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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1509.04777 | Daniele Spiga | D. Spiga, L. Raimondi | X-ray optical systems: from metrology to Point Spread Function | null | Proceedings of the SPIE, Advances in Computational Methods for
X-Ray Optics III, Vol. 9209, 92090E (2014) | 10.1117/12.2061657 | null | physics.optics astro-ph.IM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | One of the problems often encountered in X-ray mirror manufacturing is
setting proper manufacturing tolerances to guarantee an angular resolution -
often expressed in terms of Point Spread Function (PSF) - as needed by the
specific science goal. To do this, we need an accurate metrological apparatus,
covering a very broad range of spatial frequencies, and an affordable method to
compute the PSF from the metrology dataset. [...] However, the separation
between these spectral ranges is difficult do define exactly, and it is also
unclear how to affordably combine the PSFs, computed with different methods in
different spectral ranges, into a PSF expectation at a given X-ray energy. For
this reason, we have proposed a method entirely based on the Huygens-Fresnel
principle to compute the diffracted field of real Wolter-I optics, including
measured defects over a wide range of spatial frequencies. Owing to the shallow
angles at play, the computation can be simplified limiting the computation to
the longitudinal profiles, neglecting completely the effect of roundness
errors. Other authors had already proposed similar approaches in the past, but
only in far-field approximation, therefore they could not be applied to the
case of Wolter-I optics, in which two reflections occur in sequence within a
short range. The method we suggest is versatile, as it can be applied to
multiple reflection systems, at any X-ray energy, and regardless of the nominal
shape of the mirrors in the optical system. The method has been implemented in
the WISE code, successfully used to explain the measured PSFs of
multilayer-coated optics for astronomic use, and of a K-B optical system in use
at the FERMI free electron laser.
| [
{
"version": "v1",
"created": "Wed, 16 Sep 2015 00:29:46 GMT"
}
] | 2015-09-17T00:00:00 | [
[
"Spiga",
"D.",
""
],
[
"Raimondi",
"L.",
""
]
] | TITLE: X-ray optical systems: from metrology to Point Spread Function
ABSTRACT: One of the problems often encountered in X-ray mirror manufacturing is
setting proper manufacturing tolerances to guarantee an angular resolution -
often expressed in terms of Point Spread Function (PSF) - as needed by the
specific science goal. To do this, we need an accurate metrological apparatus,
covering a very broad range of spatial frequencies, and an affordable method to
compute the PSF from the metrology dataset. [...] However, the separation
between these spectral ranges is difficult do define exactly, and it is also
unclear how to affordably combine the PSFs, computed with different methods in
different spectral ranges, into a PSF expectation at a given X-ray energy. For
this reason, we have proposed a method entirely based on the Huygens-Fresnel
principle to compute the diffracted field of real Wolter-I optics, including
measured defects over a wide range of spatial frequencies. Owing to the shallow
angles at play, the computation can be simplified limiting the computation to
the longitudinal profiles, neglecting completely the effect of roundness
errors. Other authors had already proposed similar approaches in the past, but
only in far-field approximation, therefore they could not be applied to the
case of Wolter-I optics, in which two reflections occur in sequence within a
short range. The method we suggest is versatile, as it can be applied to
multiple reflection systems, at any X-ray energy, and regardless of the nominal
shape of the mirrors in the optical system. The method has been implemented in
the WISE code, successfully used to explain the measured PSFs of
multilayer-coated optics for astronomic use, and of a K-B optical system in use
at the FERMI free electron laser.
| no_new_dataset | 0.944689 |
1509.04783 | Ziming Zhang | Ziming Zhang, Yuting Chen, and Venkatesh Saligrama | Group Membership Prediction | accepted for ICCV 2015 | null | null | null | cs.CV stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The group membership prediction (GMP) problem involves predicting whether or
not a collection of instances share a certain semantic property. For instance,
in kinship verification given a collection of images, the goal is to predict
whether or not they share a {\it familial} relationship. In this context we
propose a novel probability model and introduce latent {\em view-specific} and
{\em view-shared} random variables to jointly account for the view-specific
appearance and cross-view similarities among data instances. Our model posits
that data from each view is independent conditioned on the shared variables.
This postulate leads to a parametric probability model that decomposes group
membership likelihood into a tensor product of data-independent parameters and
data-dependent factors. We propose learning the data-independent parameters in
a discriminative way with bilinear classifiers, and test our prediction
algorithm on challenging visual recognition tasks such as multi-camera person
re-identification and kinship verification. On most benchmark datasets, our
method can significantly outperform the current state-of-the-art.
| [
{
"version": "v1",
"created": "Wed, 16 Sep 2015 01:22:40 GMT"
}
] | 2015-09-17T00:00:00 | [
[
"Zhang",
"Ziming",
""
],
[
"Chen",
"Yuting",
""
],
[
"Saligrama",
"Venkatesh",
""
]
] | TITLE: Group Membership Prediction
ABSTRACT: The group membership prediction (GMP) problem involves predicting whether or
not a collection of instances share a certain semantic property. For instance,
in kinship verification given a collection of images, the goal is to predict
whether or not they share a {\it familial} relationship. In this context we
propose a novel probability model and introduce latent {\em view-specific} and
{\em view-shared} random variables to jointly account for the view-specific
appearance and cross-view similarities among data instances. Our model posits
that data from each view is independent conditioned on the shared variables.
This postulate leads to a parametric probability model that decomposes group
membership likelihood into a tensor product of data-independent parameters and
data-dependent factors. We propose learning the data-independent parameters in
a discriminative way with bilinear classifiers, and test our prediction
algorithm on challenging visual recognition tasks such as multi-camera person
re-identification and kinship verification. On most benchmark datasets, our
method can significantly outperform the current state-of-the-art.
| no_new_dataset | 0.947866 |
1509.04904 | Tshilidzi Marwala | Pramod Kumar Parida, Tshilidzi Marwala and Snehashish Chakraverty | Causal Model Analysis using Collider v-structure with Negative
Percentage Mapping | null | null | null | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A major problem of causal inference is the arrangement of dependent nodes in
a directed acyclic graph (DAG) with path coefficients and observed confounders.
