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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1508.03755 | Team Lear | Danila Potapov (LEAR), Matthijs Douze (LEAR), Jerome Revaud (LEAR),
Zaid Harchaoui (LEAR, CIMS), Cordelia Schmid (LEAR) | Beat-Event Detection in Action Movie Franchises | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | While important advances were recently made towards temporally localizing and
recognizing specific human actions or activities in videos, efficient detection
and classification of long video chunks belonging to semantically defined
categories such as "pursuit" or "romance" remains challenging.We introduce a
new dataset, Action Movie Franchises, consisting of a collection of Hollywood
action movie franchises. We define 11 non-exclusive semantic categories -
called beat-categories - that are broad enough to cover most of the movie
footage. The corresponding beat-events are annotated as groups of video shots,
possibly overlapping.We propose an approach for localizing beat-events based on
classifying shots into beat-categories and learning the temporal constraints
between shots. We show that temporal constraints significantly improve the
classification performance. We set up an evaluation protocol for beat-event
localization as well as for shot classification, depending on whether movies
from the same franchise are present or not in the training data.
| [
{
"version": "v1",
"created": "Sat, 15 Aug 2015 17:04:50 GMT"
}
] | 2015-08-18T00:00:00 | [
[
"Potapov",
"Danila",
"",
"LEAR"
],
[
"Douze",
"Matthijs",
"",
"LEAR"
],
[
"Revaud",
"Jerome",
"",
"LEAR"
],
[
"Harchaoui",
"Zaid",
"",
"LEAR, CIMS"
],
[
"Schmid",
"Cordelia",
"",
"LEAR"
]
] | TITLE: Beat-Event Detection in Action Movie Franchises
ABSTRACT: While important advances were recently made towards temporally localizing and
recognizing specific human actions or activities in videos, efficient detection
and classification of long video chunks belonging to semantically defined
categories such as "pursuit" or "romance" remains challenging.We introduce a
new dataset, Action Movie Franchises, consisting of a collection of Hollywood
action movie franchises. We define 11 non-exclusive semantic categories -
called beat-categories - that are broad enough to cover most of the movie
footage. The corresponding beat-events are annotated as groups of video shots,
possibly overlapping.We propose an approach for localizing beat-events based on
classifying shots into beat-categories and learning the temporal constraints
between shots. We show that temporal constraints significantly improve the
classification performance. We set up an evaluation protocol for beat-event
localization as well as for shot classification, depending on whether movies
from the same franchise are present or not in the training data.
| new_dataset | 0.957675 |
1508.03826 | Shaohua Li | Shaohua Li, Jun Zhu, Chunyan Miao | A Generative Word Embedding Model and its Low Rank Positive Semidefinite
Solution | Proceedings of the Conference on Empirical Methods in Natural
Language Processing (EMNLP) 2015 2015, 11 pages, 2 figures | null | null | null | cs.CL cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Most existing word embedding methods can be categorized into Neural Embedding
Models and Matrix Factorization (MF)-based methods. However some models are
opaque to probabilistic interpretation, and MF-based methods, typically solved
using Singular Value Decomposition (SVD), may incur loss of corpus information.
In addition, it is desirable to incorporate global latent factors, such as
topics, sentiments or writing styles, into the word embedding model. Since
generative models provide a principled way to incorporate latent factors, we
propose a generative word embedding model, which is easy to interpret, and can
serve as a basis of more sophisticated latent factor models. The model
inference reduces to a low rank weighted positive semidefinite approximation
problem. Its optimization is approached by eigendecomposition on a submatrix,
followed by online blockwise regression, which is scalable and avoids the
information loss in SVD. In experiments on 7 common benchmark datasets, our
vectors are competitive to word2vec, and better than other MF-based methods.
| [
{
"version": "v1",
"created": "Sun, 16 Aug 2015 14:12:17 GMT"
}
] | 2015-08-18T00:00:00 | [
[
"Li",
"Shaohua",
""
],
[
"Zhu",
"Jun",
""
],
[
"Miao",
"Chunyan",
""
]
] | TITLE: A Generative Word Embedding Model and its Low Rank Positive Semidefinite
Solution
ABSTRACT: Most existing word embedding methods can be categorized into Neural Embedding
Models and Matrix Factorization (MF)-based methods. However some models are
opaque to probabilistic interpretation, and MF-based methods, typically solved
using Singular Value Decomposition (SVD), may incur loss of corpus information.
In addition, it is desirable to incorporate global latent factors, such as
topics, sentiments or writing styles, into the word embedding model. Since
generative models provide a principled way to incorporate latent factors, we
propose a generative word embedding model, which is easy to interpret, and can
serve as a basis of more sophisticated latent factor models. The model
inference reduces to a low rank weighted positive semidefinite approximation
problem. Its optimization is approached by eigendecomposition on a submatrix,
followed by online blockwise regression, which is scalable and avoids the
information loss in SVD. In experiments on 7 common benchmark datasets, our
vectors are competitive to word2vec, and better than other MF-based methods.
| no_new_dataset | 0.942242 |
1508.03928 | Hongyang Li | Hongyang Li, Huchuan Lu, Zhe Lin, Xiaohui Shen, Brian Price | LCNN: Low-level Feature Embedded CNN for Salient Object Detection | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | In this paper, we propose a novel deep neural network framework embedded with
low-level features (LCNN) for salient object detection in complex images. We
utilise the advantage of convolutional neural networks to automatically learn
the high-level features that capture the structured information and semantic
context in the image. In order to better adapt a CNN model into the saliency
task, we redesign the network architecture based on the small-scale datasets.
Several low-level features are extracted, which can effectively capture
contrast and spatial information in the salient regions, and incorporated to
compensate with the learned high-level features at the output of the last fully
connected layer. The concatenated feature vector is further fed into a
hinge-loss SVM detector in a joint discriminative learning manner and the final
saliency score of each region within the bounding box is obtained by the linear
combination of the detector's weights. Experiments on three challenging
benchmark (MSRA-5000, PASCAL-S, ECCSD) demonstrate our algorithm to be
effective and superior than most low-level oriented state-of-the-arts in terms
of P-R curves, F-measure and mean absolute errors.
| [
{
"version": "v1",
"created": "Mon, 17 Aug 2015 05:45:12 GMT"
}
] | 2015-08-18T00:00:00 | [
[
"Li",
"Hongyang",
""
],
[
"Lu",
"Huchuan",
""
],
[
"Lin",
"Zhe",
""
],
[
"Shen",
"Xiaohui",
""
],
[
"Price",
"Brian",
""
]
] | TITLE: LCNN: Low-level Feature Embedded CNN for Salient Object Detection
ABSTRACT: In this paper, we propose a novel deep neural network framework embedded with
low-level features (LCNN) for salient object detection in complex images. We
utilise the advantage of convolutional neural networks to automatically learn
the high-level features that capture the structured information and semantic
context in the image. In order to better adapt a CNN model into the saliency
task, we redesign the network architecture based on the small-scale datasets.
Several low-level features are extracted, which can effectively capture
contrast and spatial information in the salient regions, and incorporated to
compensate with the learned high-level features at the output of the last fully
connected layer. The concatenated feature vector is further fed into a
hinge-loss SVM detector in a joint discriminative learning manner and the final
saliency score of each region within the bounding box is obtained by the linear
combination of the detector's weights. Experiments on three challenging
benchmark (MSRA-5000, PASCAL-S, ECCSD) demonstrate our algorithm to be
effective and superior than most low-level oriented state-of-the-arts in terms
of P-R curves, F-measure and mean absolute errors.
| no_new_dataset | 0.947914 |
1508.03953 | Tam Nguyen | Kang Wang, Tam V. Nguyen, Jiashi Feng, Jose Sepulveda | Sense Beyond Expressions: Cuteness | 4 pages | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the development of Internet culture, cuteness has become a popular
concept. Many people are curious about what factors making a person look cute.
However, there is rare research to answer this interesting question. In this
work, we construct a dataset of personal images with comprehensively annotated
cuteness scores and facial attributes to investigate this high-level concept in
depth. Based on this dataset, through an automatic attributes mining process,
we find several critical attributes determining the cuteness of a person. We
also develop a novel Continuous Latent Support Vector Machine (C-LSVM) method
to predict the cuteness score of one person given only his image. Extensive
evaluations validate the effectiveness of the proposed method for cuteness
prediction.
| [
{
"version": "v1",
"created": "Mon, 17 Aug 2015 08:48:54 GMT"
}
] | 2015-08-18T00:00:00 | [
[
"Wang",
"Kang",
""
],
[
"Nguyen",
"Tam V.",
""
],
[
"Feng",
"Jiashi",
""
],
[
"Sepulveda",
"Jose",
""
]
] | TITLE: Sense Beyond Expressions: Cuteness
ABSTRACT: With the development of Internet culture, cuteness has become a popular
concept. Many people are curious about what factors making a person look cute.
However, there is rare research to answer this interesting question. In this
work, we construct a dataset of personal images with comprehensively annotated
cuteness scores and facial attributes to investigate this high-level concept in
depth. Based on this dataset, through an automatic attributes mining process,
we find several critical attributes determining the cuteness of a person. We
also develop a novel Continuous Latent Support Vector Machine (C-LSVM) method
to predict the cuteness score of one person given only his image. Extensive
evaluations validate the effectiveness of the proposed method for cuteness
prediction.
| new_dataset | 0.957118 |
1508.03975 | Rossi Kamal Mr | Rossi Kamal, Choonog Seon Hong, and Mi Jung Choi | Autonomic Resilient Internet-of-Things(IoT)Management | null | null | null | null | cs.NI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In Resilient IoT, the revenue of service provider is resilient to uncertain
usage-contexts(e.g. emotion, environmental contexts) of Smart-device users.
Hence, Autonomic Resilient IoT Management problem is decomposed into two
subproblems, namely m-connectivity and k-dominance, such that m-alternations on
revenue making process is resilient to users common interests, which might be
depicted through k-1 alternations of usage-contexts. In this context, a greedy
approximation scheme Bee is proposed, which resolves aforementioned
sub-problems with five consecutive models, namely Maverick, Siren, Pigmy, Arkeo
and Augeas, respectively. Theoretical analysis justifies the problem as
NP-hard, combinatorial optimization problem, which is amenable to greedy
approximation. Moreover, Bee lays out the theoretical foundation of Resilient
Fact-finding, followed by theoretical and experimental(i.e synthetic) proof,
which show how Bee-resilience resolves acute CDS measurement problem.
Accordingly, experiments on real Social rumor dataset extract dominator and
dominate to justify how Bee resilience improves CDS measurement. Finally,
case-study and prototype development are performed on Android and Web platforms
in a Resilient IoT scenario, where service provider recommends personalized
services for Smart-device users.
| [
{
"version": "v1",
"created": "Mon, 17 Aug 2015 10:59:16 GMT"
}
] | 2015-08-18T00:00:00 | [
[
"Kamal",
"Rossi",
""
],
[
"Hong",
"Choonog Seon",
""
],
[
"Choi",
"Mi Jung",
""
]
] | TITLE: Autonomic Resilient Internet-of-Things(IoT)Management
ABSTRACT: In Resilient IoT, the revenue of service provider is resilient to uncertain
usage-contexts(e.g. emotion, environmental contexts) of Smart-device users.
Hence, Autonomic Resilient IoT Management problem is decomposed into two
subproblems, namely m-connectivity and k-dominance, such that m-alternations on
revenue making process is resilient to users common interests, which might be
depicted through k-1 alternations of usage-contexts. In this context, a greedy
approximation scheme Bee is proposed, which resolves aforementioned
sub-problems with five consecutive models, namely Maverick, Siren, Pigmy, Arkeo
and Augeas, respectively. Theoretical analysis justifies the problem as
NP-hard, combinatorial optimization problem, which is amenable to greedy
approximation. Moreover, Bee lays out the theoretical foundation of Resilient
Fact-finding, followed by theoretical and experimental(i.e synthetic) proof,
which show how Bee-resilience resolves acute CDS measurement problem.
Accordingly, experiments on real Social rumor dataset extract dominator and
dominate to justify how Bee resilience improves CDS measurement. Finally,
case-study and prototype development are performed on Android and Web platforms
in a Resilient IoT scenario, where service provider recommends personalized
services for Smart-device users.
| no_new_dataset | 0.951774 |
1508.04073 | Ali Mousavi | Ali Mousavi, Richard G. Baraniuk | An Information-Theoretic Measure of Dependency Among Variables in Large
Datasets | null | null | null | null | cs.IT math.IT stat.ME | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The maximal information coefficient (MIC), which measures the amount of
dependence between two variables, is able to detect both linear and non-linear
associations. However, computational cost grows rapidly as a function of the
dataset size. In this paper, we develop a computationally efficient
approximation to the MIC that replaces its dynamic programming step with a much
simpler technique based on the uniform partitioning of data grid. A variety of
experiments demonstrate the quality of our approximation.
| [
{
"version": "v1",
"created": "Mon, 17 Aug 2015 16:00:30 GMT"
}
] | 2015-08-18T00:00:00 | [
[
"Mousavi",
"Ali",
""
],
[
"Baraniuk",
"Richard G.",
""
]
] | TITLE: An Information-Theoretic Measure of Dependency Among Variables in Large
Datasets
ABSTRACT: The maximal information coefficient (MIC), which measures the amount of
dependence between two variables, is able to detect both linear and non-linear
associations. However, computational cost grows rapidly as a function of the
dataset size. In this paper, we develop a computationally efficient
approximation to the MIC that replaces its dynamic programming step with a much
simpler technique based on the uniform partitioning of data grid. A variety of
experiments demonstrate the quality of our approximation.
| no_new_dataset | 0.947039 |
1508.04123 | Alex Mbaziira | Alex V. Mbaziira, Ehab Abozinadah, and James H. Jones Jr | Evaluating Classifiers in Detecting 419 Scams in Bilingual Cybercriminal
Communities | 7 pages | null | null | null | cs.SI cs.CY cs.LG | http://creativecommons.org/licenses/by/4.0/ | Incidents of organized cybercrime are rising because of criminals are reaping
high financial rewards while incurring low costs to commit crime. As the
digital landscape broadens to accommodate more internet-enabled devices and
technologies like social media, more cybercriminals who are not native English
speakers are invading cyberspace to cash in on quick exploits. In this paper we
evaluate the performance of three machine learning classifiers in detecting 419
scams in a bilingual Nigerian cybercriminal community. We use three popular
classifiers in text processing namely: Na\"ive Bayes, k-nearest neighbors (IBK)
and Support Vector Machines (SVM). The preliminary results on a real world
dataset reveal the SVM significantly outperforms Na\"ive Bayes and IBK at 95%
confidence level.
| [
{
"version": "v1",
"created": "Mon, 17 Aug 2015 19:38:50 GMT"
}
] | 2015-08-18T00:00:00 | [
[
"Mbaziira",
"Alex V.",
""
],
[
"Abozinadah",
"Ehab",
""
],
[
"Jones",
"James H.",
"Jr"
]
] | TITLE: Evaluating Classifiers in Detecting 419 Scams in Bilingual Cybercriminal
Communities
ABSTRACT: Incidents of organized cybercrime are rising because of criminals are reaping
high financial rewards while incurring low costs to commit crime. As the
digital landscape broadens to accommodate more internet-enabled devices and
technologies like social media, more cybercriminals who are not native English
speakers are invading cyberspace to cash in on quick exploits. In this paper we
evaluate the performance of three machine learning classifiers in detecting 419
scams in a bilingual Nigerian cybercriminal community. We use three popular
classifiers in text processing namely: Na\"ive Bayes, k-nearest neighbors (IBK)
and Support Vector Machines (SVM). The preliminary results on a real world
dataset reveal the SVM significantly outperforms Na\"ive Bayes and IBK at 95%
confidence level.
| no_new_dataset | 0.95253 |
1508.03601 | Sanjay Singh | Ranjitha R. K. and Sanjay Singh | Is Stack Overflow Overflowing With Questions and Tags | 11 pages, 7 figures, 3 tables Presented at Third International
Symposium on Women in Computing and Informatics (WCI-2015) | null | null | null | cs.SI cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Programming question and answer (Q & A) websites, such as Quora, Stack
Overflow, and Yahoo! Answer etc. helps us to understand the programming
concepts easily and quickly in a way that has been tested and applied by many
software developers. Stack Overflow is one of the most frequently used
programming Q\&A website where the questions and answers posted are presently
analyzed manually, which requires a huge amount of time and resource. To save
the effort, we present a topic modeling based technique to analyze the words of
the original texts to discover the themes that run through them. We also
propose a method to automate the process of reviewing the quality of questions
on Stack Overflow dataset in order to avoid ballooning the stack overflow with
insignificant questions. The proposed method also recommends the appropriate
tags for the new post, which averts the creation of unnecessary tags on Stack
Overflow.
| [
{
"version": "v1",
"created": "Fri, 14 Aug 2015 18:39:18 GMT"
}
] | 2015-08-17T00:00:00 | [
[
"K.",
"Ranjitha R.",
""
],
[
"Singh",
"Sanjay",
""
]
] | TITLE: Is Stack Overflow Overflowing With Questions and Tags
ABSTRACT: Programming question and answer (Q & A) websites, such as Quora, Stack
Overflow, and Yahoo! Answer etc. helps us to understand the programming
concepts easily and quickly in a way that has been tested and applied by many
software developers. Stack Overflow is one of the most frequently used
programming Q\&A website where the questions and answers posted are presently
analyzed manually, which requires a huge amount of time and resource. To save
the effort, we present a topic modeling based technique to analyze the words of
the original texts to discover the themes that run through them. We also
propose a method to automate the process of reviewing the quality of questions
on Stack Overflow dataset in order to avoid ballooning the stack overflow with
insignificant questions. The proposed method also recommends the appropriate
tags for the new post, which averts the creation of unnecessary tags on Stack
Overflow.
| no_new_dataset | 0.950641 |
1508.03116 | Christan Grant | Christan Grant, Daisy Zhe Wang, Michael L. Wick | Query-Driven Sampling for Collective Entity Resolution | null | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Probabilistic databases play a preeminent role in the processing and
management of uncertain data. Recently, many database research efforts have
integrated probabilistic models into databases to support tasks such as
information extraction and labeling. Many of these efforts are based on batch
oriented inference which inhibits a realtime workflow. One important task is
entity resolution (ER). ER is the process of determining records (mentions) in
a database that correspond to the same real-world entity. Traditional pairwise
ER methods can lead to inconsistencies and low accuracy due to localized
decisions. Leading ER systems solve this problem by collectively resolving all
records using a probabilistic graphical model and Markov chain Monte Carlo
(MCMC) inference. However, for large datasets this is an extremely expensive
process. One key observation is that, such exhaustive ER process incurs a huge
up-front cost, which is wasteful in practice because most users are interested
in only a small subset of entities. In this paper, we advocate pay-as-you-go
entity resolution by developing a number of query-driven collective ER
techniques. We introduce two classes of SQL queries that involve ER operators
--- selection-driven ER and join-driven ER. We implement novel variations of
the MCMC Metropolis Hastings algorithm to generate biased samples and
selectivity-based scheduling algorithms to support the two classes of ER
queries. Finally, we show that query-driven ER algorithms can converge and
return results within minutes over a database populated with the extraction
from a newswire dataset containing 71 million mentions.
| [
{
"version": "v1",
"created": "Thu, 13 Aug 2015 04:23:58 GMT"
}
] | 2015-08-14T00:00:00 | [
[
"Grant",
"Christan",
""
],
[
"Wang",
"Daisy Zhe",
""
],
[
"Wick",
"Michael L.",
""
]
] | TITLE: Query-Driven Sampling for Collective Entity Resolution
ABSTRACT: Probabilistic databases play a preeminent role in the processing and
management of uncertain data. Recently, many database research efforts have
integrated probabilistic models into databases to support tasks such as
information extraction and labeling. Many of these efforts are based on batch
oriented inference which inhibits a realtime workflow. One important task is
entity resolution (ER). ER is the process of determining records (mentions) in
a database that correspond to the same real-world entity. Traditional pairwise
ER methods can lead to inconsistencies and low accuracy due to localized
decisions. Leading ER systems solve this problem by collectively resolving all
records using a probabilistic graphical model and Markov chain Monte Carlo
(MCMC) inference. However, for large datasets this is an extremely expensive
process. One key observation is that, such exhaustive ER process incurs a huge
up-front cost, which is wasteful in practice because most users are interested
in only a small subset of entities. In this paper, we advocate pay-as-you-go
entity resolution by developing a number of query-driven collective ER
techniques. We introduce two classes of SQL queries that involve ER operators
--- selection-driven ER and join-driven ER. We implement novel variations of
the MCMC Metropolis Hastings algorithm to generate biased samples and
selectivity-based scheduling algorithms to support the two classes of ER
queries. Finally, we show that query-driven ER algorithms can converge and
return results within minutes over a database populated with the extraction
from a newswire dataset containing 71 million mentions.
| no_new_dataset | 0.907926 |
1508.02968 | Djamal Belazzougui | Djamal Belazzougui and Fabio Cunial | Space-efficient detection of unusual words | arXiv admin note: text overlap with arXiv:1502.06370 | null | null | null | cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Detecting all the strings that occur in a text more frequently or less
frequently than expected according to an IID or a Markov model is a basic
problem in string mining, yet current algorithms are based on data structures
that are either space-inefficient or incur large slowdowns, and current
implementations cannot scale to genomes or metagenomes in practice. In this
paper we engineer an algorithm based on the suffix tree of a string to use just
a small data structure built on the Burrows-Wheeler transform, and a stack of
$O(\sigma^2\log^2 n)$ bits, where $n$ is the length of the string and $\sigma$
is the size of the alphabet. The size of the stack is $o(n)$ except for very
large values of $\sigma$. We further improve the algorithm by removing its time
dependency on $\sigma$, by reporting only a subset of the maximal repeats and
of the minimal rare words of the string, and by detecting and scoring candidate
under-represented strings that $\textit{do not occur}$ in the string. Our
algorithms are practical and work directly on the BWT, thus they can be
immediately applied to a number of existing datasets that are available in this
form, returning this string mining problem to a manageable scale.
| [
{
"version": "v1",
"created": "Wed, 12 Aug 2015 16:01:21 GMT"
}
] | 2015-08-13T00:00:00 | [
[
"Belazzougui",
"Djamal",
""
],
[
"Cunial",
"Fabio",
""
]
] | TITLE: Space-efficient detection of unusual words
ABSTRACT: Detecting all the strings that occur in a text more frequently or less
frequently than expected according to an IID or a Markov model is a basic
problem in string mining, yet current algorithms are based on data structures
that are either space-inefficient or incur large slowdowns, and current
implementations cannot scale to genomes or metagenomes in practice. In this
paper we engineer an algorithm based on the suffix tree of a string to use just
a small data structure built on the Burrows-Wheeler transform, and a stack of
$O(\sigma^2\log^2 n)$ bits, where $n$ is the length of the string and $\sigma$
is the size of the alphabet. The size of the stack is $o(n)$ except for very
large values of $\sigma$. We further improve the algorithm by removing its time
dependency on $\sigma$, by reporting only a subset of the maximal repeats and
of the minimal rare words of the string, and by detecting and scoring candidate
under-represented strings that $\textit{do not occur}$ in the string. Our
algorithms are practical and work directly on the BWT, thus they can be
immediately applied to a number of existing datasets that are available in this
form, returning this string mining problem to a manageable scale.
| no_new_dataset | 0.943086 |
1503.04337 | Yuekai Sun | Jason D. Lee, Yuekai Sun, Qiang Liu, Jonathan E. Taylor | Communication-efficient sparse regression: a one-shot approach | 29 pages, 3 figures | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We devise a one-shot approach to distributed sparse regression in the
high-dimensional setting. The key idea is to average "debiased" or
"desparsified" lasso estimators. We show the approach converges at the same
rate as the lasso as long as the dataset is not split across too many machines.
We also extend the approach to generalized linear models.
| [
{
"version": "v1",
"created": "Sat, 14 Mar 2015 19:43:30 GMT"
},
{
"version": "v2",
"created": "Mon, 10 Aug 2015 13:57:12 GMT"
},
{
"version": "v3",
"created": "Tue, 11 Aug 2015 17:16:01 GMT"
}
] | 2015-08-12T00:00:00 | [
[
"Lee",
"Jason D.",
""
],
[
"Sun",
"Yuekai",
""
],
[
"Liu",
"Qiang",
""
],
[
"Taylor",
"Jonathan E.",
""
]
] | TITLE: Communication-efficient sparse regression: a one-shot approach
ABSTRACT: We devise a one-shot approach to distributed sparse regression in the
high-dimensional setting. The key idea is to average "debiased" or
"desparsified" lasso estimators. We show the approach converges at the same
rate as the lasso as long as the dataset is not split across too many machines.
We also extend the approach to generalized linear models.
| no_new_dataset | 0.948632 |
1505.01728 | Yamuna Prasad | Yamuna Prasad, K. K. Biswas | Integrating K-means with Quadratic Programming Feature Selection | 17 pages, 11 figures | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Several data mining problems are characterized by data in high dimensions.
One of the popular ways to reduce the dimensionality of the data is to perform
feature selection, i.e, select a subset of relevant and non-redundant features.
Recently, Quadratic Programming Feature Selection (QPFS) has been proposed
which formulates the feature selection problem as a quadratic program. It has
been shown to outperform many of the existing feature selection methods for a
variety of applications. Though, better than many existing approaches, the
running time complexity of QPFS is cubic in the number of features, which can
be quite computationally expensive even for moderately sized datasets. In this
paper we propose a novel method for feature selection by integrating k-means
clustering with QPFS. The basic variant of our approach runs k-means to bring
down the number of features which need to be passed on to QPFS. We then enhance
this idea, wherein we gradually refine the feature space from a very coarse
clustering to a fine-grained one, by interleaving steps of QPFS with k-means
clustering. Every step of QPFS helps in identifying the clusters of irrelevant
features (which can then be thrown away), whereas every step of k-means further
refines the clusters which are potentially relevant. We show that our iterative
refinement of clusters is guaranteed to converge. We provide bounds on the
number of distance computations involved in the k-means algorithm. Further,
each QPFS run is now cubic in number of clusters, which can be much smaller
than actual number of features. Experiments on eight publicly available
datasets show that our approach gives significant computational gains (both in
time and memory), over standard QPFS as well as other state of the art feature
selection methods, even while improving the overall accuracy.
| [
{
"version": "v1",
"created": "Thu, 7 May 2015 14:45:11 GMT"
},
{
"version": "v2",
"created": "Tue, 11 Aug 2015 18:06:36 GMT"
}
] | 2015-08-12T00:00:00 | [
[
"Prasad",
"Yamuna",
""
],
[
"Biswas",
"K. K.",
""
]
] | TITLE: Integrating K-means with Quadratic Programming Feature Selection
ABSTRACT: Several data mining problems are characterized by data in high dimensions.
One of the popular ways to reduce the dimensionality of the data is to perform
feature selection, i.e, select a subset of relevant and non-redundant features.
Recently, Quadratic Programming Feature Selection (QPFS) has been proposed
which formulates the feature selection problem as a quadratic program. It has
been shown to outperform many of the existing feature selection methods for a
variety of applications. Though, better than many existing approaches, the
running time complexity of QPFS is cubic in the number of features, which can
be quite computationally expensive even for moderately sized datasets. In this
paper we propose a novel method for feature selection by integrating k-means
clustering with QPFS. The basic variant of our approach runs k-means to bring
down the number of features which need to be passed on to QPFS. We then enhance
this idea, wherein we gradually refine the feature space from a very coarse
clustering to a fine-grained one, by interleaving steps of QPFS with k-means
clustering. Every step of QPFS helps in identifying the clusters of irrelevant
features (which can then be thrown away), whereas every step of k-means further
refines the clusters which are potentially relevant. We show that our iterative
refinement of clusters is guaranteed to converge. We provide bounds on the
number of distance computations involved in the k-means algorithm. Further,
each QPFS run is now cubic in number of clusters, which can be much smaller
than actual number of features. Experiments on eight publicly available
datasets show that our approach gives significant computational gains (both in
time and memory), over standard QPFS as well as other state of the art feature
selection methods, even while improving the overall accuracy.
| no_new_dataset | 0.946597 |
1508.01951 | Besmira Nushi | Besmira Nushi, Adish Singla, Anja Gruenheid, Erfan Zamanian, Andreas
Krause, Donald Kossmann | Crowd Access Path Optimization: Diversity Matters | 10 pages, 3rd AAAI Conference on Human Computation and Crowdsourcing
(HCOMP 2015) | null | null | null | cs.LG cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Quality assurance is one the most important challenges in crowdsourcing.
Assigning tasks to several workers to increase quality through redundant
answers can be expensive if asking homogeneous sources. This limitation has
been overlooked by current crowdsourcing platforms resulting therefore in
costly solutions. In order to achieve desirable cost-quality tradeoffs it is
essential to apply efficient crowd access optimization techniques. Our work
argues that optimization needs to be aware of diversity and correlation of
information within groups of individuals so that crowdsourcing redundancy can
be adequately planned beforehand. Based on this intuitive idea, we introduce
the Access Path Model (APM), a novel crowd model that leverages the notion of
access paths as an alternative way of retrieving information. APM aggregates
answers ensuring high quality and meaningful confidence. Moreover, we devise a
greedy optimization algorithm for this model that finds a provably good
approximate plan to access the crowd. We evaluate our approach on three
crowdsourced datasets that illustrate various aspects of the problem. Our
results show that the Access Path Model combined with greedy optimization is
cost-efficient and practical to overcome common difficulties in large-scale
crowdsourcing like data sparsity and anonymity.
| [
{
"version": "v1",
"created": "Sat, 8 Aug 2015 20:36:54 GMT"
},
{
"version": "v2",
"created": "Tue, 11 Aug 2015 07:21:57 GMT"
}
] | 2015-08-12T00:00:00 | [
[
"Nushi",
"Besmira",
""
],
[
"Singla",
"Adish",
""
],
[
"Gruenheid",
"Anja",
""
],
[
"Zamanian",
"Erfan",
""
],
[
"Krause",
"Andreas",
""
],
[
"Kossmann",
"Donald",
""
]
] | TITLE: Crowd Access Path Optimization: Diversity Matters
ABSTRACT: Quality assurance is one the most important challenges in crowdsourcing.
