obtained
saliency maps which are further used to calculate saliency scores for proposal
ranking. Experiments on PASCAL VOC 2007 test dataset demonstrate the proposed
refinement approach can greatly improve existing models.
| no_new_dataset | 0.954223 |
1610.03147 | Pan Zhou Prof. | Yifan Hou, Pan Zhou, Ting Wang, Li Yu, Yuchong Hu, Dapeng Wu | Context-Aware Online Learning for Course Recommendation of MOOC Big Data | null | null | null | null | cs.LG cs.CY cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The Massive Open Online Course (MOOC) has expanded significantly in recent
years. With the widespread of MOOC, the opportunity to study the fascinating
courses for free has attracted numerous people of diverse educational
backgrounds all over the world. In the big data era, a key research topic for
MOOC is how to mine the needed courses in the massive course databases in cloud
for each individual student accurately and rapidly as the number of courses is
increasing fleetly. In this respect, the key challenge is how to realize
personalized course recommendation as well as to reduce the computing and
storage costs for the tremendous course data. In this paper, we propose a big
data-supported, context-aware online learning-based course recommender system
that could handle the dynamic and infinitely massive datasets, which recommends
courses by using personalized context information and historical statistics.
The context-awareness takes the personal preferences into consideration, making
the recommendation suitable for people with different backgrounds. Besides, the
algorithm achieves the sublinear regret performance, which means it can
gradually recommend the mostly preferred and matched courses to students. In
addition, our storage module is expanded to the distributed-connected storage
nodes, where the devised algorithm can handle massive course storage problems
from heterogeneous sources of course datasets. Comparing to existing
algorithms, our proposed algorithms achieve the linear time complexity and
space complexity. Experiment results verify the superiority of our algorithms
when comparing with existing ones in the MOOC big data setting.
| [
{
"version": "v1",
"created": "Tue, 11 Oct 2016 01:02:15 GMT"
},
{
"version": "v2",
"created": "Sun, 16 Oct 2016 03:34:37 GMT"
}
] | 2016-10-18T00:00:00 | [
[
"Hou",
"Yifan",
""
],
[
"Zhou",
"Pan",
""
],
[
"Wang",
"Ting",
""
],
[
"Yu",
"Li",
""
],
[
"Hu",
"Yuchong",
""
],
[
"Wu",
"Dapeng",
""
]
] | TITLE: Context-Aware Online Learning for Course Recommendation of MOOC Big Data
ABSTRACT: The Massive Open Online Course (MOOC) has expanded significantly in recent
years. With the widespread of MOOC, the opportunity to study the fascinating
courses for free has attracted numerous people of diverse educational
backgrounds all over the world. In the big data era, a key research topic for
MOOC is how to mine the needed courses in the massive course databases in cloud
for each individual student accurately and rapidly as the number of courses is
increasing fleetly. In this respect, the key challenge is how to realize
personalized course recommendation as well as to reduce the computing and
storage costs for the tremendous course data. In this paper, we propose a big
data-supported, context-aware online learning-based course recommender system
that could handle the dynamic and infinitely massive datasets, which recommends
courses by using personalized context information and historical statistics.
The context-awareness takes the personal preferences into consideration, making
the recommendation suitable for people with different backgrounds. Besides, the
algorithm achieves the sublinear regret performance, which means it can
gradually recommend the mostly preferred and matched courses to students. In
addition, our storage module is expanded to the distributed-connected storage
nodes, where the devised algorithm can handle massive course storage problems
from heterogeneous sources of course datasets. Comparing to existing
algorithms, our proposed algorithms achieve the linear time complexity and
space complexity. Experiment results verify the superiority of our algorithms
when comparing with existing ones in the MOOC big data setting.
| no_new_dataset | 0.951774 |
1610.04668 | Shuai Zheng | Shuai Zheng, Xiao Cai, Chris Ding, Feiping Nie, Heng Huang | A Closed Form Solution to Multi-View Low-Rank Regression | Twenty-Ninth AAAI Conference on Artificial Intelligence, AAAI, 2015 | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Real life data often includes information from different channels. For
example, in computer vision, we can describe an image using different image
features, such as pixel intensity, color, HOG, GIST feature, SIFT features,
etc.. These different aspects of the same objects are often called multi-view
(or multi-modal) data. Low-rank regression model has been proved to be an
effective learning mechanism by exploring the low-rank structure of real life
data. But previous low-rank regression model only works on single view data. In
this paper, we propose a multi-view low-rank regression model by imposing
low-rank constraints on multi-view regression model. Most importantly, we
provide a closed-form solution to the multi-view low-rank regression model.
Extensive experiments on 4 multi-view datasets show that the multi-view
low-rank regression model outperforms single-view regression model and reveals
that multi-view low-rank structure is very helpful.
| [
{
"version": "v1",
"created": "Fri, 14 Oct 2016 23:43:47 GMT"
}
] | 2016-10-18T00:00:00 | [
[
"Zheng",
"Shuai",
""
],
[
"Cai",
"Xiao",
""
],
[
"Ding",
"Chris",
""
],
[
"Nie",
"Feiping",
""
],
[
"Huang",
"Heng",
""
]
] | TITLE: A Closed Form Solution to Multi-View Low-Rank Regression
ABSTRACT: Real life data often includes information from different channels. For
example, in computer vision, we can describe an image using different image
features, such as pixel intensity, color, HOG, GIST feature, SIFT features,
etc.. These different aspects of the same objects are often called multi-view
(or multi-modal) data. Low-rank regression model has been proved to be an
effective learning mechanism by exploring the low-rank structure of real life
data. But previous low-rank regression model only works on single view data. In
this paper, we propose a multi-view low-rank regression model by imposing
low-rank constraints on multi-view regression model. Most importantly, we
provide a closed-form solution to the multi-view low-rank regression model.
Extensive experiments on 4 multi-view datasets show that the multi-view
low-rank regression model outperforms single-view regression model and reveals
that multi-view low-rank structure is very helpful.
| no_new_dataset | 0.948058 |
1610.04725 | Takoua Kefi | Takoua Kefi, Riadh Ksantini, M.Becha Kaaniche and Adel Bouhoula | Incremental One-Class Models for Data Classification | 4 pages, accepted in PhD Forum Session of the ECML-PKDD 2016 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper we outline a PhD research plan. This research contributes to
the field of one-class incremental learning and classification in case of
non-stationary environments. The goal of this PhD is to define a new
classification framework able to deal with very small learning dataset at the
beginning of the process and with abilities to adjust itself according to the
variability of the incoming data which create large scale datasets. As a
preliminary work, incremental Covariance-guided One-Class Support Vector
Machine is proposed to deal with sequentially obtained data. It is inspired
from COSVM which put more emphasis on the low variance directions while keeping
the basic formulation of incremental One-Class Support Vector Machine
untouched. The incremental procedure is introduced by controlling the possible
changes of support vectors after the addition of new data points, thanks to the
Karush-Kuhn-Tucker conditions, that have to be maintained on all previously
acquired data. Comparative experimental results with contemporary incremental
and non-incremental one-class classifiers on numerous artificial and real data
sets show that our method results in significantly better classification
performance.
| [
{
"version": "v1",
"created": "Sat, 15 Oct 2016 12:06:12 GMT"
}
] | 2016-10-18T00:00:00 | [
[
"Kefi",
"Takoua",
""
],
[
"Ksantini",
"Riadh",
""
],
[
"Kaaniche",
"M. Becha",
""
],
[
"Bouhoula",
"Adel",
""
]
] | TITLE: Incremental One-Class Models for Data Classification
ABSTRACT: In this paper we outline a PhD research plan. This research contributes to
the field of one-class incremental learning and classification in case of
non-stationary environments. The goal of this PhD is to define a new
classification framework able to deal with very small learning dataset at the
beginning of the process and with abilities to adjust itself according to the
variability of the incoming data which create large scale datasets. As a
preliminary work, incremental Covariance-guided One-Class Support Vector
Machine is proposed to deal with sequentially obtained data. It is inspired
from COSVM which put more emphasis on the low variance directions while keeping
the basic formulation of incremental One-Class Support Vector Machine
untouched. The incremental procedure is introduced by controlling the possible
changes of support vectors after the addition of new data points, thanks to the
Karush-Kuhn-Tucker conditions, that have to be maintained on all previously
acquired data. Comparative experimental results with contemporary incremental
and non-incremental one-class classifiers on numerous artificial and real data
sets show that our method results in significantly better classification
performance.
| no_new_dataset | 0.945851 |
1610.04730 | Piotr Sapiezynski | Piotr Sapiezynski, Arkadiusz Stopczynski, David Kofoed Wind, Jure
Leskovec, Sune Lehmann | Inferring Person-to-person Proximity Using WiFi Signals | null | null | null | null | cs.CY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Today's societies are enveloped in an ever-growing telecommunication
infrastructure. This infrastructure offers important opportunities for sensing
and recording a multitude of human behaviors. Human mobility patterns are a
prominent example of such a behavior which has been studied based on cell phone
towers, Bluetooth beacons, and WiFi networks as proxies for location. However,
while mobility is an important aspect of human behavior, understanding complex
social systems requires studying not only the movement of individuals, but also
their interactions. Sensing social interactions on a large scale is a technical
challenge and many commonly used approaches---including RFID badges or
Bluetooth scanning---offer only limited scalability. Here we show that it is
possible, in a scalable and robust way, to accurately infer person-to-person
physical proximity from the lists of WiFi access points measured by smartphones
carried by the two individuals. Based on a longitudinal dataset of
approximately 800 participants with ground-truth interactions collected over a
year, we show that our model performs better than the current state-of-the-art.
Our results demonstrate the value of WiFi signals in social sensing as well as
potential threats to privacy that they imply.
| [
{
"version": "v1",
"created": "Sat, 15 Oct 2016 13:02:46 GMT"
}
] | 2016-10-18T00:00:00 | [
[
"Sapiezynski",
"Piotr",
""
],
[
"Stopczynski",
"Arkadiusz",
""
],
[
"Wind",
"David Kofoed",
""
],
[
"Leskovec",
"Jure",
""
],
[
"Lehmann",
"Sune",
""
]
] | TITLE: Inferring Person-to-person Proximity Using WiFi Signals
ABSTRACT: Today's societies are enveloped in an ever-growing telecommunication
infrastructure. This infrastructure offers important opportunities for sensing
and recording a multitude of human behaviors. Human mobility patterns are a
prominent example of such a behavior which has been studied based on cell phone
towers, Bluetooth beacons, and WiFi networks as proxies for location. However,
while mobility is an important aspect of human behavior, understanding complex
social systems requires studying not only the movement of individuals, but also
their interactions. Sensing social interactions on a large scale is a technical
challenge and many commonly used approaches---including RFID badges or
Bluetooth scanning---offer only limited scalability. Here we show that it is
possible, in a scalable and robust way, to accurately infer person-to-person
physical proximity from the lists of WiFi access points measured by smartphones
carried by the two individuals. Based on a longitudinal dataset of
approximately 800 participants with ground-truth interactions collected over a
year, we show that our model performs better than the current state-of-the-art.
Our results demonstrate the value of WiFi signals in social sensing as well as
potential threats to privacy that they imply.
| new_dataset | 0.964855 |
1610.04752 | Paolo Missier | Paolo Missier and Jacek Cala and Maisha Rathi | Preserving the value of large scale data analytics over time through
selective re-computation | null | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A pervasive problem in Data Science is that the knowledge generated by
possibly expensive analytics processes is subject to decay over time, as the
data used to compute it drifts, the algorithms used in the processes are
improved, and the external knowledge embodied by reference datasets used in the
computation evolves. Deciding when such knowledge outcomes should be refreshed,
following a sequence of data change events, requires problem-specific functions
to quantify their value and its decay over time, as well as models for
estimating the cost of their re-computation. What makes this problem
challenging is the ambition to develop a decision support system for informing
data analytics re-computation decisions over time, that is both generic and
customisable. With the help of a case study from genomics, in this vision paper
we offer an initial formalisation of this problem, highlight research
challenges, and outline a possible approach based on the collection and
analysis of metadata from a history of past computations.
| [
{
"version": "v1",
"created": "Sat, 15 Oct 2016 16:08:22 GMT"
}
] | 2016-10-18T00:00:00 | [
[
"Missier",
"Paolo",
""
],
[
"Cala",
"Jacek",
""
],
[
"Rathi",
"Maisha",
""
]
] | TITLE: Preserving the value of large scale data analytics over time through
selective re-computation
ABSTRACT: A pervasive problem in Data Science is that the knowledge generated by
possibly expensive analytics processes is subject to decay over time, as the
data used to compute it drifts, the algorithms used in the processes are
improved, and the external knowledge embodied by reference datasets used in the
computation evolves. Deciding when such knowledge outcomes should be refreshed,
following a sequence of data change events, requires problem-specific functions
to quantify their value and its decay over time, as well as models for
estimating the cost of their re-computation. What makes this problem
challenging is the ambition to develop a decision support system for informing
data analytics re-computation decisions over time, that is both generic and
customisable. With the help of a case study from genomics, in this vision paper
we offer an initial formalisation of this problem, highlight research
challenges, and outline a possible approach based on the collection and
analysis of metadata from a history of past computations.
| no_new_dataset | 0.944893 |
1610.04814 | Mahamad Suhil | D S Guru and Mahamad Suhil | Term-Class-Max-Support (TCMS): A Simple Text Document Categorization
Approach Using Term-Class Relevance Measure | 4 Pages, 4 Figures; 2016 Intl. Conference on Advances in Computing,
Communications and Informatics (ICACCI), Sept. 21-24, 2016, Jaipur, India | null | null | null | cs.IR cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, a simple text categorization method using term-class relevance
measures is proposed. Initially, text documents are processed to extract
significant terms present in them. For every term extracted from a document, we
compute its importance in preserving the content of a class through a novel
term-weighting scheme known as Term_Class Relevance (TCR) measure proposed by
Guru and Suhil (2015) [1]. In this way, for every term, its relevance for all
the classes present in the corpus is computed and stored in the knowledgebase.
During testing, the terms present in the test document are extracted and the
term-class relevance of each term is obtained from the stored knowledgebase. To
achieve quick search of term weights, Btree indexing data structure has been
adapted. Finally, the class which receives maximum support in terms of
term-class relevance is decided to be the class of the given test document. The
proposed method works in logarithmic complexity in testing time and simple to
implement when compared to any other text categorization techniques available
in literature. The experiments conducted on various benchmarking datasets have
revealed that the performance of the proposed method is satisfactory and
encouraging.
| [
{
"version": "v1",
"created": "Sun, 16 Oct 2016 03:40:13 GMT"
}
] | 2016-10-18T00:00:00 | [
[
"Guru",
"D S",
""
],
[
"Suhil",
"Mahamad",
""
]
] | TITLE: Term-Class-Max-Support (TCMS): A Simple Text Document Categorization
Approach Using Term-Class Relevance Measure
ABSTRACT: In this paper, a simple text categorization method using term-class relevance
measures is proposed. Initially, text documents are processed to extract
significant terms present in them. For every term extracted from a document, we
compute its importance in preserving the content of a class through a novel
term-weighting scheme known as Term_Class Relevance (TCR) measure proposed by
Guru and Suhil (2015) [1]. In this way, for every term, its relevance for all
the classes present in the corpus is computed and stored in the knowledgebase.
During testing, the terms present in the test document are extracted and the
term-class relevance of each term is obtained from the stored knowledgebase. To
achieve quick search of term weights, Btree indexing data structure has been
adapted. Finally, the class which receives maximum support in terms of
term-class relevance is decided to be the class of the given test document. The
proposed method works in logarithmic complexity in testing time and simple to
implement when compared to any other text categorization techniques available
in literature. The experiments conducted on various benchmarking datasets have
revealed that the performance of the proposed method is satisfactory and
encouraging.
| no_new_dataset | 0.95452 |
1610.04889 | Srinath Sridhar | Srinath Sridhar, Franziska Mueller, Michael Zollh\"ofer, Dan Casas,
Antti Oulasvirta, Christian Theobalt | Real-time Joint Tracking of a Hand Manipulating an Object from RGB-D
Input | Proceedings of ECCV 2016 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Real-time simultaneous tracking of hands manipulating and interacting with
external objects has many potential applications in augmented reality, tangible
computing, and wearable computing. However, due to difficult occlusions, fast
motions, and uniform hand appearance, jointly tracking hand and object pose is
more challenging than tracking either of the two separately. Many previous
approaches resort to complex multi-camera setups to remedy the occlusion
problem and often employ expensive segmentation and optimization steps which
makes real-time tracking impossible. In this paper, we propose a real-time
solution that uses a single commodity RGB-D camera. The core of our approach is
a 3D articulated Gaussian mixture alignment strategy tailored to hand-object
tracking that allows fast pose optimization. The alignment energy uses novel
regularizers to address occlusions and hand-object contacts. For added
robustness, we guide the optimization with discriminative part classification
of the hand and segmentation of the object. We conducted extensive experiments
on several existing datasets and introduce a new annotated hand-object dataset.