Path coefficients do not provide the units to measure the strength of
information flowing from one node to the other. Here we proposed the method of
causal structure learning using collider v-structures (CVS) with Negative
Percentage Mapping (NPM) to get selective thresholds of information strength,
to direct the edges and subjective confounders in a DAG. The NPM is used to
scale the strength of information passed through nodes in units of percentage
from interval from 0 to 1. The causal structures are constructed by bottom up
approach using path coefficients, causal directions and confounders, derived
implementing collider v-structure and NPM. The method is self-sufficient to
observe all the latent confounders present in the causal model and capable of
detecting every responsible causal direction. The results are tested for
simulated datasets of non-Gaussian distributions and compared with DirectLiNGAM
and ICA-LiNGAM to check efficiency of the proposed method.
| [
{
"version": "v1",
"created": "Wed, 16 Sep 2015 12:37:30 GMT"
}
] | 2015-09-17T00:00:00 | [
[
"Parida",
"Pramod Kumar",
""
],
[
"Marwala",
"Tshilidzi",
""
],
[
"Chakraverty",
"Snehashish",
""
]
] | TITLE: Causal Model Analysis using Collider v-structure with Negative
Percentage Mapping
ABSTRACT: A major problem of causal inference is the arrangement of dependent nodes in
a directed acyclic graph (DAG) with path coefficients and observed confounders.
Path coefficients do not provide the units to measure the strength of
information flowing from one node to the other. Here we proposed the method of
causal structure learning using collider v-structures (CVS) with Negative
Percentage Mapping (NPM) to get selective thresholds of information strength,
to direct the edges and subjective confounders in a DAG. The NPM is used to
scale the strength of information passed through nodes in units of percentage
from interval from 0 to 1. The causal structures are constructed by bottom up
approach using path coefficients, causal directions and confounders, derived
implementing collider v-structure and NPM. The method is self-sufficient to
observe all the latent confounders present in the causal model and capable of
detecting every responsible causal direction. The results are tested for
simulated datasets of non-Gaussian distributions and compared with DirectLiNGAM
and ICA-LiNGAM to check efficiency of the proposed method.
| no_new_dataset | 0.94743 |
1509.04916 | Mengyang Yu | Li Liu, Mengyang Yu, Ling Shao | Projection Bank: From High-dimensional Data to Medium-length Binary
Codes | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recently, very high-dimensional feature representations, e.g., Fisher Vector,
have achieved excellent performance for visual recognition and retrieval.
However, these lengthy representations always cause extremely heavy
computational and storage costs and even become unfeasible in some large-scale
applications. A few existing techniques can transfer very high-dimensional data
into binary codes, but they still require the reduced code length to be
relatively long to maintain acceptable accuracies. To target a better balance
between computational efficiency and accuracies, in this paper, we propose a
novel embedding method called Binary Projection Bank (BPB), which can
effectively reduce the very high-dimensional representations to
medium-dimensional binary codes without sacrificing accuracies. Instead of
using conventional single linear or bilinear projections, the proposed method
learns a bank of small projections via the max-margin constraint to optimally
preserve the intrinsic data similarity. We have systematically evaluated the
proposed method on three datasets: Flickr 1M, ILSVR2010 and UCF101, showing
competitive retrieval and recognition accuracies compared with state-of-the-art
approaches, but with a significantly smaller memory footprint and lower coding
complexity.
| [
{
"version": "v1",
"created": "Wed, 16 Sep 2015 13:42:42 GMT"
}
] | 2015-09-17T00:00:00 | [
[
"Liu",
"Li",
""
],
[
"Yu",
"Mengyang",
""
],
[
"Shao",
"Ling",
""
]
] | TITLE: Projection Bank: From High-dimensional Data to Medium-length Binary
Codes
ABSTRACT: Recently, very high-dimensional feature representations, e.g., Fisher Vector,
have achieved excellent performance for visual recognition and retrieval.
However, these lengthy representations always cause extremely heavy
computational and storage costs and even become unfeasible in some large-scale
applications. A few existing techniques can transfer very high-dimensional data
into binary codes, but they still require the reduced code length to be
relatively long to maintain acceptable accuracies. To target a better balance
between computational efficiency and accuracies, in this paper, we propose a
novel embedding method called Binary Projection Bank (BPB), which can
effectively reduce the very high-dimensional representations to
medium-dimensional binary codes without sacrificing accuracies. Instead of
using conventional single linear or bilinear projections, the proposed method
learns a bank of small projections via the max-margin constraint to optimally
preserve the intrinsic data similarity. We have systematically evaluated the
proposed method on three datasets: Flickr 1M, ILSVR2010 and UCF101, showing
competitive retrieval and recognition accuracies compared with state-of-the-art
approaches, but with a significantly smaller memory footprint and lower coding
complexity.
| no_new_dataset | 0.948917 |
1509.04942 | Xu Jia | Xu Jia and Efstratios Gavves and Basura Fernando and Tinne Tuytelaars | Guiding Long-Short Term Memory for Image Caption Generation | accepted by ICCV 2015 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work we focus on the problem of image caption generation. We propose
an extension of the long short term memory (LSTM) model, which we coin gLSTM
for short. In particular, we add semantic information extracted from the image
as extra input to each unit of the LSTM block, with the aim of guiding the
model towards solutions that are more tightly coupled to the image content.