Assigning tasks to several workers to increase quality through redundant
answers can be expensive if asking homogeneous sources. This limitation has
been overlooked by current crowdsourcing platforms resulting therefore in
costly solutions. In order to achieve desirable cost-quality tradeoffs it is
essential to apply efficient crowd access optimization techniques. Our work
argues that optimization needs to be aware of diversity and correlation of
information within groups of individuals so that crowdsourcing redundancy can
be adequately planned beforehand. Based on this intuitive idea, we introduce
the Access Path Model (APM), a novel crowd model that leverages the notion of
access paths as an alternative way of retrieving information. APM aggregates
answers ensuring high quality and meaningful confidence. Moreover, we devise a
greedy optimization algorithm for this model that finds a provably good
approximate plan to access the crowd. We evaluate our approach on three
crowdsourced datasets that illustrate various aspects of the problem. Our
results show that the Access Path Model combined with greedy optimization is
cost-efficient and practical to overcome common difficulties in large-scale
crowdsourcing like data sparsity and anonymity.
| no_new_dataset | 0.951188 |
1412.8293 | Jiyan Yang | Haim Avron, Vikas Sindhwani, Jiyan Yang, Michael Mahoney | Quasi-Monte Carlo Feature Maps for Shift-Invariant Kernels | A short version of this paper has been presented in ICML 2014 | null | null | null | stat.ML cs.LG math.NA stat.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider the problem of improving the efficiency of randomized Fourier
feature maps to accelerate training and testing speed of kernel methods on
large datasets. These approximate feature maps arise as Monte Carlo
approximations to integral representations of shift-invariant kernel functions
(e.g., Gaussian kernel). In this paper, we propose to use Quasi-Monte Carlo
(QMC) approximations instead, where the relevant integrands are evaluated on a
low-discrepancy sequence of points as opposed to random point sets as in the
Monte Carlo approach. We derive a new discrepancy measure called box
discrepancy based on theoretical characterizations of the integration error
with respect to a given sequence. We then propose to learn QMC sequences
adapted to our setting based on explicit box discrepancy minimization. Our
theoretical analyses are complemented with empirical results that demonstrate
the effectiveness of classical and adaptive QMC techniques for this problem.
| [
{
"version": "v1",
"created": "Mon, 29 Dec 2014 10:00:39 GMT"
},
{
"version": "v2",
"created": "Sun, 9 Aug 2015 07:20:00 GMT"
}
] | 2015-08-11T00:00:00 | [
[
"Avron",
"Haim",
""
],
[
"Sindhwani",
"Vikas",
""
],
[
"Yang",
"Jiyan",
""
],
[
"Mahoney",
"Michael",
""
]
] | TITLE: Quasi-Monte Carlo Feature Maps for Shift-Invariant Kernels
ABSTRACT: We consider the problem of improving the efficiency of randomized Fourier
feature maps to accelerate training and testing speed of kernel methods on
large datasets. These approximate feature maps arise as Monte Carlo
approximations to integral representations of shift-invariant kernel functions
(e.g., Gaussian kernel). In this paper, we propose to use Quasi-Monte Carlo
(QMC) approximations instead, where the relevant integrands are evaluated on a
low-discrepancy sequence of points as opposed to random point sets as in the
Monte Carlo approach. We derive a new discrepancy measure called box
discrepancy based on theoretical characterizations of the integration error
with respect to a given sequence. We then propose to learn QMC sequences
adapted to our setting based on explicit box discrepancy minimization. Our
theoretical analyses are complemented with empirical results that demonstrate
the effectiveness of classical and adaptive QMC techniques for this problem.
| no_new_dataset | 0.949949 |
1508.02050 | Tahani Almanie | Tahani Almanie, Rsha Mirza and Elizabeth Lor | Crime Prediction Based On Crime Types And Using Spatial And Temporal
Criminal Hotspots | 19 pages, 18 figures, 7 tables | International Journal of Data Mining & Knowledge Management
Process (IJDKP) Vol.5, No.4, July 2015 | 10.5121/ijdkp.2015.5401 | null | cs.AI cs.CY cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper focuses on finding spatial and temporal criminal hotspots. It
analyses two different real-world crimes datasets for Denver, CO and Los
Angeles, CA and provides a comparison between the two datasets through a
statistical analysis supported by several graphs. Then, it clarifies how we
conducted Apriori algorithm to produce interesting frequent patterns for
criminal hotspots. In addition, the paper shows how we used Decision Tree
classifier and Naive Bayesian classifier in order to predict potential crime
types. To further analyse crimes datasets, the paper introduces an analysis
study by combining our findings of Denver crimes dataset with its demographics
information in order to capture the factors that might affect the safety of
neighborhoods. The results of this solution could be used to raise awareness
regarding the dangerous locations and to help agencies to predict future crimes
in a specific location within a particular time.
| [
{
"version": "v1",
"created": "Sun, 9 Aug 2015 17:15:56 GMT"
}
] | 2015-08-11T00:00:00 | [
[
"Almanie",
"Tahani",
""
],
[
"Mirza",
"Rsha",
""
],
[
"Lor",
"Elizabeth",
""
]
] | TITLE: Crime Prediction Based On Crime Types And Using Spatial And Temporal
Criminal Hotspots
ABSTRACT: This paper focuses on finding spatial and temporal criminal hotspots. It
analyses two different real-world crimes datasets for Denver, CO and Los
Angeles, CA and provides a comparison between the two datasets through a
statistical analysis supported by several graphs. Then, it clarifies how we
conducted Apriori algorithm to produce interesting frequent patterns for
criminal hotspots. In addition, the paper shows how we used Decision Tree
classifier and Naive Bayesian classifier in order to predict potential crime
types. To further analyse crimes datasets, the paper introduces an analysis
study by combining our findings of Denver crimes dataset with its demographics
information in order to capture the factors that might affect the safety of
neighborhoods. The results of this solution could be used to raise awareness
regarding the dangerous locations and to help agencies to predict future crimes
in a specific location within a particular time.
| no_new_dataset | 0.950641 |
1508.02086 | Hassan Kingravi | Hassan A. Kingravi, Harshal Maske, Girish Chowdhary | Kernel Controllers: A Systems-Theoretic Approach for Data-Driven
Modeling and Control of Spatiotemporally Evolving Processes | null | null | null | null | cs.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider the problem of modeling, estimating, and controlling the latent
state of a spatiotemporally evolving continuous function using very few sensor
measurements and actuator locations. Our solution to the problem consists of
two parts: a predictive model of functional evolution, and feedback based
estimator and controllers that can robustly recover the state of the model and
drive it to a desired function. We show that layering a dynamical systems prior
over temporal evolution of weights of a kernel model is a valid approach to
spatiotemporal modeling that leads to systems theoretic, control-usable,
predictive models. We provide sufficient conditions on the number of sensors
and actuators required to guarantee observability and controllability. The
approach is validated on a large real dataset, and in simulation for the
control of spatiotemporally evolving function.
| [
{
"version": "v1",
"created": "Sun, 9 Aug 2015 21:26:55 GMT"
}
] | 2015-08-11T00:00:00 | [
[
"Kingravi",
"Hassan A.",
""
],
[
"Maske",
"Harshal",
""
],
[
"Chowdhary",
"Girish",
""
]
] | TITLE: Kernel Controllers: A Systems-Theoretic Approach for Data-Driven
Modeling and Control of Spatiotemporally Evolving Processes
ABSTRACT: We consider the problem of modeling, estimating, and controlling the latent
state of a spatiotemporally evolving continuous function using very few sensor
measurements and actuator locations. Our solution to the problem consists of
two parts: a predictive model of functional evolution, and feedback based
estimator and controllers that can robustly recover the state of the model and
drive it to a desired function. We show that layering a dynamical systems prior
over temporal evolution of weights of a kernel model is a valid approach to
spatiotemporal modeling that leads to systems theoretic, control-usable,
predictive models. We provide sufficient conditions on the number of sensors
and actuators required to guarantee observability and controllability. The
approach is validated on a large real dataset, and in simulation for the
control of spatiotemporally evolving function.
| no_new_dataset | 0.943919 |
1508.02091 | Jack Hessel | Jack Hessel, Nicolas Savva, Michael J. Wilber | Image Representations and New Domains in Neural Image Captioning | 11 Pages, 5 Images, To appear at EMNLP 2015's Vision + Learning
workshop | null | null | null | cs.CL cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We examine the possibility that recent promising results in automatic caption
generation are due primarily to language models. By varying image
representation quality produced by a convolutional neural network, we find that
a state-of-the-art neural captioning algorithm is able to produce quality
captions even when provided with surprisingly poor image representations. We
replicate this result in a new, fine-grained, transfer learned captioning
domain, consisting of 66K recipe image/title pairs. We also provide some
experiments regarding the appropriateness of datasets for automatic captioning,
and find that having multiple captions per image is beneficial, but not an
absolute requirement.
| [
{
"version": "v1",
"created": "Sun, 9 Aug 2015 22:52:10 GMT"
}
] | 2015-08-11T00:00:00 | [
[
"Hessel",
"Jack",
""
],
[
"Savva",
"Nicolas",
""
],
[
"Wilber",
"Michael J.",
""
]
] | TITLE: Image Representations and New Domains in Neural Image Captioning
ABSTRACT: We examine the possibility that recent promising results in automatic caption
generation are due primarily to language models. By varying image
representation quality produced by a convolutional neural network, we find that
a state-of-the-art neural captioning algorithm is able to produce quality
captions even when provided with surprisingly poor image representations. We
replicate this result in a new, fine-grained, transfer learned captioning
domain, consisting of 66K recipe image/title pairs. We also provide some
experiments regarding the appropriateness of datasets for automatic captioning,
and find that having multiple captions per image is beneficial, but not an
absolute requirement.
| no_new_dataset | 0.948632 |
1508.02268 | Ning Chen | Ning Chen and Jun Zhu and Jianfei Chen and Ting Chen | Dropout Training for SVMs with Data Augmentation | 15 pages. arXiv admin note: substantial text overlap with
arXiv:1404.4171 | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Dropout and other feature noising schemes have shown promising results in
controlling over-fitting by artificially corrupting the training data. Though
extensive theoretical and empirical studies have been performed for generalized
linear models, little work has been done for support vector machines (SVMs),
one of the most successful approaches for supervised learning. This paper
presents dropout training for both linear SVMs and the nonlinear extension with
latent representation learning. For linear SVMs, to deal with the intractable
expectation of the non-smooth hinge loss under corrupting distributions, we
develop an iteratively re-weighted least square (IRLS) algorithm by exploring
data augmentation techniques. Our algorithm iteratively minimizes the
expectation of a re-weighted least square problem, where the re-weights are
analytically updated. For nonlinear latent SVMs, we consider learning one layer
of latent representations in SVMs and extend the data augmentation technique in
conjunction with first-order Taylor-expansion to deal with the intractable
expected non-smooth hinge loss and the nonlinearity of latent representations.
Finally, we apply the similar data augmentation ideas to develop a new IRLS
algorithm for the expected logistic loss under corrupting distributions, and we
further develop a non-linear extension of logistic regression by incorporating
one layer of latent representations. Our algorithms offer insights on the
connection and difference between the hinge loss and logistic loss in dropout
training. Empirical results on several real datasets demonstrate the
effectiveness of dropout training on significantly boosting the classification
accuracy of both linear and nonlinear SVMs. In addition, the nonlinear SVMs
further improve the prediction performance on several image datasets.
| [
{
"version": "v1",
"created": "Mon, 10 Aug 2015 14:57:30 GMT"
}
] | 2015-08-11T00:00:00 | [
[
"Chen",
"Ning",
""
],
[
"Zhu",
"Jun",
""
],
[
"Chen",
"Jianfei",
""
],
[
"Chen",
"Ting",
""
]
] | TITLE: Dropout Training for SVMs with Data Augmentation
ABSTRACT: Dropout and other feature noising schemes have shown promising results in
controlling over-fitting by artificially corrupting the training data. Though
extensive theoretical and empirical studies have been performed for generalized
linear models, little work has been done for support vector machines (SVMs),
one of the most successful approaches for supervised learning. This paper
presents dropout training for both linear SVMs and the nonlinear extension with
latent representation learning. For linear SVMs, to deal with the intractable
expectation of the non-smooth hinge loss under corrupting distributions, we
develop an iteratively re-weighted least square (IRLS) algorithm by exploring
data augmentation techniques. Our algorithm iteratively minimizes the
expectation of a re-weighted least square problem, where the re-weights are
analytically updated. For nonlinear latent SVMs, we consider learning one layer
of latent representations in SVMs and extend the data augmentation technique in
conjunction with first-order Taylor-expansion to deal with the intractable
expected non-smooth hinge loss and the nonlinearity of latent representations.
Finally, we apply the similar data augmentation ideas to develop a new IRLS
algorithm for the expected logistic loss under corrupting distributions, and we
further develop a non-linear extension of logistic regression by incorporating
one layer of latent representations. Our algorithms offer insights on the
connection and difference between the hinge loss and logistic loss in dropout
training. Empirical results on several real datasets demonstrate the
effectiveness of dropout training on significantly boosting the classification
accuracy of both linear and nonlinear SVMs. In addition, the nonlinear SVMs
further improve the prediction performance on several image datasets.
| no_new_dataset | 0.952574 |
1508.01534 | Jundong Liu | Bibo Shi, Jundong Liu | Nonlinear Metric Learning for kNN and SVMs through Geometric
Transformations | null | null | null | null | cs.LG cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In recent years, research efforts to extend linear metric learning models to
handle nonlinear structures have attracted great interests. In this paper, we
propose a novel nonlinear solution through the utilization of deformable
geometric models to learn spatially varying metrics, and apply the strategy to
boost the performance of both kNN and SVM classifiers. Thin-plate splines (TPS)
are chosen as the geometric model due to their remarkable versatility and
representation power in accounting for high-order deformations. By transforming
the input space through TPS, we can pull same-class neighbors closer while
pushing different-class points farther away in kNN, as well as make the input
data points more linearly separable in SVMs. Improvements in the performance of
kNN classification are demonstrated through experiments on synthetic and real
world datasets, with comparisons made with several state-of-the-art metric
learning solutions. Our SVM-based models also achieve significant improvements
over traditional linear and kernel SVMs with the same datasets.
| [
{
"version": "v1",
"created": "Thu, 6 Aug 2015 20:29:28 GMT"
}
] | 2015-08-10T00:00:00 | [
[
"Shi",
"Bibo",
""
],
[
"Liu",
"Jundong",
""
]
] | TITLE: Nonlinear Metric Learning for kNN and SVMs through Geometric
Transformations
ABSTRACT: In recent years, research efforts to extend linear metric learning models to
handle nonlinear structures have attracted great interests. In this paper, we
propose a novel nonlinear solution through the utilization of deformable
geometric models to learn spatially varying metrics, and apply the strategy to
boost the performance of both kNN and SVM classifiers. Thin-plate splines (TPS)
are chosen as the geometric model due to their remarkable versatility and
representation power in accounting for high-order deformations. By transforming
the input space through TPS, we can pull same-class neighbors closer while
pushing different-class points farther away in kNN, as well as make the input
data points more linearly separable in SVMs. Improvements in the performance of
kNN classification are demonstrated through experiments on synthetic and real
world datasets, with comparisons made with several state-of-the-art metric
learning solutions. Our SVM-based models also achieve significant improvements
over traditional linear and kernel SVMs with the same datasets.
| no_new_dataset | 0.954984 |
1508.01549 | Uday Kamath Dr. | Uday Kamath, Carlotta Domeniconi and Kenneth De Jong | Theoretical and Empirical Analysis of a Parallel Boosting Algorithm | null | null | null | null | cs.LG cs.DC | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Many real-world problems involve massive amounts of data. Under these
circumstances learning algorithms often become prohibitively expensive, making
scalability a pressing issue to be addressed. A common approach is to perform
sampling to reduce the size of the dataset and enable efficient learning.
Alternatively, one customizes learning algorithms to achieve scalability. In
either case, the key challenge is to obtain algorithmic efficiency without
compromising the quality of the results. In this paper we discuss a
meta-learning algorithm (PSBML) which combines features of parallel algorithms
with concepts from ensemble and boosting methodologies to achieve the desired
scalability property. We present both theoretical and empirical analyses which
show that PSBML preserves a critical property of boosting, specifically,
convergence to a distribution centered around the margin. We then present
additional empirical analyses showing that this meta-level algorithm provides a
general and effective framework that can be used in combination with a variety
of learning classifiers. We perform extensive experiments to investigate the
tradeoff achieved between scalability and accuracy, and robustness to noise, on
both synthetic and real-world data. These empirical results corroborate our
theoretical analysis, and demonstrate the potential of PSBML in achieving
scalability without sacrificing accuracy.
| [
{
"version": "v1",
"created": "Thu, 6 Aug 2015 21:54:34 GMT"
}
] | 2015-08-10T00:00:00 | [
[
"Kamath",
"Uday",
""
],
[
"Domeniconi",
"Carlotta",
""
],
[
"De Jong",
"Kenneth",
""
]
] | TITLE: Theoretical and Empirical Analysis of a Parallel Boosting Algorithm
ABSTRACT: Many real-world problems involve massive amounts of data. Under these
circumstances learning algorithms often become prohibitively expensive, making
scalability a pressing issue to be addressed. A common approach is to perform
sampling to reduce the size of the dataset and enable efficient learning.
Alternatively, one customizes learning algorithms to achieve scalability. In
either case, the key challenge is to obtain algorithmic efficiency without
compromising the quality of the results. In this paper we discuss a
meta-learning algorithm (PSBML) which combines features of parallel algorithms
with concepts from ensemble and boosting methodologies to achieve the desired
scalability property. We present both theoretical and empirical analyses which
show that PSBML preserves a critical property of boosting, specifically,
convergence to a distribution centered around the margin. We then present
additional empirical analyses showing that this meta-level algorithm provides a
general and effective framework that can be used in combination with a variety
of learning classifiers. We perform extensive experiments to investigate the
tradeoff achieved between scalability and accuracy, and robustness to noise, on
both synthetic and real-world data. These empirical results corroborate our
theoretical analysis, and demonstrate the potential of PSBML in achieving
scalability without sacrificing accuracy.
| no_new_dataset | 0.945951 |
1508.01571 | Humberto Corona | Humberto Corona, Michael P. O'Mahony | A Mood-based Genre Classification of Television Content | in ACM Workshop on Recommendation Systems for Television and Online
Video 2014 Foster City, California USA | null | null | null | cs.IR cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The classification of television content helps users organise and navigate
through the large list of channels and programs now available. In this paper,
we address the problem of television content classification by exploiting text
information extracted from program transcriptions. We present an analysis which
adapts a model for sentiment that has been widely and successfully applied in
other fields such as music or blog posts. We use a real-world dataset obtained
from the Boxfish API to compare the performance of classifiers trained on a
number of different feature sets. Our experiments show that, over a large
collection of television content, program genres can be represented in a
three-dimensional space of valence, arousal and dominance, and that promising
classification results can be achieved using features based on this
representation. This finding supports the use of the proposed representation of
television content as a feature space for similarity computation and
recommendation generation.
| [
{
"version": "v1",
"created": "Thu, 6 Aug 2015 23:53:30 GMT"
}
] | 2015-08-10T00:00:00 | [
[
"Corona",
"Humberto",
""
],
[
"O'Mahony",
"Michael P.",
""
]
] | TITLE: A Mood-based Genre Classification of Television Content
ABSTRACT: The classification of television content helps users organise and navigate
through the large list of channels and programs now available. In this paper,
we address the problem of television content classification by exploiting text
information extracted from program transcriptions. We present an analysis which
adapts a model for sentiment that has been widely and successfully applied in
other fields such as music or blog posts. We use a real-world dataset obtained
from the Boxfish API to compare the performance of classifiers trained on a
number of different feature sets. Our experiments show that, over a large
collection of television content, program genres can be represented in a
three-dimensional space of valence, arousal and dominance, and that promising
classification results can be achieved using features based on this
representation. This finding supports the use of the proposed representation of
television content as a feature space for similarity computation and
recommendation generation.
| no_new_dataset | 0.946892 |
1508.01667 | Limin Wang | Limin Wang, Sheng Guo, Weilin Huang, Yu Qiao | Places205-VGGNet Models for Scene Recognition | 2 pages | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | VGGNets have turned out to be effective for object recognition in still
images. However, it is unable to yield good performance by directly adapting
the VGGNet models trained on the ImageNet dataset for scene recognition. This
report describes our implementation of training the VGGNets on the large-scale
Places205 dataset. Specifically, we train three VGGNet models, namely
VGGNet-11, VGGNet-13, and VGGNet-16, by using a Multi-GPU extension of Caffe
toolbox with high computational efficiency. We verify the performance of
trained Places205-VGGNet models on three datasets: MIT67, SUN397, and
Places205. Our trained models achieve the state-of-the-art performance on these
datasets and are made public available.
| [
{
"version": "v1",
"created": "Fri, 7 Aug 2015 12:11:06 GMT"
}
] | 2015-08-10T00:00:00 | [
[
"Wang",
"Limin",
""
],
[
"Guo",
"Sheng",
""
],
[
"Huang",
"Weilin",
""
],
[
"Qiao",
"Yu",
""
]
] | TITLE: Places205-VGGNet Models for Scene Recognition
ABSTRACT: VGGNets have turned out to be effective for object recognition in still
images. However, it is unable to yield good performance by directly adapting
the VGGNet models trained on the ImageNet dataset for scene recognition. This
report describes our implementation of training the VGGNets on the large-scale
Places205 dataset. Specifically, we train three VGGNet models, namely
VGGNet-11, VGGNet-13, and VGGNet-16, by using a Multi-GPU extension of Caffe
toolbox with high computational efficiency. We verify the performance of
trained Places205-VGGNet models on three datasets: MIT67, SUN397, and
Places205. Our trained models achieve the state-of-the-art performance on these
datasets and are made public available.
| no_new_dataset | 0.951051 |
1508.01696 | Kasra Madadipouya | Kasra Madadipouya | A Location-Based Movie Recommender System Using Collaborative Filtering | 7 pages in International Journal in Foundations of Computer Science &
Technology (IJFCST), Vol.5, No.4, July 2015 | null | 10.5121/ijfcst.2015.5402 | null | cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Available recommender systems mostly provide recommendations based on the
users preferences by utilizing traditional methods such as collaborative
filtering which only relies on the similarities between users and items.
However, collaborative filtering might lead to provide poor recommendation
because it does not rely on other useful available data such as users locations
and hence the accuracy of the recommendations could be very low and
inefficient. This could be very obvious in the systems that locations would
affect users preferences highly such as movie recommender systems. In this
paper a new location-based movie recommender system based on the collaborative
filtering is introduced for enhancing the accuracy and the quality of
recommendations. In this approach, users locations have been utilized and take
in consideration in the entire processing of the recommendations and peer
selections. The potential of the proposed approach in providing novel and
better quality recommendations have been discussed through experiments in real
datasets.
| [
{
"version": "v1",
"created": "Fri, 7 Aug 2015 14:03:41 GMT"
}
] | 2015-08-10T00:00:00 | [
[
"Madadipouya",
"Kasra",
""
]
] | TITLE: A Location-Based Movie Recommender System Using Collaborative Filtering
ABSTRACT: Available recommender systems mostly provide recommendations based on the
users preferences by utilizing traditional methods such as collaborative
filtering which only relies on the similarities between users and items.
However, collaborative filtering might lead to provide poor recommendation
because it does not rely on other useful available data such as users locations
and hence the accuracy of the recommendations could be very low and
inefficient. This could be very obvious in the systems that locations would
affect users preferences highly such as movie recommender systems. In this
paper a new location-based movie recommender system based on the collaborative
filtering is introduced for enhancing the accuracy and the quality of
recommendations. In this approach, users locations have been utilized and take
in consideration in the entire processing of the recommendations and peer
selections. The potential of the proposed approach in providing novel and
better quality recommendations have been discussed through experiments in real
datasets.
| no_new_dataset | 0.950824 |
1508.01753 | Martin Marinov Mr | Martin Marinov, Nicholas Nash and David Gregg | Practical Algorithms for Finding Extremal Sets | null | null | null | null | cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The minimal sets within a collection of sets are defined as the ones which do
not have a proper subset within the collection, and the maximal sets are the
ones which do not have a proper superset within the collection. Identifying
extremal sets is a fundamental problem with a wide-range of applications in SAT
solvers, data-mining and social network analysis. In this paper, we present two
novel improvements of the high-quality extremal set identification algorithm,
\textit{AMS-Lex}, described by Bayardo and Panda. The first technique uses
memoization to improve the execution time of the single-threaded variant of the
AMS-Lex, whilst our second improvement uses parallel programming methods. In a
subset of the presented experiments our memoized algorithm executes more than
$400$ times faster than the highly efficient publicly available implementation
of AMS-Lex. Moreover, we show that our modified algorithm's speedup is not
bounded above by a constant and that it increases as the length of the common
prefixes in successive input \textit{itemsets} increases. We provide
experimental results using both real-world and synthetic data sets, and show
our multi-threaded variant algorithm out-performing AMS-Lex by $3$ to $6$
times. We find that on synthetic input datasets when executed using $16$ CPU
cores of a $32$-core machine, our multi-threaded program executes about as fast
as the state of the art parallel GPU-based program using an NVIDIA GTX 580
graphics processing unit.
| [
{
"version": "v1",
"created": "Fri, 7 Aug 2015 16:33:54 GMT"
}
] | 2015-08-10T00:00:00 | [
[
"Marinov",
"Martin",
""
],
[
"Nash",
"Nicholas",
""
],
[
"Gregg",
"David",
""
]
] | TITLE: Practical Algorithms for Finding Extremal Sets
ABSTRACT: The minimal sets within a collection of sets are defined as the ones which do
not have a proper subset within the collection, and the maximal sets are the
ones which do not have a proper superset within the collection. Identifying
extremal sets is a fundamental problem with a wide-range of applications in SAT
solvers, data-mining and social network analysis. In this paper, we present two
novel improvements of the high-quality extremal set identification algorithm,
\textit{AMS-Lex}, described by Bayardo and Panda. The first technique uses
memoization to improve the execution time of the single-threaded variant of the
AMS-Lex, whilst our second improvement uses parallel programming methods. In a
subset of the presented experiments our memoized algorithm executes more than
$400$ times faster than the highly efficient publicly available implementation
of AMS-Lex. Moreover, we show that our modified algorithm's speedup is not
bounded above by a constant and that it increases as the length of the common
prefixes in successive input \textit{itemsets} increases. We provide
experimental results using both real-world and synthetic data sets, and show
our multi-threaded variant algorithm out-performing AMS-Lex by $3$ to $6$
times. We find that on synthetic input datasets when executed using $16$ CPU
cores of a $32$-core machine, our multi-threaded program executes about as fast
as the state of the art parallel GPU-based program using an NVIDIA GTX 580
graphics processing unit.
| no_new_dataset | 0.946745 |
1508.01755 | Tsung-Hsien Wen | Tsung-Hsien Wen, Milica Gasic, Dongho Kim, Nikola Mrksic, Pei-Hao Su,
David Vandyke, Steve Young | Stochastic Language Generation in Dialogue using Recurrent Neural
Networks with Convolutional Sentence Reranking | To be appear in SigDial 2015 | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The natural language generation (NLG) component of a spoken dialogue system
(SDS) usually needs a substantial amount of handcrafting or a well-labeled
dataset to be trained on. These limitations add significantly to development
costs and make cross-domain, multi-lingual dialogue systems intractable.
Moreover, human languages are context-aware. The most natural response should
be directly learned from data rather than depending on predefined syntaxes or
rules. This paper presents a statistical language generator based on a joint
recurrent and convolutional neural network structure which can be trained on
dialogue act-utterance pairs without any semantic alignments or predefined
grammar trees. Objective metrics suggest that this new model outperforms
previous methods under the same experimental conditions. Results of an
evaluation by human judges indicate that it produces not only high quality but
linguistically varied utterances which are preferred compared to n-gram and
rule-based systems.
| [
{
"version": "v1",
"created": "Fri, 7 Aug 2015 16:34:11 GMT"
}
] | 2015-08-10T00:00:00 | [
[
"Wen",
"Tsung-Hsien",
""
],
[
"Gasic",
"Milica",
""
],
[
"Kim",
"Dongho",
""
],
[
"Mrksic",
"Nikola",
""
],
[
"Su",
"Pei-Hao",
""
],
[
"Vandyke",
"David",
""
],
[
"Young",
"Steve",
""
]
] | TITLE: Stochastic Language Generation in Dialogue using Recurrent Neural
Networks with Convolutional Sentence Reranking
ABSTRACT: The natural language generation (NLG) component of a spoken dialogue system
(SDS) usually needs a substantial amount of handcrafting or a well-labeled
dataset to be trained on. These limitations add significantly to development
costs and make cross-domain, multi-lingual dialogue systems intractable.
Moreover, human languages are context-aware. The most natural response should
be directly learned from data rather than depending on predefined syntaxes or
rules. This paper presents a statistical language generator based on a joint
recurrent and convolutional neural network structure which can be trained on
dialogue act-utterance pairs without any semantic alignments or predefined
grammar trees. Objective metrics suggest that this new model outperforms
previous methods under the same experimental conditions. Results of an
evaluation by human judges indicate that it produces not only high quality but
linguistically varied utterances which are preferred compared to n-gram and
rule-based systems.
| no_new_dataset | 0.952486 |
1508.01420 | Luis Marujo | Lu\'is Marujo, Jos\'e Port\^elo, Wang Ling, David Martins de Matos,
Jo\~ao P. Neto, Anatole Gershman, Jaime Carbonell, Isabel Trancoso, Bhiksha
Raj | Privacy-Preserving Multi-Document Summarization | 4 pages, In Proceedings of 2nd ACM SIGIR Workshop on
Privacy-Preserving Information Retrieval, August 2015. arXiv admin note: text
overlap with arXiv:1407.5416 | null | null | null | cs.IR cs.CL cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | State-of-the-art extractive multi-document summarization systems are usually
designed without any concern about privacy issues, meaning that all documents
are open to third parties. In this paper we propose a privacy-preserving
approach to multi-document summarization. Our approach enables other parties to
obtain summaries without learning anything else about the original documents'
content. We use a hashing scheme known as Secure Binary Embeddings to convert
documents representation containing key phrases and bag-of-words into bit
strings, allowing the computation of approximate distances, instead of exact
ones. Our experiments indicate that our system yields similar results to its
non-private counterpart on standard multi-document evaluation datasets.
| [
{
"version": "v1",
"created": "Thu, 6 Aug 2015 14:30:47 GMT"
}
] | 2015-08-07T00:00:00 | [
[
"Marujo",
"Luís",
""
],
[
"Portêlo",
"José",
""
],
[
"Ling",
"Wang",
""
],
[
"de Matos",
"David Martins",
""
],
[
"Neto",
"João P.",
""
],
[
"Gershman",
"Anatole",
""
],
[
"Carbonell",
"Jaime",
""
],
[
"Trancoso",
"Isabel",
""
],
[
"Raj",
"Bhiksha",
""
]
] | TITLE: Privacy-Preserving Multi-Document Summarization
ABSTRACT: State-of-the-art extractive multi-document summarization systems are usually
designed without any concern about privacy issues, meaning that all documents
are open to third parties. In this paper we propose a privacy-preserving
approach to multi-document summarization. Our approach enables other parties to
obtain summaries without learning anything else about the original documents'
content. We use a hashing scheme known as Secure Binary Embeddings to convert
documents representation containing key phrases and bag-of-words into bit
strings, allowing the computation of approximate distances, instead of exact
ones. Our experiments indicate that our system yields similar results to its
non-private counterpart on standard multi-document evaluation datasets.
| no_new_dataset | 0.946745 |
1508.01447 | Iyad AlAgha | Iyad AlAgha | Using Linguistic Analysis to Translate Arabic Natural Language Queries
to SPARQL | Journal Paper | null | null | null | cs.CL cs.AI cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The logic-based machine-understandable framework of the Semantic Web often
challenges naive users when they try to query ontology-based knowledge bases.