Quantitative and qualitative results show the key advantages of our method:
speed, accuracy, and robustness.
| [
{
"version": "v1",
"created": "Sun, 16 Oct 2016 17:11:58 GMT"
}
] | 2016-10-18T00:00:00 | [
[
"Sridhar",
"Srinath",
""
],
[
"Mueller",
"Franziska",
""
],
[
"Zollhöfer",
"Michael",
""
],
[
"Casas",
"Dan",
""
],
[
"Oulasvirta",
"Antti",
""
],
[
"Theobalt",
"Christian",
""
]
] | TITLE: Real-time Joint Tracking of a Hand Manipulating an Object from RGB-D
Input
ABSTRACT: Real-time simultaneous tracking of hands manipulating and interacting with
external objects has many potential applications in augmented reality, tangible
computing, and wearable computing. However, due to difficult occlusions, fast
motions, and uniform hand appearance, jointly tracking hand and object pose is
more challenging than tracking either of the two separately. Many previous
approaches resort to complex multi-camera setups to remedy the occlusion
problem and often employ expensive segmentation and optimization steps which
makes real-time tracking impossible. In this paper, we propose a real-time
solution that uses a single commodity RGB-D camera. The core of our approach is
a 3D articulated Gaussian mixture alignment strategy tailored to hand-object
tracking that allows fast pose optimization. The alignment energy uses novel
regularizers to address occlusions and hand-object contacts. For added
robustness, we guide the optimization with discriminative part classification
of the hand and segmentation of the object. We conducted extensive experiments
on several existing datasets and introduce a new annotated hand-object dataset.
Quantitative and qualitative results show the key advantages of our method:
speed, accuracy, and robustness.
| new_dataset | 0.958499 |
1610.04929 | Li Wang | Li Wang | Probabilistic Dimensionality Reduction via Structure Learning | 32 pages, 6 figures | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a novel probabilistic dimensionality reduction framework that can
naturally integrate the generative model and the locality information of data.
Based on this framework, we present a new model, which is able to learn a
smooth skeleton of embedding points in a low-dimensional space from
high-dimensional noisy data. The formulation of the new model can be
equivalently interpreted as two coupled learning problem, i.e., structure
learning and the learning of projection matrix. This interpretation motivates
the learning of the embedding points that can directly form an explicit graph
structure. We develop a new method to learn the embedding points that form a
spanning tree, which is further extended to obtain a discriminative and compact
feature representation for clustering problems. Unlike traditional clustering
methods, we assume that centers of clusters should be close to each other if
they are connected in a learned graph, and other cluster centers should be
distant. This can greatly facilitate data visualization and scientific
discovery in downstream analysis. Extensive experiments are performed that
demonstrate that the proposed framework is able to obtain discriminative
feature representations, and correctly recover the intrinsic structures of
various real-world datasets.
| [
{
"version": "v1",
"created": "Sun, 16 Oct 2016 23:37:26 GMT"
}
] | 2016-10-18T00:00:00 | [
[
"Wang",
"Li",
""
]
] | TITLE: Probabilistic Dimensionality Reduction via Structure Learning
ABSTRACT: We propose a novel probabilistic dimensionality reduction framework that can
naturally integrate the generative model and the locality information of data.
Based on this framework, we present a new model, which is able to learn a
smooth skeleton of embedding points in a low-dimensional space from
high-dimensional noisy data. The formulation of the new model can be
equivalently interpreted as two coupled learning problem, i.e., structure
learning and the learning of projection matrix. This interpretation motivates
the learning of the embedding points that can directly form an explicit graph
structure. We develop a new method to learn the embedding points that form a
spanning tree, which is further extended to obtain a discriminative and compact
feature representation for clustering problems. Unlike traditional clustering
methods, we assume that centers of clusters should be close to each other if
they are connected in a learned graph, and other cluster centers should be
distant. This can greatly facilitate data visualization and scientific
discovery in downstream analysis. Extensive experiments are performed that
demonstrate that the proposed framework is able to obtain discriminative
feature representations, and correctly recover the intrinsic structures of
various real-world datasets.
| no_new_dataset | 0.947962 |
1610.04957 | Arnold Wiliem | Liangchen Liu, Arnold Wiliem, Shaokang Chen, Brian C. Lovell | What is the Best Way for Extracting Meaningful Attributes from Pictures? | Submission to Pattern Recognition | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Automatic attribute discovery methods have gained in popularity to extract
sets of visual attributes from images or videos for various tasks. Despite
their good performance in some classification tasks, it is difficult to
evaluate whether the attributes discovered by these methods are meaningful and
which methods are the most appropriate to discover attributes for visual
descriptions. In its simplest form, such an evaluation can be performed by
manually verifying whether there is any consistent identifiable visual concept
distinguishing between positive and negative exemplars labelled by an
attribute. This manual checking is tedious, expensive and labour intensive. In
addition, comparisons between different methods could also be problematic as it
is not clear how one could quantitatively decide which attribute is more
meaningful than the others. In this paper, we propose a novel attribute
meaningfulness metric to address this challenging problem. With this metric,
automatic quantitative evaluation can be performed on the attribute sets; thus,
reducing the enormous effort to perform manual evaluation. The proposed metric
is applied to some recent automatic attribute discovery and hashing methods on
four attribute-labelled datasets. To further validate the efficacy of the
proposed method, we conducted a user study. In addition, we also compared our
metric with a semi-supervised attribute discover method using the mixture of
probabilistic PCA. In our evaluation, we gleaned several insights that could be
beneficial in developing new automatic attribute discovery methods.
| [
{
"version": "v1",
"created": "Mon, 17 Oct 2016 02:51:43 GMT"
}
] | 2016-10-18T00:00:00 | [
[
"Liu",
"Liangchen",
""
],
[
"Wiliem",
"Arnold",
""
],
[
"Chen",
"Shaokang",
""
],
[
"Lovell",
"Brian C.",
""
]
] | TITLE: What is the Best Way for Extracting Meaningful Attributes from Pictures?
ABSTRACT: Automatic attribute discovery methods have gained in popularity to extract
sets of visual attributes from images or videos for various tasks. Despite
their good performance in some classification tasks, it is difficult to
evaluate whether the attributes discovered by these methods are meaningful and
which methods are the most appropriate to discover attributes for visual
descriptions. In its simplest form, such an evaluation can be performed by
manually verifying whether there is any consistent identifiable visual concept
distinguishing between positive and negative exemplars labelled by an
attribute. This manual checking is tedious, expensive and labour intensive. In
addition, comparisons between different methods could also be problematic as it
is not clear how one could quantitatively decide which attribute is more
meaningful than the others. In this paper, we propose a novel attribute
meaningfulness metric to address this challenging problem. With this metric,
automatic quantitative evaluation can be performed on the attribute sets; thus,
reducing the enormous effort to perform manual evaluation. The proposed metric
is applied to some recent automatic attribute discovery and hashing methods on
four attribute-labelled datasets. To further validate the efficacy of the
proposed method, we conducted a user study. In addition, we also compared our
metric with a semi-supervised attribute discover method using the mixture of
probabilistic PCA. In our evaluation, we gleaned several insights that could be
beneficial in developing new automatic attribute discovery methods.
| no_new_dataset | 0.941493 |
1610.04963 | Hui Miao | Hui Miao, Amit Chavan, Amol Deshpande | ProvDB: A System for Lifecycle Management of Collaborative Analysis
Workflows | null | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As data-driven methods are becoming pervasive in a wide variety of
disciplines, there is an urgent need to develop scalable and sustainable tools
to simplify the process of data science, to make it easier to keep track of the
analyses being performed and datasets being generated, and to enable
introspection of the workflows. In this paper, we describe our vision of a
unified provenance and metadata management system to support lifecycle
management of complex collaborative data science workflows. We argue that a
large amount of information about the analysis processes and data artifacts
can, and should be, captured in a semi-passive manner; and we show that
querying and analyzing this information can not only simplify bookkeeping and
debugging tasks for data analysts but can also enable a rich new set of
capabilities like identifying flaws in the data science process itself. It can
also significantly reduce the time spent in fixing post-deployment problems
through automated analysis and monitoring. We have implemented an initial
prototype of our system, called ProvDB, on top of git (a version control
system) and Neo4j (a graph database), and we describe its key features and
capabilities.
| [
{
"version": "v1",
"created": "Mon, 17 Oct 2016 03:22:58 GMT"
}
] | 2016-10-18T00:00:00 | [
[
"Miao",
"Hui",
""
],
[
"Chavan",
"Amit",
""
],
[
"Deshpande",
"Amol",
""
]
] | TITLE: ProvDB: A System for Lifecycle Management of Collaborative Analysis
Workflows
ABSTRACT: As data-driven methods are becoming pervasive in a wide variety of
disciplines, there is an urgent need to develop scalable and sustainable tools
to simplify the process of data science, to make it easier to keep track of the
analyses being performed and datasets being generated, and to enable
introspection of the workflows. In this paper, we describe our vision of a
unified provenance and metadata management system to support lifecycle
management of complex collaborative data science workflows. We argue that a
large amount of information about the analysis processes and data artifacts
can, and should be, captured in a semi-passive manner; and we show that
querying and analyzing this information can not only simplify bookkeeping and
debugging tasks for data analysts but can also enable a rich new set of
capabilities like identifying flaws in the data science process itself. It can
also significantly reduce the time spent in fixing post-deployment problems
through automated analysis and monitoring. We have implemented an initial
prototype of our system, called ProvDB, on top of git (a version control
system) and Neo4j (a graph database), and we describe its key features and
capabilities.
| no_new_dataset | 0.941439 |
1610.04973 | Amine Ben Khalifa | Amine Ben Khalifa and Hichem Frigui | Multiple Instance Fuzzy Inference Neural Networks | Submitted to IEEE Transactions On Cybernetics for review | null | null | null | cs.NE cs.CV cs.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Fuzzy logic is a powerful tool to model knowledge uncertainty, measurements
imprecision, and vagueness. However, there is another type of vagueness that
arises when data have multiple forms of expression that fuzzy logic does not
address quite well. This is the case for multiple instance learning problems
(MIL). In MIL, an object is represented by a collection of instances, called a
bag. A bag is labeled negative if all of its instances are negative, and
positive if at least one of its instances is positive. Positive bags encode
ambiguity since the instances themselves are not labeled. In this paper, we
introduce fuzzy inference systems and neural networks designed to handle bags
of instances as input and capable of learning from ambiguously labeled data.
First, we introduce the Multiple Instance Sugeno style fuzzy inference
(MI-Sugeno) that extends the standard Sugeno style inference to handle
reasoning with multiple instances. Second, we use MI-Sugeno to define and
develop Multiple Instance Adaptive Neuro Fuzzy Inference System (MI-ANFIS). We
expand the architecture of the standard ANFIS to allow reasoning with bags and
derive a learning algorithm using backpropagation to identify the premise and
consequent parameters of the network. The proposed inference system is tested
and validated using synthetic and benchmark datasets suitable for MIL problems.
We also apply the proposed MI-ANFIS to fuse the output of multiple
discrimination algorithms for the purpose of landmine detection using Ground
Penetrating Radar.
| [
{
"version": "v1",
"created": "Mon, 17 Oct 2016 05:07:09 GMT"
}
] | 2016-10-18T00:00:00 | [
[
"Khalifa",
"Amine Ben",
""
],
[
"Frigui",
"Hichem",
""
]
] | TITLE: Multiple Instance Fuzzy Inference Neural Networks
ABSTRACT: Fuzzy logic is a powerful tool to model knowledge uncertainty, measurements
imprecision, and vagueness. However, there is another type of vagueness that
arises when data have multiple forms of expression that fuzzy logic does not
address quite well. This is the case for multiple instance learning problems
(MIL). In MIL, an object is represented by a collection of instances, called a
bag. A bag is labeled negative if all of its instances are negative, and
positive if at least one of its instances is positive. Positive bags encode
ambiguity since the instances themselves are not labeled. In this paper, we
introduce fuzzy inference systems and neural networks designed to handle bags
of instances as input and capable of learning from ambiguously labeled data.
First, we introduce the Multiple Instance Sugeno style fuzzy inference
(MI-Sugeno) that extends the standard Sugeno style inference to handle
reasoning with multiple instances. Second, we use MI-Sugeno to define and
develop Multiple Instance Adaptive Neuro Fuzzy Inference System (MI-ANFIS). We
expand the architecture of the standard ANFIS to allow reasoning with bags and
derive a learning algorithm using backpropagation to identify the premise and
consequent parameters of the network. The proposed inference system is tested
and validated using synthetic and benchmark datasets suitable for MIL problems.
We also apply the proposed MI-ANFIS to fuse the output of multiple
discrimination algorithms for the purpose of landmine detection using Ground
Penetrating Radar.
| no_new_dataset | 0.947284 |
1610.04989 | Jiacheng Xu | Jiacheng Xu, Danlu Chen, Xipeng Qiu and Xuangjing Huang | Cached Long Short-Term Memory Neural Networks for Document-Level
Sentiment Classification | Published as long paper of EMNLP2016 | null | null | null | cs.CL cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recently, neural networks have achieved great success on sentiment
classification due to their ability to alleviate feature engineering. However,
one of the remaining challenges is to model long texts in document-level
sentiment classification under a recurrent architecture because of the
deficiency of the memory unit. To address this problem, we present a Cached
Long Short-Term Memory neural networks (CLSTM) to capture the overall semantic
information in long texts. CLSTM introduces a cache mechanism, which divides
memory into several groups with different forgetting rates and thus enables the
network to keep sentiment information better within a recurrent unit. The
proposed CLSTM outperforms the state-of-the-art models on three publicly
available document-level sentiment analysis datasets.
| [
{
"version": "v1",
"created": "Mon, 17 Oct 2016 07:28:06 GMT"
}
] | 2016-10-18T00:00:00 | [
[
"Xu",
"Jiacheng",
""
],
[
"Chen",
"Danlu",
""
],
[
"Qiu",
"Xipeng",
""
],
[
"Huang",
"Xuangjing",
""
]
] | TITLE: Cached Long Short-Term Memory Neural Networks for Document-Level
Sentiment Classification
ABSTRACT: Recently, neural networks have achieved great success on sentiment
classification due to their ability to alleviate feature engineering. However,
one of the remaining challenges is to model long texts in document-level
sentiment classification under a recurrent architecture because of the
deficiency of the memory unit. To address this problem, we present a Cached
Long Short-Term Memory neural networks (CLSTM) to capture the overall semantic
information in long texts. CLSTM introduces a cache mechanism, which divides
memory into several groups with different forgetting rates and thus enables the
network to keep sentiment information better within a recurrent unit. The
proposed CLSTM outperforms the state-of-the-art models on three publicly
available document-level sentiment analysis datasets.
| no_new_dataset | 0.946597 |
1610.05036 | Itir Onal Ertugrul | Itir Onal Ertugrul and Mete Ozay and Fatos T. Yarman Vural | Encoding the Local Connectivity Patterns of fMRI for Cognitive State
Classification | 8 pages, 5 figures | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work, we propose a novel framework to encode the local connectivity
patterns of brain, using Fisher Vectors (FV), Vector of Locally Aggregated
Descriptors (VLAD) and Bag-of-Words (BoW) methods. We first obtain local
descriptors, called Mesh Arc Descriptors (MADs) from fMRI data, by forming
local meshes around anatomical regions, and estimating their relationship
within a neighborhood. Then, we extract a dictionary of relationships, called
\textit{brain connectivity dictionary} by fitting a generative Gaussian mixture
model (GMM) to a set of MADs, and selecting the codewords at the mean of each
component of the mixture. Codewords represent the connectivity patterns among
anatomical regions. We also encode MADs by VLAD and BoW methods using the
k-Means clustering.
We classify the cognitive states of Human Connectome Project (HCP) task fMRI
dataset, where we train support vector machines (SVM) by the encoded MADs.