Additionally, we explore different length normalization strategies for beam
search in order to prevent from favoring short sentences. On various benchmark
datasets such as Flickr8K, Flickr30K and MS COCO, we obtain results that are on
par with or even outperform the current state-of-the-art.
| [
{
"version": "v1",
"created": "Wed, 16 Sep 2015 15:02:30 GMT"
}
] | 2015-09-17T00:00:00 | [
[
"Jia",
"Xu",
""
],
[
"Gavves",
"Efstratios",
""
],
[
"Fernando",
"Basura",
""
],
[
"Tuytelaars",
"Tinne",
""
]
] | TITLE: Guiding Long-Short Term Memory for Image Caption Generation
ABSTRACT: In this work we focus on the problem of image caption generation. We propose
an extension of the long short term memory (LSTM) model, which we coin gLSTM
for short. In particular, we add semantic information extracted from the image
as extra input to each unit of the LSTM block, with the aim of guiding the
model towards solutions that are more tightly coupled to the image content.
Additionally, we explore different length normalization strategies for beam
search in order to prevent from favoring short sentences. On various benchmark
datasets such as Flickr8K, Flickr30K and MS COCO, we obtain results that are on
par with or even outperform the current state-of-the-art.
| no_new_dataset | 0.951774 |
1509.04954 | Heng Yang | Heng Yang and Renqiao Zhang and Peter Robinson | Human and Sheep Facial Landmarks Localisation by Triplet Interpolated
Features | submitted to WACV2016 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper we present a method for localisation of facial landmarks on
human and sheep. We introduce a new feature extraction scheme called
triplet-interpolated feature used at each iteration of the cascaded shape
regression framework. It is able to extract features from similar semantic
location given an estimated shape, even when head pose variations are large and
the facial landmarks are very sparsely distributed. Furthermore, we study the
impact of training data imbalance on model performance and propose a training
sample augmentation scheme that produces more initialisations for training
samples from the minority. More specifically, the augmentation number for a
training sample is made to be negatively correlated to the value of the fitted
probability density function at the sample's position. We evaluate the proposed
scheme on both human and sheep facial landmarks localisation. On the benchmark
300w human face dataset, we demonstrate the benefits of our proposed methods
and show very competitive performance when comparing to other methods. On a
newly created sheep face dataset, we get very good performance despite the fact
that we only have a limited number of training samples and a set of sparse
landmarks are annotated.
| [
{
"version": "v1",
"created": "Wed, 16 Sep 2015 15:50:01 GMT"
}
] | 2015-09-17T00:00:00 | [
[
"Yang",
"Heng",
""
],
[
"Zhang",
"Renqiao",
""
],
[
"Robinson",
"Peter",
""
]
] | TITLE: Human and Sheep Facial Landmarks Localisation by Triplet Interpolated
Features
ABSTRACT: In this paper we present a method for localisation of facial landmarks on
human and sheep. We introduce a new feature extraction scheme called
triplet-interpolated feature used at each iteration of the cascaded shape
regression framework. It is able to extract features from similar semantic
location given an estimated shape, even when head pose variations are large and
the facial landmarks are very sparsely distributed. Furthermore, we study the
impact of training data imbalance on model performance and propose a training
sample augmentation scheme that produces more initialisations for training
samples from the minority. More specifically, the augmentation number for a
training sample is made to be negatively correlated to the value of the fitted
probability density function at the sample's position. We evaluate the proposed
scheme on both human and sheep facial landmarks localisation. On the benchmark
300w human face dataset, we demonstrate the benefits of our proposed methods
and show very competitive performance when comparing to other methods. On a
newly created sheep face dataset, we get very good performance despite the fact
that we only have a limited number of training samples and a set of sparse
landmarks are annotated.
| new_dataset | 0.959269 |
1509.04397 | Suriya Gunasekar | Suriya Gunasekar, Pradeep Ravikumar, Joydeep Ghosh | Exponential Family Matrix Completion under Structural Constraints | 20 pages, 9 figures | Gunasekar, Suriya, Pradeep Ravikumar, and Joydeep Ghosh.
"Exponential family matrix completion under structural constraints".
Proceedings of The 31st International Conference on Machine Learning, pp.
1917-1925, 2014 | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider the matrix completion problem of recovering a structured matrix
from noisy and partial measurements. Recent works have proposed tractable
estimators with strong statistical guarantees for the case where the underlying
matrix is low--rank, and the measurements consist of a subset, either of the
exact individual entries, or of the entries perturbed by additive Gaussian
noise, which is thus implicitly suited for thin--tailed continuous data.
Arguably, common applications of matrix completion require estimators for (a)
heterogeneous data--types, such as skewed--continuous, count, binary, etc., (b)
for heterogeneous noise models (beyond Gaussian), which capture varied
uncertainty in the measurements, and (c) heterogeneous structural constraints
beyond low--rank, such as block--sparsity, or a superposition structure of
low--rank plus elementwise sparseness, among others. In this paper, we provide
a vastly unified framework for generalized matrix completion by considering a
matrix completion setting wherein the matrix entries are sampled from any
member of the rich family of exponential family distributions; and impose
general structural constraints on the underlying matrix, as captured by a
general regularizer $\mathcal{R}(.)$. We propose a simple convex regularized
$M$--estimator for the generalized framework, and provide a unified and novel
statistical analysis for this general class of estimators. We finally
corroborate our theoretical results on simulated datasets.
| [
{
"version": "v1",
"created": "Tue, 15 Sep 2015 04:49:57 GMT"
}
] | 2015-09-16T00:00:00 | [
[
"Gunasekar",
"Suriya",
""
],
[
"Ravikumar",
"Pradeep",
""
],
[
"Ghosh",
"Joydeep",
""
]
] | TITLE: Exponential Family Matrix Completion under Structural Constraints
ABSTRACT: We consider the matrix completion problem of recovering a structured matrix
from noisy and partial measurements. Recent works have proposed tractable
estimators with strong statistical guarantees for the case where the underlying
matrix is low--rank, and the measurements consist of a subset, either of the
exact individual entries, or of the entries perturbed by additive Gaussian
noise, which is thus implicitly suited for thin--tailed continuous data.