Existing research efforts have approached this problem by introducing Natural
Language (NL) interfaces to ontologies. These NL interfaces have the ability to
construct SPARQL queries based on NL user queries. However, most efforts were
restricted to queries expressed in English, and they often benefited from the
advancement of English NLP tools. However, little research has been done to
support querying the Arabic content on the Semantic Web by using NL queries.
This paper presents a domain-independent approach to translate Arabic NL
queries to SPARQL by leveraging linguistic analysis. Based on a special
consideration on Noun Phrases (NPs), our approach uses a language parser to
extract NPs and the relations from Arabic parse trees and match them to the
underlying ontology. It then utilizes knowledge in the ontology to group NPs
into triple-based representations. A SPARQL query is finally generated by
extracting targets and modifiers, and interpreting them into SPARQL. The
interpretation of advanced semantic features including negation, conjunctive
and disjunctive modifiers is also supported. The approach was evaluated by
using two datasets consisting of OWL test data and queries, and the obtained
results have confirmed its feasibility to translate Arabic NL queries to
SPARQL.
| [
{
"version": "v1",
"created": "Thu, 6 Aug 2015 16:10:21 GMT"
}
] | 2015-08-07T00:00:00 | [
[
"AlAgha",
"Iyad",
""
]
] | TITLE: Using Linguistic Analysis to Translate Arabic Natural Language Queries
to SPARQL
ABSTRACT: The logic-based machine-understandable framework of the Semantic Web often
challenges naive users when they try to query ontology-based knowledge bases.
Existing research efforts have approached this problem by introducing Natural
Language (NL) interfaces to ontologies. These NL interfaces have the ability to
construct SPARQL queries based on NL user queries. However, most efforts were
restricted to queries expressed in English, and they often benefited from the
advancement of English NLP tools. However, little research has been done to
support querying the Arabic content on the Semantic Web by using NL queries.
This paper presents a domain-independent approach to translate Arabic NL
queries to SPARQL by leveraging linguistic analysis. Based on a special
consideration on Noun Phrases (NPs), our approach uses a language parser to
extract NPs and the relations from Arabic parse trees and match them to the
underlying ontology. It then utilizes knowledge in the ontology to group NPs
into triple-based representations. A SPARQL query is finally generated by
extracting targets and modifiers, and interpreting them into SPARQL. The
interpretation of advanced semantic features including negation, conjunctive
and disjunctive modifiers is also supported. The approach was evaluated by
using two datasets consisting of OWL test data and queries, and the obtained
results have confirmed its feasibility to translate Arabic NL queries to
SPARQL.
| no_new_dataset | 0.944228 |
1505.03823 | Miao Fan | Miao Fan, Qiang Zhou and Thomas Fang Zheng | Distant Supervision for Entity Linking | null | null | null | null | cs.CL cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Entity linking is an indispensable operation of populating knowledge
repositories for information extraction. It studies on aligning a textual
entity mention to its corresponding disambiguated entry in a knowledge
repository. In this paper, we propose a new paradigm named distantly supervised
entity linking (DSEL), in the sense that the disambiguated entities that belong
to a huge knowledge repository (Freebase) are automatically aligned to the
corresponding descriptive webpages (Wiki pages). In this way, a large scale of
weakly labeled data can be generated without manual annotation and fed to a
classifier for linking more newly discovered entities. Compared with
traditional paradigms based on solo knowledge base, DSEL benefits more via
jointly leveraging the respective advantages of Freebase and Wikipedia.
Specifically, the proposed paradigm facilitates bridging the disambiguated
labels (Freebase) of entities and their textual descriptions (Wikipedia) for
Web-scale entities. Experiments conducted on a dataset of 140,000 items and
60,000 features achieve a baseline F1-measure of 0.517. Furthermore, we analyze
the feature performance and improve the F1-measure to 0.545.
| [
{
"version": "v1",
"created": "Thu, 14 May 2015 18:15:49 GMT"
},
{
"version": "v2",
"created": "Tue, 19 May 2015 14:45:19 GMT"
},
{
"version": "v3",
"created": "Wed, 5 Aug 2015 01:25:26 GMT"
}
] | 2015-08-06T00:00:00 | [
[
"Fan",
"Miao",
""
],
[
"Zhou",
"Qiang",
""
],
[
"Zheng",
"Thomas Fang",
""
]
] | TITLE: Distant Supervision for Entity Linking
ABSTRACT: Entity linking is an indispensable operation of populating knowledge
repositories for information extraction. It studies on aligning a textual
entity mention to its corresponding disambiguated entry in a knowledge
repository. In this paper, we propose a new paradigm named distantly supervised
entity linking (DSEL), in the sense that the disambiguated entities that belong
to a huge knowledge repository (Freebase) are automatically aligned to the
corresponding descriptive webpages (Wiki pages). In this way, a large scale of
weakly labeled data can be generated without manual annotation and fed to a
classifier for linking more newly discovered entities. Compared with
traditional paradigms based on solo knowledge base, DSEL benefits more via
jointly leveraging the respective advantages of Freebase and Wikipedia.
Specifically, the proposed paradigm facilitates bridging the disambiguated
labels (Freebase) of entities and their textual descriptions (Wikipedia) for
Web-scale entities. Experiments conducted on a dataset of 140,000 items and
60,000 features achieve a baseline F1-measure of 0.517. Furthermore, we analyze
the feature performance and improve the F1-measure to 0.545.
| no_new_dataset | 0.937038 |
1507.02206 | Won-Yong Shin | Won-Yong Shin, Bikash C. Singh, Jaehee Cho, and Andr\'e M. Everett | A New Understanding of Friendships in Space: Complex Networks Meet
Twitter | 17 pages, 5 figures, 6 tables, To appear in Journal of Information
Science (Special Issue on Recent Advances on Big Social Data) | null | null | null | cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Studies on friendships in online social networks involving geographic
distance have so far relied on the city location provided in users' profiles.
Consequently, most of the research on friendships have provided accuracy at the
city level, at best, to designate a user's location. This study analyzes a
Twitter dataset because it provides the exact geographic distance between
corresponding users. We start by introducing a strong definition of "friend" on
Twitter (i.e., a definition of bidirectional friendship), requiring
bidirectional communication. Next, we utilize geo-tagged mentions delivered by
users to determine their locations, where "@username" is contained anywhere in
the body of tweets. To provide analysis results, we first introduce a friend
counting algorithm. From the fact that Twitter users are likely to post
consecutive tweets in the static mode, we also introduce a two-stage distance
estimation algorithm. As the first of our main contributions, we verify that
the number of friends of a particular Twitter user follows a well-known
power-law distribution (i.e., a Zipf's distribution or a Pareto distribution).
Our study also provides the following newly-discovered friendship degree
related to the issue of space: The number of friends according to distance
follows a double power-law (i.e., a double Pareto law) distribution, indicating
that the probability of befriending a particular Twitter user is significantly
reduced beyond a certain geographic distance between users, termed the
separation point. Our analysis provides concrete evidence that Twitter can be a
useful platform for assigning a more accurate scalar value to the degree of
friendship between two users.
| [
{
"version": "v1",
"created": "Wed, 8 Jul 2015 16:06:47 GMT"
},
{
"version": "v2",
"created": "Sat, 18 Jul 2015 12:12:25 GMT"
},
{
"version": "v3",
"created": "Tue, 21 Jul 2015 07:18:13 GMT"
},
{
"version": "v4",
"created": "Wed, 5 Aug 2015 08:51:02 GMT"
}
] | 2015-08-06T00:00:00 | [
[
"Shin",
"Won-Yong",
""
],
[
"Singh",
"Bikash C.",
""
],
[
"Cho",
"Jaehee",
""
],
[
"Everett",
"André M.",
""
]
] | TITLE: A New Understanding of Friendships in Space: Complex Networks Meet
Twitter
ABSTRACT: Studies on friendships in online social networks involving geographic
distance have so far relied on the city location provided in users' profiles.
Consequently, most of the research on friendships have provided accuracy at the
city level, at best, to designate a user's location. This study analyzes a
Twitter dataset because it provides the exact geographic distance between
corresponding users. We start by introducing a strong definition of "friend" on
Twitter (i.e., a definition of bidirectional friendship), requiring
bidirectional communication. Next, we utilize geo-tagged mentions delivered by
users to determine their locations, where "@username" is contained anywhere in
the body of tweets. To provide analysis results, we first introduce a friend
counting algorithm. From the fact that Twitter users are likely to post
consecutive tweets in the static mode, we also introduce a two-stage distance
estimation algorithm. As the first of our main contributions, we verify that
the number of friends of a particular Twitter user follows a well-known
power-law distribution (i.e., a Zipf's distribution or a Pareto distribution).
Our study also provides the following newly-discovered friendship degree
related to the issue of space: The number of friends according to distance
follows a double power-law (i.e., a double Pareto law) distribution, indicating
that the probability of befriending a particular Twitter user is significantly
reduced beyond a certain geographic distance between users, termed the
separation point. Our analysis provides concrete evidence that Twitter can be a
useful platform for assigning a more accurate scalar value to the degree of
friendship between two users.
| no_new_dataset | 0.9434 |
1508.00966 | Yankui Sun | Yankui Sun, Tian Zhang, Yue Zhao, Yufan He | 3D Automatic Segmentation Method for Retinal Optical Coherence
Tomography Volume Data Using Boundary Surface Enhancement | 27 pages, 19 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the introduction of spectral-domain optical coherence tomography
(SDOCT), much larger image datasets are routinely acquired compared to what was
possible using the previous generation of time-domain OCT. Thus, there is a
critical need for the development of 3D segmentation methods for processing
these data. We present here a novel 3D automatic segmentation method for
retinal OCT volume data. Briefly, to segment a boundary surface, two OCT volume
datasets are obtained by using a 3D smoothing filter and a 3D differential
filter. Their linear combination is then calculated to generate new volume data
with an enhanced boundary surface, where pixel intensity, boundary position
information, and intensity changes on both sides of the boundary surface are
used simultaneously. Next, preliminary discrete boundary points are detected
from the A-Scans of the volume data. Finally, surface smoothness constraints
and a dynamic threshold are applied to obtain a smoothed boundary surface by
correcting a small number of error points. Our method can extract retinal layer
boundary surfaces sequentially with a decreasing search region of volume data.
We performed automatic segmentation on eight human OCT volume datasets acquired
from a commercial Spectralis OCT system, where each volume of data consisted of
97 OCT images with a resolution of 496 512; experimental results show that this
method can accurately segment seven layer boundary surfaces in normal as well
as some abnormal eyes.
| [
{
"version": "v1",
"created": "Wed, 5 Aug 2015 03:42:54 GMT"
}
] | 2015-08-06T00:00:00 | [
[
"Sun",
"Yankui",
""
],
[
"Zhang",
"Tian",
""
],
[
"Zhao",
"Yue",
""
],
[
"He",
"Yufan",
""
]
] | TITLE: 3D Automatic Segmentation Method for Retinal Optical Coherence
Tomography Volume Data Using Boundary Surface Enhancement
ABSTRACT: With the introduction of spectral-domain optical coherence tomography
(SDOCT), much larger image datasets are routinely acquired compared to what was
possible using the previous generation of time-domain OCT. Thus, there is a
critical need for the development of 3D segmentation methods for processing
these data. We present here a novel 3D automatic segmentation method for
retinal OCT volume data. Briefly, to segment a boundary surface, two OCT volume
datasets are obtained by using a 3D smoothing filter and a 3D differential
filter. Their linear combination is then calculated to generate new volume data
with an enhanced boundary surface, where pixel intensity, boundary position
information, and intensity changes on both sides of the boundary surface are
used simultaneously. Next, preliminary discrete boundary points are detected
from the A-Scans of the volume data. Finally, surface smoothness constraints
and a dynamic threshold are applied to obtain a smoothed boundary surface by
correcting a small number of error points. Our method can extract retinal layer
boundary surfaces sequentially with a decreasing search region of volume data.
We performed automatic segmentation on eight human OCT volume datasets acquired
from a commercial Spectralis OCT system, where each volume of data consisted of
97 OCT images with a resolution of 496 512; experimental results show that this
method can accurately segment seven layer boundary surfaces in normal as well
as some abnormal eyes.
| no_new_dataset | 0.956513 |
1508.00973 | Peixian Chen | Peixian Chen, Nevin L. Zhang, Leonard K.M. Poon, Zhourong Chen | Progressive EM for Latent Tree Models and Hierarchical Topic Detection | null | null | null | null | cs.LG cs.CL cs.IR stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Hierarchical latent tree analysis (HLTA) is recently proposed as a new method
for topic detection. It differs fundamentally from the LDA-based methods in
terms of topic definition, topic-document relationship, and learning method. It
has been shown to discover significantly more coherent topics and better topic
hierarchies. However, HLTA relies on the Expectation-Maximization (EM)
algorithm for parameter estimation and hence is not efficient enough to deal
with large datasets. In this paper, we propose a method to drastically speed up
HLTA using a technique inspired by recent advances in the moments method.
Empirical experiments show that our method greatly improves the efficiency of
HLTA. It is as efficient as the state-of-the-art LDA-based method for
hierarchical topic detection and finds substantially better topics and topic
hierarchies.
| [
{
"version": "v1",
"created": "Wed, 5 Aug 2015 05:00:32 GMT"
}
] | 2015-08-06T00:00:00 | [
[
"Chen",
"Peixian",
""
],
[
"Zhang",
"Nevin L.",
""
],
[
"Poon",
"Leonard K. M.",
""
],
[
"Chen",
"Zhourong",
""
]
] | TITLE: Progressive EM for Latent Tree Models and Hierarchical Topic Detection
ABSTRACT: Hierarchical latent tree analysis (HLTA) is recently proposed as a new method
for topic detection. It differs fundamentally from the LDA-based methods in
terms of topic definition, topic-document relationship, and learning method. It
has been shown to discover significantly more coherent topics and better topic
hierarchies. However, HLTA relies on the Expectation-Maximization (EM)
algorithm for parameter estimation and hence is not efficient enough to deal
with large datasets. In this paper, we propose a method to drastically speed up
HLTA using a technique inspired by recent advances in the moments method.
Empirical experiments show that our method greatly improves the efficiency of
HLTA. It is as efficient as the state-of-the-art LDA-based method for
hierarchical topic detection and finds substantially better topics and topic
hierarchies.
| no_new_dataset | 0.948917 |
1508.01192 | Paulo Shakarian | Andrew Stanton, Amanda Thart, Ashish Jain, Priyank Vyas, Arpan
Chatterjee, Paulo Shakarian | Mining for Causal Relationships: A Data-Driven Study of the Islamic
State | null | Final version presented at KDD 2015 | null | null | cs.CY cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The Islamic State of Iraq and al-Sham (ISIS) is a dominant insurgent group
operating in Iraq and Syria that rose to prominence when it took over Mosul in
June, 2014. In this paper, we present a data-driven approach to analyzing this
group using a dataset consisting of 2200 incidents of military activity
surrounding ISIS and the forces that oppose it (including Iraqi, Syrian, and
the American-led coalition). We combine ideas from logic programming and causal
reasoning to mine for association rules for which we present evidence of
causality. We present relationships that link ISIS vehicle-bourne improvised
explosive device (VBIED) activity in Syria with military operations in Iraq,
coalition air strikes, and ISIS IED activity, as well as rules that may serve
as indicators of spikes in indirect fire, suicide attacks, and arrests.
| [
{
"version": "v1",
"created": "Wed, 5 Aug 2015 19:50:54 GMT"
}
] | 2015-08-06T00:00:00 | [
[
"Stanton",
"Andrew",
""
],
[
"Thart",
"Amanda",
""
],
[
"Jain",
"Ashish",
""
],
[
"Vyas",
"Priyank",
""
],
[
"Chatterjee",
"Arpan",
""
],
[
"Shakarian",
"Paulo",
""
]
] | TITLE: Mining for Causal Relationships: A Data-Driven Study of the Islamic
State
ABSTRACT: The Islamic State of Iraq and al-Sham (ISIS) is a dominant insurgent group
operating in Iraq and Syria that rose to prominence when it took over Mosul in
June, 2014. In this paper, we present a data-driven approach to analyzing this
group using a dataset consisting of 2200 incidents of military activity
surrounding ISIS and the forces that oppose it (including Iraqi, Syrian, and
the American-led coalition). We combine ideas from logic programming and causal
reasoning to mine for association rules for which we present evidence of
causality. We present relationships that link ISIS vehicle-bourne improvised
explosive device (VBIED) activity in Syria with military operations in Iraq,
coalition air strikes, and ISIS IED activity, as well as rules that may serve
as indicators of spikes in indirect fire, suicide attacks, and arrests.
| new_dataset | 0.967256 |
1408.5418 | Jianping Shi | Jianping Shi, Qiong Yan, Li Xu, Jiaya Jia | Hierarchical Saliency Detection on Extended CSSD | 14 pages, 15 figures | null | null | CUHK-CSE-201408 | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Complex structures commonly exist in natural images. When an image contains
small-scale high-contrast patterns either in the background or foreground,
saliency detection could be adversely affected, resulting erroneous and
non-uniform saliency assignment. The issue forms a fundamental challenge for
prior methods. We tackle it from a scale point of view and propose a
multi-layer approach to analyze saliency cues. Different from varying patch
sizes or downsizing images, we measure region-based scales. The final saliency
values are inferred optimally combining all the saliency cues in different
scales using hierarchical inference. Through our inference model, single-scale
information is selected to obtain a saliency map. Our method improves detection
quality on many images that cannot be handled well traditionally. We also
construct an extended Complex Scene Saliency Dataset (ECSSD) to include complex
but general natural images.
| [
{
"version": "v1",
"created": "Mon, 11 Aug 2014 15:18:47 GMT"
},
{
"version": "v2",
"created": "Tue, 4 Aug 2015 07:49:43 GMT"
}
] | 2015-08-05T00:00:00 | [
[
"Shi",
"Jianping",
""
],
[
"Yan",
"Qiong",
""
],
[
"Xu",
"Li",
""
],
[
"Jia",
"Jiaya",
""
]
] | TITLE: Hierarchical Saliency Detection on Extended CSSD
ABSTRACT: Complex structures commonly exist in natural images. When an image contains
small-scale high-contrast patterns either in the background or foreground,
saliency detection could be adversely affected, resulting erroneous and
non-uniform saliency assignment. The issue forms a fundamental challenge for
prior methods. We tackle it from a scale point of view and propose a
multi-layer approach to analyze saliency cues. Different from varying patch
sizes or downsizing images, we measure region-based scales. The final saliency
values are inferred optimally combining all the saliency cues in different
scales using hierarchical inference. Through our inference model, single-scale
information is selected to obtain a saliency map. Our method improves detection
quality on many images that cannot be handled well traditionally. We also
construct an extended Complex Scene Saliency Dataset (ECSSD) to include complex
but general natural images.
| new_dataset | 0.943919 |
1502.06435 | Philip Schniter | Jeremy Vila, Philip Schniter, and Joseph Meola | Hyperspectral Unmixing via Turbo Bilinear Approximate Message Passing | null | null | null | null | cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The goal of hyperspectral unmixing is to decompose an electromagnetic
spectral dataset measured over M spectral bands and T pixels into N constituent
material spectra (or "end-members") with corresponding spatial abundances. In
this paper, we propose a novel approach to hyperspectral unmixing based on
loopy belief propagation (BP) that enables the exploitation of spectral
coherence in the endmembers and spatial coherence in the abundances. In
particular, we partition the factor graph into spectral coherence, spatial
coherence, and bilinear subgraphs, and pass messages between them using a
"turbo" approach. To perform message passing within the bilinear subgraph, we
employ the bilinear generalized approximate message passing algorithm
(BiG-AMP), a recently proposed belief-propagation-based approach to matrix
factorization. Furthermore, we propose an expectation-maximization (EM)
strategy to tune the prior parameters and a model-order selection strategy to
select the number of materials N. Numerical experiments conducted with both
synthetic and real-world data show favorable unmixing performance relative to
existing methods.
| [
{
"version": "v1",
"created": "Mon, 23 Feb 2015 14:13:01 GMT"
},
{
"version": "v2",
"created": "Tue, 4 Aug 2015 13:39:06 GMT"
}
] | 2015-08-05T00:00:00 | [
[
"Vila",
"Jeremy",
""
],
[
"Schniter",
"Philip",
""
],
[
"Meola",
"Joseph",
""
]
] | TITLE: Hyperspectral Unmixing via Turbo Bilinear Approximate Message Passing
ABSTRACT: The goal of hyperspectral unmixing is to decompose an electromagnetic
spectral dataset measured over M spectral bands and T pixels into N constituent
material spectra (or "end-members") with corresponding spatial abundances. In
this paper, we propose a novel approach to hyperspectral unmixing based on
loopy belief propagation (BP) that enables the exploitation of spectral
coherence in the endmembers and spatial coherence in the abundances. In
particular, we partition the factor graph into spectral coherence, spatial
coherence, and bilinear subgraphs, and pass messages between them using a
"turbo" approach. To perform message passing within the bilinear subgraph, we
employ the bilinear generalized approximate message passing algorithm
(BiG-AMP), a recently proposed belief-propagation-based approach to matrix
factorization. Furthermore, we propose an expectation-maximization (EM)
strategy to tune the prior parameters and a model-order selection strategy to
select the number of materials N. Numerical experiments conducted with both
synthetic and real-world data show favorable unmixing performance relative to
existing methods.
| no_new_dataset | 0.950549 |
1508.00430 | Mengyang Yu | Mengyang Yu, Li Liu, Ling Shao | Kernelized Multiview Projection | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Conventional vision algorithms adopt a single type of feature or a simple
concatenation of multiple features, which is always represented in a
high-dimensional space. In this paper, we propose a novel unsupervised spectral
embedding algorithm called Kernelized Multiview Projection (KMP) to better fuse
and embed different feature representations. Computing the kernel matrices from
different features/views, KMP can encode them with the corresponding weights to
achieve a low-dimensional and semantically meaningful subspace where the
distribution of each view is sufficiently smooth and discriminative. More
crucially, KMP is linear for the reproducing kernel Hilbert space (RKHS) and
solves the out-of-sample problem, which allows it to be competent for various
practical applications. Extensive experiments on three popular image datasets
demonstrate the effectiveness of our multiview embedding algorithm.
| [
{
"version": "v1",
"created": "Mon, 3 Aug 2015 14:33:03 GMT"
},
{
"version": "v2",
"created": "Tue, 4 Aug 2015 09:42:14 GMT"
}
] | 2015-08-05T00:00:00 | [
[
"Yu",
"Mengyang",
""
],
[
"Liu",
"Li",
""
],
[
"Shao",
"Ling",
""
]
] | TITLE: Kernelized Multiview Projection
ABSTRACT: Conventional vision algorithms adopt a single type of feature or a simple
concatenation of multiple features, which is always represented in a
high-dimensional space. In this paper, we propose a novel unsupervised spectral
embedding algorithm called Kernelized Multiview Projection (KMP) to better fuse
and embed different feature representations. Computing the kernel matrices from
different features/views, KMP can encode them with the corresponding weights to
achieve a low-dimensional and semantically meaningful subspace where the
distribution of each view is sufficiently smooth and discriminative. More
crucially, KMP is linear for the reproducing kernel Hilbert space (RKHS) and
solves the out-of-sample problem, which allows it to be competent for various
practical applications. Extensive experiments on three popular image datasets
demonstrate the effectiveness of our multiview embedding algorithm.
| no_new_dataset | 0.947332 |
1508.00749 | Indre Zliobaite | Tomas Krilavicius, Indre Zliobaite, Henrikas Simonavicius and Laimonas
Jarusevicius | Predicting respiratory motion for real-time tumour tracking in
radiotherapy | null | null | null | null | cs.AI cs.CE physics.med-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Purpose. Radiation therapy is a local treatment aimed at cells in and around
a tumor. The goal of this study is to develop an algorithmic solution for
predicting the position of a target in 3D in real time, aiming for the short
fixed calibration time for each patient at the beginning of the procedure.
Accurate predictions of lung tumor motion are expected to improve the precision
of radiation treatment by controlling the position of a couch or a beam in
order to compensate for respiratory motion during radiation treatment.
Methods. For developing the algorithmic solution, data mining techniques are
used. A model form from the family of exponential smoothing is assumed, and the
model parameters are fitted by minimizing the absolute disposition error, and
the fluctuations of the prediction signal (jitter). The predictive performance
is evaluated retrospectively on clinical datasets capturing different behavior
(being quiet, talking, laughing), and validated in real-time on a prototype
system with respiratory motion imitation.
Results. An algorithmic solution for respiratory motion prediction (called
ExSmi) is designed. ExSmi achieves good accuracy of prediction (error $4-9$
mm/s) with acceptable jitter values (5-7 mm/s), as tested on out-of-sample
data. The datasets, the code for algorithms and the experiments are openly
available for research purposes on a dedicated website.
Conclusions. The developed algorithmic solution performs well to be
prototyped and deployed in applications of radiotherapy.
| [
{
"version": "v1",
"created": "Tue, 4 Aug 2015 12:26:00 GMT"
}
] | 2015-08-05T00:00:00 | [
[
"Krilavicius",
"Tomas",
""
],
[
"Zliobaite",
"Indre",
""
],
[
"Simonavicius",
"Henrikas",
""
],
[
"Jarusevicius",
"Laimonas",
""
]
] | TITLE: Predicting respiratory motion for real-time tumour tracking in
radiotherapy
ABSTRACT: Purpose. Radiation therapy is a local treatment aimed at cells in and around
a tumor. The goal of this study is to develop an algorithmic solution for
predicting the position of a target in 3D in real time, aiming for the short
fixed calibration time for each patient at the beginning of the procedure.
Accurate predictions of lung tumor motion are expected to improve the precision
of radiation treatment by controlling the position of a couch or a beam in
order to compensate for respiratory motion during radiation treatment.
Methods. For developing the algorithmic solution, data mining techniques are
used. A model form from the family of exponential smoothing is assumed, and the
model parameters are fitted by minimizing the absolute disposition error, and
the fluctuations of the prediction signal (jitter). The predictive performance
is evaluated retrospectively on clinical datasets capturing different behavior
(being quiet, talking, laughing), and validated in real-time on a prototype
system with respiratory motion imitation.
Results. An algorithmic solution for respiratory motion prediction (called
ExSmi) is designed. ExSmi achieves good accuracy of prediction (error $4-9$
mm/s) with acceptable jitter values (5-7 mm/s), as tested on out-of-sample
data. The datasets, the code for algorithms and the experiments are openly
available for research purposes on a dedicated website.
Conclusions. The developed algorithmic solution performs well to be
prototyped and deployed in applications of radiotherapy.
| no_new_dataset | 0.949763 |
1508.00776 | Eleonora Vig | Adrien Gaidon and Eleonora Vig | Online Domain Adaptation for Multi-Object Tracking | To appear at BMVC 2015 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Automatically detecting, labeling, and tracking objects in videos depends
first and foremost on accurate category-level object detectors. These might,
however, not always be available in practice, as acquiring high-quality large
scale labeled training datasets is either too costly or impractical for all
possible real-world application scenarios. A scalable solution consists in
re-using object detectors pre-trained on generic datasets. This work is the
first to investigate the problem of on-line domain adaptation of object
detectors for causal multi-object tracking (MOT). We propose to alleviate the
dataset bias by adapting detectors from category to instances, and back: (i) we
jointly learn all target models by adapting them from the pre-trained one, and
(ii) we also adapt the pre-trained model on-line. We introduce an on-line
multi-task learning algorithm to efficiently share parameters and reduce drift,
while gradually improving recall. Our approach is applicable to any linear
object detector, and we evaluate both cheap "mini-Fisher Vectors" and expensive
"off-the-shelf" ConvNet features. We quantitatively measure the benefit of our
domain adaptation strategy on the KITTI tracking benchmark and on a new dataset
(PASCAL-to-KITTI) we introduce to study the domain mismatch problem in MOT.
| [
{
"version": "v1",
"created": "Tue, 4 Aug 2015 14:01:55 GMT"
}
] | 2015-08-05T00:00:00 | [
[
"Gaidon",
"Adrien",
""
],
[
"Vig",
"Eleonora",
""
]
] | TITLE: Online Domain Adaptation for Multi-Object Tracking
ABSTRACT: Automatically detecting, labeling, and tracking objects in videos depends
first and foremost on accurate category-level object detectors. These might,
however, not always be available in practice, as acquiring high-quality large
scale labeled training datasets is either too costly or impractical for all
possible real-world application scenarios. A scalable solution consists in
re-using object detectors pre-trained on generic datasets. This work is the
first to investigate the problem of on-line domain adaptation of object
detectors for causal multi-object tracking (MOT). We propose to alleviate the
dataset bias by adapting detectors from category to instances, and back: (i) we
jointly learn all target models by adapting them from the pre-trained one, and
(ii) we also adapt the pre-trained model on-line. We introduce an on-line
multi-task learning algorithm to efficiently share parameters and reduce drift,
while gradually improving recall. Our approach is applicable to any linear
object detector, and we evaluate both cheap "mini-Fisher Vectors" and expensive
"off-the-shelf" ConvNet features. We quantitatively measure the benefit of our
domain adaptation strategy on the KITTI tracking benchmark and on a new dataset
(PASCAL-to-KITTI) we introduce to study the domain mismatch problem in MOT.
| no_new_dataset | 0.943348 |
1508.00088 | Shashaank Sivakumar | D.S. Shashaank, V. Sruthi, M.L.S Vijayalakshimi and Jacob Shomona
Garcia | Turnover Prediction Of Shares using Data Mining techniques : A Case
Study | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Predicting the turnover of a company in the ever fluctuating Stock market has
always proved to be a precarious situation and most certainly a difficult task
in hand. Data mining is a well-known sphere of Computer Science that aims on
extracting meaningful information from large databases. However, despite the
existence of many algorithms for the purpose of predicting the future trends,
their efficiency is questionable as their predictions suffer from a high error
rate. The objective of this paper is to investigate various classification
algorithms to predict the turnover of different companies based on the Stock
price. The authorized dataset for predicting the turnover was taken from
www.bsc.com and included the stock market values of various companies over the
past 10 years. The algorithms were investigated using the "R" tool. The feature
selection algorithm, Boruta, was run on this dataset to extract the important
and influential features for classification. With these extracted features, the
Total Turnover of the company was predicted using various classification
algorithms like Random Forest, Decision Tree, SVM and Multinomial Regression.
This prediction mechanism was implemented to predict the turnover of a company
on an everyday basis and hence could help navigate through dubious stock market
trades. An accuracy rate of 95% was achieved by the above prediction process.
Moreover, the importance of stock market attributes was established as well.
| [
{
"version": "v1",
"created": "Sat, 1 Aug 2015 06:50:01 GMT"
}
] | 2015-08-04T00:00:00 | [
[
"Shashaank",
"D. S.",
""
],
[
"Sruthi",
"V.",
""
],
[
"Vijayalakshimi",
"M. L. S",
""
],
[
"Garcia",
"Jacob Shomona",
""
]
] | TITLE: Turnover Prediction Of Shares using Data Mining techniques : A Case
Study
ABSTRACT: Predicting the turnover of a company in the ever fluctuating Stock market has
always proved to be a precarious situation and most certainly a difficult task
in hand. Data mining is a well-known sphere of Computer Science that aims on
extracting meaningful information from large databases. However, despite the
existence of many algorithms for the purpose of predicting the future trends,
their efficiency is questionable as their predictions suffer from a high error
rate. The objective of this paper is to investigate various classification
algorithms to predict the turnover of different companies based on the Stock
price. The authorized dataset for predicting the turnover was taken from
www.bsc.com and included the stock market values of various companies over the
past 10 years. The algorithms were investigated using the "R" tool. The feature
selection algorithm, Boruta, was run on this dataset to extract the important
and influential features for classification. With these extracted features, the
Total Turnover of the company was predicted using various classification
algorithms like Random Forest, Decision Tree, SVM and Multinomial Regression.