Results demonstrate that, FV encoding of MADs can be successfully employed for
classification of cognitive tasks, and outperform the VLAD and BoW
representations. Moreover, we identify the significant Gaussians in mixture
models by computing energy of their corresponding FV parts, and analyze their
effect on classification accuracy. Finally, we suggest a new method to
visualize the codewords of brain connectivity dictionary.
| [
{
"version": "v1",
"created": "Mon, 17 Oct 2016 10:08:09 GMT"
}
] | 2016-10-18T00:00:00 | [
[
"Ertugrul",
"Itir Onal",
""
],
[
"Ozay",
"Mete",
""
],
[
"Vural",
"Fatos T. Yarman",
""
]
] | TITLE: Encoding the Local Connectivity Patterns of fMRI for Cognitive State
Classification
ABSTRACT: In this work, we propose a novel framework to encode the local connectivity
patterns of brain, using Fisher Vectors (FV), Vector of Locally Aggregated
Descriptors (VLAD) and Bag-of-Words (BoW) methods. We first obtain local
descriptors, called Mesh Arc Descriptors (MADs) from fMRI data, by forming
local meshes around anatomical regions, and estimating their relationship
within a neighborhood. Then, we extract a dictionary of relationships, called
\textit{brain connectivity dictionary} by fitting a generative Gaussian mixture
model (GMM) to a set of MADs, and selecting the codewords at the mean of each
component of the mixture. Codewords represent the connectivity patterns among
anatomical regions. We also encode MADs by VLAD and BoW methods using the
k-Means clustering.
We classify the cognitive states of Human Connectome Project (HCP) task fMRI
dataset, where we train support vector machines (SVM) by the encoded MADs.
Results demonstrate that, FV encoding of MADs can be successfully employed for
classification of cognitive tasks, and outperform the VLAD and BoW
representations. Moreover, we identify the significant Gaussians in mixture
models by computing energy of their corresponding FV parts, and analyze their
effect on classification accuracy. Finally, we suggest a new method to
visualize the codewords of brain connectivity dictionary.
| no_new_dataset | 0.950686 |
1610.05045 | Luisa Cutillo | Annamaria Carissimo and Luisa Cutillo and Italia Defeis | Validation of community robustness | arXiv admin note: text overlap with arXiv:0908.1062,
arXiv:cond-mat/0610077 by other authors | null | null | null | cs.SI cs.DS physics.soc-ph stat.ME | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The large amount of work on community detection and its applications leaves
unaddressed one important question: the statistical validation of the results.
In this paper we present a methodology able to clearly detect if the community
structure found by some algorithms is statistically significant or is a result
of chance, merely due to edge positions in the network. Given a community
detection method and a network of interest, our proposal examines the stability
of the partition recovered against random perturbations of the original graph
structure. To address this issue, we specify a perturbation strategy and a null
model to build a set of procedures based on a special measure of clustering
distance, namely Variation of Information, using tools set up for functional
data analysis. The procedures determine whether the obtained clustering departs
significantly from the null model. This strongly supports the robustness
against perturbation of the algorithm used to identify the community structure.
We show the results obtained with the proposed technique on simulated and real
datasets.
| [
{
"version": "v1",
"created": "Mon, 17 Oct 2016 11:16:18 GMT"
}
] | 2016-10-18T00:00:00 | [
[
"Carissimo",
"Annamaria",
""
],
[
"Cutillo",
"Luisa",
""
],
[
"Defeis",
"Italia",
""
]
] | TITLE: Validation of community robustness
ABSTRACT: The large amount of work on community detection and its applications leaves
unaddressed one important question: the statistical validation of the results.
In this paper we present a methodology able to clearly detect if the community
structure found by some algorithms is statistically significant or is a result
of chance, merely due to edge positions in the network. Given a community
detection method and a network of interest, our proposal examines the stability
of the partition recovered against random perturbations of the original graph
structure. To address this issue, we specify a perturbation strategy and a null
model to build a set of procedures based on a special measure of clustering
distance, namely Variation of Information, using tools set up for functional
data analysis. The procedures determine whether the obtained clustering departs
significantly from the null model. This strongly supports the robustness
against perturbation of the algorithm used to identify the community structure.
We show the results obtained with the proposed technique on simulated and real
datasets.
| no_new_dataset | 0.946051 |
1610.05112 | Harishchandra Dubey | Harishchandra Dubey, Ramdas Kumaresan, Kunal Mankodiya | Harmonic Sum-based Method for Heart Rate Estimation using PPG Signals
Affected with Motion Artifacts | 14 Pages, 11 Figures, 2 Tables, 27 Equations, Journal of Ambient
Intelligence and Humanized Computing, Oct. 2016 | null | null | null | cs.CY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Wearable photoplethysmography (WPPG) has recently become a common technology
in heart rate (HR) monitoring. General observation is that the motion artifacts
change the statistics of the acquired PPG signal. Consequently, estimation of
HR from such a corrupted PPG signal is challenging. However, if an
accelerometer is also used to acquire the acceleration signal simultaneously,
it can provide helpful information that can be used to reduce the motion
artifacts in the PPG signal. By dint of repetitive movements of the subjects
hands while running, the accelerometer signal is found to be quasi-periodic.
Over short-time intervals, it can be modeled by a finite harmonic sum (HSUM).
Using the harmonic sum (HSUM) model, we obtain an estimate of the instantaneous
fundamental frequency of the accelerometer signal. Since the PPG signal is a
composite of the heart rate information (that is also quasi-periodic) and the
motion artifact, we fit a joint harmonic sum (HSUM) model to the PPG signal.
One of the harmonic sums corresponds to the heart-beat component in PPG and the
other models the motion artifact. However, the fundamental frequency of the
motion artifact has already been determined from the accelerometer signal.
Subsequently, the HR is estimated from the joint HSUM model. The mean absolute
error in HR estimates was 0.7359 beats per minute (BPM) with a standard
deviation of 0.8328 BPM for 2015 IEEE Signal Processing (SP) cup data. The
ground-truth HR was obtained from the simultaneously acquired ECG for
validating the accuracy of the proposed method. The proposed method is compared
with four methods that were recently developed and evaluated on the same
dataset.
| [
{
"version": "v1",
"created": "Mon, 3 Oct 2016 17:52:09 GMT"
}
] | 2016-10-18T00:00:00 | [
[
"Dubey",
"Harishchandra",
""
],
[
"Kumaresan",
"Ramdas",
""
],
[
"Mankodiya",
"Kunal",
""
]
] | TITLE: Harmonic Sum-based Method for Heart Rate Estimation using PPG Signals
Affected with Motion Artifacts
ABSTRACT: Wearable photoplethysmography (WPPG) has recently become a common technology
in heart rate (HR) monitoring. General observation is that the motion artifacts
change the statistics of the acquired PPG signal. Consequently, estimation of
HR from such a corrupted PPG signal is challenging. However, if an
accelerometer is also used to acquire the acceleration signal simultaneously,
it can provide helpful information that can be used to reduce the motion
artifacts in the PPG signal. By dint of repetitive movements of the subjects
hands while running, the accelerometer signal is found to be quasi-periodic.
Over short-time intervals, it can be modeled by a finite harmonic sum (HSUM).
Using the harmonic sum (HSUM) model, we obtain an estimate of the instantaneous
fundamental frequency of the accelerometer signal. Since the PPG signal is a
composite of the heart rate information (that is also quasi-periodic) and the
motion artifact, we fit a joint harmonic sum (HSUM) model to the PPG signal.
One of the harmonic sums corresponds to the heart-beat component in PPG and the
other models the motion artifact. However, the fundamental frequency of the
motion artifact has already been determined from the accelerometer signal.
Subsequently, the HR is estimated from the joint HSUM model. The mean absolute
error in HR estimates was 0.7359 beats per minute (BPM) with a standard
deviation of 0.8328 BPM for 2015 IEEE Signal Processing (SP) cup data. The
ground-truth HR was obtained from the simultaneously acquired ECG for
validating the accuracy of the proposed method. The proposed method is compared
with four methods that were recently developed and evaluated on the same
dataset.
| no_new_dataset | 0.945551 |
1610.05116 | Sameh Shohdy | Sameh Shohdy, Abhinav Vishnu, Gagan Agrawal | Fault Tolerant Frequent Pattern Mining | 10 Pages, High Performance Computing Conference (HIPC 2016) | null | null | null | cs.DC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | FP-Growth algorithm is a Frequent Pattern Min- ing (FPM) algorithm that has
been extensively used to study correlations and patterns in large scale
datasets. While several researchers have designed distributed memory FP-Growth
algorithms, it is pivotal to consider fault tolerant FP-Growth, which can
address the increasing fault rates in large scale systems. In this work, we
propose a novel parallel, algorithm-level fault-tolerant FP-Growth algorithm.
We leverage algorithmic properties and MPI advanced features to guarantee an
O(1) space complexity, achieved by using the dataset memory space itself for
checkpointing. We also propose a recovery algorithm that can use in-memory and
disk-based checkpointing, though in many cases the recovery can be completed
without any disk access, and incurring no memory overhead for checkpointing. We
evaluate our FT algorithm on a large scale InfiniBand cluster with several
large datasets using up to 2K cores. Our evaluation demonstrates excellent
efficiency for checkpointing and recovery in comparison to the disk-based
approach. We have also observed 20x average speed-up in comparison to Spark,
establishing that a well designed algorithm can easily outperform a solution
based on a general fault-tolerant programming model.
| [
{
"version": "v1",
"created": "Mon, 17 Oct 2016 13:54:53 GMT"
}
] | 2016-10-18T00:00:00 | [
[
"Shohdy",
"Sameh",
""
],
[
"Vishnu",
"Abhinav",
""
],
[
"Agrawal",
"Gagan",
""
]
] | TITLE: Fault Tolerant Frequent Pattern Mining
ABSTRACT: FP-Growth algorithm is a Frequent Pattern Min- ing (FPM) algorithm that has
been extensively used to study correlations and patterns in large scale
datasets. While several researchers have designed distributed memory FP-Growth
algorithms, it is pivotal to consider fault tolerant FP-Growth, which can
address the increasing fault rates in large scale systems. In this work, we
propose a novel parallel, algorithm-level fault-tolerant FP-Growth algorithm.
We leverage algorithmic properties and MPI advanced features to guarantee an
O(1) space complexity, achieved by using the dataset memory space itself for
checkpointing. We also propose a recovery algorithm that can use in-memory and
disk-based checkpointing, though in many cases the recovery can be completed
without any disk access, and incurring no memory overhead for checkpointing. We
evaluate our FT algorithm on a large scale InfiniBand cluster with several
large datasets using up to 2K cores. Our evaluation demonstrates excellent
efficiency for checkpointing and recovery in comparison to the disk-based
approach. We have also observed 20x average speed-up in comparison to Spark,
establishing that a well designed algorithm can easily outperform a solution
based on a general fault-tolerant programming model.
| no_new_dataset | 0.944689 |
1610.05174 | Aznul Qalid Md Sabri | Aznul Qalid Md Sabri, Jacques Boonaert, Erma Rahayu Mohd Faizal
Abdullah and Ali Mohammed Mansoor | Spatio-temporal Co-Occurrence Characterizations for Human Action
Classification | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The human action classification task is a widely researched topic and is
still an open problem. Many state-of-the-arts approaches involve the usage of
bag-of-video-words with spatio-temporal local features to construct
characterizations for human actions. In order to improve beyond this standard
approach, we investigate the usage of co-occurrences between local features. We
propose the usage of co-occurrences information to characterize human actions.
A trade-off factor is used to define an optimal trade-off between vocabulary
size and classification rate. Next, a spatio-temporal co-occurrence technique
is applied to extract co-occurrence information between labeled local features.
Novel characterizations for human actions are then constructed. These include a
vector quantized correlogram-elements vector, a highly discriminative PCA
(Principal Components Analysis) co-occurrence vector and a Haralick texture
vector. Multi-channel kernel SVM (support vector machine) is utilized for
classification. For evaluation, the well known KTH as well as the challenging
UCF-Sports action datasets are used. We obtained state-of-the-arts
classification performance. We also demonstrated that we are able to fully
utilize co-occurrence information, and improve the standard bag-of-video-words
approach.
| [
{
"version": "v1",
"created": "Tue, 2 Aug 2016 02:22:42 GMT"
}
] | 2016-10-18T00:00:00 | [
[
"Sabri",
"Aznul Qalid Md",
""
],
[
"Boonaert",
"Jacques",
""
],
[
"Abdullah",
"Erma Rahayu Mohd Faizal",
""
],
[
"Mansoor",
"Ali Mohammed",
""
]
] | TITLE: Spatio-temporal Co-Occurrence Characterizations for Human Action
Classification
ABSTRACT: The human action classification task is a widely researched topic and is
still an open problem. Many state-of-the-arts approaches involve the usage of
bag-of-video-words with spatio-temporal local features to construct
characterizations for human actions. In order to improve beyond this standard
approach, we investigate the usage of co-occurrences between local features. We
propose the usage of co-occurrences information to characterize human actions.
A trade-off factor is used to define an optimal trade-off between vocabulary
size and classification rate. Next, a spatio-temporal co-occurrence technique
is applied to extract co-occurrence information between labeled local features.
Novel characterizations for human actions are then constructed. These include a
vector quantized correlogram-elements vector, a highly discriminative PCA
(Principal Components Analysis) co-occurrence vector and a Haralick texture
vector. Multi-channel kernel SVM (support vector machine) is utilized for
classification. For evaluation, the well known KTH as well as the challenging
UCF-Sports action datasets are used. We obtained state-of-the-arts
classification performance. We also demonstrated that we are able to fully
utilize co-occurrence information, and improve the standard bag-of-video-words
approach.
| no_new_dataset | 0.947235 |
1610.04416 | Dimitri Kartsaklis | Dimitri Kartsaklis, Mehrnoosh Sadrzadeh | Distributional Inclusion Hypothesis for Tensor-based Composition | To appear in COLING 2016 | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | According to the distributional inclusion hypothesis, entailment between
words can be measured via the feature inclusions of their distributional
vectors. In recent work, we showed how this hypothesis can be extended from
words to phrases and sentences in the setting of compositional distributional
semantics. This paper focuses on inclusion properties of tensors; its main
contribution is a theoretical and experimental analysis of how feature
inclusion works in different concrete models of verb tensors. We present
results for relational, Frobenius, projective, and holistic methods and compare
them to the simple vector addition, multiplication, min, and max models. The
degrees of entailment thus obtained are evaluated via a variety of existing
word-based measures, such as Weed's and Clarke's, KL-divergence, APinc,
balAPinc, and two of our previously proposed metrics at the phrase/sentence
level. We perform experiments on three entailment datasets, investigating which
version of tensor-based composition achieves the highest performance when
combined with the sentence-level measures.
| [
{
"version": "v1",
"created": "Fri, 14 Oct 2016 11:52:19 GMT"
}
] | 2016-10-17T00:00:00 | [
[
"Kartsaklis",
"Dimitri",
""
],
[
"Sadrzadeh",
"Mehrnoosh",
""
]
] | TITLE: Distributional Inclusion Hypothesis for Tensor-based Composition
ABSTRACT: According to the distributional inclusion hypothesis, entailment between
words can be measured via the feature inclusions of their distributional
vectors. In recent work, we showed how this hypothesis can be extended from
words to phrases and sentences in the setting of compositional distributional
semantics. This paper focuses on inclusion properties of tensors; its main
contribution is a theoretical and experimental analysis of how feature
inclusion works in different concrete models of verb tensors. We present
results for relational, Frobenius, projective, and holistic methods and compare
them to the simple vector addition, multiplication, min, and max models. The
degrees of entailment thus obtained are evaluated via a variety of existing
word-based measures, such as Weed's and Clarke's, KL-divergence, APinc,
balAPinc, and two of our previously proposed metrics at the phrase/sentence
level. We perform experiments on three entailment datasets, investigating which
version of tensor-based composition achieves the highest performance when
combined with the sentence-level measures.
| no_new_dataset | 0.947817 |
1610.04533 | Issa Atoum | Issa Atoum, Ahmed Otoom and Narayanan Kulathuramaiyer | A Comprehensive Comparative Study of Word and Sentence Similarity
Measures | 7 pages,4 figures | International Journal of Computer Applications,2016,135(1),
Foundation of Computer Science (FCS), NY, USA | 10.5120/ijca2016908259 | null | cs.IR cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Sentence similarity is considered the basis of many natural language tasks
such as information retrieval, question answering and text summarization. The
semantic meaning between compared text fragments is based on the words semantic
features and their relationships. This article reviews a set of word and
sentence similarity measures and compares them on benchmark datasets. On the
studied datasets, results showed that hybrid semantic measures perform better
than both knowledge and corpus based measures.