Arguably, common applications of matrix completion require estimators for (a)
heterogeneous data--types, such as skewed--continuous, count, binary, etc., (b)
for heterogeneous noise models (beyond Gaussian), which capture varied
uncertainty in the measurements, and (c) heterogeneous structural constraints
beyond low--rank, such as block--sparsity, or a superposition structure of
low--rank plus elementwise sparseness, among others. In this paper, we provide
a vastly unified framework for generalized matrix completion by considering a
matrix completion setting wherein the matrix entries are sampled from any
member of the rich family of exponential family distributions; and impose
general structural constraints on the underlying matrix, as captured by a
general regularizer $\mathcal{R}(.)$. We propose a simple convex regularized
$M$--estimator for the generalized framework, and provide a unified and novel
statistical analysis for this general class of estimators. We finally
corroborate our theoretical results on simulated datasets.
| no_new_dataset | 0.946001 |
1509.04581 | Zhen Liu | Zhen Liu | Kernelized Deep Convolutional Neural Network for Describing Complex
Images | 9 pages | null | null | null | cs.CV cs.AI cs.IR cs.MM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the impressive capability to capture visual content, deep convolutional
neural networks (CNN) have demon- strated promising performance in various
vision-based ap- plications, such as classification, recognition, and objec- t
detection. However, due to the intrinsic structure design of CNN, for images
with complex content, it achieves lim- ited capability on invariance to
translation, rotation, and re-sizing changes, which is strongly emphasized in
the s- cenario of content-based image retrieval. In this paper, to address this
problem, we proposed a new kernelized deep convolutional neural network. We
first discuss our motiva- tion by an experimental study to demonstrate the
sensitivi- ty of the global CNN feature to the basic geometric trans-
formations. Then, we propose to represent visual content with approximate
invariance to the above geometric trans- formations from a kernelized
perspective. We extract CNN features on the detected object-like patches and
aggregate these patch-level CNN features to form a vectorial repre- sentation
with the Fisher vector model. The effectiveness of our proposed algorithm is
demonstrated on image search application with three benchmark datasets.
| [
{
"version": "v1",
"created": "Tue, 15 Sep 2015 14:35:11 GMT"
}
] | 2015-09-16T00:00:00 | [
[
"Liu",
"Zhen",
""
]
] | TITLE: Kernelized Deep Convolutional Neural Network for Describing Complex
Images
ABSTRACT: With the impressive capability to capture visual content, deep convolutional
neural networks (CNN) have demon- strated promising performance in various
vision-based ap- plications, such as classification, recognition, and objec- t
detection. However, due to the intrinsic structure design of CNN, for images
with complex content, it achieves lim- ited capability on invariance to
translation, rotation, and re-sizing changes, which is strongly emphasized in
the s- cenario of content-based image retrieval. In this paper, to address this
problem, we proposed a new kernelized deep convolutional neural network. We
first discuss our motiva- tion by an experimental study to demonstrate the
sensitivi- ty of the global CNN feature to the basic geometric trans-
formations. Then, we propose to represent visual content with approximate
invariance to the above geometric trans- formations from a kernelized
perspective. We extract CNN features on the detected object-like patches and
aggregate these patch-level CNN features to form a vectorial repre- sentation
with the Fisher vector model. The effectiveness of our proposed algorithm is
demonstrated on image search application with three benchmark datasets.
| no_new_dataset | 0.948775 |
1308.0271 | Qiang Qiu | Qiang Qiu, Rama Chellappa | Compositional Dictionaries for Domain Adaptive Face Recognition | Transactions on Image Processing, 2015 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a dictionary learning approach to compensate for the
transformation of faces due to changes in view point, illumination, resolution,
etc. The key idea of our approach is to force domain-invariant sparse coding,
i.e., design a consistent sparse representation of the same face in different
domains. In this way, classifiers trained on the sparse codes in the source
domain consisting of frontal faces for example can be applied to the target
domain (consisting of faces in different poses, illumination conditions, etc)
without much loss in recognition accuracy. The approach is to first learn a
domain base dictionary, and then describe each domain shift (identity, pose,
illumination) using a sparse representation over the base dictionary. The
dictionary adapted to each domain is expressed as sparse linear combinations of
the base dictionary. In the context of face recognition, with the proposed
compositional dictionary approach, a face image can be decomposed into sparse
representations for a given subject, pose and illumination respectively. This
approach has three advantages: first, the extracted sparse representation for a
subject is consistent across domains and enables pose and illumination
insensitive face recognition. Second, sparse representations for pose and
illumination can subsequently be used to estimate the pose and illumination
condition of a face image. Finally, by composing sparse representations for
subject and the different domains, we can also perform pose alignment and
illumination normalization. Extensive experiments using two public face
datasets are presented to demonstrate the effectiveness of our approach for
face recognition.
| [
{
"version": "v1",
"created": "Thu, 1 Aug 2013 17:27:31 GMT"
},
{
"version": "v2",
"created": "Sat, 12 Sep 2015 20:55:51 GMT"
}
] | 2015-09-15T00:00:00 | [
[
"Qiu",
"Qiang",
""
],
[
"Chellappa",
"Rama",
""
]
] | TITLE: Compositional Dictionaries for Domain Adaptive Face Recognition
ABSTRACT: We present a dictionary learning approach to compensate for the
transformation of faces due to changes in view point, illumination, resolution,
etc. The key idea of our approach is to force domain-invariant sparse coding,
i.e., design a consistent sparse representation of the same face in different
domains. In this way, classifiers trained on the sparse codes in the source
domain consisting of frontal faces for example can be applied to the target
domain (consisting of faces in different poses, illumination conditions, etc)
without much loss in recognition accuracy. The approach is to first learn a
domain base dictionary, and then describe each domain shift (identity, pose,
illumination) using a sparse representation over the base dictionary. The
dictionary adapted to each domain is expressed as sparse linear combinations of
the base dictionary. In the context of face recognition, with the proposed
compositional dictionary approach, a face image can be decomposed into sparse
representations for a given subject, pose and illumination respectively. This
approach has three advantages: first, the extracted sparse representation for a
subject is consistent across domains and enables pose and illumination
insensitive face recognition. Second, sparse representations for pose and
illumination can subsequently be used to estimate the pose and illumination
condition of a face image. Finally, by composing sparse representations for
subject and the different domains, we can also perform pose alignment and
illumination normalization. Extensive experiments using two public face
datasets are presented to demonstrate the effectiveness of our approach for
face recognition.