This prediction mechanism was implemented to predict the turnover of a company
on an everyday basis and hence could help navigate through dubious stock market
trades. An accuracy rate of 95% was achieved by the above prediction process.
Moreover, the importance of stock market attributes was established as well.
| no_new_dataset | 0.947817 |
1508.00092 | Giovanni Poggi | Marco Castelluccio, Giovanni Poggi, Carlo Sansone, Luisa Verdoliva | Land Use Classification in Remote Sensing Images by Convolutional Neural
Networks | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We explore the use of convolutional neural networks for the semantic
classification of remote sensing scenes. Two recently proposed architectures,
CaffeNet and GoogLeNet, are adopted, with three different learning modalities.
Besides conventional training from scratch, we resort to pre-trained networks
that are only fine-tuned on the target data, so as to avoid overfitting
problems and reduce design time. Experiments on two remote sensing datasets,
with markedly different characteristics, testify on the effectiveness and wide
applicability of the proposed solution, which guarantees a significant
performance improvement over all state-of-the-art references.
| [
{
"version": "v1",
"created": "Sat, 1 Aug 2015 07:15:19 GMT"
}
] | 2015-08-04T00:00:00 | [
[
"Castelluccio",
"Marco",
""
],
[
"Poggi",
"Giovanni",
""
],
[
"Sansone",
"Carlo",
""
],
[
"Verdoliva",
"Luisa",
""
]
] | TITLE: Land Use Classification in Remote Sensing Images by Convolutional Neural
Networks
ABSTRACT: We explore the use of convolutional neural networks for the semantic
classification of remote sensing scenes. Two recently proposed architectures,
CaffeNet and GoogLeNet, are adopted, with three different learning modalities.
Besides conventional training from scratch, we resort to pre-trained networks
that are only fine-tuned on the target data, so as to avoid overfitting
problems and reduce design time. Experiments on two remote sensing datasets,
with markedly different characteristics, testify on the effectiveness and wide
applicability of the proposed solution, which guarantees a significant
performance improvement over all state-of-the-art references.
| no_new_dataset | 0.947721 |
1508.00192 | Ling Chen | Ling Chen, Ting Yu, Rada Chirkova | WaveCluster with Differential Privacy | null | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | WaveCluster is an important family of grid-based clustering algorithms that
are capable of finding clusters of arbitrary shapes. In this paper, we
investigate techniques to perform WaveCluster while ensuring differential
privacy. Our goal is to develop a general technique for achieving differential
privacy on WaveCluster that accommodates different wavelet transforms. We show
that straightforward techniques based on synthetic data generation and
introduction of random noise when quantizing the data, though generally
preserving the distribution of data, often introduce too much noise to preserve
useful clusters. We then propose two optimized techniques, PrivTHR and
PrivTHREM, which can significantly reduce data distortion during two key steps
of WaveCluster: the quantization step and the significant grid identification
step. We conduct extensive experiments based on four datasets that are
particularly interesting in the context of clustering, and show that PrivTHR
and PrivTHREM achieve high utility when privacy budgets are properly allocated.
| [
{
"version": "v1",
"created": "Sun, 2 Aug 2015 04:41:51 GMT"
}
] | 2015-08-04T00:00:00 | [
[
"Chen",
"Ling",
""
],
[
"Yu",
"Ting",
""
],
[
"Chirkova",
"Rada",
""
]
] | TITLE: WaveCluster with Differential Privacy
ABSTRACT: WaveCluster is an important family of grid-based clustering algorithms that
are capable of finding clusters of arbitrary shapes. In this paper, we
investigate techniques to perform WaveCluster while ensuring differential
privacy. Our goal is to develop a general technique for achieving differential
privacy on WaveCluster that accommodates different wavelet transforms. We show
that straightforward techniques based on synthetic data generation and
introduction of random noise when quantizing the data, though generally
preserving the distribution of data, often introduce too much noise to preserve
useful clusters. We then propose two optimized techniques, PrivTHR and
PrivTHREM, which can significantly reduce data distortion during two key steps
of WaveCluster: the quantization step and the significant grid identification
step. We conduct extensive experiments based on four datasets that are
particularly interesting in the context of clustering, and show that PrivTHR
and PrivTHREM achieve high utility when privacy budgets are properly allocated.
| no_new_dataset | 0.952926 |
1508.00305 | Panupong Pasupat | Panupong Pasupat, Percy Liang | Compositional Semantic Parsing on Semi-Structured Tables | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Two important aspects of semantic parsing for question answering are the
breadth of the knowledge source and the depth of logical compositionality.
While existing work trades off one aspect for another, this paper
simultaneously makes progress on both fronts through a new task: answering
complex questions on semi-structured tables using question-answer pairs as
supervision. The central challenge arises from two compounding factors: the
broader domain results in an open-ended set of relations, and the deeper
compositionality results in a combinatorial explosion in the space of logical
forms. We propose a logical-form driven parsing algorithm guided by strong
typing constraints and show that it obtains significant improvements over
natural baselines. For evaluation, we created a new dataset of 22,033 complex
questions on Wikipedia tables, which is made publicly available.
| [
{
"version": "v1",
"created": "Mon, 3 Aug 2015 02:53:01 GMT"
}
] | 2015-08-04T00:00:00 | [
[
"Pasupat",
"Panupong",
""
],
[
"Liang",
"Percy",
""
]
] | TITLE: Compositional Semantic Parsing on Semi-Structured Tables
ABSTRACT: Two important aspects of semantic parsing for question answering are the
breadth of the knowledge source and the depth of logical compositionality.
While existing work trades off one aspect for another, this paper
simultaneously makes progress on both fronts through a new task: answering
complex questions on semi-structured tables using question-answer pairs as
supervision. The central challenge arises from two compounding factors: the
broader domain results in an open-ended set of relations, and the deeper
compositionality results in a combinatorial explosion in the space of logical
forms. We propose a logical-form driven parsing algorithm guided by strong
typing constraints and show that it obtains significant improvements over
natural baselines. For evaluation, we created a new dataset of 22,033 complex
questions on Wikipedia tables, which is made publicly available.
| new_dataset | 0.954393 |
1508.00307 | Weilin Huang | Sheng Guo and Weilin Huang and Yu Qiao | Local Color Contrastive Descriptor for Image Classification | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Image representation and classification are two fundamental tasks towards
multimedia content retrieval and understanding. The idea that shape and texture
information (e.g. edge or orientation) are the key features for visual
representation is ingrained and dominated in current multimedia and computer
vision communities. A number of low-level features have been proposed by
computing local gradients (e.g. SIFT, LBP and HOG), and have achieved great
successes on numerous multimedia applications. In this paper, we present a
simple yet efficient local descriptor for image classification, referred as
Local Color Contrastive Descriptor (LCCD), by leveraging the neural mechanisms
of color contrast. The idea originates from the observation in neural science
that color and shape information are linked inextricably in visual cortical
processing. The color contrast yields key information for visual color
perception and provides strong linkage between color and shape. We propose a
novel contrastive mechanism to compute the color contrast in both spatial
location and multiple channels. The color contrast is computed by measuring
\emph{f}-divergence between the color distributions of two regions. Our
descriptor enriches local image representation with both color and contrast
information. We verified experimentally that it can compensate strongly for the
shape based descriptor (e.g. SIFT), while keeping computationally simple.
Extensive experimental results on image classification show that our descriptor
improves the performance of SIFT substantially by combinations, and achieves
the state-of-the-art performance on three challenging benchmark datasets. It
improves recent Deep Learning model (DeCAF) [1] largely from the accuracy of
40.94% to 49.68% in the large scale SUN397 database. Codes for the LCCD will be
available.
| [
{
"version": "v1",
"created": "Mon, 3 Aug 2015 03:29:50 GMT"
}
] | 2015-08-04T00:00:00 | [
[
"Guo",
"Sheng",
""
],
[
"Huang",
"Weilin",
""
],
[
"Qiao",
"Yu",
""
]
] | TITLE: Local Color Contrastive Descriptor for Image Classification
ABSTRACT: Image representation and classification are two fundamental tasks towards
multimedia content retrieval and understanding. The idea that shape and texture
information (e.g. edge or orientation) are the key features for visual
representation is ingrained and dominated in current multimedia and computer
vision communities. A number of low-level features have been proposed by
computing local gradients (e.g. SIFT, LBP and HOG), and have achieved great
successes on numerous multimedia applications. In this paper, we present a
simple yet efficient local descriptor for image classification, referred as
Local Color Contrastive Descriptor (LCCD), by leveraging the neural mechanisms
of color contrast. The idea originates from the observation in neural science
that color and shape information are linked inextricably in visual cortical
processing. The color contrast yields key information for visual color
perception and provides strong linkage between color and shape. We propose a
novel contrastive mechanism to compute the color contrast in both spatial
location and multiple channels. The color contrast is computed by measuring
\emph{f}-divergence between the color distributions of two regions. Our
descriptor enriches local image representation with both color and contrast
information. We verified experimentally that it can compensate strongly for the
shape based descriptor (e.g. SIFT), while keeping computationally simple.
Extensive experimental results on image classification show that our descriptor
improves the performance of SIFT substantially by combinations, and achieves
the state-of-the-art performance on three challenging benchmark datasets. It
improves recent Deep Learning model (DeCAF) [1] largely from the accuracy of
40.94% to 49.68% in the large scale SUN397 database. Codes for the LCCD will be
available.
| no_new_dataset | 0.950869 |
1508.00317 | Roni Mittelman Roni Mittelman | Roni Mittelman | Time-series modeling with undecimated fully convolutional neural
networks | null | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a new convolutional neural network-based time-series model.
Typical convolutional neural network (CNN) architectures rely on the use of
max-pooling operators in between layers, which leads to reduced resolution at
the top layers. Instead, in this work we consider a fully convolutional network
(FCN) architecture that uses causal filtering operations, and allows for the
rate of the output signal to be the same as that of the input signal. We
furthermore propose an undecimated version of the FCN, which we refer to as the
undecimated fully convolutional neural network (UFCNN), and is motivated by the
undecimated wavelet transform. Our experimental results verify that using the
undecimated version of the FCN is necessary in order to allow for effective
time-series modeling. The UFCNN has several advantages compared to other
time-series models such as the recurrent neural network (RNN) and long
short-term memory (LSTM), since it does not suffer from either the vanishing or
exploding gradients problems, and is therefore easier to train. Convolution
operations can also be implemented more efficiently compared to the recursion
that is involved in RNN-based models. We evaluate the performance of our model
in a synthetic target tracking task using bearing only measurements generated
from a state-space model, a probabilistic modeling of polyphonic music
sequences problem, and a high frequency trading task using a time-series of
ask/bid quotes and their corresponding volumes. Our experimental results using
synthetic and real datasets verify the significant advantages of the UFCNN
compared to the RNN and LSTM baselines.
| [
{
"version": "v1",
"created": "Mon, 3 Aug 2015 05:58:52 GMT"
}
] | 2015-08-04T00:00:00 | [
[
"Mittelman",
"Roni",
""
]
] | TITLE: Time-series modeling with undecimated fully convolutional neural
networks
ABSTRACT: We present a new convolutional neural network-based time-series model.
Typical convolutional neural network (CNN) architectures rely on the use of
max-pooling operators in between layers, which leads to reduced resolution at
the top layers. Instead, in this work we consider a fully convolutional network
(FCN) architecture that uses causal filtering operations, and allows for the
rate of the output signal to be the same as that of the input signal. We
furthermore propose an undecimated version of the FCN, which we refer to as the
undecimated fully convolutional neural network (UFCNN), and is motivated by the
undecimated wavelet transform. Our experimental results verify that using the
undecimated version of the FCN is necessary in order to allow for effective
time-series modeling. The UFCNN has several advantages compared to other
time-series models such as the recurrent neural network (RNN) and long
short-term memory (LSTM), since it does not suffer from either the vanishing or
exploding gradients problems, and is therefore easier to train. Convolution
operations can also be implemented more efficiently compared to the recursion
that is involved in RNN-based models. We evaluate the performance of our model
in a synthetic target tracking task using bearing only measurements generated
from a state-space model, a probabilistic modeling of polyphonic music
sequences problem, and a high frequency trading task using a time-series of
ask/bid quotes and their corresponding volumes. Our experimental results using
synthetic and real datasets verify the significant advantages of the UFCNN
compared to the RNN and LSTM baselines.
| no_new_dataset | 0.952662 |
1508.00507 | Tameem Adel | Tameem Adel, Alexander Wong, Daniel Stashuk | A Weakly Supervised Learning Approach based on Spectral Graph-Theoretic
Grouping | Submitted to IEEE Access | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this study, a spectral graph-theoretic grouping strategy for weakly
supervised classification is introduced, where a limited number of labelled
samples and a larger set of unlabelled samples are used to construct a larger
annotated training set composed of strongly labelled and weakly labelled
samples. The inherent relationship between the set of strongly labelled samples
and the set of unlabelled samples is established via spectral grouping, with
the unlabelled samples subsequently weakly annotated based on the strongly
labelled samples within the associated spectral groups. A number of similarity
graph models for spectral grouping, including two new similarity graph models
introduced in this study, are explored to investigate their performance in the
context of weakly supervised classification in handling different types of
data. Experimental results using benchmark datasets as well as real EMG
datasets demonstrate that the proposed approach to weakly supervised
classification can provide noticeable improvements in classification
performance, and that the proposed similarity graph models can lead to ultimate
learning results that are either better than or on a par with existing
similarity graph models in the context of spectral grouping for weakly
supervised classification.
| [
{
"version": "v1",
"created": "Mon, 3 Aug 2015 18:08:04 GMT"
}
] | 2015-08-04T00:00:00 | [
[
"Adel",
"Tameem",
""
],
[
"Wong",
"Alexander",
""
],
[
"Stashuk",
"Daniel",
""
]
] | TITLE: A Weakly Supervised Learning Approach based on Spectral Graph-Theoretic
Grouping
ABSTRACT: In this study, a spectral graph-theoretic grouping strategy for weakly
supervised classification is introduced, where a limited number of labelled
samples and a larger set of unlabelled samples are used to construct a larger
annotated training set composed of strongly labelled and weakly labelled
samples. The inherent relationship between the set of strongly labelled samples
and the set of unlabelled samples is established via spectral grouping, with
the unlabelled samples subsequently weakly annotated based on the strongly
labelled samples within the associated spectral groups. A number of similarity
graph models for spectral grouping, including two new similarity graph models
introduced in this study, are explored to investigate their performance in the
context of weakly supervised classification in handling different types of
data. Experimental results using benchmark datasets as well as real EMG
datasets demonstrate that the proposed approach to weakly supervised
classification can provide noticeable improvements in classification
performance, and that the proposed similarity graph models can lead to ultimate
learning results that are either better than or on a par with existing
similarity graph models in the context of spectral grouping for weakly
supervised classification.
| no_new_dataset | 0.951639 |
1508.00537 | Rodrigo Nogueira | Rodrigo Frassetto Nogueira, Roberto de Alencar Lotufo, Rubens Campos
Machado | Evaluating software-based fingerprint liveness detection using
Convolutional Networks and Local Binary Patterns | arXiv admin note: text overlap with arXiv:1301.3557 by other authors | Biometric Measurements and Systems for Security and Medical
Applications (BIOMS) Proceedings, 2014 IEEE Workshop on | 10.1109/BIOMS.2014.6951531 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the growing use of biometric authentication systems in the past years,
spoof fingerprint detection has become increasingly important. In this work, we
implement and evaluate two different feature extraction techniques for
software-based fingerprint liveness detection: Convolutional Networks with
random weights and Local Binary Patterns. Both techniques were used in
conjunction with a Support Vector Machine (SVM) classifier. Dataset
Augmentation was used to increase classifier's performance and a variety of
preprocessing operations were tested, such as frequency filtering, contrast
equalization, and region of interest filtering. The experiments were made on
the datasets used in The Liveness Detection Competition of years 2009, 2011 and
2013, which comprise almost 50,000 real and fake fingerprints' images. Our best
method achieves an overall rate of 95.2% of correctly classified samples - an
improvement of 35% in test error when compared with the best previously
published results.
| [
{
"version": "v1",
"created": "Mon, 3 Aug 2015 19:21:03 GMT"
}
] | 2015-08-04T00:00:00 | [
[
"Nogueira",
"Rodrigo Frassetto",
""
],
[
"Lotufo",
"Roberto de Alencar",
""
],
[
"Machado",
"Rubens Campos",
""
]
] | TITLE: Evaluating software-based fingerprint liveness detection using
Convolutional Networks and Local Binary Patterns
ABSTRACT: With the growing use of biometric authentication systems in the past years,
spoof fingerprint detection has become increasingly important. In this work, we
implement and evaluate two different feature extraction techniques for
software-based fingerprint liveness detection: Convolutional Networks with
random weights and Local Binary Patterns. Both techniques were used in
conjunction with a Support Vector Machine (SVM) classifier. Dataset
Augmentation was used to increase classifier's performance and a variety of
preprocessing operations were tested, such as frequency filtering, contrast
equalization, and region of interest filtering. The experiments were made on
the datasets used in The Liveness Detection Competition of years 2009, 2011 and
2013, which comprise almost 50,000 real and fake fingerprints' images. Our best
method achieves an overall rate of 95.2% of correctly classified samples - an
improvement of 35% in test error when compared with the best previously
published results.
| no_new_dataset | 0.955194 |
1502.01097 | Gui-Song Xia | Jingwen Hu, Gui-Song Xia, Fan Hu, Liangpei Zhang | Dense v.s. Sparse: A Comparative Study of Sampling Analysis in Scene
Classification of High-Resolution Remote Sensing Imagery | This paper has been withdrawn by the author due to the submission
requirement of a journal | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Scene classification is a key problem in the interpretation of
high-resolution remote sensing imagery. Many state-of-the-art methods, e.g.
bag-of-visual-words model and its variants, the topic models as well as deep
learning-based approaches, share similar procedures: patch sampling, feature
description/learning and classification. Patch sampling is the first and a key
procedure which has a great influence on the results. In the literature, many
different sampling strategies have been used, {e.g. dense sampling, random
sampling, keypoint-based sampling and saliency-based sampling, etc. However, it
is still not clear which sampling strategy is suitable for the scene
classification of high-resolution remote sensing images. In this paper, we
comparatively study the effects of different sampling strategies under the
scenario of scene classification of high-resolution remote sensing images. We
divide the existing sampling methods into two types: dense sampling and sparse
sampling, the later of which includes random sampling, keypoint-based sampling
and various saliency-based sampling proposed recently. In order to compare
their performances, we rely on a standard bag-of-visual-words model to
construct our testing scheme, owing to their simplicity, robustness and
efficiency. The experimental results on two commonly used datasets show that
dense sampling has the best performance among all the strategies but with high
spatial and computational complexity, random sampling gives better or
comparable results than other sparse sampling methods, like the sophisticated
multi-scale key-point operators and the saliency-based methods which are
intensively studied and commonly used recently.
| [
{
"version": "v1",
"created": "Wed, 4 Feb 2015 05:34:31 GMT"
},
{
"version": "v2",
"created": "Fri, 31 Jul 2015 07:02:30 GMT"
}
] | 2015-08-03T00:00:00 | [
[
"Hu",
"Jingwen",
""
],
[
"Xia",
"Gui-Song",
""
],
[
"Hu",
"Fan",
""
],
[
"Zhang",
"Liangpei",
""
]
] | TITLE: Dense v.s. Sparse: A Comparative Study of Sampling Analysis in Scene
Classification of High-Resolution Remote Sensing Imagery
ABSTRACT: Scene classification is a key problem in the interpretation of
high-resolution remote sensing imagery. Many state-of-the-art methods, e.g.
bag-of-visual-words model and its variants, the topic models as well as deep
learning-based approaches, share similar procedures: patch sampling, feature
description/learning and classification. Patch sampling is the first and a key
procedure which has a great influence on the results. In the literature, many
different sampling strategies have been used, {e.g. dense sampling, random
sampling, keypoint-based sampling and saliency-based sampling, etc. However, it
is still not clear which sampling strategy is suitable for the scene
classification of high-resolution remote sensing images. In this paper, we
comparatively study the effects of different sampling strategies under the
scenario of scene classification of high-resolution remote sensing images. We
divide the existing sampling methods into two types: dense sampling and sparse
sampling, the later of which includes random sampling, keypoint-based sampling
and various saliency-based sampling proposed recently. In order to compare
their performances, we rely on a standard bag-of-visual-words model to
construct our testing scheme, owing to their simplicity, robustness and
efficiency. The experimental results on two commonly used datasets show that
dense sampling has the best performance among all the strategies but with high
spatial and computational complexity, random sampling gives better or
comparable results than other sparse sampling methods, like the sophisticated
multi-scale key-point operators and the saliency-based methods which are
intensively studied and commonly used recently.
| no_new_dataset | 0.953794 |
1507.08761 | Amir Shahroudy | Amir Shahroudy, Gang Wang, Tian-Tsong Ng, Qingxiong Yang | Multimodal Multipart Learning for Action Recognition in Depth Videos | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The articulated and complex nature of human actions makes the task of action
recognition difficult. One approach to handle this complexity is dividing it to
the kinetics of body parts and analyzing the actions based on these partial
descriptors. We propose a joint sparse regression based learning method which
utilizes the structured sparsity to model each action as a combination of
multimodal features from a sparse set of body parts. To represent dynamics and
appearance of parts, we employ a heterogeneous set of depth and skeleton based
features. The proper structure of multimodal multipart features are formulated
into the learning framework via the proposed hierarchical mixed norm, to
regularize the structured features of each part and to apply sparsity between
them, in favor of a group feature selection. Our experimental results expose
the effectiveness of the proposed learning method in which it outperforms other
methods in all three tested datasets while saturating one of them by achieving
perfect accuracy.
| [
{
"version": "v1",
"created": "Fri, 31 Jul 2015 06:02:56 GMT"
}
] | 2015-08-03T00:00:00 | [
[
"Shahroudy",
"Amir",
""
],
[
"Wang",
"Gang",
""
],
[
"Ng",
"Tian-Tsong",
""
],
[
"Yang",
"Qingxiong",
""
]
] | TITLE: Multimodal Multipart Learning for Action Recognition in Depth Videos
ABSTRACT: The articulated and complex nature of human actions makes the task of action
recognition difficult. One approach to handle this complexity is dividing it to
the kinetics of body parts and analyzing the actions based on these partial
descriptors. We propose a joint sparse regression based learning method which
utilizes the structured sparsity to model each action as a combination of
multimodal features from a sparse set of body parts. To represent dynamics and
appearance of parts, we employ a heterogeneous set of depth and skeleton based
features. The proper structure of multimodal multipart features are formulated
into the learning framework via the proposed hierarchical mixed norm, to
regularize the structured features of each part and to apply sparsity between
them, in favor of a group feature selection. Our experimental results expose
the effectiveness of the proposed learning method in which it outperforms other
methods in all three tested datasets while saturating one of them by achieving
perfect accuracy.
| no_new_dataset | 0.944125 |
1409.3821 | Andrea Montanari | Andrea Montanari | Computational Implications of Reducing Data to Sufficient Statistics | 20 pages | null | null | null | stat.CO cs.IT cs.LG math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Given a large dataset and an estimation task, it is common to pre-process the
data by reducing them to a set of sufficient statistics. This step is often
regarded as straightforward and advantageous (in that it simplifies statistical
analysis). I show that -on the contrary- reducing data to sufficient statistics
can change a computationally tractable estimation problem into an intractable
one. I discuss connections with recent work in theoretical computer science,
and implications for some techniques to estimate graphical models.
| [
{
"version": "v1",
"created": "Fri, 12 Sep 2014 18:57:01 GMT"
},
{
"version": "v2",
"created": "Mon, 15 Sep 2014 16:39:26 GMT"
},
{
"version": "v3",
"created": "Thu, 30 Jul 2015 19:35:44 GMT"
}
] | 2015-07-31T00:00:00 | [
[
"Montanari",
"Andrea",
""
]
] | TITLE: Computational Implications of Reducing Data to Sufficient Statistics
ABSTRACT: Given a large dataset and an estimation task, it is common to pre-process the
data by reducing them to a set of sufficient statistics. This step is often
regarded as straightforward and advantageous (in that it simplifies statistical
analysis). I show that -on the contrary- reducing data to sufficient statistics
can change a computationally tractable estimation problem into an intractable
one. I discuss connections with recent work in theoretical computer science,
and implications for some techniques to estimate graphical models.
| no_new_dataset | 0.951051 |
1506.00711 | Babak Saleh | Ahmed Elgammal and Babak Saleh | Quantifying Creativity in Art Networks | This paper will be published in the sixth International Conference on
Computational Creativity (ICCC) June 29-July 2nd 2015, Park City, Utah, USA.
This arXiv version is an extended version of the conference paper | null | null | null | cs.AI cs.CV cs.CY cs.MM cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Can we develop a computer algorithm that assesses the creativity of a
painting given its context within art history? This paper proposes a novel
computational framework for assessing the creativity of creative products, such
as paintings, sculptures, poetry, etc. We use the most common definition of
creativity, which emphasizes the originality of the product and its influential
value. The proposed computational framework is based on constructing a network
between creative products and using this network to infer about the originality
and influence of its nodes. Through a series of transformations, we construct a
Creativity Implication Network. We show that inference about creativity in this
network reduces to a variant of network centrality problems which can be solved
efficiently. We apply the proposed framework to the task of quantifying
creativity of paintings (and sculptures). We experimented on two datasets with
over 62K paintings to illustrate the behavior of the proposed framework. We
also propose a methodology for quantitatively validating the results of the
proposed algorithm, which we call the "time machine experiment".
| [
{
"version": "v1",
"created": "Tue, 2 Jun 2015 00:20:54 GMT"
}
] | 2015-07-31T00:00:00 | [
[
"Elgammal",
"Ahmed",
""
],
[
"Saleh",
"Babak",
""
]
] | TITLE: Quantifying Creativity in Art Networks
ABSTRACT: Can we develop a computer algorithm that assesses the creativity of a
painting given its context within art history? This paper proposes a novel
computational framework for assessing the creativity of creative products, such
as paintings, sculptures, poetry, etc. We use the most common definition of
creativity, which emphasizes the originality of the product and its influential
value. The proposed computational framework is based on constructing a network
between creative products and using this network to infer about the originality
and influence of its nodes. Through a series of transformations, we construct a
Creativity Implication Network. We show that inference about creativity in this
network reduces to a variant of network centrality problems which can be solved
efficiently. We apply the proposed framework to the task of quantifying
creativity of paintings (and sculptures). We experimented on two datasets with
over 62K paintings to illustrate the behavior of the proposed framework. We
also propose a methodology for quantitatively validating the results of the
proposed algorithm, which we call the "time machine experiment".
| no_new_dataset | 0.950503 |
1507.08286 | David Held | David Held, Sebastian Thrun, Silvio Savarese | Deep Learning for Single-View Instance Recognition | 16 pages, 15 figures | null | null | null | cs.CV cs.LG cs.NE cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep learning methods have typically been trained on large datasets in which
many training examples are available. However, many real-world product datasets
have only a small number of images available for each product. We explore the
use of deep learning methods for recognizing object instances when we have only
a single training example per class. We show that feedforward neural networks
outperform state-of-the-art methods for recognizing objects from novel
viewpoints even when trained from just a single image per object. To further
improve our performance on this task, we propose to take advantage of a
supplementary dataset in which we observe a separate set of objects from
multiple viewpoints. We introduce a new approach for training deep learning
methods for instance recognition with limited training data, in which we use an
auxiliary multi-view dataset to train our network to be robust to viewpoint
changes. We find that this approach leads to a more robust classifier for
recognizing objects from novel viewpoints, outperforming previous
state-of-the-art approaches including keypoint-matching, template-based
techniques, and sparse coding.
| [
{
"version": "v1",
"created": "Wed, 29 Jul 2015 20:11:12 GMT"
}
] | 2015-07-31T00:00:00 | [
[
"Held",
"David",
""
],
[
"Thrun",
"Sebastian",
""
],
[
"Savarese",
"Silvio",
""
]
] | TITLE: Deep Learning for Single-View Instance Recognition
ABSTRACT: Deep learning methods have typically been trained on large datasets in which
many training examples are available. However, many real-world product datasets
have only a small number of images available for each product. We explore the
use of deep learning methods for recognizing object instances when we have only
a single training example per class. We show that feedforward neural networks
outperform state-of-the-art methods for recognizing objects from novel
viewpoints even when trained from just a single image per object. To further
improve our performance on this task, we propose to take advantage of a
supplementary dataset in which we observe a separate set of objects from
multiple viewpoints. We introduce a new approach for training deep learning
methods for instance recognition with limited training data, in which we use an
auxiliary multi-view dataset to train our network to be robust to viewpoint
changes. We find that this approach leads to a more robust classifier for
recognizing objects from novel viewpoints, outperforming previous
state-of-the-art approaches including keypoint-matching, template-based
techniques, and sparse coding.
| no_new_dataset | 0.948202 |
1507.08363 | Massimo Piccardi | Shaukat Abidi, Massimo Piccardi, Mary-Anne Williams | Action recognition in still images by latent superpixel classification | To appear in the Proceedings of the IEEE International Conference on
Image Processing. Copyright 2015 IEEE. Please be aware of your obligations
with respect to copyrighted material | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Action recognition from still images is an important task of computer vision
applications such as image annotation, robotic navigation, video surveillance
and several others. Existing approaches mainly rely on either bag-of-feature
representations or articulated body-part models. However, the relationship
between the action and the image segments is still substantially unexplored.
For this reason, in this paper we propose to approach action recognition by
leveraging an intermediate layer of "superpixels" whose latent classes can act
as attributes of the action. In the proposed approach, the action class is
predicted by a structural model(learnt by Latent Structural SVM) based on
measurements from the image superpixels and their latent classes. Experimental
results over the challenging Stanford 40 Actions dataset report a significant
average accuracy of 74.06% for the positive class and 88.50% for the negative
class, giving evidence to the performance of the proposed approach.
| [
{
"version": "v1",
"created": "Thu, 30 Jul 2015 03:05:47 GMT"
}
] | 2015-07-31T00:00:00 | [
[
"Abidi",
"Shaukat",
""
],
[
"Piccardi",
"Massimo",
""
],
[
"Williams",
"Mary-Anne",
""
]
] | TITLE: Action recognition in still images by latent superpixel classification
ABSTRACT: Action recognition from still images is an important task of computer vision
applications such as image annotation, robotic navigation, video surveillance
and several others. Existing approaches mainly rely on either bag-of-feature
representations or articulated body-part models. However, the relationship
between the action and the image segments is still substantially unexplored.