| [
{
"version": "v1",
"created": "Wed, 17 Feb 2016 19:33:47 GMT"
}
] | 2016-10-17T00:00:00 | [
[
"Atoum",
"Issa",
""
],
[
"Otoom",
"Ahmed",
""
],
[
"Kulathuramaiyer",
"Narayanan",
""
]
] | TITLE: A Comprehensive Comparative Study of Word and Sentence Similarity
Measures
ABSTRACT: Sentence similarity is considered the basis of many natural language tasks
such as information retrieval, question answering and text summarization. The
semantic meaning between compared text fragments is based on the words semantic
features and their relationships. This article reviews a set of word and
sentence similarity measures and compares them on benchmark datasets. On the
studied datasets, results showed that hybrid semantic measures perform better
than both knowledge and corpus based measures.
| no_new_dataset | 0.949763 |
1610.04577 | Kratika Tyagi | Kratika Tyagi, Prof. Sanjeev Thakur | A Survey on Various Data Mining Techniques for ECG Meta Analysis | null | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Data Mining is the process of examining the information from different point
of view and compressing it for the relevant data. This data can also be
utilized to build the incomes. Data Mining is also known as Data or Knowledge
Discovery. The basic purpose of data mining is to search patterns which have
minimal user inputs and efforts. Data Mining plays a very crucial role in the
various fields. There are various data mining procedures which can be connected
in different fields of innovation. By using data mining techniques, it is
observed that less time is taken for the prediction of any disease with more
accuracy. In this paper we would review various data mining techniques which
are categorized under classification, regression and clustering and apply these
algorithms over an ECG dataset. The purpose of this work is to determine the
most suitable data mining technique and use it to improve the accuracy of
analyzing ECG data for better decision making.
| [
{
"version": "v1",
"created": "Fri, 22 Apr 2016 11:41:23 GMT"
}
] | 2016-10-17T00:00:00 | [
[
"Tyagi",
"Kratika",
""
],
[
"Thakur",
"Prof. Sanjeev",
""
]
] | TITLE: A Survey on Various Data Mining Techniques for ECG Meta Analysis
ABSTRACT: Data Mining is the process of examining the information from different point
of view and compressing it for the relevant data. This data can also be
utilized to build the incomes. Data Mining is also known as Data or Knowledge
Discovery. The basic purpose of data mining is to search patterns which have
minimal user inputs and efforts. Data Mining plays a very crucial role in the
various fields. There are various data mining procedures which can be connected
in different fields of innovation. By using data mining techniques, it is
observed that less time is taken for the prediction of any disease with more
accuracy. In this paper we would review various data mining techniques which
are categorized under classification, regression and clustering and apply these
algorithms over an ECG dataset. The purpose of this work is to determine the
most suitable data mining technique and use it to improve the accuracy of
analyzing ECG data for better decision making.
| no_new_dataset | 0.951908 |
1506.04364 | Yunwen Lei | Yunwen Lei and Alexander Binder and \"Ur\"un Dogan and Marius Kloft | Localized Multiple Kernel Learning---A Convex Approach | to appear in ACML 2016 | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a localized approach to multiple kernel learning that can be
formulated as a convex optimization problem over a given cluster structure. For
which we obtain generalization error guarantees and derive an optimization
algorithm based on the Fenchel dual representation. Experiments on real-world
datasets from the application domains of computational biology and computer
vision show that convex localized multiple kernel learning can achieve higher
prediction accuracies than its global and non-convex local counterparts.
| [
{
"version": "v1",
"created": "Sun, 14 Jun 2015 09:11:13 GMT"
},
{
"version": "v2",
"created": "Thu, 13 Oct 2016 00:54:24 GMT"
}
] | 2016-10-14T00:00:00 | [
[
"Lei",
"Yunwen",
""
],
[
"Binder",
"Alexander",
""
],
[
"Dogan",
"Ürün",
""
],
[
"Kloft",
"Marius",
""
]
] | TITLE: Localized Multiple Kernel Learning---A Convex Approach
ABSTRACT: We propose a localized approach to multiple kernel learning that can be
formulated as a convex optimization problem over a given cluster structure. For
which we obtain generalization error guarantees and derive an optimization
algorithm based on the Fenchel dual representation. Experiments on real-world
datasets from the application domains of computational biology and computer
vision show that convex localized multiple kernel learning can achieve higher
prediction accuracies than its global and non-convex local counterparts.
| no_new_dataset | 0.9455 |
1609.07299 | Tatiana Alessandra Bubba | Tatiana A. Bubba, Andreas Hauptmann, Simo Huotari, Juho Rimpel\"ainen
and Samuli Siltanen | Tomographic X-ray data of a lotus root filled with attenuating objects | arXiv admin note: substantial text overlap with arXiv:1502.04064 | null | null | null | physics.data-an physics.med-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This is the documentation of the tomographic X-ray data of a lotus root,
filled with four different attenuating objects, of different sizes. Data are
available at www.fips.fi/dataset.php, and can be freely used for scientific
purposes with appropriate references to them, and to this document in
http://arxiv.org/arXiv. The data set consists of (1) the X-ray sinogram of a
single 2D slice of the lotus root with two different resolutions and (2) the
corresponding measurement matrices modeling the linear operation of the X-ray
transform. Each of these sinograms was obtained from a measured 360-projection
fan-beam sinogram by down-sampling and taking logarithms. The original
(measured) sinogram is also provided in its original form and resolution.
| [
{
"version": "v1",
"created": "Fri, 23 Sep 2016 10:15:31 GMT"
},
{
"version": "v2",
"created": "Thu, 13 Oct 2016 13:42:33 GMT"
}
] | 2016-10-14T00:00:00 | [
[
"Bubba",
"Tatiana A.",
""
],
[
"Hauptmann",
"Andreas",
""
],
[
"Huotari",
"Simo",
""
],
[
"Rimpeläinen",
"Juho",
""
],
[
"Siltanen",
"Samuli",
""
]
] | TITLE: Tomographic X-ray data of a lotus root filled with attenuating objects
ABSTRACT: This is the documentation of the tomographic X-ray data of a lotus root,
filled with four different attenuating objects, of different sizes. Data are
available at www.fips.fi/dataset.php, and can be freely used for scientific
purposes with appropriate references to them, and to this document in
http://arxiv.org/arXiv. The data set consists of (1) the X-ray sinogram of a
single 2D slice of the lotus root with two different resolutions and (2) the
corresponding measurement matrices modeling the linear operation of the X-ray
transform. Each of these sinograms was obtained from a measured 360-projection
fan-beam sinogram by down-sampling and taking logarithms. The original
(measured) sinogram is also provided in its original form and resolution.
| no_new_dataset | 0.919715 |
1609.08431 | Kaustubh Beedkar | Kaustubh Beedkar and Rainer Gemulla | DESQ: Frequent Sequence Mining with Subsequence Constraints | Long version of the paper accepted at the IEEE ICDM 2016 conference | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Frequent sequence mining methods often make use of constraints to control
which subsequences should be mined. A variety of such subsequence constraints
has been studied in the literature, including length, gap, span,
regular-expression, and hierarchy constraints. In this paper, we show that many
subsequence constraints---including and beyond those considered in the
literature---can be unified in a single framework. A unified treatment allows
researchers to study jointly many types of subsequence constraints (instead of
each one individually) and helps to improve usability of pattern mining systems
for practitioners. In more detail, we propose a set of simple and intuitive
"pattern expressions" to describe subsequence constraints and explore
algorithms for efficiently mining frequent subsequences under such general
constraints. Our algorithms translate pattern expressions to compressed finite
state transducers, which we use as computational model, and simulate these
transducers in a way suitable for frequent sequence mining. Our experimental
study on real-world datasets indicates that our algorithms---although more
general---are competitive to existing state-of-the-art algorithms.
| [
{
"version": "v1",
"created": "Tue, 27 Sep 2016 13:34:25 GMT"
},
{
"version": "v2",
"created": "Thu, 13 Oct 2016 17:20:05 GMT"
}
] | 2016-10-14T00:00:00 | [
[
"Beedkar",
"Kaustubh",
""
],
[
"Gemulla",
"Rainer",
""
]
] | TITLE: DESQ: Frequent Sequence Mining with Subsequence Constraints
ABSTRACT: Frequent sequence mining methods often make use of constraints to control
which subsequences should be mined. A variety of such subsequence constraints
has been studied in the literature, including length, gap, span,
regular-expression, and hierarchy constraints. In this paper, we show that many
subsequence constraints---including and beyond those considered in the
literature---can be unified in a single framework. A unified treatment allows
researchers to study jointly many types of subsequence constraints (instead of
each one individually) and helps to improve usability of pattern mining systems
for practitioners. In more detail, we propose a set of simple and intuitive
"pattern expressions" to describe subsequence constraints and explore
algorithms for efficiently mining frequent subsequences under such general
constraints. Our algorithms translate pattern expressions to compressed finite
state transducers, which we use as computational model, and simulate these
transducers in a way suitable for frequent sequence mining. Our experimental
study on real-world datasets indicates that our algorithms---although more
general---are competitive to existing state-of-the-art algorithms.
| no_new_dataset | 0.950869 |
1610.03098 | Aaditya Prakash | Aaditya Prakash, Sadid A. Hasan, Kathy Lee, Vivek Datla, Ashequl
Qadir, Joey Liu, Oladimeji Farri | Neural Paraphrase Generation with Stacked Residual LSTM Networks | COLING 2016 | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we propose a novel neural approach for paraphrase generation.
Conventional para- phrase generation methods either leverage hand-written rules
and thesauri-based alignments, or use statistical machine learning principles.
To the best of our knowledge, this work is the first to explore deep learning
models for paraphrase generation. Our primary contribution is a stacked
residual LSTM network, where we add residual connections between LSTM layers.
This allows for efficient training of deep LSTMs. We evaluate our model and
other state-of-the-art deep learning models on three different datasets: PPDB,
WikiAnswers and MSCOCO. Evaluation results demonstrate that our model
outperforms sequence to sequence, attention-based and bi- directional LSTM
models on BLEU, METEOR, TER and an embedding-based sentence similarity metric.
| [
{
"version": "v1",
"created": "Mon, 10 Oct 2016 21:01:00 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Oct 2016 15:02:02 GMT"
},
{
"version": "v3",
"created": "Thu, 13 Oct 2016 00:37:33 GMT"
}
] | 2016-10-14T00:00:00 | [
[
"Prakash",
"Aaditya",
""
],
[
"Hasan",
"Sadid A.",
""
],
[
"Lee",
"Kathy",
""
],
[
"Datla",
"Vivek",
""
],
[
"Qadir",
"Ashequl",
""
],
[
"Liu",
"Joey",
""
],
[
"Farri",
"Oladimeji",
""
]
] | TITLE: Neural Paraphrase Generation with Stacked Residual LSTM Networks
ABSTRACT: In this paper, we propose a novel neural approach for paraphrase generation.
Conventional para- phrase generation methods either leverage hand-written rules
and thesauri-based alignments, or use statistical machine learning principles.
To the best of our knowledge, this work is the first to explore deep learning
models for paraphrase generation. Our primary contribution is a stacked
residual LSTM network, where we add residual connections between LSTM layers.
This allows for efficient training of deep LSTMs. We evaluate our model and
other state-of-the-art deep learning models on three different datasets: PPDB,
WikiAnswers and MSCOCO. Evaluation results demonstrate that our model
outperforms sequence to sequence, attention-based and bi- directional LSTM
models on BLEU, METEOR, TER and an embedding-based sentence similarity metric.
| no_new_dataset | 0.949669 |
1610.04062 | Dong Zhang | Amir Mazaheri, Dong Zhang, Mubarak Shah | Video Fill in the Blank with Merging LSTMs | for Large Scale Movie Description and Understanding Challenge (LSMDC)
2016, "Movie fill-in-the-blank" Challenge, UCF_CRCV | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Given a video and its incomplete textural description with missing words, the
Video-Fill-in-the-Blank (ViFitB) task is to automatically find the missing
word. The contextual information of the sentences are important to infer the
missing words; the visual cues are even more crucial to get a more accurate
inference. In this paper, we presents a new method which intuitively takes
advantage of the structure of the sentences and employs merging LSTMs (to merge
two LSTMs) to tackle the problem with embedded textural and visual cues. In the
experiments, we have demonstrated the superior performance of the proposed
method on the challenging "Movie Fill-in-the-Blank" dataset.
| [
{
"version": "v1",
"created": "Thu, 13 Oct 2016 13:05:41 GMT"
}
] | 2016-10-14T00:00:00 | [
[
"Mazaheri",
"Amir",
""
],
[
"Zhang",
"Dong",
""
],
[
"Shah",
"Mubarak",
""
]
] | TITLE: Video Fill in the Blank with Merging LSTMs
ABSTRACT: Given a video and its incomplete textural description with missing words, the
Video-Fill-in-the-Blank (ViFitB) task is to automatically find the missing
word. The contextual information of the sentences are important to infer the
missing words; the visual cues are even more crucial to get a more accurate
inference. In this paper, we presents a new method which intuitively takes
advantage of the structure of the sentences and employs merging LSTMs (to merge
two LSTMs) to tackle the problem with embedded textural and visual cues. In the
experiments, we have demonstrated the superior performance of the proposed
method on the challenging "Movie Fill-in-the-Blank" dataset.
| no_new_dataset | 0.941115 |
1408.5405 | Khalid Raza | Khalid Raza and Mansaf Alam | Recurrent Neural Network Based Hybrid Model of Gene Regulatory Network | 18 pages, 9 figures and 4 tables | Computational Biology and Chemistry, 64: 322-334, 2016 | 10.1016/j.compbiolchem.2016.08.002 | null | cs.NE cs.CE q-bio.MN | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Systems biology is an emerging interdisciplinary area of research that
focuses on study of complex interactions in a biological system, such as gene
regulatory networks. The discovery of gene regulatory networks leads to a wide
range of applications, such as pathways related to a disease that can unveil in
what way the disease acts and provide novel tentative drug targets. In
addition, the development of biological models from discovered networks or
pathways can help to predict the responses to disease and can be much useful
for the novel drug development and treatments. The inference of regulatory
networks from biological data is still in its infancy stage. This paper
proposes a recurrent neural network (RNN) based gene regulatory network (GRN)
model hybridized with generalized extended Kalman filter for weight update in
backpropagation through time training algorithm. The RNN is a complex neural
network that gives a better settlement between the biological closeness and
mathematical flexibility to model GRN. The RNN is able to capture complex,
non-linear and dynamic relationship among variables. Gene expression data are
inherently noisy and Kalman filter performs well for estimation even in noisy
data. Hence, non-linear version of Kalman filter, i.e., generalized extended
Kalman filter has been applied for weight update during network training. The
developed model has been applied on DNA SOS repair network, IRMA network, and
two synthetic networks from DREAM Challenge. We compared our results with other
state-of-the-art techniques that show superiority of our model. Further, 5%
Gaussian noise has been added in the dataset and result of the proposed model
shows negligible effect of noise on the results.
| [
{
"version": "v1",
"created": "Fri, 22 Aug 2014 18:35:27 GMT"
},
{
"version": "v2",
"created": "Fri, 13 Nov 2015 12:38:53 GMT"
}
] | 2016-10-13T00:00:00 | [
[
"Raza",
"Khalid",
""
],
[
"Alam",
"Mansaf",
""
]
] | TITLE: Recurrent Neural Network Based Hybrid Model of Gene Regulatory Network
ABSTRACT: Systems biology is an emerging interdisciplinary area of research that
focuses on study of complex interactions in a biological system, such as gene
regulatory networks. The discovery of gene regulatory networks leads to a wide
range of applications, such as pathways related to a disease that can unveil in
what way the disease acts and provide novel tentative drug targets. In
addition, the development of biological models from discovered networks or
pathways can help to predict the responses to disease and can be much useful
for the novel drug development and treatments. The inference of regulatory
networks from biological data is still in its infancy stage. This paper
proposes a recurrent neural network (RNN) based gene regulatory network (GRN)
model hybridized with generalized extended Kalman filter for weight update in
backpropagation through time training algorithm. The RNN is a complex neural
network that gives a better settlement between the biological closeness and
mathematical flexibility to model GRN. The RNN is able to capture complex,
non-linear and dynamic relationship among variables. Gene expression data are
inherently noisy and Kalman filter performs well for estimation even in noisy
data. Hence, non-linear version of Kalman filter, i.e., generalized extended
Kalman filter has been applied for weight update during network training. The
developed model has been applied on DNA SOS repair network, IRMA network, and
two synthetic networks from DREAM Challenge. We compared our results with other
state-of-the-art techniques that show superiority of our model. Further, 5%
Gaussian noise has been added in the dataset and result of the proposed model
shows negligible effect of noise on the results.
| no_new_dataset | 0.948346 |
1501.01361 | Changchang Liu | Changhchang Liu and Prateek Mittal | LinkMirage: How to Anonymize Links in Dynamic Social Systems | 19 pages, 20 figures | null | null | null | cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Social network based trust relationships present a critical foundation for
designing trustworthy systems, such as Sybil defenses, secure routing, and
anonymous/censorshipresilient communications. A key issue in the design of such
systems, is the revelation of users' trusted social contacts to an
adversary-information that is considered sensitive in today's society.