| no_new_dataset | 0.94887 |
1409.4988 | Filippo Maria Bianchi | Filippo Maria Bianchi, Enrico Maiorino, Lorenzo Livi, Antonello Rizzi
and Alireza Sadeghian | An Agent-Based Algorithm exploiting Multiple Local Dissimilarities for
Clusters Mining and Knowledge Discovery | null | null | 10.1007/s00500-015-1876-1 | null | cs.LG cs.DC cs.MA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a multi-agent algorithm able to automatically discover relevant
regularities in a given dataset, determining at the same time the set of
configurations of the adopted parametric dissimilarity measure yielding compact
and separated clusters. Each agent operates independently by performing a
Markovian random walk on a suitable weighted graph representation of the input
dataset. Such a weighted graph representation is induced by the specific
parameter configuration of the dissimilarity measure adopted by the agent,
which searches and takes decisions autonomously for one cluster at a time.
Results show that the algorithm is able to discover parameter configurations
that yield a consistent and interpretable collection of clusters. Moreover, we
demonstrate that our algorithm shows comparable performances with other similar
state-of-the-art algorithms when facing specific clustering problems.
| [
{
"version": "v1",
"created": "Wed, 17 Sep 2014 14:39:37 GMT"
}
] | 2015-09-15T00:00:00 | [
[
"Bianchi",
"Filippo Maria",
""
],
[
"Maiorino",
"Enrico",
""
],
[
"Livi",
"Lorenzo",
""
],
[
"Rizzi",
"Antonello",
""
],
[
"Sadeghian",
"Alireza",
""
]
] | TITLE: An Agent-Based Algorithm exploiting Multiple Local Dissimilarities for
Clusters Mining and Knowledge Discovery
ABSTRACT: We propose a multi-agent algorithm able to automatically discover relevant
regularities in a given dataset, determining at the same time the set of
configurations of the adopted parametric dissimilarity measure yielding compact
and separated clusters. Each agent operates independently by performing a
Markovian random walk on a suitable weighted graph representation of the input
dataset. Such a weighted graph representation is induced by the specific
parameter configuration of the dissimilarity measure adopted by the agent,
which searches and takes decisions autonomously for one cluster at a time.
Results show that the algorithm is able to discover parameter configurations
that yield a consistent and interpretable collection of clusters. Moreover, we
demonstrate that our algorithm shows comparable performances with other similar
state-of-the-art algorithms when facing specific clustering problems.
| no_new_dataset | 0.950319 |
1509.02587 | Bardia Yousefi | Bardia Yousefi, C.K. Loo | A Dual Fast and Slow Feature Interaction in Biologically Inspired Visual
Recognition of Human Action | This paper has been withdrawn by the author due to a mistake in file | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Computational neuroscience studies that have examined human visual system
through functional magnetic resonance imaging (fMRI) have identified a model
where the mammalian brain pursues two distinct pathways (for recognition of
biological movement tasks). In the brain, dorsal stream analyzes the
information of motion (optical flow), which is the fast features, and ventral
stream (form pathway) analyzes form information (through active basis model
based incremental slow feature analysis ) as slow features. The proposed
approach suggests the motion perception of the human visual system composes of
fast and slow feature interactions that identifies biological movements. Form
features in the visual system biologically follows the application of active
basis model with incremental slow feature analysis for the extraction of the
slowest form features of human objects movements in the ventral stream.
Applying incremental slow feature analysis provides an opportunity to use the
action prototypes. To extract the slowest features episodic observation is
required but the fast features updates the processing of motion information in
every frames. Experimental results have shown promising accuracy for the
proposed model and good performance with two datasets (KTH and Weizmann).
| [
{
"version": "v1",
"created": "Wed, 9 Sep 2015 00:31:53 GMT"
},
{
"version": "v2",
"created": "Sat, 12 Sep 2015 23:44:42 GMT"
}
] | 2015-09-15T00:00:00 | [
[
"Yousefi",
"Bardia",
""
],
[
"Loo",
"C. K.",
""
]
] | TITLE: A Dual Fast and Slow Feature Interaction in Biologically Inspired Visual
Recognition of Human Action
ABSTRACT: Computational neuroscience studies that have examined human visual system
through functional magnetic resonance imaging (fMRI) have identified a model
where the mammalian brain pursues two distinct pathways (for recognition of
biological movement tasks). In the brain, dorsal stream analyzes the
information of motion (optical flow), which is the fast features, and ventral
stream (form pathway) analyzes form information (through active basis model
based incremental slow feature analysis ) as slow features. The proposed
approach suggests the motion perception of the human visual system composes of
fast and slow feature interactions that identifies biological movements. Form
features in the visual system biologically follows the application of active
basis model with incremental slow feature analysis for the extraction of the
slowest form features of human objects movements in the ventral stream.