For this reason, in this paper we propose to approach action recognition by
leveraging an intermediate layer of "superpixels" whose latent classes can act
as attributes of the action. In the proposed approach, the action class is
predicted by a structural model(learnt by Latent Structural SVM) based on
measurements from the image superpixels and their latent classes. Experimental
results over the challenging Stanford 40 Actions dataset report a significant
average accuracy of 74.06% for the positive class and 88.50% for the negative
class, giving evidence to the performance of the proposed approach.
| no_new_dataset | 0.951142 |
1507.08373 | Mehrtash Harandi | Mehrtash Harandi, Mathieu Salzmann, Fatih Porikli | When VLAD met Hilbert | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Vectors of Locally Aggregated Descriptors (VLAD) have emerged as powerful
image/video representations that compete with or even outperform
state-of-the-art approaches on many challenging visual recognition tasks. In
this paper, we address two fundamental limitations of VLAD: its requirement for
the local descriptors to have vector form and its restriction to linear
classifiers due to its high-dimensionality. To this end, we introduce a
kernelized version of VLAD. This not only lets us inherently exploit more
sophisticated classification schemes, but also enables us to efficiently
aggregate non-vector descriptors (e.g., tensors) in the VLAD framework.
Furthermore, we propose three approximate formulations that allow us to
accelerate the coding process while still benefiting from the properties of
kernel VLAD. Our experiments demonstrate the effectiveness of our approach at
handling manifold-valued data, such as covariance descriptors, on several
classification tasks. Our results also evidence the benefits of our nonlinear
VLAD descriptors against the linear ones in Euclidean space using several
standard benchmark datasets.
| [
{
"version": "v1",
"created": "Thu, 30 Jul 2015 04:17:02 GMT"
}
] | 2015-07-31T00:00:00 | [
[
"Harandi",
"Mehrtash",
""
],
[
"Salzmann",
"Mathieu",
""
],
[
"Porikli",
"Fatih",
""
]
] | TITLE: When VLAD met Hilbert
ABSTRACT: Vectors of Locally Aggregated Descriptors (VLAD) have emerged as powerful
image/video representations that compete with or even outperform
state-of-the-art approaches on many challenging visual recognition tasks. In
this paper, we address two fundamental limitations of VLAD: its requirement for
the local descriptors to have vector form and its restriction to linear
classifiers due to its high-dimensionality. To this end, we introduce a
kernelized version of VLAD. This not only lets us inherently exploit more
sophisticated classification schemes, but also enables us to efficiently
aggregate non-vector descriptors (e.g., tensors) in the VLAD framework.
Furthermore, we propose three approximate formulations that allow us to
accelerate the coding process while still benefiting from the properties of
kernel VLAD. Our experiments demonstrate the effectiveness of our approach at
handling manifold-valued data, such as covariance descriptors, on several
classification tasks. Our results also evidence the benefits of our nonlinear
VLAD descriptors against the linear ones in Euclidean space using several
standard benchmark datasets.
| no_new_dataset | 0.947672 |
1507.08429 | Shuchang Zhou | Shuchang Zhou, Yuxin Wu | Multilinear Map Layer: Prediction Regularization by Structural
Constraint | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper we propose and study a technique to impose structural
constraints on the output of a neural network, which can reduce amount of
computation and number of parameters besides improving prediction accuracy when
the output is known to approximately conform to the low-rankness prior. The
technique proceeds by replacing the output layer of neural network with the
so-called MLM layers, which forces the output to be the result of some
Multilinear Map, like a hybrid-Kronecker-dot product or Kronecker Tensor
Product. In particular, given an "autoencoder" model trained on SVHN dataset,
we can construct a new model with MLM layer achieving 62\% reduction in total
number of parameters and reduction of $\ell_2$ reconstruction error from 0.088
to 0.004. Further experiments on other autoencoder model variants trained on
SVHN datasets also demonstrate the efficacy of MLM layers.
| [
{
"version": "v1",
"created": "Thu, 30 Jul 2015 09:34:30 GMT"
}
] | 2015-07-31T00:00:00 | [
[
"Zhou",
"Shuchang",
""
],
[
"Wu",
"Yuxin",
""
]
] | TITLE: Multilinear Map Layer: Prediction Regularization by Structural
Constraint
ABSTRACT: In this paper we propose and study a technique to impose structural
constraints on the output of a neural network, which can reduce amount of
computation and number of parameters besides improving prediction accuracy when
the output is known to approximately conform to the low-rankness prior. The
technique proceeds by replacing the output layer of neural network with the
so-called MLM layers, which forces the output to be the result of some
Multilinear Map, like a hybrid-Kronecker-dot product or Kronecker Tensor
Product. In particular, given an "autoencoder" model trained on SVHN dataset,
we can construct a new model with MLM layer achieving 62\% reduction in total
number of parameters and reduction of $\ell_2$ reconstruction error from 0.088
to 0.004. Further experiments on other autoencoder model variants trained on
SVHN datasets also demonstrate the efficacy of MLM layers.
| no_new_dataset | 0.948346 |
1507.08445 | Ankan Bansal | Ankan Bansal and K.S. Venkatesh | People Counting in High Density Crowds from Still Images | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a method of estimating the number of people in high density crowds
from still images. The method estimates counts by fusing information from
multiple sources. Most of the existing work on crowd counting deals with very
small crowds (tens of individuals) and use temporal information from videos.
Our method uses only still images to estimate the counts in high density images
(hundreds to thousands of individuals). At this scale, we cannot rely on only
one set of features for count estimation. We, therefore, use multiple sources,
viz. interest points (SIFT), Fourier analysis, wavelet decomposition, GLCM
features and low confidence head detections, to estimate the counts. Each of
these sources gives a separate estimate of the count along with confidences and
other statistical measures which are then combined to obtain the final
estimate. We test our method on an existing dataset of fifty images containing
over 64000 individuals. Further, we added another fifty annotated images of
crowds and tested on the complete dataset of hundred images containing over
87000 individuals. The counts per image range from 81 to 4633. We report the
performance in terms of mean absolute error, which is a measure of accuracy of
the method, and mean normalised absolute error, which is a measure of the
robustness.
| [
{
"version": "v1",
"created": "Thu, 30 Jul 2015 10:47:31 GMT"
}
] | 2015-07-31T00:00:00 | [
[
"Bansal",
"Ankan",
""
],
[
"Venkatesh",
"K. S.",
""
]
] | TITLE: People Counting in High Density Crowds from Still Images
ABSTRACT: We present a method of estimating the number of people in high density crowds
from still images. The method estimates counts by fusing information from
multiple sources. Most of the existing work on crowd counting deals with very
small crowds (tens of individuals) and use temporal information from videos.
Our method uses only still images to estimate the counts in high density images
(hundreds to thousands of individuals). At this scale, we cannot rely on only
one set of features for count estimation. We, therefore, use multiple sources,
viz. interest points (SIFT), Fourier analysis, wavelet decomposition, GLCM
features and low confidence head detections, to estimate the counts. Each of
these sources gives a separate estimate of the count along with confidences and
other statistical measures which are then combined to obtain the final
estimate. We test our method on an existing dataset of fifty images containing
over 64000 individuals. Further, we added another fifty annotated images of
crowds and tested on the complete dataset of hundred images containing over
87000 individuals. The counts per image range from 81 to 4633. We report the
performance in terms of mean absolute error, which is a measure of accuracy of
the method, and mean normalised absolute error, which is a measure of the
robustness.
| new_dataset | 0.893216 |
1501.04158 | Niko S\"underhauf | Niko S\"underhauf, Feras Dayoub, Sareh Shirazi, Ben Upcroft, and
Michael Milford | On the Performance of ConvNet Features for Place Recognition | null | null | null | null | cs.RO cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | After the incredible success of deep learning in the computer vision domain,
there has been much interest in applying Convolutional Network (ConvNet)
features in robotic fields such as visual navigation and SLAM. Unfortunately,
there are fundamental differences and challenges involved. Computer vision
datasets are very different in character to robotic camera data, real-time
performance is essential, and performance priorities can be different. This
paper comprehensively evaluates and compares the utility of three
state-of-the-art ConvNets on the problems of particular relevance to navigation
for robots; viewpoint-invariance and condition-invariance, and for the first
time enables real-time place recognition performance using ConvNets with large
maps by integrating a variety of existing (locality-sensitive hashing) and
novel (semantic search space partitioning) optimization techniques. We present
extensive experiments on four real world datasets cultivated to evaluate each
of the specific challenges in place recognition. The results demonstrate that
speed-ups of two orders of magnitude can be achieved with minimal accuracy
degradation, enabling real-time performance. We confirm that networks trained
for semantic place categorization also perform better at (specific) place
recognition when faced with severe appearance changes and provide a reference
for which networks and layers are optimal for different aspects of the place
recognition problem.
| [
{
"version": "v1",
"created": "Sat, 17 Jan 2015 05:16:12 GMT"
},
{
"version": "v2",
"created": "Tue, 7 Jul 2015 11:35:10 GMT"
},
{
"version": "v3",
"created": "Wed, 29 Jul 2015 01:56:54 GMT"
}
] | 2015-07-30T00:00:00 | [
[
"Sünderhauf",
"Niko",
""
],
[
"Dayoub",
"Feras",
""
],
[
"Shirazi",
"Sareh",
""
],
[
"Upcroft",
"Ben",
""
],
[
"Milford",
"Michael",
""
]
] | TITLE: On the Performance of ConvNet Features for Place Recognition
ABSTRACT: After the incredible success of deep learning in the computer vision domain,
there has been much interest in applying Convolutional Network (ConvNet)
features in robotic fields such as visual navigation and SLAM. Unfortunately,
there are fundamental differences and challenges involved. Computer vision
datasets are very different in character to robotic camera data, real-time
performance is essential, and performance priorities can be different. This
paper comprehensively evaluates and compares the utility of three
state-of-the-art ConvNets on the problems of particular relevance to navigation
for robots; viewpoint-invariance and condition-invariance, and for the first
time enables real-time place recognition performance using ConvNets with large
maps by integrating a variety of existing (locality-sensitive hashing) and
novel (semantic search space partitioning) optimization techniques. We present
extensive experiments on four real world datasets cultivated to evaluate each
of the specific challenges in place recognition. The results demonstrate that
speed-ups of two orders of magnitude can be achieved with minimal accuracy
degradation, enabling real-time performance. We confirm that networks trained
for semantic place categorization also perform better at (specific) place
recognition when faced with severe appearance changes and provide a reference
for which networks and layers are optimal for different aspects of the place
recognition problem.
| no_new_dataset | 0.947672 |
1502.06108 | Xiao Lin | Xiao Lin, Devi Parikh | Don't Just Listen, Use Your Imagination: Leveraging Visual Common Sense
for Non-Visual Tasks | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Artificial agents today can answer factual questions. But they fall short on
questions that require common sense reasoning. Perhaps this is because most
existing common sense databases rely on text to learn and represent knowledge.
But much of common sense knowledge is unwritten - partly because it tends not
to be interesting enough to talk about, and partly because some common sense is
unnatural to articulate in text. While unwritten, it is not unseen. In this
paper we leverage semantic common sense knowledge learned from images - i.e.
visual common sense - in two textual tasks: fill-in-the-blank and visual
paraphrasing. We propose to "imagine" the scene behind the text, and leverage
visual cues from the "imagined" scenes in addition to textual cues while
answering these questions. We imagine the scenes as a visual abstraction. Our
approach outperforms a strong text-only baseline on these tasks. Our proposed
tasks can serve as benchmarks to quantitatively evaluate progress in solving
tasks that go "beyond recognition". Our code and datasets are publicly
available.
| [
{
"version": "v1",
"created": "Sat, 21 Feb 2015 15:25:40 GMT"
},
{
"version": "v2",
"created": "Tue, 5 May 2015 18:54:05 GMT"
},
{
"version": "v3",
"created": "Wed, 29 Jul 2015 03:04:19 GMT"
}
] | 2015-07-30T00:00:00 | [
[
"Lin",
"Xiao",
""
],
[
"Parikh",
"Devi",
""
]
] | TITLE: Don't Just Listen, Use Your Imagination: Leveraging Visual Common Sense
for Non-Visual Tasks
ABSTRACT: Artificial agents today can answer factual questions. But they fall short on
questions that require common sense reasoning. Perhaps this is because most
existing common sense databases rely on text to learn and represent knowledge.
But much of common sense knowledge is unwritten - partly because it tends not
to be interesting enough to talk about, and partly because some common sense is
unnatural to articulate in text. While unwritten, it is not unseen. In this
paper we leverage semantic common sense knowledge learned from images - i.e.
visual common sense - in two textual tasks: fill-in-the-blank and visual
paraphrasing. We propose to "imagine" the scene behind the text, and leverage
visual cues from the "imagined" scenes in addition to textual cues while
answering these questions. We imagine the scenes as a visual abstraction. Our
approach outperforms a strong text-only baseline on these tasks. Our proposed
tasks can serve as benchmarks to quantitatively evaluate progress in solving
tasks that go "beyond recognition". Our code and datasets are publicly
available.
| no_new_dataset | 0.943034 |
1506.04723 | Ming-Yu Liu | Ming-Yu Liu, Shuoxin Lin, Srikumar Ramalingam, Oncel Tuzel | Layered Interpretation of Street View Images | The paper will be presented in the 2015 Robotics: Science and Systems
Conference (RSS) | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a layered street view model to encode both depth and semantic
information on street view images for autonomous driving. Recently, stixels,
stix-mantics, and tiered scene labeling methods have been proposed to model
street view images. We propose a 4-layer street view model, a compact
representation over the recently proposed stix-mantics model. Our layers encode
semantic classes like ground, pedestrians, vehicles, buildings, and sky in
addition to the depths. The only input to our algorithm is a pair of stereo
images. We use a deep neural network to extract the appearance features for
semantic classes. We use a simple and an efficient inference algorithm to
jointly estimate both semantic classes and layered depth values. Our method
outperforms other competing approaches in Daimler urban scene segmentation
dataset. Our algorithm is massively parallelizable, allowing a GPU
implementation with a processing speed about 9 fps.
| [
{
"version": "v1",
"created": "Mon, 15 Jun 2015 19:38:59 GMT"
},
{
"version": "v2",
"created": "Wed, 29 Jul 2015 15:38:28 GMT"
}
] | 2015-07-30T00:00:00 | [
[
"Liu",
"Ming-Yu",
""
],
[
"Lin",
"Shuoxin",
""
],
[
"Ramalingam",
"Srikumar",
""
],
[
"Tuzel",
"Oncel",
""
]
] | TITLE: Layered Interpretation of Street View Images
ABSTRACT: We propose a layered street view model to encode both depth and semantic
information on street view images for autonomous driving. Recently, stixels,
stix-mantics, and tiered scene labeling methods have been proposed to model
street view images. We propose a 4-layer street view model, a compact
representation over the recently proposed stix-mantics model. Our layers encode
semantic classes like ground, pedestrians, vehicles, buildings, and sky in
addition to the depths. The only input to our algorithm is a pair of stereo
images. We use a deep neural network to extract the appearance features for
semantic classes. We use a simple and an efficient inference algorithm to
jointly estimate both semantic classes and layered depth values. Our method
outperforms other competing approaches in Daimler urban scene segmentation
dataset. Our algorithm is massively parallelizable, allowing a GPU
implementation with a processing speed about 9 fps.
| no_new_dataset | 0.949716 |
1507.07242 | Dayong Wang | Dayong Wang and Charles Otto and Anil K. Jain | Face Search at Scale: 80 Million Gallery | 14 pages, 16 figures | null | null | MSU TECHNICAL REPORT MSU-CSE-15-11, JULY 24, 2015 | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Due to the prevalence of social media websites, one challenge facing computer
vision researchers is to devise methods to process and search for persons of
interest among the billions of shared photos on these websites. Facebook
revealed in a 2013 white paper that its users have uploaded more than 250
billion photos, and are uploading 350 million new photos each day. Due to this
humongous amount of data, large-scale face search for mining web images is both
important and challenging. Despite significant progress in face recognition,
searching a large collection of unconstrained face images has not been
adequately addressed. To address this challenge, we propose a face search
system which combines a fast search procedure, coupled with a state-of-the-art
commercial off the shelf (COTS) matcher, in a cascaded framework. Given a probe
face, we first filter the large gallery of photos to find the top-k most
similar faces using deep features generated from a convolutional neural
network. The k candidates are re-ranked by combining similarities from deep
features and the COTS matcher. We evaluate the proposed face search system on a
gallery containing 80 million web-downloaded face images. Experimental results
demonstrate that the deep features are competitive with state-of-the-art
methods on unconstrained face recognition benchmarks (LFW and IJB-A). Further,
the proposed face search system offers an excellent trade-off between accuracy
and scalability on datasets consisting of millions of images. Additionally, in
an experiment involving searching for face images of the Tsarnaev brothers,
convicted of the Boston Marathon bombing, the proposed face search system could
find the younger brother's (Dzhokhar Tsarnaev) photo at rank 1 in 1 second on a
5M gallery and at rank 8 in 7 seconds on an 80M gallery.
| [
{
"version": "v1",
"created": "Sun, 26 Jul 2015 20:06:43 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Jul 2015 22:09:17 GMT"
}
] | 2015-07-30T00:00:00 | [
[
"Wang",
"Dayong",
""
],
[
"Otto",
"Charles",
""
],
[
"Jain",
"Anil K.",
""
]
] | TITLE: Face Search at Scale: 80 Million Gallery
ABSTRACT: Due to the prevalence of social media websites, one challenge facing computer
vision researchers is to devise methods to process and search for persons of
interest among the billions of shared photos on these websites. Facebook
revealed in a 2013 white paper that its users have uploaded more than 250
billion photos, and are uploading 350 million new photos each day. Due to this
humongous amount of data, large-scale face search for mining web images is both
important and challenging. Despite significant progress in face recognition,
searching a large collection of unconstrained face images has not been
adequately addressed. To address this challenge, we propose a face search
system which combines a fast search procedure, coupled with a state-of-the-art
commercial off the shelf (COTS) matcher, in a cascaded framework. Given a probe
face, we first filter the large gallery of photos to find the top-k most
similar faces using deep features generated from a convolutional neural
network. The k candidates are re-ranked by combining similarities from deep
features and the COTS matcher. We evaluate the proposed face search system on a
gallery containing 80 million web-downloaded face images. Experimental results
demonstrate that the deep features are competitive with state-of-the-art
methods on unconstrained face recognition benchmarks (LFW and IJB-A). Further,
the proposed face search system offers an excellent trade-off between accuracy
and scalability on datasets consisting of millions of images. Additionally, in
an experiment involving searching for face images of the Tsarnaev brothers,
convicted of the Boston Marathon bombing, the proposed face search system could
find the younger brother's (Dzhokhar Tsarnaev) photo at rank 1 in 1 second on a
5M gallery and at rank 8 in 7 seconds on an 80M gallery.
| no_new_dataset | 0.93835 |
1507.08030 | Anthony Cazanoves Mr | Anthony Cazasnoves and Fanny Buyens and Sylvie Sevestre | Adapted sampling for 3D X-ray computed tomography | The 13th International Meeting on Fully Three-Dimensional Image
Reconstruction in Radiology and Nuclear Medicine | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we introduce a method to build an adapted mesh representation
of a 3D object for X-Ray tomography reconstruction. Using this representation,
we provide means to reduce the computational cost of reconstruction by way of
iterative algorithms. The adapted sampling of the reconstruction space is
directly obtained from the projection dataset and prior to any reconstruction.
It is built following two stages : firstly, 2D structural information is
extracted from the projection images and is secondly merged in 3D to obtain a
3D pointcloud sampling the interfaces of the object. A relevant mesh is then
built from this cloud by way of tetrahedralization. Critical parameters
selections have been automatized through a statistical framework, thus avoiding
dependence on users expertise. Applying this approach on geometrical shapes and
on a 3D Shepp-Logan phantom, we show the relevance of such a sampling -
obtained in a few seconds - and the drastic decrease in cells number to be
estimated during reconstruction when compared to the usual regular voxel
lattice. A first iterative reconstruction of the Shepp-Logan using this kind of
sampling shows the relevant advantages in terms of low dose or sparse
acquisition sampling contexts. The method can also prove useful for other
applications such as finite element method computations.
| [
{
"version": "v1",
"created": "Wed, 29 Jul 2015 06:30:04 GMT"
}
] | 2015-07-30T00:00:00 | [
[
"Cazasnoves",
"Anthony",
""
],
[
"Buyens",
"Fanny",
""
],
[
"Sevestre",
"Sylvie",
""
]
] | TITLE: Adapted sampling for 3D X-ray computed tomography
ABSTRACT: In this paper, we introduce a method to build an adapted mesh representation
of a 3D object for X-Ray tomography reconstruction. Using this representation,
we provide means to reduce the computational cost of reconstruction by way of
iterative algorithms. The adapted sampling of the reconstruction space is
directly obtained from the projection dataset and prior to any reconstruction.
It is built following two stages : firstly, 2D structural information is
extracted from the projection images and is secondly merged in 3D to obtain a
3D pointcloud sampling the interfaces of the object. A relevant mesh is then
built from this cloud by way of tetrahedralization. Critical parameters
selections have been automatized through a statistical framework, thus avoiding
dependence on users expertise. Applying this approach on geometrical shapes and
on a 3D Shepp-Logan phantom, we show the relevance of such a sampling -
obtained in a few seconds - and the drastic decrease in cells number to be
estimated during reconstruction when compared to the usual regular voxel
lattice. A first iterative reconstruction of the Shepp-Logan using this kind of
sampling shows the relevant advantages in terms of low dose or sparse
acquisition sampling contexts. The method can also prove useful for other
applications such as finite element method computations.
| no_new_dataset | 0.945601 |
1507.08074 | Sergey Novoselov | Sergey Novoselov, Alexandr Kozlov, Galina Lavrentyeva, Konstantin
Simonchik, Vadim Shchemelinin | STC Anti-spoofing Systems for the ASVspoof 2015 Challenge | 5 pages, 8 figures, 3 tables | null | null | null | cs.SD cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents the Speech Technology Center (STC) systems submitted to
Automatic Speaker Verification Spoofing and Countermeasures (ASVspoof)
Challenge 2015. In this work we investigate different acoustic feature spaces
to determine reliable and robust countermeasures against spoofing attacks. In
addition to the commonly used front-end MFCC features we explored features
derived from phase spectrum and features based on applying the multiresolution
wavelet transform. Similar to state-of-the-art ASV systems, we used the
standard TV-JFA approach for probability modelling in spoofing detection
systems. Experiments performed on the development and evaluation datasets of
the Challenge demonstrate that the use of phase-related and wavelet-based
features provides a substantial input into the efficiency of the resulting STC
systems. In our research we also focused on the comparison of the linear (SVM)
and nonlinear (DBN) classifiers.
| [
{
"version": "v1",
"created": "Wed, 29 Jul 2015 09:22:58 GMT"
}
] | 2015-07-30T00:00:00 | [
[
"Novoselov",
"Sergey",
""
],
[
"Kozlov",
"Alexandr",
""
],
[
"Lavrentyeva",
"Galina",
""
],
[
"Simonchik",
"Konstantin",
""
],
[
"Shchemelinin",
"Vadim",
""
]
] | TITLE: STC Anti-spoofing Systems for the ASVspoof 2015 Challenge
ABSTRACT: This paper presents the Speech Technology Center (STC) systems submitted to
Automatic Speaker Verification Spoofing and Countermeasures (ASVspoof)
Challenge 2015. In this work we investigate different acoustic feature spaces
to determine reliable and robust countermeasures against spoofing attacks. In
addition to the commonly used front-end MFCC features we explored features
derived from phase spectrum and features based on applying the multiresolution
wavelet transform. Similar to state-of-the-art ASV systems, we used the
standard TV-JFA approach for probability modelling in spoofing detection
systems. Experiments performed on the development and evaluation datasets of
the Challenge demonstrate that the use of phase-related and wavelet-based
features provides a substantial input into the efficiency of the resulting STC
systems. In our research we also focused on the comparison of the linear (SVM)
and nonlinear (DBN) classifiers.
| no_new_dataset | 0.944587 |
1507.08104 | Brian McWilliams | Barbora Micenkov\'a, Brian McWilliams, Ira Assent | Learning Representations for Outlier Detection on a Budget | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The problem of detecting a small number of outliers in a large dataset is an
important task in many fields from fraud detection to high-energy physics. Two
approaches have emerged to tackle this problem: unsupervised and supervised.
Supervised approaches require a sufficient amount of labeled data and are
challenged by novel types of outliers and inherent class imbalance, whereas
unsupervised methods do not take advantage of available labeled training
examples and often exhibit poorer predictive performance. We propose BORE (a
Bagged Outlier Representation Ensemble) which uses unsupervised outlier scoring
functions (OSFs) as features in a supervised learning framework. BORE is able
to adapt to arbitrary OSF feature representations, to the imbalance in labeled
data as well as to prediction-time constraints on computational cost. We
demonstrate the good performance of BORE compared to a variety of competing
methods in the non-budgeted and the budgeted outlier detection problem on 12
real-world datasets.
| [
{
"version": "v1",
"created": "Wed, 29 Jul 2015 11:28:41 GMT"
}
] | 2015-07-30T00:00:00 | [
[
"Micenková",
"Barbora",
""
],
[
"McWilliams",
"Brian",
""
],
[
"Assent",
"Ira",
""
]
] | TITLE: Learning Representations for Outlier Detection on a Budget
ABSTRACT: The problem of detecting a small number of outliers in a large dataset is an
important task in many fields from fraud detection to high-energy physics. Two
approaches have emerged to tackle this problem: unsupervised and supervised.
Supervised approaches require a sufficient amount of labeled data and are
challenged by novel types of outliers and inherent class imbalance, whereas
unsupervised methods do not take advantage of available labeled training
examples and often exhibit poorer predictive performance. We propose BORE (a
Bagged Outlier Representation Ensemble) which uses unsupervised outlier scoring
functions (OSFs) as features in a supervised learning framework. BORE is able
to adapt to arbitrary OSF feature representations, to the imbalance in labeled
data as well as to prediction-time constraints on computational cost. We
demonstrate the good performance of BORE compared to a variety of competing
methods in the non-budgeted and the budgeted outlier detection problem on 12
real-world datasets.
| no_new_dataset | 0.947332 |
1507.08155 | Teng Qiu | Teng Qiu, Yongjie Li | IT-Dendrogram: A New Member of the In-Tree (IT) Clustering Family | 13 pages, 6 figures. IT-Dendrogram: An Effective Method to Visualize
the In-Tree structure by Dendrogram | null | null | null | stat.ML cs.CV cs.LG stat.ME | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Previously, we proposed a physically-inspired method to construct data points
into an effective in-tree (IT) structure, in which the underlying cluster
structure in the dataset is well revealed. Although there are some edges in the
IT structure requiring to be removed, such undesired edges are generally
distinguishable from other edges and thus are easy to be determined. For
instance, when the IT structures for the 2-dimensional (2D) datasets are
graphically presented, those undesired edges can be easily spotted and
interactively determined. However, in practice, there are many datasets that do
not lie in the 2D Euclidean space, thus their IT structures cannot be
graphically presented. But if we can effectively map those IT structures into a
visualized space in which the salient features of those undesired edges are
preserved, then the undesired edges in the IT structures can still be visually
determined in a visualization environment. Previously, this purpose was reached
by our method called IT-map. The outstanding advantage of IT-map is that
clusters can still be found even with the so-called crowding problem in the
embedding.
In this paper, we propose another method, called IT-Dendrogram, to achieve
the same goal through an effective combination of the IT structure and the
single link hierarchical clustering (SLHC) method. Like IT-map, IT-Dendrogram
can also effectively represent the IT structures in a visualization
environment, whereas using another form, called the Dendrogram. IT-Dendrogram
can serve as another visualization method to determine the undesired edges in
the IT structures and thus benefit the IT-based clustering analysis. This was
demonstrated on several datasets with different shapes, dimensions, and
attributes. Unlike IT-map, IT-Dendrogram can always avoid the crowding problem,
which could help users make more reliable cluster analysis in certain problems.
| [
{
"version": "v1",
"created": "Wed, 29 Jul 2015 14:22:13 GMT"
}
] | 2015-07-30T00:00:00 | [
[
"Qiu",
"Teng",
""
],
[
"Li",
"Yongjie",
""
]
] | TITLE: IT-Dendrogram: A New Member of the In-Tree (IT) Clustering Family
ABSTRACT: Previously, we proposed a physically-inspired method to construct data points
into an effective in-tree (IT) structure, in which the underlying cluster
structure in the dataset is well revealed. Although there are some edges in the
IT structure requiring to be removed, such undesired edges are generally
distinguishable from other edges and thus are easy to be determined. For
instance, when the IT structures for the 2-dimensional (2D) datasets are
graphically presented, those undesired edges can be easily spotted and
interactively determined. However, in practice, there are many datasets that do
not lie in the 2D Euclidean space, thus their IT structures cannot be
graphically presented. But if we can effectively map those IT structures into a
visualized space in which the salient features of those undesired edges are
preserved, then the undesired edges in the IT structures can still be visually
determined in a visualization environment. Previously, this purpose was reached
by our method called IT-map. The outstanding advantage of IT-map is that
clusters can still be found even with the so-called crowding problem in the
embedding.