In this work, we focus on the challenge of preserving the privacy of users'
social contacts, while still enabling the design of social trust based
applications. First, we propose LinkMirage, a community detection based
algorithm for anonymizing links in social network topologies; LinkMirage
preserves community structures in the social topology while anonymizing links
within the communities. LinkMirage considers the evolution of the social
network topologies, and minimizes privacy leakage due to temporal dynamics of
the system.
Second, we define metrics for quantifying the privacy and utility of a time
series of social topologies with anonymized links. We analyze the privacy and
utility provided by LinkMirage both theoretically, as well as using real world
social network topologies: a Facebook dataset with 870K links and a large-scale
Google+ dataset with 940M links. We find that our approach significantly
outperforms the existing state-of-art.
Finally, we demonstrate the applicability of LinkMirage in real-world
applications such as Sybil defenses, reputation systems, anonymity systems and
vertex anonymity. We also prototype LinkMirage as a Facebook application such
that real world systems can bootstrap privacy-preserving trust relationships
without the cooperation of the OSN operators.
| [
{
"version": "v1",
"created": "Wed, 7 Jan 2015 03:34:52 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Oct 2016 01:36:35 GMT"
}
] | 2016-10-13T00:00:00 | [
[
"Liu",
"Changhchang",
""
],
[
"Mittal",
"Prateek",
""
]
] | TITLE: LinkMirage: How to Anonymize Links in Dynamic Social Systems
ABSTRACT: Social network based trust relationships present a critical foundation for
designing trustworthy systems, such as Sybil defenses, secure routing, and
anonymous/censorshipresilient communications. A key issue in the design of such
systems, is the revelation of users' trusted social contacts to an
adversary-information that is considered sensitive in today's society.
In this work, we focus on the challenge of preserving the privacy of users'
social contacts, while still enabling the design of social trust based
applications. First, we propose LinkMirage, a community detection based
algorithm for anonymizing links in social network topologies; LinkMirage
preserves community structures in the social topology while anonymizing links
within the communities. LinkMirage considers the evolution of the social
network topologies, and minimizes privacy leakage due to temporal dynamics of
the system.
Second, we define metrics for quantifying the privacy and utility of a time
series of social topologies with anonymized links. We analyze the privacy and
utility provided by LinkMirage both theoretically, as well as using real world
social network topologies: a Facebook dataset with 870K links and a large-scale
Google+ dataset with 940M links. We find that our approach significantly
outperforms the existing state-of-art.
Finally, we demonstrate the applicability of LinkMirage in real-world
applications such as Sybil defenses, reputation systems, anonymity systems and
vertex anonymity. We also prototype LinkMirage as a Facebook application such
that real world systems can bootstrap privacy-preserving trust relationships
without the cooperation of the OSN operators.
| no_new_dataset | 0.938294 |
1510.08012 | Guofeng Zhang | Guofeng Zhang, Haomin Liu, Zilong Dong, Jiaya Jia, Tien-Tsin Wong and
Hujun Bao | ENFT: Efficient Non-Consecutive Feature Tracking for Robust
Structure-from-Motion | 15 pages, 12 figures | null | 10.1109/TIP.2016.2607425 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Structure-from-motion (SfM) largely relies on feature tracking. In image
sequences, if disjointed tracks caused by objects moving in and out of the
field of view, occasional occlusion, or image noise, are not handled well,
corresponding SfM could be affected. This problem becomes severer for
large-scale scenes, which typically requires to capture multiple sequences to
cover the whole scene. In this paper, we propose an efficient non-consecutive
feature tracking (ENFT) framework to match interrupted tracks distributed in
different subsequences or even in different videos. Our framework consists of
steps of solving the feature `dropout' problem when indistinctive structures,
noise or large image distortion exists, and of rapidly recognizing and joining
common features located in different subsequences. In addition, we contribute
an effective segment-based coarse-to-fine SfM algorithm for robustly handling
large datasets. Experimental results on challenging video data demonstrate the
effectiveness of the proposed system.
| [
{
"version": "v1",
"created": "Tue, 27 Oct 2015 18:00:42 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Oct 2016 11:29:33 GMT"
}
] | 2016-10-13T00:00:00 | [
[
"Zhang",
"Guofeng",
""
],
[
"Liu",
"Haomin",
""
],
[
"Dong",
"Zilong",
""
],
[
"Jia",
"Jiaya",
""
],
[
"Wong",
"Tien-Tsin",
""
],
[
"Bao",
"Hujun",
""
]
] | TITLE: ENFT: Efficient Non-Consecutive Feature Tracking for Robust
Structure-from-Motion
ABSTRACT: Structure-from-motion (SfM) largely relies on feature tracking. In image
sequences, if disjointed tracks caused by objects moving in and out of the
field of view, occasional occlusion, or image noise, are not handled well,
corresponding SfM could be affected. This problem becomes severer for
large-scale scenes, which typically requires to capture multiple sequences to
cover the whole scene. In this paper, we propose an efficient non-consecutive
feature tracking (ENFT) framework to match interrupted tracks distributed in
different subsequences or even in different videos. Our framework consists of
steps of solving the feature `dropout' problem when indistinctive structures,
noise or large image distortion exists, and of rapidly recognizing and joining
common features located in different subsequences. In addition, we contribute
an effective segment-based coarse-to-fine SfM algorithm for robustly handling
large datasets. Experimental results on challenging video data demonstrate the
effectiveness of the proposed system.
| no_new_dataset | 0.948965 |
1603.08270 | Steven Esser | Steven K. Esser, Paul A. Merolla, John V. Arthur, Andrew S. Cassidy,
Rathinakumar Appuswamy, Alexander Andreopoulos, David J. Berg, Jeffrey L.
McKinstry, Timothy Melano, Davis R. Barch, Carmelo di Nolfo, Pallab Datta,
Arnon Amir, Brian Taba, Myron D. Flickner, and Dharmendra S. Modha | Convolutional Networks for Fast, Energy-Efficient Neuromorphic Computing | 7 pages, 6 figures | PNAS 113 (2016) 11441-11446 | 10.1073/pnas.1604850113 | null | cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep networks are now able to achieve human-level performance on a broad
spectrum of recognition tasks. Independently, neuromorphic computing has now
demonstrated unprecedented energy-efficiency through a new chip architecture
based on spiking neurons, low precision synapses, and a scalable communication
network. Here, we demonstrate that neuromorphic computing, despite its novel
architectural primitives, can implement deep convolution networks that i)
approach state-of-the-art classification accuracy across 8 standard datasets,
encompassing vision and speech, ii) perform inference while preserving the
hardware's underlying energy-efficiency and high throughput, running on the
aforementioned datasets at between 1200 and 2600 frames per second and using
between 25 and 275 mW (effectively > 6000 frames / sec / W) and iii) can be
specified and trained using backpropagation with the same ease-of-use as
contemporary deep learning. For the first time, the algorithmic power of deep
learning can be merged with the efficiency of neuromorphic processors, bringing
the promise of embedded, intelligent, brain-inspired computing one step closer.
| [
{
"version": "v1",
"created": "Mon, 28 Mar 2016 00:15:35 GMT"
},
{
"version": "v2",
"created": "Tue, 24 May 2016 18:46:56 GMT"
}
] | 2016-10-13T00:00:00 | [
[
"Esser",
"Steven K.",
""
],
[
"Merolla",
"Paul A.",
""
],
[
"Arthur",
"John V.",
""
],
[
"Cassidy",
"Andrew S.",
""
],
[
"Appuswamy",
"Rathinakumar",
""
],
[
"Andreopoulos",
"Alexander",
""
],
[
"Berg",
"David J.",
""
],
[
"McKinstry",
"Jeffrey L.",
""
],
[
"Melano",
"Timothy",
""
],
[
"Barch",
"Davis R.",
""
],
[
"di Nolfo",
"Carmelo",
""
],
[
"Datta",
"Pallab",
""
],
[
"Amir",
"Arnon",
""
],
[
"Taba",
"Brian",
""
],
[
"Flickner",
"Myron D.",
""
],
[
"Modha",
"Dharmendra S.",
""
]
] | TITLE: Convolutional Networks for Fast, Energy-Efficient Neuromorphic Computing
ABSTRACT: Deep networks are now able to achieve human-level performance on a broad
spectrum of recognition tasks. Independently, neuromorphic computing has now
demonstrated unprecedented energy-efficiency through a new chip architecture
based on spiking neurons, low precision synapses, and a scalable communication
network. Here, we demonstrate that neuromorphic computing, despite its novel
architectural primitives, can implement deep convolution networks that i)
approach state-of-the-art classification accuracy across 8 standard datasets,
encompassing vision and speech, ii) perform inference while preserving the
hardware's underlying energy-efficiency and high throughput, running on the
aforementioned datasets at between 1200 and 2600 frames per second and using
between 25 and 275 mW (effectively > 6000 frames / sec / W) and iii) can be
specified and trained using backpropagation with the same ease-of-use as
contemporary deep learning. For the first time, the algorithmic power of deep
learning can be merged with the efficiency of neuromorphic processors, bringing
the promise of embedded, intelligent, brain-inspired computing one step closer.
| no_new_dataset | 0.944638 |
1610.03628 | Carlos Ciller Mr. | Stefanos Apostolopoulos, Carlos Ciller, Sandro I. De Zanet, Sebastian
Wolf and Raphael Sznitman | RetiNet: Automatic AMD identification in OCT volumetric data | 14 pages, 10 figures, Code available | null | null | null | cs.CV cs.LG cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Optical Coherence Tomography (OCT) provides a unique ability to image the eye
retina in 3D at micrometer resolution and gives ophthalmologist the ability to
visualize retinal diseases such as Age-Related Macular Degeneration (AMD).
While visual inspection of OCT volumes remains the main method for AMD
identification, doing so is time consuming as each cross-section within the
volume must be inspected individually by the clinician. In much the same way,
acquiring ground truth information for each cross-section is expensive and time
consuming. This fact heavily limits the ability to acquire large amounts of
ground truth, which subsequently impacts the performance of learning-based
methods geared at automatic pathology identification. To avoid this burden, we
propose a novel strategy for automatic analysis of OCT volumes where only
volume labels are needed. That is, we train a classifier in a semi-supervised
manner to conduct this task. Our approach uses a novel Convolutional Neural
Network (CNN) architecture, that only needs volume-level labels to be trained
to automatically asses whether an OCT volume is healthy or contains AMD. Our
architecture involves first learning a cross-section pathology classifier using
pseudo-labels that could be corrupted and then leverage these towards a more
accurate volume-level classification. We then show that our approach provides
excellent performances on a publicly available dataset and outperforms a number
of existing automatic techniques.
| [
{
"version": "v1",
"created": "Wed, 12 Oct 2016 07:56:24 GMT"
}
] | 2016-10-13T00:00:00 | [
[
"Apostolopoulos",
"Stefanos",
""
],
[
"Ciller",
"Carlos",
""
],
[
"De Zanet",
"Sandro I.",
""
],
[
"Wolf",
"Sebastian",
""
],
[
"Sznitman",
"Raphael",
""
]
] | TITLE: RetiNet: Automatic AMD identification in OCT volumetric data
ABSTRACT: Optical Coherence Tomography (OCT) provides a unique ability to image the eye
retina in 3D at micrometer resolution and gives ophthalmologist the ability to
visualize retinal diseases such as Age-Related Macular Degeneration (AMD).
While visual inspection of OCT volumes remains the main method for AMD
identification, doing so is time consuming as each cross-section within the
volume must be inspected individually by the clinician. In much the same way,
acquiring ground truth information for each cross-section is expensive and time
consuming. This fact heavily limits the ability to acquire large amounts of
ground truth, which subsequently impacts the performance of learning-based
methods geared at automatic pathology identification. To avoid this burden, we
propose a novel strategy for automatic analysis of OCT volumes where only
volume labels are needed. That is, we train a classifier in a semi-supervised
manner to conduct this task. Our approach uses a novel Convolutional Neural
Network (CNN) architecture, that only needs volume-level labels to be trained
to automatically asses whether an OCT volume is healthy or contains AMD. Our
architecture involves first learning a cross-section pathology classifier using
pseudo-labels that could be corrupted and then leverage these towards a more
accurate volume-level classification. We then show that our approach provides
excellent performances on a publicly available dataset and outperforms a number
of existing automatic techniques.
| no_new_dataset | 0.951818 |
1610.03713 | Jesse Krijthe | Jesse H. Krijthe and Marco Loog | Optimistic Semi-supervised Least Squares Classification | 6 pages, 6 figures. International Conference on Pattern Recognition
(ICPR) 2016, Cancun, Mexico | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The goal of semi-supervised learning is to improve supervised classifiers by
using additional unlabeled training examples. In this work we study a simple
self-learning approach to semi-supervised learning applied to the least squares
classifier. We show that a soft-label and a hard-label variant of self-learning
can be derived by applying block coordinate descent to two related but slightly
different objective functions. The resulting soft-label approach is related to
an idea about dealing with missing data that dates back to the 1930s. We show
that the soft-label variant typically outperforms the hard-label variant on
benchmark datasets and partially explain this behaviour by studying the
relative difficulty of finding good local minima for the corresponding
objective functions.
| [
{
"version": "v1",
"created": "Wed, 12 Oct 2016 13:52:07 GMT"
}
] | 2016-10-13T00:00:00 | [
[
"Krijthe",
"Jesse H.",
""
],
[
"Loog",
"Marco",
""
]
] | TITLE: Optimistic Semi-supervised Least Squares Classification
ABSTRACT: The goal of semi-supervised learning is to improve supervised classifiers by
using additional unlabeled training examples. In this work we study a simple
self-learning approach to semi-supervised learning applied to the least squares
classifier. We show that a soft-label and a hard-label variant of self-learning
can be derived by applying block coordinate descent to two related but slightly
different objective functions. The resulting soft-label approach is related to
an idea about dealing with missing data that dates back to the 1930s. We show
that the soft-label variant typically outperforms the hard-label variant on
benchmark datasets and partially explain this behaviour by studying the
relative difficulty of finding good local minima for the corresponding
objective functions.
| no_new_dataset | 0.946349 |
1610.03771 | Marzieh Saeidi Marzieh Saeidi | Marzieh Saeidi, Guillaume Bouchard, Maria Liakata, Sebastian Riedel | SentiHood: Targeted Aspect Based Sentiment Analysis Dataset for Urban
Neighbourhoods | Accepted at COLING 2016 | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we introduce the task of targeted aspect-based sentiment
analysis. The goal is to extract fine-grained information with respect to
entities mentioned in user comments. This work extends both aspect-based
sentiment analysis that assumes a single entity per document and targeted
sentiment analysis that assumes a single sentiment towards a target entity. In
particular, we identify the sentiment towards each aspect of one or more
entities. As a testbed for this task, we introduce the SentiHood dataset,
extracted from a question answering (QA) platform where urban neighbourhoods
are discussed by users. In this context units of text often mention several
aspects of one or more neighbourhoods. This is the first time that a generic
social media platform in this case a QA platform, is used for fine-grained
opinion mining. Text coming from QA platforms is far less constrained compared
to text from review specific platforms which current datasets are based on. We
develop several strong baselines, relying on logistic regression and
state-of-the-art recurrent neural networks.
| [
{
"version": "v1",
"created": "Wed, 12 Oct 2016 16:23:11 GMT"
}
] | 2016-10-13T00:00:00 | [
[
"Saeidi",
"Marzieh",
""
],
[
"Bouchard",
"Guillaume",
""
],
[
"Liakata",
"Maria",
""
],
[
"Riedel",
"Sebastian",
""
]
] | TITLE: SentiHood: Targeted Aspect Based Sentiment Analysis Dataset for Urban
Neighbourhoods
ABSTRACT: In this paper, we introduce the task of targeted aspect-based sentiment
analysis. The goal is to extract fine-grained information with respect to
entities mentioned in user comments. This work extends both aspect-based
sentiment analysis that assumes a single entity per document and targeted
sentiment analysis that assumes a single sentiment towards a target entity. In
particular, we identify the sentiment towards each aspect of one or more
entities. As a testbed for this task, we introduce the SentiHood dataset,
extracted from a question answering (QA) platform where urban neighbourhoods
are discussed by users. In this context units of text often mention several
aspects of one or more neighbourhoods. This is the first time that a generic
social media platform in this case a QA platform, is used for fine-grained
opinion mining. Text coming from QA platforms is far less constrained compared
to text from review specific platforms which current datasets are based on. We
develop several strong baselines, relying on logistic regression and
state-of-the-art recurrent neural networks.
| new_dataset | 0.965964 |
1610.03772 | Peter Dugan Dr | Peter J. Dugan, Holger Klinck, Marie A. Roch and Tyler A. Helble | RAVEN X High Performance Data Mining Toolbox for Bioacoustic Data
Analysis | null | null | null | N00014-16-1-3156 | cs.SD | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Objective of this work is to integrate high performance computing (HPC)
technologies and bioacoustics data-mining capabilities by offering a
MATLAB-based toolbox called Raven-X. Raven-X will provide a
hardware-independent solution, for processing large acoustic datasets - the
toolkit will be available to the community at no cost. This goal will be
achieved by leveraging prior work done which successfully deployed MATLAB based
HPC tools within Cornell University's Bioacoustics Research Program (BRP).