Applying incremental slow feature analysis provides an opportunity to use the
action prototypes. To extract the slowest features episodic observation is
required but the fast features updates the processing of motion information in
every frames. Experimental results have shown promising accuracy for the
proposed model and good performance with two datasets (KTH and Weizmann).
| no_new_dataset | 0.954223 |
1509.02730 | Rangeet Mitra | Rangeet Mitra and Vimal Bhatia | Finite Dictionary Variants of the Diffusion KLMS Algorithm | null | null | null | null | cs.SY cs.DC cs.IT cs.LG math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The diffusion based distributed learning approaches have been found to be a
viable solution for learning over linearly separable datasets over a network.
However, approaches till date are suitable for linearly separable datasets and
need to be extended to scenarios in which we need to learn a non-linearity. In
such scenarios, the recently proposed diffusion kernel least mean squares
(KLMS) has been found to be performing better than diffusion least mean squares
(LMS). The drawback of diffusion KLMS is that it requires infinite storage for
observations (also called dictionary). This paper formulates the diffusion KLMS
in a fixed budget setting such that the storage requirement is curtailed while
maintaining appreciable performance in terms of convergence. Simulations have
been carried out to validate the two newly proposed algorithms named as
quantised diffusion KLMS (QDKLMS) and fixed budget diffusion KLMS (FBDKLMS)
against KLMS, which indicate that both the proposed algorithms deliver better
performance as compared to the KLMS while reducing the dictionary size storage
requirement.
| [
{
"version": "v1",
"created": "Wed, 9 Sep 2015 11:38:01 GMT"
}
] | 2015-09-15T00:00:00 | [
[
"Mitra",
"Rangeet",
""
],
[
"Bhatia",
"Vimal",
""
]
] | TITLE: Finite Dictionary Variants of the Diffusion KLMS Algorithm
ABSTRACT: The diffusion based distributed learning approaches have been found to be a
viable solution for learning over linearly separable datasets over a network.
However, approaches till date are suitable for linearly separable datasets and
need to be extended to scenarios in which we need to learn a non-linearity. In
such scenarios, the recently proposed diffusion kernel least mean squares
(KLMS) has been found to be performing better than diffusion least mean squares
(LMS). The drawback of diffusion KLMS is that it requires infinite storage for
observations (also called dictionary). This paper formulates the diffusion KLMS
in a fixed budget setting such that the storage requirement is curtailed while
maintaining appreciable performance in terms of convergence. Simulations have
been carried out to validate the two newly proposed algorithms named as
quantised diffusion KLMS (QDKLMS) and fixed budget diffusion KLMS (FBDKLMS)
against KLMS, which indicate that both the proposed algorithms deliver better
performance as compared to the KLMS while reducing the dictionary size storage
requirement.
| no_new_dataset | 0.949012 |
1509.03302 | Matt Barnes | Matt Barnes, Kyle Miller, Artur Dubrawski | Performance Bounds for Pairwise Entity Resolution | null | null | null | null | stat.ML cs.CY cs.DB cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | One significant challenge to scaling entity resolution algorithms to massive
datasets is understanding how performance changes after moving beyond the realm
of small, manually labeled reference datasets. Unlike traditional machine
learning tasks, when an entity resolution algorithm performs well on small
hold-out datasets, there is no guarantee this performance holds on larger
hold-out datasets. We prove simple bounding properties between the performance
of a match function on a small validation set and the performance of a pairwise
entity resolution algorithm on arbitrarily sized datasets. Thus, our approach
enables optimization of pairwise entity resolution algorithms for large
datasets, using a small set of labeled data.
| [
{
"version": "v1",
"created": "Thu, 10 Sep 2015 19:58:44 GMT"
}
] | 2015-09-15T00:00:00 | [
[
"Barnes",
"Matt",
""
],
[
"Miller",
"Kyle",
""
],
[
"Dubrawski",
"Artur",
""
]
] | TITLE: Performance Bounds for Pairwise Entity Resolution
ABSTRACT: One significant challenge to scaling entity resolution algorithms to massive
datasets is understanding how performance changes after moving beyond the realm
of small, manually labeled reference datasets. Unlike traditional machine
learning tasks, when an entity resolution algorithm performs well on small
hold-out datasets, there is no guarantee this performance holds on larger
hold-out datasets. We prove simple bounding properties between the performance
of a match function on a small validation set and the performance of a pairwise
entity resolution algorithm on arbitrarily sized datasets. Thus, our approach
enables optimization of pairwise entity resolution algorithms for large
datasets, using a small set of labeled data.
| no_new_dataset | 0.949435 |
1509.03844 | Yiannis Andreopoulos | Alhabib Abbas, Nikos Deligiannis and Yiannis Andreopoulos | Vectors of Locally Aggregated Centers for Compact Video Representation | Proc. IEEE International Conference on Multimedia and Expo, ICME
2015, Torino, Italy | null | 10.1109/ICME.2015.7177501 | null | cs.MM cs.CV cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a novel vector aggregation technique for compact video
representation, with application in accurate similarity detection within large
video datasets. The current state-of-the-art in visual search is formed by the
vector of locally aggregated descriptors (VLAD) of Jegou et. al. VLAD generates
compact video representations based on scale-invariant feature transform (SIFT)
vectors (extracted per frame) and local feature centers computed over a
training set. With the aim to increase robustness to visual distortions, we
propose a new approach that operates at a coarser level in the feature
representation. We create vectors of locally aggregated centers (VLAC) by first
clustering SIFT features to obtain local feature centers (LFCs) and then
encoding the latter with respect to given centers of local feature centers
(CLFCs), extracted from a training set. The sum-of-differences between the LFCs
and the CLFCs are aggregated to generate an extremely-compact video description
used for accurate video segment similarity detection. Experimentation using a
video dataset, comprising more than 1000 minutes of content from the Open Video
Project, shows that VLAC obtains substantial gains in terms of mean Average
Precision (mAP) against VLAD and the hyper-pooling method of Douze et. al.,
under the same compaction factor and the same set of distortions.