In this paper, we propose another method, called IT-Dendrogram, to achieve
the same goal through an effective combination of the IT structure and the
single link hierarchical clustering (SLHC) method. Like IT-map, IT-Dendrogram
can also effectively represent the IT structures in a visualization
environment, whereas using another form, called the Dendrogram. IT-Dendrogram
can serve as another visualization method to determine the undesired edges in
the IT structures and thus benefit the IT-based clustering analysis. This was
demonstrated on several datasets with different shapes, dimensions, and
attributes. Unlike IT-map, IT-Dendrogram can always avoid the crowding problem,
which could help users make more reliable cluster analysis in certain problems.
| no_new_dataset | 0.946001 |
1507.07508 | Peng Sun | Peng Sun, Haoyin Zhou, Devon Lundine, James K. Min, Guanglei Xiong | Fast Segmentation of Left Ventricle in CT Images by Explicit Shape
Regression using Random Pixel Difference Features | 8 pages, link to a video demo | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recently, machine learning has been successfully applied to model-based left
ventricle (LV) segmentation. The general framework involves two stages, which
starts with LV localization and is followed by boundary delineation. Both are
driven by supervised learning techniques. When compared to previous
non-learning-based methods, several advantages have been shown, including full
automation and improved accuracy. However, the speed is still slow, in the
order of several seconds, for applications involving a large number of cases or
case loads requiring real-time performance. In this paper, we propose a fast LV
segmentation algorithm by joint localization and boundary delineation via
training explicit shape regressor with random pixel difference features. Tested
on 3D cardiac computed tomography (CT) image volumes, the average running time
of the proposed algorithm is 1.2 milliseconds per case. On a dataset consisting
of 139 CT volumes, a 5-fold cross validation shows the segmentation error is
$1.21 \pm 0.11$ for LV endocardium and $1.23 \pm 0.11$ millimeters for
epicardium. Compared with previous work, the proposed method is more stable
(lower standard deviation) without significant compromise to the accuracy.
| [
{
"version": "v1",
"created": "Mon, 27 Jul 2015 18:17:55 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Jul 2015 14:07:05 GMT"
}
] | 2015-07-29T00:00:00 | [
[
"Sun",
"Peng",
""
],
[
"Zhou",
"Haoyin",
""
],
[
"Lundine",
"Devon",
""
],
[
"Min",
"James K.",
""
],
[
"Xiong",
"Guanglei",
""
]
] | TITLE: Fast Segmentation of Left Ventricle in CT Images by Explicit Shape
Regression using Random Pixel Difference Features
ABSTRACT: Recently, machine learning has been successfully applied to model-based left
ventricle (LV) segmentation. The general framework involves two stages, which
starts with LV localization and is followed by boundary delineation. Both are
driven by supervised learning techniques. When compared to previous
non-learning-based methods, several advantages have been shown, including full
automation and improved accuracy. However, the speed is still slow, in the
order of several seconds, for applications involving a large number of cases or
case loads requiring real-time performance. In this paper, we propose a fast LV
segmentation algorithm by joint localization and boundary delineation via
training explicit shape regressor with random pixel difference features. Tested
on 3D cardiac computed tomography (CT) image volumes, the average running time
of the proposed algorithm is 1.2 milliseconds per case. On a dataset consisting
of 139 CT volumes, a 5-fold cross validation shows the segmentation error is
$1.21 \pm 0.11$ for LV endocardium and $1.23 \pm 0.11$ millimeters for
epicardium. Compared with previous work, the proposed method is more stable
(lower standard deviation) without significant compromise to the accuracy.
| no_new_dataset | 0.946498 |
1507.07760 | Konrad Simon | Konrad Simon, Sameer Sheorey, David Jacobs and Ronen Basri | A Hyperelastic Two-Scale Optimization Model for Shape Matching | null | null | null | null | cs.CG cs.CV cs.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We suggest a novel shape matching algorithm for three-dimensional surface
meshes of disk or sphere topology. The method is based on the physical theory
of nonlinear elasticity and can hence handle large rotations and deformations.
Deformation boundary conditions that supplement the underlying equations are
usually unknown. Given an initial guess, these are optimized such that the
mechanical boundary forces that are responsible for the deformation are of a
simple nature. We show a heuristic way to approximate the nonlinear
optimization problem by a sequence of convex problems using finite elements.
The deformation cost, i.e, the forces, is measured on a coarse scale while
ICP-like matching is done on the fine scale. We demonstrate the plausibility of
our algorithm on examples taken from different datasets.
| [
{
"version": "v1",
"created": "Tue, 28 Jul 2015 13:27:51 GMT"
}
] | 2015-07-29T00:00:00 | [
[
"Simon",
"Konrad",
""
],
[
"Sheorey",
"Sameer",
""
],
[
"Jacobs",
"David",
""
],
[
"Basri",
"Ronen",
""
]
] | TITLE: A Hyperelastic Two-Scale Optimization Model for Shape Matching
ABSTRACT: We suggest a novel shape matching algorithm for three-dimensional surface
meshes of disk or sphere topology. The method is based on the physical theory
of nonlinear elasticity and can hence handle large rotations and deformations.
Deformation boundary conditions that supplement the underlying equations are
usually unknown. Given an initial guess, these are optimized such that the
mechanical boundary forces that are responsible for the deformation are of a
simple nature. We show a heuristic way to approximate the nonlinear
optimization problem by a sequence of convex problems using finite elements.
The deformation cost, i.e, the forces, is measured on a coarse scale while
ICP-like matching is done on the fine scale. We demonstrate the plausibility of
our algorithm on examples taken from different datasets.
| no_new_dataset | 0.951278 |
1507.07815 | Svebor Karaman | Giuseppe Lisanti and Svebor Karaman and Daniele Pezzatini and Alberto
Del Bimbo | A Multi-Camera Image Processing and Visualization System for Train
Safety Assessment | 11 pages | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper we present a machine vision system to efficiently monitor,
analyze and present visual data acquired with a railway overhead gantry
equipped with multiple cameras. This solution aims to improve the safety of
daily life railway transportation in a two- fold manner: (1) by providing
automatic algorithms that can process large imagery of trains (2) by helping
train operators to keep attention on any possible malfunction. The system is
designed with the latest cutting edge, high-rate visible and thermal cameras
that ob- serve a train passing under an railway overhead gantry. The machine
vision system is composed of three principal modules: (1) an automatic wagon
identification system, recognizing the wagon ID according to the UIC
classification of railway coaches; (2) a temperature monitoring system; (3) a
system for the detection, localization and visualization of the pantograph of
the train. These three machine vision modules process batch trains sequences
and their resulting analysis are presented to an operator using a multitouch
user interface. We detail all technical aspects of our multi-camera portal: the
hardware requirements, the software developed to deal with the high-frame rate
cameras and ensure reliable acquisition, the algorithms proposed to solve each
computer vision task, and the multitouch interaction and visualization
interface. We evaluate each component of our system on a dataset recorded in an
ad-hoc railway test-bed, showing the potential of our proposed portal for train
safety assessment.
| [
{
"version": "v1",
"created": "Tue, 28 Jul 2015 15:36:24 GMT"
}
] | 2015-07-29T00:00:00 | [
[
"Lisanti",
"Giuseppe",
""
],
[
"Karaman",
"Svebor",
""
],
[
"Pezzatini",
"Daniele",
""
],
[
"Del Bimbo",
"Alberto",
""
]
] | TITLE: A Multi-Camera Image Processing and Visualization System for Train
Safety Assessment
ABSTRACT: In this paper we present a machine vision system to efficiently monitor,
analyze and present visual data acquired with a railway overhead gantry
equipped with multiple cameras. This solution aims to improve the safety of
daily life railway transportation in a two- fold manner: (1) by providing
automatic algorithms that can process large imagery of trains (2) by helping
train operators to keep attention on any possible malfunction. The system is
designed with the latest cutting edge, high-rate visible and thermal cameras
that ob- serve a train passing under an railway overhead gantry. The machine
vision system is composed of three principal modules: (1) an automatic wagon
identification system, recognizing the wagon ID according to the UIC
classification of railway coaches; (2) a temperature monitoring system; (3) a
system for the detection, localization and visualization of the pantograph of
the train. These three machine vision modules process batch trains sequences
and their resulting analysis are presented to an operator using a multitouch
user interface. We detail all technical aspects of our multi-camera portal: the
hardware requirements, the software developed to deal with the high-frame rate
cameras and ensure reliable acquisition, the algorithms proposed to solve each
computer vision task, and the multitouch interaction and visualization
interface. We evaluate each component of our system on a dataset recorded in an
ad-hoc railway test-bed, showing the potential of our proposed portal for train
safety assessment.
| no_new_dataset | 0.944995 |
1507.07882 | Samarth Manoj Brahmbhatt | Samarth Brahmbhatt, Heni Ben Amor and Henrik Christensen | Occlusion-Aware Object Localization, Segmentation and Pose Estimation | British Machine Vision Conference 2015 (poster) | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a learning approach for localization and segmentation of objects
in an image in a manner that is robust to partial occlusion. Our algorithm
produces a bounding box around the full extent of the object and labels pixels
in the interior that belong to the object. Like existing segmentation aware
detection approaches, we learn an appearance model of the object and consider
regions that do not fit this model as potential occlusions. However, in
addition to the established use of pairwise potentials for encouraging local
consistency, we use higher order potentials which capture information at the
level of im- age segments. We also propose an efficient loss function that
targets both localization and segmentation performance. Our algorithm achieves
13.52% segmentation error and 0.81 area under the false-positive per image vs.
recall curve on average over the challenging CMU Kitchen Occlusion Dataset.
This is a 42.44% decrease in segmentation error and a 16.13% increase in
localization performance compared to the state-of-the-art. Finally, we show
that the visibility labelling produced by our algorithm can make full 3D pose
estimation from a single image robust to occlusion.
| [
{
"version": "v1",
"created": "Mon, 27 Jul 2015 18:16:35 GMT"
}
] | 2015-07-29T00:00:00 | [
[
"Brahmbhatt",
"Samarth",
""
],
[
"Amor",
"Heni Ben",
""
],
[
"Christensen",
"Henrik",
""
]
] | TITLE: Occlusion-Aware Object Localization, Segmentation and Pose Estimation
ABSTRACT: We present a learning approach for localization and segmentation of objects
in an image in a manner that is robust to partial occlusion. Our algorithm
produces a bounding box around the full extent of the object and labels pixels
in the interior that belong to the object. Like existing segmentation aware
detection approaches, we learn an appearance model of the object and consider
regions that do not fit this model as potential occlusions. However, in
addition to the established use of pairwise potentials for encouraging local
consistency, we use higher order potentials which capture information at the
level of im- age segments. We also propose an efficient loss function that
targets both localization and segmentation performance. Our algorithm achieves
13.52% segmentation error and 0.81 area under the false-positive per image vs.
recall curve on average over the challenging CMU Kitchen Occlusion Dataset.
This is a 42.44% decrease in segmentation error and a 16.13% increase in
localization performance compared to the state-of-the-art. Finally, we show
that the visibility labelling produced by our algorithm can make full 3D pose
estimation from a single image robust to occlusion.
| no_new_dataset | 0.946001 |
1507.07908 | Luiz Capretz Dr. | Arif Raza, Luiz Fernando Capretz | Addressing User Requirements in Opens Source Software: The Role of
Online Forums | null | Journal of Computing Science and Engineering, 8(1):57-63, 2014 | null | null | cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | User satisfaction has always been important in the success of software,
regardless of whether it is closed and proprietary or open source software
(OSS). OSS users are geographically distributed and include technical as well
as novice users. However, it is generally believed that if OSS was more usable,
its popularity would increase tremendously. Hence, users and their requirements
need to be addressed in the priorities of an OSS environment. Online public
forums are a major medium of communication for the OSS community. The research
model of this work studies the relationship between user requirements in open
source software and online public forums. To conduct this research, we used a
dataset consisting of 100 open source software projects in different
categories. The results show that online forums play a significant role in
identifying user requirements and addressing their requests in open source
software.
| [
{
"version": "v1",
"created": "Fri, 24 Jul 2015 15:48:05 GMT"
}
] | 2015-07-29T00:00:00 | [
[
"Raza",
"Arif",
""
],
[
"Capretz",
"Luiz Fernando",
""
]
] | TITLE: Addressing User Requirements in Opens Source Software: The Role of
Online Forums
ABSTRACT: User satisfaction has always been important in the success of software,
regardless of whether it is closed and proprietary or open source software
(OSS). OSS users are geographically distributed and include technical as well
as novice users. However, it is generally believed that if OSS was more usable,
its popularity would increase tremendously. Hence, users and their requirements
need to be addressed in the priorities of an OSS environment. Online public
forums are a major medium of communication for the OSS community. The research
model of this work studies the relationship between user requirements in open
source software and online public forums. To conduct this research, we used a
dataset consisting of 100 open source software projects in different
categories. The results show that online forums play a significant role in
identifying user requirements and addressing their requests in open source
software.
| no_new_dataset | 0.868437 |
1501.06202 | Yanwei Fu | Yanwei Fu, Timothy M. Hospedales, Tao Xiang, Jiechao Xiong, Shaogang
Gong, Yizhou Wang, and Yuan Yao | Robust Subjective Visual Property Prediction from Crowdsourced Pairwise
Labels | 14 pages, accepted by IEEE TPAMI | null | 10.1109/TPAMI.2015.2456887 | null | cs.CV cs.LG cs.MM cs.SI math.ST stat.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The problem of estimating subjective visual properties from image and video
has attracted increasing interest. A subjective visual property is useful
either on its own (e.g. image and video interestingness) or as an intermediate
representation for visual recognition (e.g. a relative attribute). Due to its
ambiguous nature, annotating the value of a subjective visual property for
learning a prediction model is challenging. To make the annotation more
reliable, recent studies employ crowdsourcing tools to collect pairwise
comparison labels because human annotators are much better at ranking two
images/videos (e.g. which one is more interesting) than giving an absolute
value to each of them separately. However, using crowdsourced data also
introduces outliers. Existing methods rely on majority voting to prune the
annotation outliers/errors. They thus require large amount of pairwise labels
to be collected. More importantly as a local outlier detection method, majority
voting is ineffective in identifying outliers that can cause global ranking
inconsistencies. In this paper, we propose a more principled way to identify
annotation outliers by formulating the subjective visual property prediction
task as a unified robust learning to rank problem, tackling both the outlier
detection and learning to rank jointly. Differing from existing methods, the
proposed method integrates local pairwise comparison labels together to
minimise a cost that corresponds to global inconsistency of ranking order. This
not only leads to better detection of annotation outliers but also enables
learning with extremely sparse annotations. Extensive experiments on various
benchmark datasets demonstrate that our new approach significantly outperforms
state-of-the-arts alternatives.
| [
{
"version": "v1",
"created": "Sun, 25 Jan 2015 20:02:45 GMT"
},
{
"version": "v2",
"created": "Fri, 30 Jan 2015 05:13:45 GMT"
},
{
"version": "v3",
"created": "Fri, 24 Jul 2015 18:40:56 GMT"
},
{
"version": "v4",
"created": "Mon, 27 Jul 2015 14:42:17 GMT"
}
] | 2015-07-28T00:00:00 | [
[
"Fu",
"Yanwei",
""
],
[
"Hospedales",
"Timothy M.",
""
],
[
"Xiang",
"Tao",
""
],
[
"Xiong",
"Jiechao",
""
],
[
"Gong",
"Shaogang",
""
],
[
"Wang",
"Yizhou",
""
],
[
"Yao",
"Yuan",
""
]
] | TITLE: Robust Subjective Visual Property Prediction from Crowdsourced Pairwise
Labels
ABSTRACT: The problem of estimating subjective visual properties from image and video
has attracted increasing interest. A subjective visual property is useful
either on its own (e.g. image and video interestingness) or as an intermediate
representation for visual recognition (e.g. a relative attribute). Due to its
ambiguous nature, annotating the value of a subjective visual property for
learning a prediction model is challenging. To make the annotation more
reliable, recent studies employ crowdsourcing tools to collect pairwise
comparison labels because human annotators are much better at ranking two
images/videos (e.g. which one is more interesting) than giving an absolute
value to each of them separately. However, using crowdsourced data also
introduces outliers. Existing methods rely on majority voting to prune the
annotation outliers/errors. They thus require large amount of pairwise labels
to be collected. More importantly as a local outlier detection method, majority
voting is ineffective in identifying outliers that can cause global ranking
inconsistencies. In this paper, we propose a more principled way to identify
annotation outliers by formulating the subjective visual property prediction
task as a unified robust learning to rank problem, tackling both the outlier
detection and learning to rank jointly. Differing from existing methods, the
proposed method integrates local pairwise comparison labels together to
minimise a cost that corresponds to global inconsistency of ranking order. This
not only leads to better detection of annotation outliers but also enables
learning with extremely sparse annotations. Extensive experiments on various
benchmark datasets demonstrate that our new approach significantly outperforms
state-of-the-arts alternatives.
| no_new_dataset | 0.951953 |
1507.06763 | Rina Okada | Rina Okada, Kazuto Fukuchi, Kazuya Kakizaki and Jun Sakuma | Differentially Private Analysis of Outliers | null | null | null | null | stat.ML cs.CR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper investigates differentially private analysis of distance-based
outliers. The problem of outlier detection is to find a small number of
instances that are apparently distant from the remaining instances. On the
other hand, the objective of differential privacy is to conceal presence (or
absence) of any particular instance. Outlier detection and privacy protection
are thus intrinsically conflicting tasks. In this paper, instead of reporting
outliers detected, we present two types of differentially private queries that
help to understand behavior of outliers. One is the query to count outliers,
which reports the number of outliers that appear in a given subspace. Our
formal analysis on the exact global sensitivity of outlier counts reveals that
regular global sensitivity based method can make the outputs too noisy,
particularly when the dimensionality of the given subspace is high. Noting that
the counts of outliers are typically expected to be relatively small compared
to the number of data, we introduce a mechanism based on the smooth upper bound
of the local sensitivity. The other is the query to discovery top-$h$ subspaces
containing a large number of outliers. This task can be naively achieved by
issuing count queries to each subspace in turn. However, the variation of
subspaces can grow exponentially in the data dimensionality. This can cause
serious consumption of the privacy budget. For this task, we propose an
exponential mechanism with a customized score function for subspace discovery.
To the best of our knowledge, this study is the first trial to ensure
differential privacy for distance-based outlier analysis. We demonstrated our
methods with synthesized datasets and real datasets. The experimental results
show that out method achieve better utility compared to the global sensitivity
based methods.
| [
{
"version": "v1",
"created": "Fri, 24 Jul 2015 07:30:49 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Jul 2015 02:19:15 GMT"
}
] | 2015-07-28T00:00:00 | [
[
"Okada",
"Rina",
""
],
[
"Fukuchi",
"Kazuto",
""
],
[
"Kakizaki",
"Kazuya",
""
],
[
"Sakuma",
"Jun",
""
]
] | TITLE: Differentially Private Analysis of Outliers
ABSTRACT: This paper investigates differentially private analysis of distance-based
outliers. The problem of outlier detection is to find a small number of
instances that are apparently distant from the remaining instances. On the
other hand, the objective of differential privacy is to conceal presence (or
absence) of any particular instance. Outlier detection and privacy protection
are thus intrinsically conflicting tasks. In this paper, instead of reporting
outliers detected, we present two types of differentially private queries that
help to understand behavior of outliers. One is the query to count outliers,
which reports the number of outliers that appear in a given subspace. Our
formal analysis on the exact global sensitivity of outlier counts reveals that
regular global sensitivity based method can make the outputs too noisy,
particularly when the dimensionality of the given subspace is high. Noting that
the counts of outliers are typically expected to be relatively small compared
to the number of data, we introduce a mechanism based on the smooth upper bound
of the local sensitivity. The other is the query to discovery top-$h$ subspaces
containing a large number of outliers. This task can be naively achieved by
issuing count queries to each subspace in turn. However, the variation of
subspaces can grow exponentially in the data dimensionality. This can cause
serious consumption of the privacy budget. For this task, we propose an
exponential mechanism with a customized score function for subspace discovery.
To the best of our knowledge, this study is the first trial to ensure
differential privacy for distance-based outlier analysis. We demonstrated our
methods with synthesized datasets and real datasets. The experimental results
show that out method achieve better utility compared to the global sensitivity
based methods.
| no_new_dataset | 0.944995 |
1507.06692 | Yasmen Wahba | Yasmen Wahba, Ehab ElSalamouny and Ghada ElTaweel | Improving the Performance of Multi-class Intrusion Detection Systems
using Feature Reduction | null | null | null | null | cs.NI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Intrusion detection systems (IDS) are widely studied by researchers nowadays
due to the dramatic growth in network-based technologies. Policy violations and
unauthorized access is in turn increasing which makes intrusion detection
systems of great importance. Existing approaches to improve intrusion detection
systems focus on feature selection or reduction since some features are
irrelevant or redundant which when removed improve the accuracy as well as the
learning time. In this paper we propose a hybrid feature selection method using
Correlation-based Feature Selection and Information Gain. In our work we apply
adaptive boosting using na\"ive Bayes as the weak (base) classifier. The key
point in our research is that we are able to improve the detection accuracy
with a reduced number of features while precisely determining the attack.
Experimental results showed that our proposed method achieved high accuracy
compared to methods using only 5-class problem. Correlation is done using
Greedy search strategy and na\"ive Bayes as the classifier on the reduced
NSL-KDD dataset.
| [
{
"version": "v1",
"created": "Thu, 23 Jul 2015 22:18:45 GMT"
}
] | 2015-07-27T00:00:00 | [
[
"Wahba",
"Yasmen",
""
],
[
"ElSalamouny",
"Ehab",
""
],
[
"ElTaweel",
"Ghada",
""
]
] | TITLE: Improving the Performance of Multi-class Intrusion Detection Systems
using Feature Reduction
ABSTRACT: Intrusion detection systems (IDS) are widely studied by researchers nowadays
due to the dramatic growth in network-based technologies. Policy violations and
unauthorized access is in turn increasing which makes intrusion detection
systems of great importance. Existing approaches to improve intrusion detection
systems focus on feature selection or reduction since some features are
irrelevant or redundant which when removed improve the accuracy as well as the
learning time. In this paper we propose a hybrid feature selection method using
Correlation-based Feature Selection and Information Gain. In our work we apply
adaptive boosting using na\"ive Bayes as the weak (base) classifier. The key
point in our research is that we are able to improve the detection accuracy
with a reduced number of features while precisely determining the attack.
Experimental results showed that our proposed method achieved high accuracy
compared to methods using only 5-class problem. Correlation is done using
Greedy search strategy and na\"ive Bayes as the classifier on the reduced
NSL-KDD dataset.
| no_new_dataset | 0.949389 |
1507.06802 | Jesse Krijthe | Jesse H. Krijthe and Marco Loog | Implicitly Constrained Semi-Supervised Least Squares Classification | 12 pages, 2 figures, 1 table. The Fourteenth International Symposium
on Intelligent Data Analysis (2015), Saint-Etienne, France | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We introduce a novel semi-supervised version of the least squares classifier.
This implicitly constrained least squares (ICLS) classifier minimizes the
squared loss on the labeled data among the set of parameters implied by all
possible labelings of the unlabeled data. Unlike other discriminative
semi-supervised methods, our approach does not introduce explicit additional
assumptions into the objective function, but leverages implicit assumptions
already present in the choice of the supervised least squares classifier. We
show this approach can be formulated as a quadratic programming problem and its
solution can be found using a simple gradient descent procedure. We prove that,
in a certain way, our method never leads to performance worse than the
supervised classifier. Experimental results corroborate this theoretical result
in the multidimensional case on benchmark datasets, also in terms of the error
rate.
| [
{
"version": "v1",
"created": "Fri, 24 Jul 2015 10:39:44 GMT"
}
] | 2015-07-27T00:00:00 | [
[
"Krijthe",
"Jesse H.",
""
],
[
"Loog",
"Marco",
""
]
] | TITLE: Implicitly Constrained Semi-Supervised Least Squares Classification
ABSTRACT: We introduce a novel semi-supervised version of the least squares classifier.
This implicitly constrained least squares (ICLS) classifier minimizes the
squared loss on the labeled data among the set of parameters implied by all
possible labelings of the unlabeled data. Unlike other discriminative
semi-supervised methods, our approach does not introduce explicit additional
assumptions into the objective function, but leverages implicit assumptions
already present in the choice of the supervised least squares classifier. We
show this approach can be formulated as a quadratic programming problem and its
solution can be found using a simple gradient descent procedure. We prove that,
in a certain way, our method never leads to performance worse than the
supervised classifier. Experimental results corroborate this theoretical result
in the multidimensional case on benchmark datasets, also in terms of the error
rate.
| no_new_dataset | 0.948537 |
1507.06841 | Jiawei Zhang | Jiawei Zhang, Philip S. Yu, Yuanhua Lv | Organizational Chart Inference | 10 pages, 9 figures, 1 table. The paper is accepted by KDD 2015 | null | null | null | cs.SI cs.CY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Nowadays, to facilitate the communication and cooperation among employees, a
new family of online social networks has been adopted in many companies, which
are called the "enterprise social networks" (ESNs). ESNs can provide employees
with various professional services to help them deal with daily work issues.
Meanwhile, employees in companies are usually organized into different
hierarchies according to the relative ranks of their positions. The company
internal management structure can be outlined with the organizational chart
visually, which is normally confidential to the public out of the privacy and
security concerns. In this paper, we want to study the IOC (Inference of
Organizational Chart) problem to identify company internal organizational chart
based on the heterogeneous online ESN launched in it. IOC is very challenging
to address as, to guarantee smooth operations, the internal organizational
charts of companies need to meet certain structural requirements (about its
depth and width). To solve the IOC problem, a novel unsupervised method Create
(ChArT REcovEr) is proposed in this paper, which consists of 3 steps: (1)
social stratification of ESN users into different social classes, (2)
supervision link inference from managers to subordinates, and (3) consecutive
social classes matching to prune the redundant supervision links. Extensive
experiments conducted on real-world online ESN dataset demonstrate that Create
can perform very well in addressing the IOC problem.
| [
{
"version": "v1",
"created": "Fri, 24 Jul 2015 13:32:30 GMT"
}
] | 2015-07-27T00:00:00 | [
[
"Zhang",
"Jiawei",
""
],
[
"Yu",
"Philip S.",
""
],
[
"Lv",
"Yuanhua",
""
]
] | TITLE: Organizational Chart Inference
ABSTRACT: Nowadays, to facilitate the communication and cooperation among employees, a
new family of online social networks has been adopted in many companies, which
are called the "enterprise social networks" (ESNs). ESNs can provide employees
with various professional services to help them deal with daily work issues.
Meanwhile, employees in companies are usually organized into different
hierarchies according to the relative ranks of their positions. The company
internal management structure can be outlined with the organizational chart
visually, which is normally confidential to the public out of the privacy and
security concerns. In this paper, we want to study the IOC (Inference of
Organizational Chart) problem to identify company internal organizational chart
based on the heterogeneous online ESN launched in it. IOC is very challenging
to address as, to guarantee smooth operations, the internal organizational
charts of companies need to meet certain structural requirements (about its
depth and width). To solve the IOC problem, a novel unsupervised method Create
(ChArT REcovEr) is proposed in this paper, which consists of 3 steps: (1)
social stratification of ESN users into different social classes, (2)
supervision link inference from managers to subordinates, and (3) consecutive
social classes matching to prune the redundant supervision links. Extensive
experiments conducted on real-world online ESN dataset demonstrate that Create
can perform very well in addressing the IOC problem.
| no_new_dataset | 0.942188 |
1507.06927 | Luiz Capretz Dr. | Faheem Ahmed, Piers Campbell, Ahmad Jaffar, Luiz Fernando Capretz | Myths and Realities about Online Forums in Open Source Software
Development: An Empirical Study | null | The Open Software Engineering Journal, Volume 4, 52-63, 2010 | 10.2174/1875107X01004010052 | null | cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The use of free and open source software (OSS) is gaining momentum due to the
ever increasing availability and use of the Internet. Organizations are also
now adopting open source software, despite some reservations, in particular
regarding the provision and availability of support. Some of the biggest
concerns about free and open source software are post release software defects
and their rectification, management of dynamic requirements and support to the
users. A common belief is that there is no appropriate support available for
this class of software. A contradictory argument is that due to the active
involvement of Internet users in online forums, there is in fact a large
resource available that communicates and manages the provision of support. The
research model of this empirical investigation examines the evidence available
to assess whether this commonly held belief is based on facts given the current
developments in OSS or simply a myth, which has developed around OSS
development. We analyzed a dataset consisting of 1880 open source software
projects covering a broad range of categories in this investigation. The
results show that online forums play a significant role in managing software
defects, implementation of new requirements and providing support to the users
in open source software and have become a major source of assistance in
maintenance of the open source projects.
| [
{
"version": "v1",
"created": "Fri, 24 Jul 2015 17:39:03 GMT"
}
] | 2015-07-27T00:00:00 | [
[
"Ahmed",
"Faheem",
""
],
[
"Campbell",
"Piers",
""
],
[
"Jaffar",
"Ahmad",
""
],
[
"Capretz",
"Luiz Fernando",
""
]
] | TITLE: Myths and Realities about Online Forums in Open Source Software
Development: An Empirical Study
ABSTRACT: The use of free and open source software (OSS) is gaining momentum due to the
ever increasing availability and use of the Internet. Organizations are also
now adopting open source software, despite some reservations, in particular
regarding the provision and availability of support. Some of the biggest
concerns about free and open source software are post release software defects
and their rectification, management of dynamic requirements and support to the
users. A common belief is that there is no appropriate support available for
this class of software. A contradictory argument is that due to the active
involvement of Internet users in online forums, there is in fact a large
resource available that communicates and manages the provision of support. The
research model of this empirical investigation examines the evidence available
to assess whether this commonly held belief is based on facts given the current
developments in OSS or simply a myth, which has developed around OSS
development. We analyzed a dataset consisting of 1880 open source software
projects covering a broad range of categories in this investigation. The
results show that online forums play a significant role in managing software
defects, implementation of new requirements and providing support to the users
in open source software and have become a major source of assistance in
maintenance of the open source projects.
| no_new_dataset | 0.906653 |
1202.4044 | Michael McCoy | Gilad Lerman, Michael McCoy, Joel A. Tropp, and Teng Zhang | Robust computation of linear models by convex relaxation | Formerly titled "Robust computation of linear models, or How to find
a needle in a haystack" | Foundations of Computational Mathematics, April 2015, Volume 15,
Issue 2, pp 363-410 | 10.1007/s10208-014-9221-0 | null | cs.IT math.IT stat.CO stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Consider a dataset of vector-valued observations that consists of noisy
inliers, which are explained well by a low-dimensional subspace, along with
some number of outliers. This work describes a convex optimization problem,
called REAPER, that can reliably fit a low-dimensional model to this type of
data. This approach parameterizes linear subspaces using orthogonal projectors,
and it uses a relaxation of the set of orthogonal projectors to reach the
convex formulation. The paper provides an efficient algorithm for solving the
REAPER problem, and it documents numerical experiments which confirm that
REAPER can dependably find linear structure in synthetic and natural data. In
addition, when the inliers lie near a low-dimensional subspace, there is a
rigorous theory that describes when REAPER can approximate this subspace.
| [
{
"version": "v1",
"created": "Sat, 18 Feb 2012 00:47:22 GMT"
},
{
"version": "v2",
"created": "Mon, 11 Aug 2014 19:19:28 GMT"
}
] | 2015-07-24T00:00:00 | [
[
"Lerman",
"Gilad",
""
],
[
"McCoy",
"Michael",
""
],
[
"Tropp",
"Joel A.",
""
],
[
"Zhang",
"Teng",
""
]
] | TITLE: Robust computation of linear models by convex relaxation
ABSTRACT: Consider a dataset of vector-valued observations that consists of noisy
inliers, which are explained well by a low-dimensional subspace, along with
some number of outliers. This work describes a convex optimization problem,
called REAPER, that can reliably fit a low-dimensional model to this type of
data. This approach parameterizes linear subspaces using orthogonal projectors,
and it uses a relaxation of the set of orthogonal projectors to reach the
convex formulation. The paper provides an efficient algorithm for solving the
REAPER problem, and it documents numerical experiments which confirm that
REAPER can dependably find linear structure in synthetic and natural data. In
addition, when the inliers lie near a low-dimensional subspace, there is a
rigorous theory that describes when REAPER can approximate this subspace.
| no_new_dataset | 0.93196 |
1505.00393 | Francesco Visin | Francesco Visin and Kyle Kastner and Kyunghyun Cho and Matteo
Matteucci and Aaron Courville and Yoshua Bengio | ReNet: A Recurrent Neural Network Based Alternative to Convolutional
Networks | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we propose a deep neural network architecture for object
recognition based on recurrent neural networks. The proposed network, called
ReNet, replaces the ubiquitous convolution+pooling layer of the deep
convolutional neural network with four recurrent neural networks that sweep
horizontally and vertically in both directions across the image. We evaluate
the proposed ReNet on three widely-used benchmark datasets; MNIST, CIFAR-10 and
SVHN. The result suggests that ReNet is a viable alternative to the deep
convolutional neural network, and that further investigation is needed.