These tools enabled commonly available multi-core computers to process data at
accelerated rates to detect and classify whale sounds in large multi-channel
sound archives. Through this collaboration, we will expand on this effort which
was featured through Mathworks research and industry forums incorporate new
cutting-edge detectors and classifiers, and disseminate Raven-X to the broader
bioacoustics community.
| [
{
"version": "v1",
"created": "Wed, 12 Oct 2016 16:24:54 GMT"
}
] | 2016-10-13T00:00:00 | [
[
"Dugan",
"Peter J.",
""
],
[
"Klinck",
"Holger",
""
],
[
"Roch",
"Marie A.",
""
],
[
"Helble",
"Tyler A.",
""
]
] | TITLE: RAVEN X High Performance Data Mining Toolbox for Bioacoustic Data
Analysis
ABSTRACT: Objective of this work is to integrate high performance computing (HPC)
technologies and bioacoustics data-mining capabilities by offering a
MATLAB-based toolbox called Raven-X. Raven-X will provide a
hardware-independent solution, for processing large acoustic datasets - the
toolkit will be available to the community at no cost. This goal will be
achieved by leveraging prior work done which successfully deployed MATLAB based
HPC tools within Cornell University's Bioacoustics Research Program (BRP).
These tools enabled commonly available multi-core computers to process data at
accelerated rates to detect and classify whale sounds in large multi-channel
sound archives. Through this collaboration, we will expand on this effort which
was featured through Mathworks research and industry forums incorporate new
cutting-edge detectors and classifiers, and disseminate Raven-X to the broader
bioacoustics community.
| no_new_dataset | 0.9455 |
1511.02680 | Alex Kendall | Alex Kendall and Vijay Badrinarayanan and Roberto Cipolla | Bayesian SegNet: Model Uncertainty in Deep Convolutional Encoder-Decoder
Architectures for Scene Understanding | null | null | null | null | cs.CV cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a deep learning framework for probabilistic pixel-wise semantic
segmentation, which we term Bayesian SegNet. Semantic segmentation is an
important tool for visual scene understanding and a meaningful measure of
uncertainty is essential for decision making. Our contribution is a practical
system which is able to predict pixel-wise class labels with a measure of model
uncertainty. We achieve this by Monte Carlo sampling with dropout at test time
to generate a posterior distribution of pixel class labels. In addition, we
show that modelling uncertainty improves segmentation performance by 2-3%
across a number of state of the art architectures such as SegNet, FCN and
Dilation Network, with no additional parametrisation. We also observe a
significant improvement in performance for smaller datasets where modelling
uncertainty is more effective. We benchmark Bayesian SegNet on the indoor SUN
Scene Understanding and outdoor CamVid driving scenes datasets.
| [
{
"version": "v1",
"created": "Mon, 9 Nov 2015 14:00:21 GMT"
},
{
"version": "v2",
"created": "Mon, 10 Oct 2016 22:04:21 GMT"
}
] | 2016-10-12T00:00:00 | [
[
"Kendall",
"Alex",
""
],
[
"Badrinarayanan",
"Vijay",
""
],
[
"Cipolla",
"Roberto",
""
]
] | TITLE: Bayesian SegNet: Model Uncertainty in Deep Convolutional Encoder-Decoder
Architectures for Scene Understanding
ABSTRACT: We present a deep learning framework for probabilistic pixel-wise semantic
segmentation, which we term Bayesian SegNet. Semantic segmentation is an
important tool for visual scene understanding and a meaningful measure of
uncertainty is essential for decision making. Our contribution is a practical
system which is able to predict pixel-wise class labels with a measure of model
uncertainty. We achieve this by Monte Carlo sampling with dropout at test time
to generate a posterior distribution of pixel class labels. In addition, we
show that modelling uncertainty improves segmentation performance by 2-3%
across a number of state of the art architectures such as SegNet, FCN and
Dilation Network, with no additional parametrisation. We also observe a
significant improvement in performance for smaller datasets where modelling
uncertainty is more effective. We benchmark Bayesian SegNet on the indoor SUN
Scene Understanding and outdoor CamVid driving scenes datasets.
| no_new_dataset | 0.949012 |
1604.04333 | Miao Sun | Miao Sun, Tony X. Han, Xun Xu, Ming-Chang Liu, Ahmad
Khodayari-Rostamabad | Latent Model Ensemble with Auto-localization | International Conference on Pattern Recognition (ICPR) 2016 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep Convolutional Neural Networks (CNN) have exhibited superior performance
in many visual recognition tasks including image classification, object
detection, and scene label- ing, due to their large learning capacity and
resistance to overfit. For the image classification task, most of the current
deep CNN- based approaches take the whole size-normalized image as input and
have achieved quite promising results. Compared with the previously dominating
approaches based on feature extraction, pooling, and classification, the deep
CNN-based approaches mainly rely on the learning capability of deep CNN to
achieve superior results: the burden of minimizing intra-class variation while
maximizing inter-class difference is entirely dependent on the implicit feature
learning component of deep CNN; we rely upon the implicitly learned filters and
pooling component to select the discriminative regions, which correspond to the
activated neurons. However, if the irrelevant regions constitute a large
portion of the image of interest, the classification performance of the deep
CNN, which takes the whole image as input, can be heavily affected. To solve
this issue, we propose a novel latent CNN framework, which treats the most
discriminate region as a latent variable. We can jointly learn the global CNN
with the latent CNN to avoid the aforementioned big irrelevant region issue,
and our experimental results show the evident advantage of the proposed latent
CNN over traditional deep CNN: latent CNN outperforms the state-of-the-art
performance of deep CNN on standard benchmark datasets including the CIFAR-10,
CIFAR- 100, MNIST and PASCAL VOC 2007 Classification dataset.
| [
{
"version": "v1",
"created": "Fri, 15 Apr 2016 02:07:42 GMT"
},
{
"version": "v2",
"created": "Tue, 11 Oct 2016 01:57:19 GMT"
}
] | 2016-10-12T00:00:00 | [
[
"Sun",
"Miao",
""
],
[
"Han",
"Tony X.",
""
],
[
"Xu",
"Xun",
""
],
[
"Liu",
"Ming-Chang",
""
],
[
"Khodayari-Rostamabad",
"Ahmad",
""
]
] | TITLE: Latent Model Ensemble with Auto-localization
ABSTRACT: Deep Convolutional Neural Networks (CNN) have exhibited superior performance
in many visual recognition tasks including image classification, object
detection, and scene label- ing, due to their large learning capacity and
resistance to overfit. For the image classification task, most of the current
deep CNN- based approaches take the whole size-normalized image as input and
have achieved quite promising results. Compared with the previously dominating
approaches based on feature extraction, pooling, and classification, the deep
CNN-based approaches mainly rely on the learning capability of deep CNN to
achieve superior results: the burden of minimizing intra-class variation while
maximizing inter-class difference is entirely dependent on the implicit feature
learning component of deep CNN; we rely upon the implicitly learned filters and
pooling component to select the discriminative regions, which correspond to the
activated neurons. However, if the irrelevant regions constitute a large
portion of the image of interest, the classification performance of the deep
CNN, which takes the whole image as input, can be heavily affected. To solve
this issue, we propose a novel latent CNN framework, which treats the most
discriminate region as a latent variable. We can jointly learn the global CNN
with the latent CNN to avoid the aforementioned big irrelevant region issue,
and our experimental results show the evident advantage of the proposed latent
CNN over traditional deep CNN: latent CNN outperforms the state-of-the-art
performance of deep CNN on standard benchmark datasets including the CIFAR-10,
CIFAR- 100, MNIST and PASCAL VOC 2007 Classification dataset.
| no_new_dataset | 0.947769 |
1605.07586 | Mohsen Kheirandishfard | Fariba Zohrizadeh, Mohsen Kheirandishfard, and Farhad Kamangar | Natural Scene Image Segmentation Based on Multi-Layer Feature Extraction | This paper has been withdrawn by the author due to the fact that the
contents need further research | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper addresses the problem of natural image segmentation by extracting
information from a multi-layer array which is constructed based on color,
gradient, and statistical properties of the local neighborhoods in an image. A
Gaussian Mixture Model (GMM) is used to improve the effectiveness of local
spectral histogram features. Grouping these features leads to forming a rough
initial over-segmented layer which contains coherent regions of pixels. The
regions are merged by using two proposed functions for calculating the distance
between two neighboring regions and making decisions about their merging.
Extensive experiments are performed on the Berkeley Segmentation Dataset to
evaluate the performance of our proposed method and compare the results with
the recent state-of-the-art methods. The experimental results indicate that our
method achieves higher level of accuracy for natural images compared to recent
methods.
| [
{
"version": "v1",
"created": "Tue, 24 May 2016 19:03:54 GMT"
},
{
"version": "v2",
"created": "Tue, 11 Oct 2016 03:58:34 GMT"
}
] | 2016-10-12T00:00:00 | [
[
"Zohrizadeh",
"Fariba",
""
],
[
"Kheirandishfard",
"Mohsen",
""
],
[
"Kamangar",
"Farhad",
""
]
] | TITLE: Natural Scene Image Segmentation Based on Multi-Layer Feature Extraction
ABSTRACT: This paper addresses the problem of natural image segmentation by extracting
information from a multi-layer array which is constructed based on color,
gradient, and statistical properties of the local neighborhoods in an image. A
Gaussian Mixture Model (GMM) is used to improve the effectiveness of local
spectral histogram features. Grouping these features leads to forming a rough
initial over-segmented layer which contains coherent regions of pixels. The
regions are merged by using two proposed functions for calculating the distance
between two neighboring regions and making decisions about their merging.
Extensive experiments are performed on the Berkeley Segmentation Dataset to
evaluate the performance of our proposed method and compare the results with
the recent state-of-the-art methods. The experimental results indicate that our
method achieves higher level of accuracy for natural images compared to recent
methods.
| no_new_dataset | 0.952926 |
1606.03126 | Jason Weston | Alexander Miller, Adam Fisch, Jesse Dodge, Amir-Hossein Karimi,
Antoine Bordes, Jason Weston | Key-Value Memory Networks for Directly Reading Documents | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Directly reading documents and being able to answer questions from them is an
unsolved challenge. To avoid its inherent difficulty, question answering (QA)
has been directed towards using Knowledge Bases (KBs) instead, which has proven
effective. Unfortunately KBs often suffer from being too restrictive, as the
schema cannot support certain types of answers, and too sparse, e.g. Wikipedia
contains much more information than Freebase. In this work we introduce a new
method, Key-Value Memory Networks, that makes reading documents more viable by
utilizing different encodings in the addressing and output stages of the memory
read operation. To compare using KBs, information extraction or Wikipedia
documents directly in a single framework we construct an analysis tool,
WikiMovies, a QA dataset that contains raw text alongside a preprocessed KB, in
the domain of movies. Our method reduces the gap between all three settings. It
also achieves state-of-the-art results on the existing WikiQA benchmark.
| [
{
"version": "v1",
"created": "Thu, 9 Jun 2016 21:33:55 GMT"
},
{
"version": "v2",
"created": "Mon, 10 Oct 2016 20:14:10 GMT"
}
] | 2016-10-12T00:00:00 | [
[
"Miller",
"Alexander",
""
],
[
"Fisch",
"Adam",
""
],
[
"Dodge",
"Jesse",
""
],
[
"Karimi",
"Amir-Hossein",
""
],
[
"Bordes",
"Antoine",
""
],
[
"Weston",
"Jason",
""
]
] | TITLE: Key-Value Memory Networks for Directly Reading Documents
ABSTRACT: Directly reading documents and being able to answer questions from them is an
unsolved challenge. To avoid its inherent difficulty, question answering (QA)
has been directed towards using Knowledge Bases (KBs) instead, which has proven
effective. Unfortunately KBs often suffer from being too restrictive, as the
schema cannot support certain types of answers, and too sparse, e.g. Wikipedia
contains much more information than Freebase. In this work we introduce a new
method, Key-Value Memory Networks, that makes reading documents more viable by
utilizing different encodings in the addressing and output stages of the memory
read operation. To compare using KBs, information extraction or Wikipedia
documents directly in a single framework we construct an analysis tool,
WikiMovies, a QA dataset that contains raw text alongside a preprocessed KB, in
the domain of movies. Our method reduces the gap between all three settings. It
also achieves state-of-the-art results on the existing WikiQA benchmark.
| new_dataset | 0.958343 |
1606.05250 | Pranav Rajpurkar | Pranav Rajpurkar, Jian Zhang, Konstantin Lopyrev, Percy Liang | SQuAD: 100,000+ Questions for Machine Comprehension of Text | To appear in Proceedings of the 2016 Conference on Empirical Methods
in Natural Language Processing (EMNLP) | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present the Stanford Question Answering Dataset (SQuAD), a new reading
comprehension dataset consisting of 100,000+ questions posed by crowdworkers on
a set of Wikipedia articles, where the answer to each question is a segment of
text from the corresponding reading passage. We analyze the dataset to
understand the types of reasoning required to answer the questions, leaning
heavily on dependency and constituency trees. We build a strong logistic
regression model, which achieves an F1 score of 51.0%, a significant
improvement over a simple baseline (20%). However, human performance (86.8%) is
much higher, indicating that the dataset presents a good challenge problem for
future research.
The dataset is freely available at https://stanford-qa.com
| [
{
"version": "v1",
"created": "Thu, 16 Jun 2016 16:36:00 GMT"
},
{
"version": "v2",
"created": "Fri, 7 Oct 2016 03:48:29 GMT"
},
{
"version": "v3",
"created": "Tue, 11 Oct 2016 02:42:36 GMT"
}
] | 2016-10-12T00:00:00 | [
[
"Rajpurkar",
"Pranav",
""
],
[
"Zhang",
"Jian",
""
],
[
"Lopyrev",
"Konstantin",
""
],
[
"Liang",
"Percy",
""
]
] | TITLE: SQuAD: 100,000+ Questions for Machine Comprehension of Text
ABSTRACT: We present the Stanford Question Answering Dataset (SQuAD), a new reading
comprehension dataset consisting of 100,000+ questions posed by crowdworkers on
a set of Wikipedia articles, where the answer to each question is a segment of
text from the corresponding reading passage. We analyze the dataset to
understand the types of reasoning required to answer the questions, leaning
heavily on dependency and constituency trees. We build a strong logistic
regression model, which achieves an F1 score of 51.0%, a significant
improvement over a simple baseline (20%). However, human performance (86.8%) is
much higher, indicating that the dataset presents a good challenge problem for
future research.