| [
{
"version": "v1",
"created": "Sun, 13 Sep 2015 13:06:36 GMT"
}
] | 2015-09-15T00:00:00 | [
[
"Abbas",
"Alhabib",
""
],
[
"Deligiannis",
"Nikos",
""
],
[
"Andreopoulos",
"Yiannis",
""
]
] | TITLE: Vectors of Locally Aggregated Centers for Compact Video Representation
ABSTRACT: We propose a novel vector aggregation technique for compact video
representation, with application in accurate similarity detection within large
video datasets. The current state-of-the-art in visual search is formed by the
vector of locally aggregated descriptors (VLAD) of Jegou et. al. VLAD generates
compact video representations based on scale-invariant feature transform (SIFT)
vectors (extracted per frame) and local feature centers computed over a
training set. With the aim to increase robustness to visual distortions, we
propose a new approach that operates at a coarser level in the feature
representation. We create vectors of locally aggregated centers (VLAC) by first
clustering SIFT features to obtain local feature centers (LFCs) and then
encoding the latter with respect to given centers of local feature centers
(CLFCs), extracted from a training set. The sum-of-differences between the LFCs
and the CLFCs are aggregated to generate an extremely-compact video description
used for accurate video segment similarity detection. Experimentation using a
video dataset, comprising more than 1000 minutes of content from the Open Video
Project, shows that VLAC obtains substantial gains in terms of mean Average
Precision (mAP) against VLAD and the hyper-pooling method of Douze et. al.,
under the same compaction factor and the same set of distortions.
| no_new_dataset | 0.947721 |
1509.03936 | Zhanpeng Zhang | Zhanpeng Zhang, Ping Luo, Chen Change Loy, Xiaoou Tang | Learning Social Relation Traits from Face Images | To appear in International Conference on Computer Vision (ICCV) 2015 | null | null | null | cs.CV cs.CY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Social relation defines the association, e.g, warm, friendliness, and
dominance, between two or more people. Motivated by psychological studies, we
investigate if such fine-grained and high-level relation traits can be
characterised and quantified from face images in the wild. To address this
challenging problem we propose a deep model that learns a rich face
representation to capture gender, expression, head pose, and age-related
attributes, and then performs pairwise-face reasoning for relation prediction.
To learn from heterogeneous attribute sources, we formulate a new network
architecture with a bridging layer to leverage the inherent correspondences
among these datasets. It can also cope with missing target attribute labels.
Extensive experiments show that our approach is effective for fine-grained
social relation learning in images and videos.
| [
{
"version": "v1",
"created": "Mon, 14 Sep 2015 03:02:36 GMT"
}
] | 2015-09-15T00:00:00 | [
[
"Zhang",
"Zhanpeng",
""
],
[
"Luo",
"Ping",
""
],
[
"Loy",
"Chen Change",
""
],
[
"Tang",
"Xiaoou",
""
]
] | TITLE: Learning Social Relation Traits from Face Images
ABSTRACT: Social relation defines the association, e.g, warm, friendliness, and
dominance, between two or more people. Motivated by psychological studies, we
investigate if such fine-grained and high-level relation traits can be
characterised and quantified from face images in the wild. To address this
challenging problem we propose a deep model that learns a rich face
representation to capture gender, expression, head pose, and age-related
attributes, and then performs pairwise-face reasoning for relation prediction.
To learn from heterogeneous attribute sources, we formulate a new network
architecture with a bridging layer to leverage the inherent correspondences
among these datasets. It can also cope with missing target attribute labels.
Extensive experiments show that our approach is effective for fine-grained
social relation learning in images and videos.
| no_new_dataset | 0.949716 |
1509.03956 | Francesco Solera | Francesco Solera, Simone Calderara and Rita Cucchiara | Learning to Divide and Conquer for Online Multi-Target Tracking | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Online Multiple Target Tracking (MTT) is often addressed within the
tracking-by-detection paradigm. Detections are previously extracted
independently in each frame and then objects trajectories are built by
maximizing specifically designed coherence functions. Nevertheless, ambiguities
arise in presence of occlusions or detection errors. In this paper we claim
that the ambiguities in tracking could be solved by a selective use of the
features, by working with more reliable features if possible and exploiting a
deeper representation of the target only if necessary. To this end, we propose
an online divide and conquer tracker for static camera scenes, which partitions
the assignment problem in local subproblems and solves them by selectively
choosing and combining the best features. The complete framework is cast as a
structural learning task that unifies these phases and learns tracker
parameters from examples. Experiments on two different datasets highlights a
significant improvement of tracking performances (MOTA +10%) over the state of
the art.
| [
{
"version": "v1",
"created": "Mon, 14 Sep 2015 05:25:52 GMT"
}
] | 2015-09-15T00:00:00 | [
[
"Solera",
"Francesco",
""
],
[
"Calderara",
"Simone",
""
],
[
"Cucchiara",
"Rita",
""
]
] | TITLE: Learning to Divide and Conquer for Online Multi-Target Tracking
ABSTRACT: Online Multiple Target Tracking (MTT) is often addressed within the
tracking-by-detection paradigm. Detections are previously extracted
independently in each frame and then objects trajectories are built by
maximizing specifically designed coherence functions. Nevertheless, ambiguities
arise in presence of occlusions or detection errors. In this paper we claim
that the ambiguities in tracking could be solved by a selective use of the
features, by working with more reliable features if possible and exploiting a
deeper representation of the target only if necessary. To this end, we propose
an online divide and conquer tracker for static camera scenes, which partitions
the assignment problem in local subproblems and solves them by selectively
choosing and combining the best features. The complete framework is cast as a
structural learning task that unifies these phases and learns tracker
parameters from examples. Experiments on two different datasets highlights a
significant improvement of tracking performances (MOTA +10%) over the state of
the art.