| [
{
"version": "v1",
"created": "Sun, 3 May 2015 04:58:53 GMT"
},
{
"version": "v2",
"created": "Fri, 3 Jul 2015 11:31:53 GMT"
},
{
"version": "v3",
"created": "Thu, 23 Jul 2015 17:11:04 GMT"
}
] | 2015-07-24T00:00:00 | [
[
"Visin",
"Francesco",
""
],
[
"Kastner",
"Kyle",
""
],
[
"Cho",
"Kyunghyun",
""
],
[
"Matteucci",
"Matteo",
""
],
[
"Courville",
"Aaron",
""
],
[
"Bengio",
"Yoshua",
""
]
] | TITLE: ReNet: A Recurrent Neural Network Based Alternative to Convolutional
Networks
ABSTRACT: In this paper, we propose a deep neural network architecture for object
recognition based on recurrent neural networks. The proposed network, called
ReNet, replaces the ubiquitous convolution+pooling layer of the deep
convolutional neural network with four recurrent neural networks that sweep
horizontally and vertically in both directions across the image. We evaluate
the proposed ReNet on three widely-used benchmark datasets; MNIST, CIFAR-10 and
SVHN. The result suggests that ReNet is a viable alternative to the deep
convolutional neural network, and that further investigation is needed.
| no_new_dataset | 0.9549 |
1506.07310 | Jingtuo Liu | Jingtuo Liu and Yafeng Deng and Tao Bai and Zhengping Wei and Chang
Huang | Targeting Ultimate Accuracy: Face Recognition via Deep Embedding | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Face Recognition has been studied for many decades. As opposed to traditional
hand-crafted features such as LBP and HOG, much more sophisticated features can
be learned automatically by deep learning methods in a data-driven way. In this
paper, we propose a two-stage approach that combines a multi-patch deep CNN and
deep metric learning, which extracts low dimensional but very discriminative
features for face verification and recognition. Experiments show that this
method outperforms other state-of-the-art methods on LFW dataset, achieving
99.77% pair-wise verification accuracy and significantly better accuracy under
other two more practical protocols. This paper also discusses the importance of
data size and the number of patches, showing a clear path to practical
high-performance face recognition systems in real world.
| [
{
"version": "v1",
"created": "Wed, 24 Jun 2015 10:36:26 GMT"
},
{
"version": "v2",
"created": "Thu, 25 Jun 2015 03:05:20 GMT"
},
{
"version": "v3",
"created": "Fri, 26 Jun 2015 03:05:45 GMT"
},
{
"version": "v4",
"created": "Thu, 23 Jul 2015 02:34:29 GMT"
}
] | 2015-07-24T00:00:00 | [
[
"Liu",
"Jingtuo",
""
],
[
"Deng",
"Yafeng",
""
],
[
"Bai",
"Tao",
""
],
[
"Wei",
"Zhengping",
""
],
[
"Huang",
"Chang",
""
]
] | TITLE: Targeting Ultimate Accuracy: Face Recognition via Deep Embedding
ABSTRACT: Face Recognition has been studied for many decades. As opposed to traditional
hand-crafted features such as LBP and HOG, much more sophisticated features can
be learned automatically by deep learning methods in a data-driven way. In this
paper, we propose a two-stage approach that combines a multi-patch deep CNN and
deep metric learning, which extracts low dimensional but very discriminative
features for face verification and recognition. Experiments show that this
method outperforms other state-of-the-art methods on LFW dataset, achieving
99.77% pair-wise verification accuracy and significantly better accuracy under
other two more practical protocols. This paper also discusses the importance of
data size and the number of patches, showing a clear path to practical
high-performance face recognition systems in real world.
| no_new_dataset | 0.952486 |
1507.06429 | Albert Gordo | Albert Gordo and Adrien Gaidon and Florent Perronnin | Deep Fishing: Gradient Features from Deep Nets | To appear at BMVC 2015 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Convolutional Networks (ConvNets) have recently improved image recognition
performance thanks to end-to-end learning of deep feed-forward models from raw
pixels. Deep learning is a marked departure from the previous state of the art,
the Fisher Vector (FV), which relied on gradient-based encoding of local
hand-crafted features. In this paper, we discuss a novel connection between
these two approaches. First, we show that one can derive gradient
representations from ConvNets in a similar fashion to the FV. Second, we show
that this gradient representation actually corresponds to a structured matrix
that allows for efficient similarity computation. We experimentally study the
benefits of transferring this representation over the outputs of ConvNet
layers, and find consistent improvements on the Pascal VOC 2007 and 2012
datasets.
| [
{
"version": "v1",
"created": "Thu, 23 Jul 2015 10:01:45 GMT"
}
] | 2015-07-24T00:00:00 | [
[
"Gordo",
"Albert",
""
],
[
"Gaidon",
"Adrien",
""
],
[
"Perronnin",
"Florent",
""
]
] | TITLE: Deep Fishing: Gradient Features from Deep Nets
ABSTRACT: Convolutional Networks (ConvNets) have recently improved image recognition
performance thanks to end-to-end learning of deep feed-forward models from raw
pixels. Deep learning is a marked departure from the previous state of the art,
the Fisher Vector (FV), which relied on gradient-based encoding of local
hand-crafted features. In this paper, we discuss a novel connection between
these two approaches. First, we show that one can derive gradient
representations from ConvNets in a similar fashion to the FV. Second, we show
that this gradient representation actually corresponds to a structured matrix
that allows for efficient similarity computation. We experimentally study the
benefits of transferring this representation over the outputs of ConvNet
layers, and find consistent improvements on the Pascal VOC 2007 and 2012
datasets.
| no_new_dataset | 0.949763 |
1507.06452 | Amin Mantrach | Robin Devooght and Nicolas Kourtellis and Amin Mantrach | Dynamic Matrix Factorization with Priors on Unknown Values | in the Proceedings of 21st ACM SIGKDD Conference on Knowledge
Discovery and Data Mining 2015 | null | null | null | stat.ML cs.IR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Advanced and effective collaborative filtering methods based on explicit
feedback assume that unknown ratings do not follow the same model as the
observed ones (\emph{not missing at random}). In this work, we build on this
assumption, and introduce a novel dynamic matrix factorization framework that
allows to set an explicit prior on unknown values. When new ratings, users, or
items enter the system, we can update the factorization in time independent of
the size of data (number of users, items and ratings). Hence, we can quickly
recommend items even to very recent users. We test our methods on three large
datasets, including two very sparse ones, in static and dynamic conditions. In
each case, we outrank state-of-the-art matrix factorization methods that do not
use a prior on unknown ratings.
| [
{
"version": "v1",
"created": "Thu, 23 Jul 2015 11:39:58 GMT"
}
] | 2015-07-24T00:00:00 | [
[
"Devooght",
"Robin",
""
],
[
"Kourtellis",
"Nicolas",
""
],
[
"Mantrach",
"Amin",
""
]
] | TITLE: Dynamic Matrix Factorization with Priors on Unknown Values
ABSTRACT: Advanced and effective collaborative filtering methods based on explicit
feedback assume that unknown ratings do not follow the same model as the
observed ones (\emph{not missing at random}). In this work, we build on this
assumption, and introduce a novel dynamic matrix factorization framework that
allows to set an explicit prior on unknown values. When new ratings, users, or
items enter the system, we can update the factorization in time independent of
the size of data (number of users, items and ratings). Hence, we can quickly
recommend items even to very recent users. We test our methods on three large
datasets, including two very sparse ones, in static and dynamic conditions. In
each case, we outrank state-of-the-art matrix factorization methods that do not
use a prior on unknown ratings.
| no_new_dataset | 0.949012 |
1507.06477 | Takayuki Mizuno | Takayuki Mizuno, Takaaki Ohnishi, Tsutomu Watanabe | Novel and topical business news and their impact on stock market
activities | 8 pages, 6 figures, 2 tables | null | null | null | q-fin.ST cs.CY physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose an indicator to measure the degree to which a particular news
article is novel, as well as an indicator to measure the degree to which a
particular news item attracts attention from investors. The novelty measure is
obtained by comparing the extent to which a particular news article is similar
to earlier news articles, and an article is regarded as novel if there was no
similar article before it. On the other hand, we say a news item receives a lot
of attention and thus is highly topical if it is simultaneously reported by
many news agencies and read by many investors who receive news from those
agencies. The topicality measure for a news item is obtained by counting the
number of news articles whose content is similar to an original news article
but which are delivered by other news agencies. To check the performance of the
indicators, we empirically examine how these indicators are correlated with
intraday financial market indicators such as the number of transactions and
price volatility. Specifically, we use a dataset consisting of over 90 million
business news articles reported in English and a dataset consisting of
minute-by-minute stock prices on the New York Stock Exchange and the NASDAQ
Stock Market from 2003 to 2014, and show that stock prices and transaction
volumes exhibited a significant response to a news article when it is novel and
topical.
| [
{
"version": "v1",
"created": "Thu, 23 Jul 2015 12:54:35 GMT"
}
] | 2015-07-24T00:00:00 | [
[
"Mizuno",
"Takayuki",
""
],
[
"Ohnishi",
"Takaaki",
""
],
[
"Watanabe",
"Tsutomu",
""
]
] | TITLE: Novel and topical business news and their impact on stock market
activities
ABSTRACT: We propose an indicator to measure the degree to which a particular news
article is novel, as well as an indicator to measure the degree to which a
particular news item attracts attention from investors. The novelty measure is
obtained by comparing the extent to which a particular news article is similar
to earlier news articles, and an article is regarded as novel if there was no
similar article before it. On the other hand, we say a news item receives a lot
of attention and thus is highly topical if it is simultaneously reported by
many news agencies and read by many investors who receive news from those
agencies. The topicality measure for a news item is obtained by counting the
number of news articles whose content is similar to an original news article
but which are delivered by other news agencies. To check the performance of the
indicators, we empirically examine how these indicators are correlated with
intraday financial market indicators such as the number of transactions and
price volatility. Specifically, we use a dataset consisting of over 90 million
business news articles reported in English and a dataset consisting of
minute-by-minute stock prices on the New York Stock Exchange and the NASDAQ
Stock Market from 2003 to 2014, and show that stock prices and transaction
volumes exhibited a significant response to a news article when it is novel and
topical.
| new_dataset | 0.967132 |
1411.2749 | Tobias Kuhn | Tobias Kuhn, Christine Chichester, Michael Krauthammer, Michel
Dumontier | Publishing without Publishers: a Decentralized Approach to
Dissemination, Retrieval, and Archiving of Data | In Proceedings of the 14th International Semantic Web Conference
(ISWC) 2015 | null | null | null | cs.DL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Making available and archiving scientific results is for the most part still
considered the task of classical publishing companies, despite the fact that
classical forms of publishing centered around printed narrative articles no
longer seem well-suited in the digital age. In particular, there exist
currently no efficient, reliable, and agreed-upon methods for publishing
scientific datasets, which have become increasingly important for science. Here
we propose to design scientific data publishing as a Web-based bottom-up
process, without top-down control of central authorities such as publishing
companies. Based on a novel combination of existing concepts and technologies,
we present a server network to decentrally store and archive data in the form
of nanopublications, an RDF-based format to represent scientific data. We show
how this approach allows researchers to publish, retrieve, verify, and
recombine datasets of nanopublications in a reliable and trustworthy manner,
and we argue that this architecture could be used for the Semantic Web in
general. Evaluation of the current small network shows that this system is
efficient and reliable.
| [
{
"version": "v1",
"created": "Tue, 11 Nov 2014 10:09:15 GMT"
},
{
"version": "v2",
"created": "Wed, 22 Jul 2015 08:25:11 GMT"
}
] | 2015-07-23T00:00:00 | [
[
"Kuhn",
"Tobias",
""
],
[
"Chichester",
"Christine",
""
],
[
"Krauthammer",
"Michael",
""
],
[
"Dumontier",
"Michel",
""
]
] | TITLE: Publishing without Publishers: a Decentralized Approach to
Dissemination, Retrieval, and Archiving of Data
ABSTRACT: Making available and archiving scientific results is for the most part still
considered the task of classical publishing companies, despite the fact that
classical forms of publishing centered around printed narrative articles no
longer seem well-suited in the digital age. In particular, there exist
currently no efficient, reliable, and agreed-upon methods for publishing
scientific datasets, which have become increasingly important for science. Here
we propose to design scientific data publishing as a Web-based bottom-up
process, without top-down control of central authorities such as publishing
companies. Based on a novel combination of existing concepts and technologies,
we present a server network to decentrally store and archive data in the form
of nanopublications, an RDF-based format to represent scientific data. We show
how this approach allows researchers to publish, retrieve, verify, and
recombine datasets of nanopublications in a reliable and trustworthy manner,
and we argue that this architecture could be used for the Semantic Web in
general. Evaluation of the current small network shows that this system is
efficient and reliable.
| no_new_dataset | 0.944536 |
1506.05869 | Oriol Vinyals | Oriol Vinyals, Quoc Le | A Neural Conversational Model | ICML Deep Learning Workshop 2015 | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Conversational modeling is an important task in natural language
understanding and machine intelligence. Although previous approaches exist,
they are often restricted to specific domains (e.g., booking an airline ticket)
and require hand-crafted rules. In this paper, we present a simple approach for
this task which uses the recently proposed sequence to sequence framework. Our
model converses by predicting the next sentence given the previous sentence or
sentences in a conversation. The strength of our model is that it can be
trained end-to-end and thus requires much fewer hand-crafted rules. We find
that this straightforward model can generate simple conversations given a large
conversational training dataset. Our preliminary results suggest that, despite
optimizing the wrong objective function, the model is able to converse well. It
is able extract knowledge from both a domain specific dataset, and from a
large, noisy, and general domain dataset of movie subtitles. On a
domain-specific IT helpdesk dataset, the model can find a solution to a
technical problem via conversations. On a noisy open-domain movie transcript
dataset, the model can perform simple forms of common sense reasoning. As
expected, we also find that the lack of consistency is a common failure mode of
our model.
| [
{
"version": "v1",
"created": "Fri, 19 Jun 2015 02:52:23 GMT"
},
{
"version": "v2",
"created": "Tue, 23 Jun 2015 22:12:47 GMT"
},
{
"version": "v3",
"created": "Wed, 22 Jul 2015 03:29:47 GMT"
}
] | 2015-07-23T00:00:00 | [
[
"Vinyals",
"Oriol",
""
],
[
"Le",
"Quoc",
""
]
] | TITLE: A Neural Conversational Model
ABSTRACT: Conversational modeling is an important task in natural language
understanding and machine intelligence. Although previous approaches exist,
they are often restricted to specific domains (e.g., booking an airline ticket)
and require hand-crafted rules. In this paper, we present a simple approach for
this task which uses the recently proposed sequence to sequence framework. Our
model converses by predicting the next sentence given the previous sentence or
sentences in a conversation. The strength of our model is that it can be
trained end-to-end and thus requires much fewer hand-crafted rules. We find
that this straightforward model can generate simple conversations given a large
conversational training dataset. Our preliminary results suggest that, despite
optimizing the wrong objective function, the model is able to converse well. It
is able extract knowledge from both a domain specific dataset, and from a
large, noisy, and general domain dataset of movie subtitles. On a
domain-specific IT helpdesk dataset, the model can find a solution to a
technical problem via conversations. On a noisy open-domain movie transcript
dataset, the model can perform simple forms of common sense reasoning. As
expected, we also find that the lack of consistency is a common failure mode of
our model.
| no_new_dataset | 0.941815 |
1507.05775 | Shuchang Zhou | Shuchang Zhou, Jia-Nan Wu | Compression of Fully-Connected Layer in Neural Network by Kronecker
Product | null | null | null | null | cs.NE cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper we propose and study a technique to reduce the number of
parameters and computation time in fully-connected layers of neural networks
using Kronecker product, at a mild cost of the prediction quality. The
technique proceeds by replacing Fully-Connected layers with so-called Kronecker
Fully-Connected layers, where the weight matrices of the FC layers are
approximated by linear combinations of multiple Kronecker products of smaller
matrices. In particular, given a model trained on SVHN dataset, we are able to
construct a new KFC model with 73\% reduction in total number of parameters,
while the error only rises mildly. In contrast, using low-rank method can only
achieve 35\% reduction in total number of parameters given similar quality
degradation allowance. If we only compare the KFC layer with its counterpart
fully-connected layer, the reduction in the number of parameters exceeds 99\%.
The amount of computation is also reduced as we replace matrix product of the
large matrices in FC layers with matrix products of a few smaller matrices in
KFC layers. Further experiments on MNIST, SVHN and some Chinese Character
recognition models also demonstrate effectiveness of our technique.
| [
{
"version": "v1",
"created": "Tue, 21 Jul 2015 10:29:11 GMT"
},
{
"version": "v2",
"created": "Wed, 22 Jul 2015 11:59:08 GMT"
}
] | 2015-07-23T00:00:00 | [
[
"Zhou",
"Shuchang",
""
],
[
"Wu",
"Jia-Nan",
""
]
] | TITLE: Compression of Fully-Connected Layer in Neural Network by Kronecker
Product
ABSTRACT: In this paper we propose and study a technique to reduce the number of
parameters and computation time in fully-connected layers of neural networks
using Kronecker product, at a mild cost of the prediction quality. The
technique proceeds by replacing Fully-Connected layers with so-called Kronecker
Fully-Connected layers, where the weight matrices of the FC layers are
approximated by linear combinations of multiple Kronecker products of smaller
matrices. In particular, given a model trained on SVHN dataset, we are able to
construct a new KFC model with 73\% reduction in total number of parameters,
while the error only rises mildly. In contrast, using low-rank method can only
achieve 35\% reduction in total number of parameters given similar quality
degradation allowance. If we only compare the KFC layer with its counterpart
fully-connected layer, the reduction in the number of parameters exceeds 99\%.
The amount of computation is also reduced as we replace matrix product of the
large matrices in FC layers with matrix products of a few smaller matrices in
KFC layers. Further experiments on MNIST, SVHN and some Chinese Character
recognition models also demonstrate effectiveness of our technique.
| no_new_dataset | 0.951639 |
1501.07873 | Yongxin Yang | Qian Yu, Yongxin Yang, Yi-Zhe Song, Tao Xiang and Timothy Hospedales | Sketch-a-Net that Beats Humans | Accepted to BMVC 2015 (oral) | null | null | null | cs.CV cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a multi-scale multi-channel deep neural network framework that,
for the first time, yields sketch recognition performance surpassing that of
humans. Our superior performance is a result of explicitly embedding the unique
characteristics of sketches in our model: (i) a network architecture designed
for sketch rather than natural photo statistics, (ii) a multi-channel
generalisation that encodes sequential ordering in the sketching process, and
(iii) a multi-scale network ensemble with joint Bayesian fusion that accounts
for the different levels of abstraction exhibited in free-hand sketches. We
show that state-of-the-art deep networks specifically engineered for photos of
natural objects fail to perform well on sketch recognition, regardless whether
they are trained using photo or sketch. Our network on the other hand not only
delivers the best performance on the largest human sketch dataset to date, but
also is small in size making efficient training possible using just CPUs.
| [
{
"version": "v1",
"created": "Fri, 30 Jan 2015 18:35:59 GMT"
},
{
"version": "v2",
"created": "Wed, 27 May 2015 18:59:06 GMT"
},
{
"version": "v3",
"created": "Tue, 21 Jul 2015 15:59:05 GMT"
}
] | 2015-07-22T00:00:00 | [
[
"Yu",
"Qian",
""
],
[
"Yang",
"Yongxin",
""
],
[
"Song",
"Yi-Zhe",
""
],
[
"Xiang",
"Tao",
""
],
[
"Hospedales",
"Timothy",
""
]
] | TITLE: Sketch-a-Net that Beats Humans
ABSTRACT: We propose a multi-scale multi-channel deep neural network framework that,
for the first time, yields sketch recognition performance surpassing that of
humans. Our superior performance is a result of explicitly embedding the unique
characteristics of sketches in our model: (i) a network architecture designed
for sketch rather than natural photo statistics, (ii) a multi-channel
generalisation that encodes sequential ordering in the sketching process, and
(iii) a multi-scale network ensemble with joint Bayesian fusion that accounts
for the different levels of abstraction exhibited in free-hand sketches. We
show that state-of-the-art deep networks specifically engineered for photos of
natural objects fail to perform well on sketch recognition, regardless whether
they are trained using photo or sketch. Our network on the other hand not only
delivers the best performance on the largest human sketch dataset to date, but
also is small in size making efficient training possible using just CPUs.
| no_new_dataset | 0.951233 |
1503.00688 | Zhilin Zhang | Zhilin Zhang | Photoplethysmography-Based Heart Rate Monitoring in Physical Activities
via Joint Sparse Spectrum Reconstruction | Published in IEEE Transactions on Biomedical Engineering, Vol. 62,
No. 8, PP. 1902-1910, August 2015 | IEEE Transactions on Biomedical Engineering, Vol. 62, No. 8, PP.
1902-1910, August 2015 | 10.1109/TBME.2015.2406332 | null | cs.OH cs.CY stat.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Goal: A new method for heart rate monitoring using photoplethysmography (PPG)
during physical activities is proposed. Methods: It jointly estimates spectra
of PPG signals and simultaneous acceleration signals, utilizing the multiple
measurement vector model in sparse signal recovery. Due to a common sparsity
constraint on spectral coefficients, the method can easily identify and remove
spectral peaks of motion artifact (MA) in PPG spectra. Thus, it does not need
any extra signal processing modular to remove MA as in some other algorithms.
Furthermore, seeking spectral peaks associated with heart rate is simplified.
Results: Experimental results on 12 PPG datasets sampled at 25 Hz and recorded
during subjects' fast running showed that it had high performance. The average
absolute estimation error was 1.28 beat per minute and the standard deviation
was 2.61 beat per minute. Conclusion and Significance: These results show that
the method has great potential to be used for PPG-based heart rate monitoring
in wearable devices for fitness tracking and health monitoring.
| [
{
"version": "v1",
"created": "Sat, 21 Feb 2015 01:48:20 GMT"
},
{
"version": "v2",
"created": "Tue, 21 Jul 2015 06:21:23 GMT"
}
] | 2015-07-22T00:00:00 | [
[
"Zhang",
"Zhilin",
""
]
] | TITLE: Photoplethysmography-Based Heart Rate Monitoring in Physical Activities
via Joint Sparse Spectrum Reconstruction
ABSTRACT: Goal: A new method for heart rate monitoring using photoplethysmography (PPG)
during physical activities is proposed. Methods: It jointly estimates spectra
of PPG signals and simultaneous acceleration signals, utilizing the multiple
measurement vector model in sparse signal recovery. Due to a common sparsity
constraint on spectral coefficients, the method can easily identify and remove
spectral peaks of motion artifact (MA) in PPG spectra. Thus, it does not need
any extra signal processing modular to remove MA as in some other algorithms.
Furthermore, seeking spectral peaks associated with heart rate is simplified.
Results: Experimental results on 12 PPG datasets sampled at 25 Hz and recorded
during subjects' fast running showed that it had high performance. The average
absolute estimation error was 1.28 beat per minute and the standard deviation
was 2.61 beat per minute. Conclusion and Significance: These results show that
the method has great potential to be used for PPG-based heart rate monitoring
in wearable devices for fitness tracking and health monitoring.
| no_new_dataset | 0.945801 |
1507.05717 | Cong Yao | Baoguang Shi and Xiang Bai and Cong Yao | An End-to-End Trainable Neural Network for Image-based Sequence
Recognition and Its Application to Scene Text Recognition | 5 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Image-based sequence recognition has been a long-standing research topic in
computer vision. In this paper, we investigate the problem of scene text
recognition, which is among the most important and challenging tasks in
image-based sequence recognition. A novel neural network architecture, which
integrates feature extraction, sequence modeling and transcription into a
unified framework, is proposed. Compared with previous systems for scene text
recognition, the proposed architecture possesses four distinctive properties:
(1) It is end-to-end trainable, in contrast to most of the existing algorithms
whose components are separately trained and tuned. (2) It naturally handles
sequences in arbitrary lengths, involving no character segmentation or
horizontal scale normalization. (3) It is not confined to any predefined
lexicon and achieves remarkable performances in both lexicon-free and
lexicon-based scene text recognition tasks. (4) It generates an effective yet
much smaller model, which is more practical for real-world application
scenarios. The experiments on standard benchmarks, including the IIIT-5K,
Street View Text and ICDAR datasets, demonstrate the superiority of the
proposed algorithm over the prior arts. Moreover, the proposed algorithm
performs well in the task of image-based music score recognition, which
evidently verifies the generality of it.
| [
{
"version": "v1",
"created": "Tue, 21 Jul 2015 06:26:32 GMT"
}
] | 2015-07-22T00:00:00 | [
[
"Shi",
"Baoguang",
""
],
[
"Bai",
"Xiang",
""
],
[
"Yao",
"Cong",
""
]
] | TITLE: An End-to-End Trainable Neural Network for Image-based Sequence
Recognition and Its Application to Scene Text Recognition
ABSTRACT: Image-based sequence recognition has been a long-standing research topic in
computer vision. In this paper, we investigate the problem of scene text
recognition, which is among the most important and challenging tasks in
image-based sequence recognition. A novel neural network architecture, which
integrates feature extraction, sequence modeling and transcription into a
unified framework, is proposed. Compared with previous systems for scene text
recognition, the proposed architecture possesses four distinctive properties:
(1) It is end-to-end trainable, in contrast to most of the existing algorithms
whose components are separately trained and tuned. (2) It naturally handles
sequences in arbitrary lengths, involving no character segmentation or
horizontal scale normalization. (3) It is not confined to any predefined
lexicon and achieves remarkable performances in both lexicon-free and
lexicon-based scene text recognition tasks. (4) It generates an effective yet
much smaller model, which is more practical for real-world application
scenarios. The experiments on standard benchmarks, including the IIIT-5K,
Street View Text and ICDAR datasets, demonstrate the superiority of the
proposed algorithm over the prior arts. Moreover, the proposed algorithm
performs well in the task of image-based music score recognition, which
evidently verifies the generality of it.
| no_new_dataset | 0.951051 |
1507.05860 | Naoshi Tobita | Naoshi Tobita, Shunsuke Honda, Kazuhiko Hara, Wataru Aoyagi, Yasuo
Arai, Toshinobu Miyoshi, Ikuo Kurachi, Takaki Hatsui, Togo Kudo, Kazuo
Kobayashi | Compensation for TID Damage in SOI Pixel Devices | null | null | null | null | physics.ins-det | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We are investigating adaption of SOI pixel devices for future high energy
physic(HEP) experiments. The pixel sensors are required to be operational in
very severe radiation environment. Most challenging issue in the adoption is
the TID (total ionizing dose) damage where holes trapped in oxide layers affect
the operation of nearby transistors. We have introduced a second SOI layer -
SOI2 beneath the BOX (Buried OXide) layer - in order to compensate for the TID
effect by applying a negative voltage to this electrode to cancel the effect
caused by accumulated positive holes. In this paper, the TID effects caused by
Co gamma-ray irradiation are presented based on the transistor characteristics
measurements. The irradiation was carried out in various biasing conditions to
investigate hole accumulation dependence on the potential configurations. We
also compare the data with samples irradiated with X-ray. Since we observed a
fair agreement between the two irradiation datasets, the TID effects have been
investigated in a wide dose range from 100~Gy to 2~MGy.
| [
{
"version": "v1",
"created": "Tue, 21 Jul 2015 14:49:13 GMT"
}
] | 2015-07-22T00:00:00 | [
[
"Tobita",
"Naoshi",
""
],
[
"Honda",
"Shunsuke",
""
],
[
"Hara",
"Kazuhiko",
""
],
[
"Aoyagi",
"Wataru",
""
],
[
"Arai",
"Yasuo",
""
],
[
"Miyoshi",
"Toshinobu",
""
],
[
"Kurachi",
"Ikuo",
""
],
[
"Hatsui",
"Takaki",
""
],
[
"Kudo",
"Togo",
""
],
[
"Kobayashi",
"Kazuo",
""
]
] | TITLE: Compensation for TID Damage in SOI Pixel Devices
ABSTRACT: We are investigating adaption of SOI pixel devices for future high energy
physic(HEP) experiments. The pixel sensors are required to be operational in
very severe radiation environment. Most challenging issue in the adoption is
the TID (total ionizing dose) damage where holes trapped in oxide layers affect
the operation of nearby transistors. We have introduced a second SOI layer -
SOI2 beneath the BOX (Buried OXide) layer - in order to compensate for the TID
effect by applying a negative voltage to this electrode to cancel the effect
caused by accumulated positive holes. In this paper, the TID effects caused by
Co gamma-ray irradiation are presented based on the transistor characteristics
measurements. The irradiation was carried out in various biasing conditions to
investigate hole accumulation dependence on the potential configurations. We
also compare the data with samples irradiated with X-ray. Since we observed a
fair agreement between the two irradiation datasets, the TID effects have been
investigated in a wide dose range from 100~Gy to 2~MGy.
| no_new_dataset | 0.946843 |
1411.7718 | Dacheng Tao | Tongliang Liu and Dacheng Tao | Classification with Noisy Labels by Importance Reweighting | null | null | 10.1109/TPAMI.2015.2456899 | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we study a classification problem in which sample labels are
randomly corrupted. In this scenario, there is an unobservable sample with
noise-free labels. However, before being observed, the true labels are
independently flipped with a probability $\rho\in[0,0.5)$, and the random label
noise can be class-conditional. Here, we address two fundamental problems
raised by this scenario. The first is how to best use the abundant surrogate
loss functions designed for the traditional classification problem when there
is label noise. We prove that any surrogate loss function can be used for
classification with noisy labels by using importance reweighting, with
consistency assurance that the label noise does not ultimately hinder the
search for the optimal classifier of the noise-free sample. The other is the
open problem of how to obtain the noise rate $\rho$. We show that the rate is
upper bounded by the conditional probability $P(y|x)$ of the noisy sample.
Consequently, the rate can be estimated, because the upper bound can be easily
reached in classification problems. Experimental results on synthetic and real
datasets confirm the efficiency of our methods.
| [
{
"version": "v1",
"created": "Thu, 27 Nov 2014 23:18:51 GMT"
},
{
"version": "v2",
"created": "Sat, 18 Jul 2015 04:03:44 GMT"
}
] | 2015-07-21T00:00:00 | [
[
"Liu",
"Tongliang",
""
],
[
"Tao",
"Dacheng",
""
]
] | TITLE: Classification with Noisy Labels by Importance Reweighting
ABSTRACT: In this paper, we study a classification problem in which sample labels are
randomly corrupted. In this scenario, there is an unobservable sample with
noise-free labels. However, before being observed, the true labels are
independently flipped with a probability $\rho\in[0,0.5)$, and the random label
noise can be class-conditional. Here, we address two fundamental problems
raised by this scenario. The first is how to best use the abundant surrogate
loss functions designed for the traditional classification problem when there
is label noise. We prove that any surrogate loss function can be used for
classification with noisy labels by using importance reweighting, with
consistency assurance that the label noise does not ultimately hinder the
search for the optimal classifier of the noise-free sample. The other is the
open problem of how to obtain the noise rate $\rho$. We show that the rate is
upper bounded by the conditional probability $P(y|x)$ of the noisy sample.