The dataset is freely available at https://stanford-qa.com
| new_dataset | 0.954393 |
1609.09028 | Arkaitz Zubiaga | Arkaitz Zubiaga, Elena Kochkina, Maria Liakata, Rob Procter, Michal
Lukasik | Stance Classification in Rumours as a Sequential Task Exploiting the
Tree Structure of Social Media Conversations | COLING 2016 | null | null | null | cs.CL cs.SI | http://creativecommons.org/licenses/by/4.0/ | Rumour stance classification, the task that determines if each tweet in a
collection discussing a rumour is supporting, denying, questioning or simply
commenting on the rumour, has been attracting substantial interest. Here we
introduce a novel approach that makes use of the sequence of transitions
observed in tree-structured conversation threads in Twitter. The conversation
threads are formed by harvesting users' replies to one another, which results
in a nested tree-like structure. Previous work addressing the stance
classification task has treated each tweet as a separate unit. Here we analyse
tweets by virtue of their position in a sequence and test two sequential
classifiers, Linear-Chain CRF and Tree CRF, each of which makes different
assumptions about the conversational structure. We experiment with eight
Twitter datasets, collected during breaking news, and show that exploiting the
sequential structure of Twitter conversations achieves significant improvements
over the non-sequential methods. Our work is the first to model Twitter
conversations as a tree structure in this manner, introducing a novel way of
tackling NLP tasks on Twitter conversations.
| [
{
"version": "v1",
"created": "Wed, 28 Sep 2016 18:24:12 GMT"
},
{
"version": "v2",
"created": "Tue, 11 Oct 2016 11:54:36 GMT"
}
] | 2016-10-12T00:00:00 | [
[
"Zubiaga",
"Arkaitz",
""
],
[
"Kochkina",
"Elena",
""
],
[
"Liakata",
"Maria",
""
],
[
"Procter",
"Rob",
""
],
[
"Lukasik",
"Michal",
""
]
] | TITLE: Stance Classification in Rumours as a Sequential Task Exploiting the
Tree Structure of Social Media Conversations
ABSTRACT: Rumour stance classification, the task that determines if each tweet in a
collection discussing a rumour is supporting, denying, questioning or simply
commenting on the rumour, has been attracting substantial interest. Here we
introduce a novel approach that makes use of the sequence of transitions
observed in tree-structured conversation threads in Twitter. The conversation
threads are formed by harvesting users' replies to one another, which results
in a nested tree-like structure. Previous work addressing the stance
classification task has treated each tweet as a separate unit. Here we analyse
tweets by virtue of their position in a sequence and test two sequential
classifiers, Linear-Chain CRF and Tree CRF, each of which makes different
assumptions about the conversational structure. We experiment with eight
Twitter datasets, collected during breaking news, and show that exploiting the
sequential structure of Twitter conversations achieves significant improvements
over the non-sequential methods. Our work is the first to model Twitter
conversations as a tree structure in this manner, introducing a novel way of
tackling NLP tasks on Twitter conversations.
| no_new_dataset | 0.950457 |
1610.02831 | Ahilan Kanagasundaram Dr | Ahilan Kanagasundaram, David Dean, Sridha Sridharan and Clinton Fookes | Domain adaptation based Speaker Recognition on Short Utterances | null | null | null | null | cs.SD | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper explores how the in- and out-domain probabilistic linear
discriminant analysis (PLDA) speaker verification behave when enrolment and
verification lengths are reduced. Experiment studies have found that when
full-length utterance is used for evaluation, in-domain PLDA approach shows
more than 28% improvement in EER and DCF values over out-domain PLDA approach
and when short utterances are used for evaluation, the performance gain of
in-domain speaker verification reduces at an increasing rate. Novel modified
inter dataset variability (IDV) compensation is used to compensate the mismatch
between in- and out-domain data and IDV-compensated out-domain PLDA shows
respectively 26% and 14% improvement over out-domain PLDA speaker verification
when SWB and NIST data are respectively used for S normalization. When the
evaluation utterance length is reduced, the performance gain by IDV also
reduces as short utterance evaluation data i-vectors have more variations due
to phonetic variations when compared to the dataset mismatch between in- and
out-domain data.
| [
{
"version": "v1",
"created": "Mon, 10 Oct 2016 10:09:49 GMT"
},
{
"version": "v2",
"created": "Tue, 11 Oct 2016 05:10:07 GMT"
}
] | 2016-10-12T00:00:00 | [
[
"Kanagasundaram",
"Ahilan",
""
],
[
"Dean",
"David",
""
],
[
"Sridharan",
"Sridha",
""
],
[
"Fookes",
"Clinton",
""
]
] | TITLE: Domain adaptation based Speaker Recognition on Short Utterances
ABSTRACT: This paper explores how the in- and out-domain probabilistic linear
discriminant analysis (PLDA) speaker verification behave when enrolment and
verification lengths are reduced. Experiment studies have found that when
full-length utterance is used for evaluation, in-domain PLDA approach shows
more than 28% improvement in EER and DCF values over out-domain PLDA approach
and when short utterances are used for evaluation, the performance gain of
in-domain speaker verification reduces at an increasing rate. Novel modified
inter dataset variability (IDV) compensation is used to compensate the mismatch
between in- and out-domain data and IDV-compensated out-domain PLDA shows
respectively 26% and 14% improvement over out-domain PLDA speaker verification
when SWB and NIST data are respectively used for S normalization. When the
evaluation utterance length is reduced, the performance gain by IDV also
reduces as short utterance evaluation data i-vectors have more variations due
to phonetic variations when compared to the dataset mismatch between in- and
out-domain data.
| no_new_dataset | 0.948489 |
1610.03105 | Yadu Babuji | Yadu N. Babuji, Kyle Chard, Aaron Gerow, Eamon Duede | A Secure Data Enclave and Analytics Platform for Social Scientists | Forthcoming eScience 2016 | null | null | null | cs.DC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Data-driven research is increasingly ubiquitous and data itself is a defining
asset for researchers, particularly in the computational social sciences and
humanities. Entire careers and research communities are built around valuable,
proprietary or sensitive datasets. However, many existing computation resources
fail to support secure and cost-effective storage of data while also enabling
secure and flexible analysis of the data. To address these needs we present
CLOUD KOTTA, a cloud-based architecture for the secure management and analysis
of social science data. CLOUD KOTTA leverages reliable, secure, and scalable
cloud resources to deliver capabilities to users, and removes the need for
users to manage complicated infrastructure. CLOUD KOTTA implements automated,
cost-aware models for efficiently provisioning tiered storage and automatically
scaled compute resources. CLOUD KOTTA has been used in production for several
months and currently manages approximately 10TB of data and has been used to
process more than 5TB of data with over 75,000 CPU hours. It has been used for
a broad variety of text analysis workflows, matrix factorization, and various
machine learning algorithms, and more broadly, it supports fast, secure and
cost-effective research.
| [
{
"version": "v1",
"created": "Mon, 10 Oct 2016 21:44:12 GMT"
}
] | 2016-10-12T00:00:00 | [
[
"Babuji",
"Yadu N.",
""
],
[
"Chard",
"Kyle",
""
],
[
"Gerow",
"Aaron",
""
],
[
"Duede",
"Eamon",
""
]
] | TITLE: A Secure Data Enclave and Analytics Platform for Social Scientists
ABSTRACT: Data-driven research is increasingly ubiquitous and data itself is a defining
asset for researchers, particularly in the computational social sciences and
humanities. Entire careers and research communities are built around valuable,
proprietary or sensitive datasets. However, many existing computation resources
fail to support secure and cost-effective storage of data while also enabling
secure and flexible analysis of the data. To address these needs we present
CLOUD KOTTA, a cloud-based architecture for the secure management and analysis
of social science data. CLOUD KOTTA leverages reliable, secure, and scalable
cloud resources to deliver capabilities to users, and removes the need for
users to manage complicated infrastructure. CLOUD KOTTA implements automated,
cost-aware models for efficiently provisioning tiered storage and automatically
scaled compute resources. CLOUD KOTTA has been used in production for several
months and currently manages approximately 10TB of data and has been used to
process more than 5TB of data with over 75,000 CPU hours. It has been used for
a broad variety of text analysis workflows, matrix factorization, and various
machine learning algorithms, and more broadly, it supports fast, secure and
cost-effective research.
| no_new_dataset | 0.937153 |
1610.03106 | Hussam Hamdan | Hussam Hamdan, Patrice Bellot, Frederic Bechet | Supervised Term Weighting Metrics for Sentiment Analysis in Short Text | null | null | null | null | cs.CL cs.IR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Term weighting metrics assign weights to terms in order to discriminate the
important terms from the less crucial ones. Due to this characteristic, these
metrics have attracted growing attention in text classification and recently in
sentiment analysis. Using the weights given by such metrics could lead to more
accurate document representation which may improve the performance of the
classification. While previous studies have focused on proposing or comparing
different weighting metrics at two-classes document level sentiment analysis,
this study propose to analyse the results given by each metric in order to find
out the characteristics of good and bad weighting metrics. Therefore we present
an empirical study of fifteen global supervised weighting metrics with four
local weighting metrics adopted from information retrieval, we also give an
analysis to understand the behavior of each metric by observing and analysing
how each metric distributes the terms and deduce some characteristics which may
distinguish the good and bad metrics. The evaluation has been done using
Support Vector Machine on three different datasets: Twitter, restaurant and
laptop reviews.
| [
{
"version": "v1",
"created": "Mon, 10 Oct 2016 21:52:47 GMT"
}
] | 2016-10-12T00:00:00 | [
[
"Hamdan",
"Hussam",
""
],
[
"Bellot",
"Patrice",
""
],
[
"Bechet",
"Frederic",
""
]
] | TITLE: Supervised Term Weighting Metrics for Sentiment Analysis in Short Text
ABSTRACT: Term weighting metrics assign weights to terms in order to discriminate the
important terms from the less crucial ones. Due to this characteristic, these
metrics have attracted growing attention in text classification and recently in
sentiment analysis. Using the weights given by such metrics could lead to more
accurate document representation which may improve the performance of the
classification. While previous studies have focused on proposing or comparing
different weighting metrics at two-classes document level sentiment analysis,
this study propose to analyse the results given by each metric in order to find
out the characteristics of good and bad weighting metrics. Therefore we present
an empirical study of fifteen global supervised weighting metrics with four
local weighting metrics adopted from information retrieval, we also give an
analysis to understand the behavior of each metric by observing and analysing
how each metric distributes the terms and deduce some characteristics which may
distinguish the good and bad metrics. The evaluation has been done using
Support Vector Machine on three different datasets: Twitter, restaurant and
laptop reviews.
| no_new_dataset | 0.9463 |
1610.03124 | Martin Treiber | Valentina Kurtc and Martin Treiber | Calibrating the Local and Platoon Dynamics of Car-following Models on
the Reconstructed NGSIM Data | 8 pages, accepted at the Proceedings of Traffic and Granular Flow '15 | null | null | null | physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The NGSIM trajectory data are used to calibrate two car-following models -
the IDM and the FVDM. We used the I80 dataset which has already been
reconstructed to eliminate outliers, unphysical data, and internal and platoon
inconsistencies contained in the original data.We extract from the data
leader-follower pairs and platoons of up to five consecutive vehicles thereby
eliminating all trajectories that are too short or contain lane changes. Four
error measures based on speed and gap deviations are considered. Furthermore,
we apply three calibration methods: local or direct calibration, global
calibration, and platoon calibration. The last approach means that a platoon of
several vehicles following a data-driven leader is simulated and compared to
the observed dynamics.
| [
{
"version": "v1",
"created": "Mon, 10 Oct 2016 23:13:16 GMT"
}
] | 2016-10-12T00:00:00 | [
[
"Kurtc",
"Valentina",
""
],
[
"Treiber",
"Martin",
""
]
] | TITLE: Calibrating the Local and Platoon Dynamics of Car-following Models on
the Reconstructed NGSIM Data
ABSTRACT: The NGSIM trajectory data are used to calibrate two car-following models -
the IDM and the FVDM. We used the I80 dataset which has already been
reconstructed to eliminate outliers, unphysical data, and internal and platoon
inconsistencies contained in the original data.We extract from the data
leader-follower pairs and platoons of up to five consecutive vehicles thereby
eliminating all trajectories that are too short or contain lane changes. Four
error measures based on speed and gap deviations are considered. Furthermore,
we apply three calibration methods: local or direct calibration, global
calibration, and platoon calibration. The last approach means that a platoon of
several vehicles following a data-driven leader is simulated and compared to
the observed dynamics.
| no_new_dataset | 0.93744 |
1610.03155 | Miao Sun | Miao Sun, Tony X. Han, Ming-Chang Liu and Ahmad Khodayari-Rostamabad | Multiple Instance Learning Convolutional Neural Networks for Object
Recognition | International Conference on Pattern Recognition(ICPR) 2016, Oral
paper | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Convolutional Neural Networks (CNN) have demon- strated its successful
applications in computer vision, speech recognition, and natural language
processing. For object recog- nition, CNNs might be limited by its strict label
requirement and an implicit assumption that images are supposed to be target-
object-dominated for optimal solutions. However, the labeling procedure,
necessitating laying out the locations of target ob- jects, is very tedious,
making high-quality large-scale dataset prohibitively expensive. Data
augmentation schemes are widely used when deep networks suffer the insufficient
training data problem. All the images produced through data augmentation share
the same label, which may be problematic since not all data augmentation
methods are label-preserving. In this paper, we propose a weakly supervised CNN
framework named Multiple Instance Learning Convolutional Neural Networks
(MILCNN) to solve this problem. We apply MILCNN framework to object recognition
and report state-of-the-art performance on three benchmark datasets: CIFAR10,
CIFAR100 and ILSVRC2015 classification dataset.
| [
{
"version": "v1",
"created": "Tue, 11 Oct 2016 02:02:16 GMT"
}
] | 2016-10-12T00:00:00 | [
[
"Sun",
"Miao",
""
],
[
"Han",
"Tony X.",
""
],
[
"Liu",
"Ming-Chang",
""
],
[
"Khodayari-Rostamabad",
"Ahmad",
""
]
] | TITLE: Multiple Instance Learning Convolutional Neural Networks for Object
Recognition
ABSTRACT: Convolutional Neural Networks (CNN) have demon- strated its successful
applications in computer vision, speech recognition, and natural language
processing. For object recog- nition, CNNs might be limited by its strict label
requirement and an implicit assumption that images are supposed to be target-
object-dominated for optimal solutions. However, the labeling procedure,
necessitating laying out the locations of target ob- jects, is very tedious,
making high-quality large-scale dataset prohibitively expensive. Data
augmentation schemes are widely used when deep networks suffer the insufficient
training data problem. All the images produced through data augmentation share
the same label, which may be problematic since not all data augmentation
methods are label-preserving. In this paper, we propose a weakly supervised CNN
framework named Multiple Instance Learning Convolutional Neural Networks
(MILCNN) to solve this problem. We apply MILCNN framework to object recognition
and report state-of-the-art performance on three benchmark datasets: CIFAR10,
CIFAR100 and ILSVRC2015 classification dataset.
| no_new_dataset | 0.950088 |
1610.03164 | Andrea Daniele | Andrea F. Daniele and Mohit Bansal and Matthew R. Walter | Navigational Instruction Generation as Inverse Reinforcement Learning
with Neural Machine Translation | null | null | null | null | cs.RO cs.AI cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Modern robotics applications that involve human-robot interaction require
robots to be able to communicate with humans seamlessly and effectively.
Natural language provides a flexible and efficient medium through which robots
can exchange information with their human partners. Significant advancements
have been made in developing robots capable of interpreting free-form
instructions, but less attention has been devoted to endowing robots with the
ability to generate natural language. We propose a navigational guide model
that enables robots to generate natural language instructions that allow humans
to navigate a priori unknown environments. We first decide which information to
share with the user according to their preferences, using a policy trained from
human demonstrations via inverse reinforcement learning. We then "translate"
this information into a natural language instruction using a neural
sequence-to-sequence model that learns to generate free-form instructions from
natural language corpora. We evaluate our method on a benchmark route
instruction dataset and achieve a BLEU score of 72.18% when compared to
human-generated reference instructions. We additionally conduct navigation
experiments with human participants that demonstrate that our method generates
instructions that people follow as accurately and easily as those produced by
humans.
| [
{
"version": "v1",
"created": "Tue, 11 Oct 2016 02:47:09 GMT"
}
] | 2016-10-12T00:00:00 | [
[
"Daniele",
"Andrea F.",
""
],
[
"Bansal",
"Mohit",
""
],
[
"Walter",
"Matthew R.",
""
]
] | TITLE: Navigational Instruction Generation as Inverse Reinforcement Learning
with Neural Machine Translation
ABSTRACT: Modern robotics applications that involve human-robot interaction require
robots to be able to communicate with humans seamlessly and effectively.
Natural language provides a flexible and efficient medium through which robots
can exchange information with their human partners. Significant advancements
have been made in developing robots capable of interpreting free-form
instructions, but less attention has been devoted to endowing robots with the
ability to generate natural language. We propose a navigational guide model
that enables robots to generate natural language instructions that allow humans
to navigate a priori unknown environments. We first decide which information to
share with the user according to their preferences, using a policy trained from
human demonstrations via inverse reinforcement learning. We then "translate"
this information into a natural language instruction using a neural
sequence-to-sequence model that learns to generate free-form instructions from
natural language corpora. We evaluate our method on a benchmark route
instruction dataset and achieve a BLEU score of 72.18% when compared to
human-generated reference instructions. We additionally conduct navigation
experiments with human participants that demonstrate that our method generates
instructions that people follow as accurately and easily as those produced by
humans.
| no_new_dataset | 0.949763 |
1606.04190 | Carlos Caminha | Carlos Caminha, Vasco Furtado, Vl\'adia Pinheiro e Caio Ponte | Micro-interventions in urban transport from pattern discovery on the
flow of passengers and on the bus network | arXiv admin note: substantial text overlap with arXiv:1606.03737 | null | 10.1109/ISC2.2016.7580776 | null | cs.AI cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we describe a case study in a big metropolis, in which from
data collected by digital sensors, we tried to understand mobility patterns of
persons using buses and how this can generate knowledge to suggest
interventions that are applied incrementally into the transportation network in
use. We have first estimated an Origin-Destination matrix of buses users from
datasets about the ticket validation and GPS positioning of buses. Then we
represent the supply of buses with their routes through bus stops as a complex
network, which allowed us to understand the bottlenecks of the current scenario
and, in particular, applying community discovery techniques, to identify
clusters that the service supply infrastructure has. Finally, from the
superimposing of the flow of people represented in the OriginDestination matrix
in the supply network, we exemplify how micro-interventions can be prospected
by means of an example of the introduction of express routes.
| [
{
"version": "v1",
"created": "Tue, 14 Jun 2016 01:44:16 GMT"
}
] | 2016-10-11T00:00:00 | [
[
"Caminha",
"Carlos",
""
],
[
"Furtado",
"Vasco",
""
],
[
"Ponte",
"Vládia Pinheiro e Caio",
""
]
] | TITLE: Micro-interventions in urban transport from pattern discovery on the
flow of passengers and on the bus network
ABSTRACT: In this paper, we describe a case study in a big metropolis, in which from
data collected by digital sensors, we tried to understand mobility patterns of
persons using buses and how this can generate knowledge to suggest
interventions that are applied incrementally into the transportation network in
use. We have first estimated an Origin-Destination matrix of buses users from
datasets about the ticket validation and GPS positioning of buses. Then we
represent the supply of buses with their routes through bus stops as a complex
network, which allowed us to understand the bottlenecks of the current scenario
and, in particular, applying community discovery techniques, to identify
clusters that the service supply infrastructure has. Finally, from the
superimposing of the flow of people represented in the OriginDestination matrix
in the supply network, we exemplify how micro-interventions can be prospected
by means of an example of the introduction of express routes.
| no_new_dataset | 0.944331 |
1607.02555 | Jakob Engel | Jakob Engel and Vladyslav Usenko and Daniel Cremers | A Photometrically Calibrated Benchmark For Monocular Visual Odometry | * Corrected a bug in the evaluation setup, which caused the real-time
results for ORB-SLAM (dashed lines in Figure 8) to be much worse than they
should be. * https://vision.in.tum.de/data/datasets/mono-dataset | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a dataset for evaluating the tracking accuracy of monocular visual
odometry and SLAM methods. It contains 50 real-world sequences comprising more
than 100 minutes of video, recorded across dozens of different environments --
ranging from narrow indoor corridors to wide outdoor scenes. All sequences
contain mostly exploring camera motion, starting and ending at the same
position. This allows to evaluate tracking accuracy via the accumulated drift
from start to end, without requiring ground truth for the full sequence. In
contrast to existing datasets, all sequences are photometrically calibrated. We
provide exposure times for each frame as reported by the sensor, the camera
response function, and dense lens attenuation factors. We also propose a novel,
simple approach to non-parametric vignette calibration, which requires minimal
set-up and is easy to reproduce. Finally, we thoroughly evaluate two existing
methods (ORB-SLAM and DSO) on the dataset, including an analysis of the effect
of image resolution, camera field of view, and the camera motion direction.
| [
{
"version": "v1",
"created": "Sat, 9 Jul 2016 00:11:14 GMT"
},
{
"version": "v2",
"created": "Sat, 8 Oct 2016 20:06:10 GMT"
}
] | 2016-10-11T00:00:00 | [
[
"Engel",
"Jakob",
""
],
[
"Usenko",
"Vladyslav",
""
],
[
"Cremers",
"Daniel",
""
]
] | TITLE: A Photometrically Calibrated Benchmark For Monocular Visual Odometry
ABSTRACT: We present a dataset for evaluating the tracking accuracy of monocular visual
odometry and SLAM methods. It contains 50 real-world sequences comprising more
than 100 minutes of video, recorded across dozens of different environments --
ranging from narrow indoor corridors to wide outdoor scenes. All sequences
contain mostly exploring camera motion, starting and ending at the same
position. This allows to evaluate tracking accuracy via the accumulated drift
from start to end, without requiring ground truth for the full sequence. In
contrast to existing datasets, all sequences are photometrically calibrated. We
provide exposure times for each frame as reported by the sensor, the camera
response function, and dense lens attenuation factors. We also propose a novel,
simple approach to non-parametric vignette calibration, which requires minimal
set-up and is easy to reproduce. Finally, we thoroughly evaluate two existing
methods (ORB-SLAM and DSO) on the dataset, including an analysis of the effect
of image resolution, camera field of view, and the camera motion direction.
| new_dataset | 0.95877 |
1607.03611 | Weishan Dong | Weishan Dong, Jian Li, Renjie Yao, Changsheng Li, Ting Yuan, Lanjun
Wang | Characterizing Driving Styles with Deep Learning | null | null | null | null | cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Characterizing driving styles of human drivers using vehicle sensor data,
e.g., GPS, is an interesting research problem and an important real-world
requirement from automotive industries. A good representation of driving
features can be highly valuable for autonomous driving, auto insurance, and
many other application scenarios. However, traditional methods mainly rely on
handcrafted features, which limit machine learning algorithms to achieve a
better performance. In this paper, we propose a novel deep learning solution to
this problem, which could be the first attempt of extending deep learning to
driving behavior analysis based on GPS data. The proposed approach can
effectively extract high level and interpretable features describing complex
driving patterns. It also requires significantly less human experience and
work. The power of the learned driving style representations are validated
through the driver identification problem using a large real dataset.
| [
{
"version": "v1",
"created": "Wed, 13 Jul 2016 07:15:30 GMT"
},
{
"version": "v2",
"created": "Sat, 8 Oct 2016 05:21:00 GMT"
}
] | 2016-10-11T00:00:00 | [
[
"Dong",
"Weishan",
""
],
[
"Li",
"Jian",
""
],
[
"Yao",
"Renjie",
""
],
[
"Li",
"Changsheng",
""
],
[
"Yuan",
"Ting",
""
],
[
"Wang",
"Lanjun",
""
]
] | TITLE: Characterizing Driving Styles with Deep Learning
ABSTRACT: Characterizing driving styles of human drivers using vehicle sensor data,
e.g., GPS, is an interesting research problem and an important real-world
requirement from automotive industries. A good representation of driving
features can be highly valuable for autonomous driving, auto insurance, and
many other application scenarios. However, traditional methods mainly rely on
handcrafted features, which limit machine learning algorithms to achieve a
better performance. In this paper, we propose a novel deep learning solution to
this problem, which could be the first attempt of extending deep learning to
driving behavior analysis based on GPS data. The proposed approach can
effectively extract high level and interpretable features describing complex
driving patterns. It also requires significantly less human experience and
work. The power of the learned driving style representations are validated
through the driver identification problem using a large real dataset.
| no_new_dataset | 0.943086 |
1608.00466 | Madhusudan Lakshmana | Madhusudan Lakshmana, Sundararajan Sellamanickam, Shirish Shevade,
Keerthi Selvaraj | Learning Semantically Coherent and Reusable Kernels in Convolution
Neural Nets for Sentence Classification | null | null | null | null | cs.CL cs.LG cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The state-of-the-art CNN models give good performance on sentence
classification tasks. The purpose of this work is to empirically study
desirable properties such as semantic coherence, attention mechanism and
reusability of CNNs in these tasks. Semantically coherent kernels are
preferable as they are a lot more interpretable for explaining the decision of
the learned CNN model. We observe that the learned kernels do not have semantic
coherence. Motivated by this observation, we propose to learn kernels with
semantic coherence using clustering scheme combined with Word2Vec
representation and domain knowledge such as SentiWordNet. We suggest a
technique to visualize attention mechanism of CNNs for decision explanation
purpose. Reusable property enables kernels learned on one problem to be used in
another problem. This helps in efficient learning as only a few additional
domain specific filters may have to be learned. We demonstrate the efficacy of
our core ideas of learning semantically coherent kernels and leveraging
reusable kernels for efficient learning on several benchmark datasets.
Experimental results show the usefulness of our approach by achieving
performance close to the state-of-the-art methods but with semantic and
reusable properties.
| [
{
"version": "v1",
"created": "Mon, 1 Aug 2016 15:14:08 GMT"
},
{
"version": "v2",
"created": "Mon, 10 Oct 2016 03:57:26 GMT"
}
] | 2016-10-11T00:00:00 | [
[
"Lakshmana",
"Madhusudan",
""
],
[
"Sellamanickam",
"Sundararajan",
""
],
[
"Shevade",
"Shirish",
""
],
[
"Selvaraj",
"Keerthi",
""
]
] | TITLE: Learning Semantically Coherent and Reusable Kernels in Convolution
Neural Nets for Sentence Classification
ABSTRACT: The state-of-the-art CNN models give good performance on sentence
classification tasks. The purpose of this work is to empirically study
desirable properties such as semantic coherence, attention mechanism and
reusability of CNNs in these tasks. Semantically coherent kernels are
preferable as they are a lot more interpretable for explaining the decision of
the learned CNN model. We observe that the learned kernels do not have semantic
coherence. Motivated by this observation, we propose to learn kernels with
semantic coherence using clustering scheme combined with Word2Vec
representation and domain knowledge such as SentiWordNet. We suggest a
technique to visualize attention mechanism of CNNs for decision explanation
purpose. Reusable property enables kernels learned on one problem to be used in
another problem. This helps in efficient learning as only a few additional
domain specific filters may have to be learned. We demonstrate the efficacy of
our core ideas of learning semantically coherent kernels and leveraging
reusable kernels for efficient learning on several benchmark datasets.
Experimental results show the usefulness of our approach by achieving
performance close to the state-of-the-art methods but with semantic and
reusable properties.
| no_new_dataset | 0.949248 |
1609.06575 | M. Ros\'ario Oliveira | Cl\'audia Pascoal, M. Ros\'ario Oliveira, Ant\'onio Pacheco, and Rui
Valadas | Theoretical Evaluation of Feature Selection Methods based on Mutual
Information | null | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Feature selection methods are usually evaluated by wrapping specific
classifiers and datasets in the evaluation process, resulting very often in
unfair comparisons between methods. In this work, we develop a theoretical
framework that allows obtaining the true feature ordering of two-dimensional
sequential forward feature selection methods based on mutual information, which
is independent of entropy or mutual information estimation methods,
classifiers, or datasets, and leads to an undoubtful comparison of the methods.
Moreover, the theoretical framework unveils problems intrinsic to some methods
that are otherwise difficult to detect, namely inconsistencies in the
construction of the objective function used to select the candidate features,
due to various types of indeterminations and to the possibility of the entropy
of continuous random variables taking null and negative values.
| [
{
"version": "v1",
"created": "Wed, 21 Sep 2016 14:23:15 GMT"
},
{
"version": "v2",
"created": "Fri, 7 Oct 2016 22:51:50 GMT"
}
] | 2016-10-11T00:00:00 | [
[
"Pascoal",
"Cláudia",
""
],
[
"Oliveira",
"M. Rosário",
""
],
[
"Pacheco",
"António",
""
],
[
"Valadas",
"Rui",
""
]
] | TITLE: Theoretical Evaluation of Feature Selection Methods based on Mutual
Information
ABSTRACT: Feature selection methods are usually evaluated by wrapping specific
classifiers and datasets in the evaluation process, resulting very often in
unfair comparisons between methods. In this work, we develop a theoretical
framework that allows obtaining the true feature ordering of two-dimensional
sequential forward feature selection methods based on mutual information, which
is independent of entropy or mutual information estimation methods,
classifiers, or datasets, and leads to an undoubtful comparison of the methods.
Moreover, the theoretical framework unveils problems intrinsic to some methods
that are otherwise difficult to detect, namely inconsistencies in the
construction of the objective function used to select the candidate features,
due to various types of indeterminations and to the possibility of the entropy
of continuous random variables taking null and negative values.
| no_new_dataset | 0.952794 |
1609.06582 | Emiliano De Cristofaro | Apostolos Pyrgelis and Emiliano De Cristofaro and Gordon Ross | Privacy-Friendly Mobility Analytics using Aggregate Location Data | Published at ACM SIGSPATIAL 2016 | null | null | null | cs.CR cs.CY cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Location data can be extremely useful to study commuting patterns and
disruptions, as well as to predict real-time traffic volumes. At the same time,
however, the fine-grained collection of user locations raises serious privacy
concerns, as this can reveal sensitive information about the users, such as,
life style, political and religious inclinations, or even identities. In this
paper, we study the feasibility of crowd-sourced mobility analytics over
aggregate location information: users periodically report their location, using
a privacy-preserving aggregation protocol, so that the server can only recover
aggregates -- i.e., how many, but not which, users are in a region at a given
time. We experiment with real-world mobility datasets obtained from the
Transport For London authority and the San Francisco Cabs network, and present
a novel methodology based on time series modeling that is geared to forecast
traffic volumes in regions of interest and to detect mobility anomalies in
them. In the presence of anomalies, we also make enhanced traffic volume
predictions by feeding our model with additional information from correlated
regions. Finally, we present and evaluate a mobile app prototype, called
Mobility Data Donors (MDD), in terms of computation, communication, and energy
overhead, demonstrating the real-world deployability of our techniques.
| [
{
"version": "v1",
"created": "Wed, 21 Sep 2016 14:31:15 GMT"
},
{
"version": "v2",
"created": "Sun, 9 Oct 2016 15:58:06 GMT"
}
] | 2016-10-11T00:00:00 | [
[
"Pyrgelis",
"Apostolos",
""
],
[
"De Cristofaro",
"Emiliano",
""
],
[
"Ross",
"Gordon",
""
]
] | TITLE: Privacy-Friendly Mobility Analytics using Aggregate Location Data
ABSTRACT: Location data can be extremely useful to study commuting patterns and
disruptions, as well as to predict real-time traffic volumes. At the same time,
however, the fine-grained collection of user locations raises serious privacy
concerns, as this can reveal sensitive information about the users, such as,
life style, political and religious inclinations, or even identities. In this
paper, we study the feasibility of crowd-sourced mobility analytics over
aggregate location information: users periodically report their location, using
a privacy-preserving aggregation protocol, so that the server can only recover
aggregates -- i.e., how many, but not which, users are in a region at a given
time. We experiment with real-world mobility datasets obtained from the
Transport For London authority and the San Francisco Cabs network, and present
a novel methodology based on time series modeling that is geared to forecast
traffic volumes in regions of interest and to detect mobility anomalies in
them. In the presence of anomalies, we also make enhanced traffic volume
predictions by feeding our model with additional information from correlated
regions. Finally, we present and evaluate a mobile app prototype, called
Mobility Data Donors (MDD), in terms of computation, communication, and energy
overhead, demonstrating the real-world deployability of our techniques.
| no_new_dataset | 0.942718 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.