Consequently, the rate can be estimated, because the upper bound can be easily
reached in classification problems. Experimental results on synthetic and real
datasets confirm the efficiency of our methods.
| no_new_dataset | 0.941007 |
1501.00102 | Natalia Neverova | Natalia Neverova and Christian Wolf and Graham W. Taylor and Florian
Nebout | ModDrop: adaptive multi-modal gesture recognition | 14 pages, 7 figures | null | null | null | cs.CV cs.HC cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a method for gesture detection and localisation based on
multi-scale and multi-modal deep learning. Each visual modality captures
spatial information at a particular spatial scale (such as motion of the upper
body or a hand), and the whole system operates at three temporal scales. Key to
our technique is a training strategy which exploits: i) careful initialization
of individual modalities; and ii) gradual fusion involving random dropping of
separate channels (dubbed ModDrop) for learning cross-modality correlations
while preserving uniqueness of each modality-specific representation. We
present experiments on the ChaLearn 2014 Looking at People Challenge gesture
recognition track, in which we placed first out of 17 teams. Fusing multiple
modalities at several spatial and temporal scales leads to a significant
increase in recognition rates, allowing the model to compensate for errors of
the individual classifiers as well as noise in the separate channels.
Futhermore, the proposed ModDrop training technique ensures robustness of the
classifier to missing signals in one or several channels to produce meaningful
predictions from any number of available modalities. In addition, we
demonstrate the applicability of the proposed fusion scheme to modalities of
arbitrary nature by experiments on the same dataset augmented with audio.
| [
{
"version": "v1",
"created": "Wed, 31 Dec 2014 09:55:43 GMT"
},
{
"version": "v2",
"created": "Sat, 6 Jun 2015 14:46:33 GMT"
}
] | 2015-07-21T00:00:00 | [
[
"Neverova",
"Natalia",
""
],
[
"Wolf",
"Christian",
""
],
[
"Taylor",
"Graham W.",
""
],
[
"Nebout",
"Florian",
""
]
] | TITLE: ModDrop: adaptive multi-modal gesture recognition
ABSTRACT: We present a method for gesture detection and localisation based on
multi-scale and multi-modal deep learning. Each visual modality captures
spatial information at a particular spatial scale (such as motion of the upper
body or a hand), and the whole system operates at three temporal scales. Key to
our technique is a training strategy which exploits: i) careful initialization
of individual modalities; and ii) gradual fusion involving random dropping of
separate channels (dubbed ModDrop) for learning cross-modality correlations
while preserving uniqueness of each modality-specific representation. We
present experiments on the ChaLearn 2014 Looking at People Challenge gesture
recognition track, in which we placed first out of 17 teams. Fusing multiple
modalities at several spatial and temporal scales leads to a significant
increase in recognition rates, allowing the model to compensate for errors of
the individual classifiers as well as noise in the separate channels.
Futhermore, the proposed ModDrop training technique ensures robustness of the
classifier to missing signals in one or several channels to produce meaningful
predictions from any number of available modalities. In addition, we
demonstrate the applicability of the proposed fusion scheme to modalities of
arbitrary nature by experiments on the same dataset augmented with audio.
| no_new_dataset | 0.947088 |
1502.04754 | Cosimo Rubino | Cosimo Rubino and Marco Crocco and Alessandro Perina and Vittorio
Murino and Alessio Del Bue | 3D Pose from Detections | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a novel method to infer, in closed-form, a general 3D spatial
occupancy and orientation of a collection of rigid objects given 2D image
detections from a sequence of images. In particular, starting from 2D ellipses
fitted to bounding boxes, this novel multi-view problem can be reformulated as
the estimation of a quadric (ellipsoid) in 3D. We show that an efficient
solution exists in the dual-space using a minimum of three views while a
solution with two views is possible through the use of regularization. However,
this algebraic solution can be negatively affected in the presence of gross
inaccuracies in the bounding boxes estimation. To this end, we also propose a
robust ellipse fitting algorithm able to improve performance in the presence of
errors in the detected objects. Results on synthetic tests and on different
real datasets, involving real challenging scenarios, demonstrate the
applicability and potential of our method.
| [
{
"version": "v1",
"created": "Tue, 17 Feb 2015 00:11:41 GMT"
},
{
"version": "v2",
"created": "Wed, 22 Apr 2015 23:40:57 GMT"
},
{
"version": "v3",
"created": "Mon, 20 Jul 2015 18:27:38 GMT"
}
] | 2015-07-21T00:00:00 | [
[
"Rubino",
"Cosimo",
""
],
[
"Crocco",
"Marco",
""
],
[
"Perina",
"Alessandro",
""
],
[
"Murino",
"Vittorio",
""
],
[
"Del Bue",
"Alessio",
""
]
] | TITLE: 3D Pose from Detections
ABSTRACT: We present a novel method to infer, in closed-form, a general 3D spatial
occupancy and orientation of a collection of rigid objects given 2D image
detections from a sequence of images. In particular, starting from 2D ellipses
fitted to bounding boxes, this novel multi-view problem can be reformulated as
the estimation of a quadric (ellipsoid) in 3D. We show that an efficient
solution exists in the dual-space using a minimum of three views while a
solution with two views is possible through the use of regularization. However,
this algebraic solution can be negatively affected in the presence of gross
inaccuracies in the bounding boxes estimation. To this end, we also propose a
robust ellipse fitting algorithm able to improve performance in the presence of
errors in the detected objects. Results on synthetic tests and on different
real datasets, involving real challenging scenarios, demonstrate the
applicability and potential of our method.
| no_new_dataset | 0.944638 |
1507.03148 | Heng Yang | Heng Yang and Wenxuan Mou and Yichi Zhang and Ioannis Patras and
Hatice Gunes and Peter Robinson | Face Alignment Assisted by Head Pose Estimation | Accepted by BMVC2015 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper we propose a supervised initialization scheme for cascaded face
alignment based on explicit head pose estimation. We first investigate the
failure cases of most state of the art face alignment approaches and observe
that these failures often share one common global property, i.e. the head pose
variation is usually large. Inspired by this, we propose a deep convolutional
network model for reliable and accurate head pose estimation. Instead of using
a mean face shape, or randomly selected shapes for cascaded face alignment
initialisation, we propose two schemes for generating initialisation: the first
one relies on projecting a mean 3D face shape (represented by 3D facial
landmarks) onto 2D image under the estimated head pose; the second one searches
nearest neighbour shapes from the training set according to head pose distance.
By doing so, the initialisation gets closer to the actual shape, which enhances
the possibility of convergence and in turn improves the face alignment
performance. We demonstrate the proposed method on the benchmark 300W dataset
and show very competitive performance in both head pose estimation and face
alignment.
| [
{
"version": "v1",
"created": "Sat, 11 Jul 2015 20:07:51 GMT"
},
{
"version": "v2",
"created": "Sat, 18 Jul 2015 12:36:58 GMT"
}
] | 2015-07-21T00:00:00 | [
[
"Yang",
"Heng",
""
],
[
"Mou",
"Wenxuan",
""
],
[
"Zhang",
"Yichi",
""
],
[
"Patras",
"Ioannis",
""
],
[
"Gunes",
"Hatice",
""
],
[
"Robinson",
"Peter",
""
]
] | TITLE: Face Alignment Assisted by Head Pose Estimation
ABSTRACT: In this paper we propose a supervised initialization scheme for cascaded face
alignment based on explicit head pose estimation. We first investigate the
failure cases of most state of the art face alignment approaches and observe
that these failures often share one common global property, i.e. the head pose
variation is usually large. Inspired by this, we propose a deep convolutional
network model for reliable and accurate head pose estimation. Instead of using
a mean face shape, or randomly selected shapes for cascaded face alignment
initialisation, we propose two schemes for generating initialisation: the first
one relies on projecting a mean 3D face shape (represented by 3D facial
landmarks) onto 2D image under the estimated head pose; the second one searches
nearest neighbour shapes from the training set according to head pose distance.
By doing so, the initialisation gets closer to the actual shape, which enhances
the possibility of convergence and in turn improves the face alignment
performance. We demonstrate the proposed method on the benchmark 300W dataset
and show very competitive performance in both head pose estimation and face
alignment.
| no_new_dataset | 0.950778 |
1507.05143 | Paul Bendich | Christopher J. Tralie and Paul Bendich | Cover Song Identification with Timbral Shape Sequences | null | null | null | null | cs.SD | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We introduce a novel low level feature for identifying cover songs which
quantifies the relative changes in the smoothed frequency spectrum of a song.
Our key insight is that a sliding window representation of a chunk of audio can
be viewed as a time-ordered point cloud in high dimensions. For corresponding
chunks of audio between different versions of the same song, these point clouds
are approximately rotated, translated, and scaled copies of each other. If we
treat MFCC embeddings as point clouds and cast the problem as a relative shape
sequence, we are able to correctly identify 42/80 cover songs in the "Covers
80" dataset. By contrast, all other work to date on cover songs exclusively
relies on matching note sequences from Chroma derived features.
| [
{
"version": "v1",
"created": "Sat, 18 Jul 2015 03:55:50 GMT"
}
] | 2015-07-21T00:00:00 | [
[
"Tralie",
"Christopher J.",
""
],
[
"Bendich",
"Paul",
""
]
] | TITLE: Cover Song Identification with Timbral Shape Sequences
ABSTRACT: We introduce a novel low level feature for identifying cover songs which
quantifies the relative changes in the smoothed frequency spectrum of a song.
Our key insight is that a sliding window representation of a chunk of audio can
be viewed as a time-ordered point cloud in high dimensions. For corresponding
chunks of audio between different versions of the same song, these point clouds
are approximately rotated, translated, and scaled copies of each other. If we
treat MFCC embeddings as point clouds and cast the problem as a relative shape
sequence, we are able to correctly identify 42/80 cover songs in the "Covers
80" dataset. By contrast, all other work to date on cover songs exclusively
relies on matching note sequences from Chroma derived features.
| no_new_dataset | 0.946547 |
1507.05181 | Matej Balog | Matej Balog and Yee Whye Teh | The Mondrian Process for Machine Learning | null | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This report is concerned with the Mondrian process and its applications in
machine learning. The Mondrian process is a guillotine-partition-valued
stochastic process that possesses an elegant self-consistency property. The
first part of the report uses simple concepts from applied probability to
define the Mondrian process and explore its properties.
The Mondrian process has been used as the main building block of a clever
online random forest classification algorithm that turns out to be equivalent
to its batch counterpart. We outline a slight adaptation of this algorithm to
regression, as the remainder of the report uses regression as a case study of
how Mondrian processes can be utilized in machine learning. In particular, the
Mondrian process will be used to construct a fast approximation to the
computationally expensive kernel ridge regression problem with a Laplace
kernel.
The complexity of random guillotine partitions generated by a Mondrian
process and hence the complexity of the resulting regression models is
controlled by a lifetime hyperparameter. It turns out that these models can be
efficiently trained and evaluated for all lifetimes in a given range at once,
without needing to retrain them from scratch for each lifetime value. This
leads to an efficient procedure for determining the right model complexity for
a dataset at hand.
The limitation of having a single lifetime hyperparameter will motivate the
final Mondrian grid model, in which each input dimension is endowed with its
own lifetime parameter. In this model we preserve the property that its
hyperparameters can be tweaked without needing to retrain the modified model
from scratch.
| [
{
"version": "v1",
"created": "Sat, 18 Jul 2015 12:58:11 GMT"
}
] | 2015-07-21T00:00:00 | [
[
"Balog",
"Matej",
""
],
[
"Teh",
"Yee Whye",
""
]
] | TITLE: The Mondrian Process for Machine Learning
ABSTRACT: This report is concerned with the Mondrian process and its applications in
machine learning. The Mondrian process is a guillotine-partition-valued
stochastic process that possesses an elegant self-consistency property. The
first part of the report uses simple concepts from applied probability to
define the Mondrian process and explore its properties.
The Mondrian process has been used as the main building block of a clever
online random forest classification algorithm that turns out to be equivalent
to its batch counterpart. We outline a slight adaptation of this algorithm to
regression, as the remainder of the report uses regression as a case study of
how Mondrian processes can be utilized in machine learning. In particular, the
Mondrian process will be used to construct a fast approximation to the
computationally expensive kernel ridge regression problem with a Laplace
kernel.
The complexity of random guillotine partitions generated by a Mondrian
process and hence the complexity of the resulting regression models is
controlled by a lifetime hyperparameter. It turns out that these models can be
efficiently trained and evaluated for all lifetimes in a given range at once,
without needing to retrain them from scratch for each lifetime value. This
leads to an efficient procedure for determining the right model complexity for
a dataset at hand.
The limitation of having a single lifetime hyperparameter will motivate the
final Mondrian grid model, in which each input dimension is endowed with its
own lifetime parameter. In this model we preserve the property that its
hyperparameters can be tweaked without needing to retrain the modified model
from scratch.
| no_new_dataset | 0.947235 |
1507.05245 | Gautam Thakur | Gautam S. Thakur, Budhendra L. Bhaduri, Jesse O. Piburn, Kelly M.
Sims, Robert N. Stewart, Marie L. Urban | PlanetSense: A Real-time Streaming and Spatio-temporal Analytics
Platform for Gathering Geo-spatial Intelligence from Open Source Data | null | null | null | null | cs.CY cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Geospatial intelligence has traditionally relied on the use of archived and
unvarying data for planning and exploration purposes. In consequence, the tools
and methods that are architected to provide insight and generate projections
only rely on such datasets. Albeit, if this approach has proven effective in
several cases, such as land use identification and route mapping, it has
severely restricted the ability of researchers to inculcate current information
in their work. This approach is inadequate in scenarios requiring real-time
information to act and to adjust in ever changing dynamic environments, such as
evacuation and rescue missions. In this work, we propose PlanetSense, a
platform for geospatial intelligence that is built to harness the existing
power of archived data and add to that, the dynamics of real-time streams,
seamlessly integrated with sophisticated data mining algorithms and analytics
tools for generating operational intelligence on the fly. The platform has four
main components - i. GeoData Cloud - a data architecture for storing and
managing disparate datasets; ii. Mechanism to harvest real-time streaming data;
iii. Data analytics framework; iv. Presentation and visualization through web
interface and RESTful services. Using two case studies, we underpin the
necessity of our platform in modeling ambient population and building occupancy
at scale.
| [
{
"version": "v1",
"created": "Sun, 19 Jul 2015 03:19:03 GMT"
}
] | 2015-07-21T00:00:00 | [
[
"Thakur",
"Gautam S.",
""
],
[
"Bhaduri",
"Budhendra L.",
""
],
[
"Piburn",
"Jesse O.",
""
],
[
"Sims",
"Kelly M.",
""
],
[
"Stewart",
"Robert N.",
""
],
[
"Urban",
"Marie L.",
""
]
] | TITLE: PlanetSense: A Real-time Streaming and Spatio-temporal Analytics
Platform for Gathering Geo-spatial Intelligence from Open Source Data
ABSTRACT: Geospatial intelligence has traditionally relied on the use of archived and
unvarying data for planning and exploration purposes. In consequence, the tools
and methods that are architected to provide insight and generate projections
only rely on such datasets. Albeit, if this approach has proven effective in
several cases, such as land use identification and route mapping, it has
severely restricted the ability of researchers to inculcate current information
in their work. This approach is inadequate in scenarios requiring real-time
information to act and to adjust in ever changing dynamic environments, such as
evacuation and rescue missions. In this work, we propose PlanetSense, a
platform for geospatial intelligence that is built to harness the existing
power of archived data and add to that, the dynamics of real-time streams,
seamlessly integrated with sophisticated data mining algorithms and analytics
tools for generating operational intelligence on the fly. The platform has four
main components - i. GeoData Cloud - a data architecture for storing and
managing disparate datasets; ii. Mechanism to harvest real-time streaming data;
iii. Data analytics framework; iv. Presentation and visualization through web
interface and RESTful services. Using two case studies, we underpin the
necessity of our platform in modeling ambient population and building occupancy
at scale.
| no_new_dataset | 0.948489 |
1507.05275 | Swakkhar Shatabda | Shanjida Khatun, Hasib Ul Alam and Swakkhar Shatabda | An Efficient Genetic Algorithm for Discovering Diverse-Frequent Patterns | 2015 International Conference on Electrical Engineering and
Information Communication Technology (ICEEICT) | null | null | null | cs.AI | http://creativecommons.org/publicdomain/zero/1.0/ | Working with exhaustive search on large dataset is infeasible for several
reasons. Recently, developed techniques that made pattern set mining feasible
by a general solver with long execution time that supports heuristic search and
are limited to small datasets only. In this paper, we investigate an approach
which aims to find diverse set of patterns using genetic algorithm to mine
diverse frequent patterns. We propose a fast heuristic search algorithm that
outperforms state-of-the-art methods on a standard set of benchmarks and
capable to produce satisfactory results within a short period of time. Our
proposed algorithm uses a relative encoding scheme for the patterns and an
effective twin removal technique to ensure diversity throughout the search.
| [
{
"version": "v1",
"created": "Sun, 19 Jul 2015 10:55:09 GMT"
}
] | 2015-07-21T00:00:00 | [
[
"Khatun",
"Shanjida",
""
],
[
"Alam",
"Hasib Ul",
""
],
[
"Shatabda",
"Swakkhar",
""
]
] | TITLE: An Efficient Genetic Algorithm for Discovering Diverse-Frequent Patterns
ABSTRACT: Working with exhaustive search on large dataset is infeasible for several
reasons. Recently, developed techniques that made pattern set mining feasible
by a general solver with long execution time that supports heuristic search and
are limited to small datasets only. In this paper, we investigate an approach
which aims to find diverse set of patterns using genetic algorithm to mine
diverse frequent patterns. We propose a fast heuristic search algorithm that
outperforms state-of-the-art methods on a standard set of benchmarks and
capable to produce satisfactory results within a short period of time. Our
proposed algorithm uses a relative encoding scheme for the patterns and an
effective twin removal technique to ensure diversity throughout the search.
| no_new_dataset | 0.951729 |
1507.05348 | Zhaowei Cai | Zhaowei Cai, Mohammad Saberian, Nuno Vasconcelos | Learning Complexity-Aware Cascades for Deep Pedestrian Detection | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The design of complexity-aware cascaded detectors, combining features of very
different complexities, is considered. A new cascade design procedure is
introduced, by formulating cascade learning as the Lagrangian optimization of a
risk that accounts for both accuracy and complexity. A boosting algorithm,
denoted as complexity aware cascade training (CompACT), is then derived to
solve this optimization. CompACT cascades are shown to seek an optimal
trade-off between accuracy and complexity by pushing features of higher
complexity to the later cascade stages, where only a few difficult candidate
patches remain to be classified. This enables the use of features of vastly
different complexities in a single detector. In result, the feature pool can be
expanded to features previously impractical for cascade design, such as the
responses of a deep convolutional neural network (CNN). This is demonstrated
through the design of a pedestrian detector with a pool of features whose
complexities span orders of magnitude. The resulting cascade generalizes the
combination of a CNN with an object proposal mechanism: rather than a
pre-processing stage, CompACT cascades seamlessly integrate CNNs in their
stages. This enables state of the art performance on the Caltech and KITTI
datasets, at fairly fast speeds.
| [
{
"version": "v1",
"created": "Sun, 19 Jul 2015 22:31:01 GMT"
}
] | 2015-07-21T00:00:00 | [
[
"Cai",
"Zhaowei",
""
],
[
"Saberian",
"Mohammad",
""
],
[
"Vasconcelos",
"Nuno",
""
]
] | TITLE: Learning Complexity-Aware Cascades for Deep Pedestrian Detection
ABSTRACT: The design of complexity-aware cascaded detectors, combining features of very
different complexities, is considered. A new cascade design procedure is
introduced, by formulating cascade learning as the Lagrangian optimization of a
risk that accounts for both accuracy and complexity. A boosting algorithm,
denoted as complexity aware cascade training (CompACT), is then derived to
solve this optimization. CompACT cascades are shown to seek an optimal
trade-off between accuracy and complexity by pushing features of higher
complexity to the later cascade stages, where only a few difficult candidate
patches remain to be classified. This enables the use of features of vastly
different complexities in a single detector. In result, the feature pool can be
expanded to features previously impractical for cascade design, such as the
responses of a deep convolutional neural network (CNN). This is demonstrated
through the design of a pedestrian detector with a pool of features whose
complexities span orders of magnitude. The resulting cascade generalizes the
combination of a CNN with an object proposal mechanism: rather than a
pre-processing stage, CompACT cascades seamlessly integrate CNNs in their
stages. This enables state of the art performance on the Caltech and KITTI
datasets, at fairly fast speeds.
| no_new_dataset | 0.946646 |
1507.05408 | Tobias Kuhn | Juan M. Banda and Tobias Kuhn and Nigam H. Shah and Michel Dumontier | Provenance-Centered Dataset of Drug-Drug Interactions | In Proceedings of the 14th International Semantic Web Conference
(ISWC) 2015 | null | null | null | cs.CY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Over the years several studies have demonstrated the ability to identify
potential drug-drug interactions via data mining from the literature (MEDLINE),
electronic health records, public databases (Drugbank), etc. While each one of
these approaches is properly statistically validated, they do not take into
consideration the overlap between them as one of their decision making
variables. In this paper we present LInked Drug-Drug Interactions (LIDDI), a
public nanopublication-based RDF dataset with trusty URIs that encompasses some
of the most cited prediction methods and sources to provide researchers a
resource for leveraging the work of others into their prediction methods. As
one of the main issues to overcome the usage of external resources is their
mappings between drug names and identifiers used, we also provide the set of
mappings we curated to be able to compare the multiple sources we aggregate in
our dataset.
| [
{
"version": "v1",
"created": "Mon, 20 Jul 2015 07:53:56 GMT"
}
] | 2015-07-21T00:00:00 | [
[
"Banda",
"Juan M.",
""
],
[
"Kuhn",
"Tobias",
""
],
[
"Shah",
"Nigam H.",
""
],
[
"Dumontier",
"Michel",
""
]
] | TITLE: Provenance-Centered Dataset of Drug-Drug Interactions
ABSTRACT: Over the years several studies have demonstrated the ability to identify
potential drug-drug interactions via data mining from the literature (MEDLINE),
electronic health records, public databases (Drugbank), etc. While each one of
these approaches is properly statistically validated, they do not take into
consideration the overlap between them as one of their decision making
variables. In this paper we present LInked Drug-Drug Interactions (LIDDI), a
public nanopublication-based RDF dataset with trusty URIs that encompasses some
of the most cited prediction methods and sources to provide researchers a
resource for leveraging the work of others into their prediction methods. As
one of the main issues to overcome the usage of external resources is their
mappings between drug names and identifiers used, we also provide the set of
mappings we curated to be able to compare the multiple sources we aggregate in
our dataset.
| new_dataset | 0.960063 |
1507.05489 | Andrea Romanoni | Andrea Romanoni, Matteo Matteucci | Efficient moving point handling for incremental 3D manifold
reconstruction | Accepted in International Conference on Image Analysis and Processing
(ICIAP 2015) | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As incremental Structure from Motion algorithms become effective, a good
sparse point cloud representing the map of the scene becomes available
frame-by-frame. From the 3D Delaunay triangulation of these points,
state-of-the-art algorithms build a manifold rough model of the scene. These
algorithms integrate incrementally new points to the 3D reconstruction only if
their position estimate does not change. Indeed, whenever a point moves in a 3D
Delaunay triangulation, for instance because its estimation gets refined, a set
of tetrahedra have to be removed and replaced with new ones to maintain the
Delaunay property; the management of the manifold reconstruction becomes thus
complex and it entails a potentially big overhead. In this paper we investigate
different approaches and we propose an efficient policy to deal with moving
points in the manifold estimation process. We tested our approach with four
sequences of the KITTI dataset and we show the effectiveness of our proposal in
comparison with state-of-the-art approaches.
| [
{
"version": "v1",
"created": "Mon, 20 Jul 2015 13:38:02 GMT"
}
] | 2015-07-21T00:00:00 | [
[
"Romanoni",
"Andrea",
""
],
[
"Matteucci",
"Matteo",
""
]
] | TITLE: Efficient moving point handling for incremental 3D manifold
reconstruction
ABSTRACT: As incremental Structure from Motion algorithms become effective, a good
sparse point cloud representing the map of the scene becomes available
frame-by-frame. From the 3D Delaunay triangulation of these points,
state-of-the-art algorithms build a manifold rough model of the scene. These
algorithms integrate incrementally new points to the 3D reconstruction only if
their position estimate does not change. Indeed, whenever a point moves in a 3D
Delaunay triangulation, for instance because its estimation gets refined, a set
of tetrahedra have to be removed and replaced with new ones to maintain the
Delaunay property; the management of the manifold reconstruction becomes thus
complex and it entails a potentially big overhead. In this paper we investigate
different approaches and we propose an efficient policy to deal with moving
points in the manifold estimation process. We tested our approach with four
sequences of the KITTI dataset and we show the effectiveness of our proposal in
comparison with state-of-the-art approaches.
| no_new_dataset | 0.949623 |
1507.05497 | Dmitry Ignatov | Dmitry I. Ignatov and Denis Kornilov | RAPS: A Recommender Algorithm Based on Pattern Structures | The paper presented at FCA4AI 2015 in conjunction with IJCAI 2015 | null | null | null | cs.IR cs.AI cs.DM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a new algorithm for recommender systems with numeric ratings which
is based on Pattern Structures (RAPS). As the input the algorithm takes rating
matrix, e.g., such that it contains movies rated by users. For a target user,
the algorithm returns a rated list of items (movies) based on its previous
ratings and ratings of other users. We compare the results of the proposed
algorithm in terms of precision and recall measures with Slope One, one of the
state-of-the-art item-based algorithms, on Movie Lens dataset and RAPS
demonstrates the best or comparable quality.
| [
{
"version": "v1",
"created": "Mon, 20 Jul 2015 13:58:30 GMT"
}
] | 2015-07-21T00:00:00 | [
[
"Ignatov",
"Dmitry I.",
""
],
[
"Kornilov",
"Denis",
""
]
] | TITLE: RAPS: A Recommender Algorithm Based on Pattern Structures
ABSTRACT: We propose a new algorithm for recommender systems with numeric ratings which
is based on Pattern Structures (RAPS). As the input the algorithm takes rating
matrix, e.g., such that it contains movies rated by users. For a target user,
the algorithm returns a rated list of items (movies) based on its previous
ratings and ratings of other users. We compare the results of the proposed
algorithm in terms of precision and recall measures with Slope One, one of the
state-of-the-art item-based algorithms, on Movie Lens dataset and RAPS
demonstrates the best or comparable quality.
| no_new_dataset | 0.948965 |
1507.05578 | Anant Raj | Anant Raj, Vinay P. Namboodiri and Tinne Tuytelaars | Subspace Alignment Based Domain Adaptation for RCNN Detector | 26th British Machine Vision Conference, Swansea, UK | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we propose subspace alignment based domain adaptation of the
state of the art RCNN based object detector. The aim is to be able to achieve
high quality object detection in novel, real world target scenarios without
requiring labels from the target domain. While, unsupervised domain adaptation
has been studied in the case of object classification, for object detection it
has been relatively unexplored. In subspace based domain adaptation for
objects, we need access to source and target subspaces for the bounding box
features. The absence of supervision (labels and bounding boxes are absent)
makes the task challenging. In this paper, we show that we can still adapt sub-
spaces that are localized to the object by obtaining detections from the RCNN
detector trained on source and applied on target. Then we form localized
subspaces from the detections and show that subspace alignment based adaptation
between these subspaces yields improved object detection. This evaluation is
done by considering challenging real world datasets of PASCAL VOC as source and
validation set of Microsoft COCO dataset as target for various categories.
| [
{
"version": "v1",
"created": "Mon, 20 Jul 2015 18:23:54 GMT"
}
] | 2015-07-21T00:00:00 | [
[
"Raj",
"Anant",
""
],
[
"Namboodiri",
"Vinay P.",
""
],
[
"Tuytelaars",
"Tinne",
""
]
] | TITLE: Subspace Alignment Based Domain Adaptation for RCNN Detector
ABSTRACT: In this paper, we propose subspace alignment based domain adaptation of the
state of the art RCNN based object detector. The aim is to be able to achieve
high quality object detection in novel, real world target scenarios without
requiring labels from the target domain. While, unsupervised domain adaptation
has been studied in the case of object classification, for object detection it
has been relatively unexplored. In subspace based domain adaptation for
objects, we need access to source and target subspaces for the bounding box
features. The absence of supervision (labels and bounding boxes are absent)
makes the task challenging. In this paper, we show that we can still adapt sub-
spaces that are localized to the object by obtaining detections from the RCNN
detector trained on source and applied on target. Then we form localized
subspaces from the detections and show that subspace alignment based adaptation
between these subspaces yields improved object detection. This evaluation is
done by considering challenging real world datasets of PASCAL VOC as source and
validation set of Microsoft COCO dataset as target for various categories.
| no_new_dataset | 0.949856 |
1501.06814 | Luca Rossi | Luca Rossi, James Walker, Mirco Musolesi | Spatio-Temporal Techniques for User Identification by means of GPS
Mobility Data | 11 pages, 8 figures | null | null | null | cs.CR cs.CY physics.data-an | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | One of the greatest concerns related to the popularity of GPS-enabled devices
and applications is the increasing availability of the personal location
information generated by them and shared with application and service
providers. Moreover, people tend to have regular routines and be characterized
by a set of "significant places", thus making it possible to identify a user
from his/her mobility data.
In this paper we present a series of techniques for identifying individuals
from their GPS movements. More specifically, we study the uniqueness of GPS
information for three popular datasets, and we provide a detailed analysis of
the discriminatory power of speed, direction and distance of travel. Most
importantly, we present a simple yet effective technique for the identification
of users from location information that are not included in the original
dataset used for training, thus raising important privacy concerns for the
management of location datasets.
| [
{
"version": "v1",
"created": "Tue, 27 Jan 2015 16:42:03 GMT"
},
{
"version": "v2",
"created": "Sat, 31 Jan 2015 10:41:03 GMT"
},
{
"version": "v3",
"created": "Fri, 17 Jul 2015 15:28:46 GMT"
}
] | 2015-07-20T00:00:00 | [
[
"Rossi",
"Luca",
""
],
[
"Walker",
"James",
""
],
[
"Musolesi",
"Mirco",
""
]
] | TITLE: Spatio-Temporal Techniques for User Identification by means of GPS
Mobility Data
ABSTRACT: One of the greatest concerns related to the popularity of GPS-enabled devices
and applications is the increasing availability of the personal location
information generated by them and shared with application and service
providers. Moreover, people tend to have regular routines and be characterized
by a set of "significant places", thus making it possible to identify a user
from his/her mobility data.
In this paper we present a series of techniques for identifying individuals
from their GPS movements. More specifically, we study the uniqueness of GPS
information for three popular datasets, and we provide a detailed analysis of
the discriminatory power of speed, direction and distance of travel. Most
importantly, we present a simple yet effective technique for the identification
of users from location information that are not included in the original
dataset used for training, thus raising important privacy concerns for the
management of location datasets.
| no_new_dataset | 0.952175 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.