| no_new_dataset | 0.94545 |
1509.03413 | Saikat Basu | Saikat Basu, Manohar Karki, Sangram Ganguly, Robert DiBiano, Supratik
Mukhopadhyay and Ramakrishna Nemani | Learning Sparse Feature Representations using Probabilistic Quadtrees
and Deep Belief Nets | Published in the European Symposium on Artificial Neural Networks,
ESANN 2015 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Learning sparse feature representations is a useful instrument for solving an
unsupervised learning problem. In this paper, we present three labeled
handwritten digit datasets, collectively called n-MNIST. Then, we propose a
novel framework for the classification of handwritten digits that learns sparse
representations using probabilistic quadtrees and Deep Belief Nets. On the
MNIST and n-MNIST datasets, our framework shows promising results and
significantly outperforms traditional Deep Belief Networks.
| [
{
"version": "v1",
"created": "Fri, 11 Sep 2015 08:13:35 GMT"
}
] | 2015-09-14T00:00:00 | [
[
"Basu",
"Saikat",
""
],
[
"Karki",
"Manohar",
""
],
[
"Ganguly",
"Sangram",
""
],
[
"DiBiano",
"Robert",
""
],
[
"Mukhopadhyay",
"Supratik",
""
],
[
"Nemani",
"Ramakrishna",
""
]
] | TITLE: Learning Sparse Feature Representations using Probabilistic Quadtrees
and Deep Belief Nets
ABSTRACT: Learning sparse feature representations is a useful instrument for solving an
unsupervised learning problem. In this paper, we present three labeled
handwritten digit datasets, collectively called n-MNIST. Then, we propose a
novel framework for the classification of handwritten digits that learns sparse
representations using probabilistic quadtrees and Deep Belief Nets. On the
MNIST and n-MNIST datasets, our framework shows promising results and
significantly outperforms traditional Deep Belief Networks.
| new_dataset | 0.944689 |
1509.03456 | Abdeslam El Harraj | Abdeslam El Harraj, Naoufal Raissouni | OCR accuracy improvement on document images through a novel
pre-processing approach | null | Signal & Image Processing : An International Journal (SIPIJ)
Vol.6, No.4, August 2015 | 10.5121/sipij.2015.6401 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Digital camera and mobile document image acquisition are new trends arising
in the world of Optical Character Recognition and text detection. In some
cases, such process integrates many distortions and produces poorly scanned
text or text-photo images and natural images, leading to an unreliable OCR
digitization. In this paper, we present a novel nonparametric and unsupervised
method to compensate for undesirable document image distortions aiming to
optimally improve OCR accuracy. Our approach relies on a very efficient stack
of document image enhancing techniques to recover deformation of the entire
document image. First, we propose a local brightness and contrast adjustment
method to effectively handle lighting variations and the irregular distribution
of image illumination. Second, we use an optimized greyscale conversion
algorithm to transform our document image to greyscale level. Third, we sharpen
the useful information in the resulting greyscale image using Un-sharp Masking
method. Finally, an optimal global binarization approach is used to prepare the
final document image to OCR recognition. The proposed approach can
significantly improve text detection rate and optical character recognition
accuracy. To demonstrate the efficiency of our approach, an exhaustive
experimentation on a standard dataset is presented.
| [
{
"version": "v1",
"created": "Fri, 11 Sep 2015 10:52:52 GMT"
}
] | 2015-09-14T00:00:00 | [
[
"Harraj",
"Abdeslam El",
""
],
[
"Raissouni",
"Naoufal",
""
]
] | TITLE: OCR accuracy improvement on document images through a novel
pre-processing approach
ABSTRACT: Digital camera and mobile document image acquisition are new trends arising
in the world of Optical Character Recognition and text detection. In some
cases, such process integrates many distortions and produces poorly scanned
text or text-photo images and natural images, leading to an unreliable OCR
digitization. In this paper, we present a novel nonparametric and unsupervised
method to compensate for undesirable document image distortions aiming to
optimally improve OCR accuracy. Our approach relies on a very efficient stack
of document image enhancing techniques to recover deformation of the entire
document image. First, we propose a local brightness and contrast adjustment
method to effectively handle lighting variations and the irregular distribution
of image illumination. Second, we use an optimized greyscale conversion
algorithm to transform our document image to greyscale level. Third, we sharpen
the useful information in the resulting greyscale image using Un-sharp Masking
method. Finally, an optimal global binarization approach is used to prepare the
final document image to OCR recognition. The proposed approach can
significantly improve text detection rate and optical character recognition
accuracy. To demonstrate the efficiency of our approach, an exhaustive
experimentation on a standard dataset is presented.
| new_dataset | 0.966851 |
1509.03602 | Saikat Basu | Saikat Basu, Sangram Ganguly, Supratik Mukhopadhyay, Robert DiBiano,
Manohar Karki and Ramakrishna Nemani | DeepSat - A Learning framework for Satellite Imagery | Paper was accepted at ACM SIGSPATIAL 2015 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Satellite image classification is a challenging problem that lies at the
crossroads of remote sensing, computer vision, and machine learning. Due to the
high variability inherent in satellite data, most of the current object
classification approaches are not suitable for handling satellite datasets. The
progress of satellite image analytics has also been inhibited by the lack of a
single labeled high-resolution dataset with multiple class labels. The
contributions of this paper are twofold - (1) first, we present two new
satellite datasets called SAT-4 and SAT-6, and (2) then, we propose a
classification framework that extracts features from an input image, normalizes
them and feeds the normalized feature vectors to a Deep Belief Network for
classification. On the SAT-4 dataset, our best network produces a
classification accuracy of 97.95% and outperforms three state-of-the-art object
recognition algorithms, namely - Deep Belief Networks, Convolutional Neural
Networks and Stacked Denoising Autoencoders by ~11%. On SAT-6, it produces a
classification accuracy of 93.9% and outperforms the other algorithms by ~15%.
Comparative studies with a Random Forest classifier show the advantage of an
unsupervised learning approach over traditional supervised learning techniques.
A statistical analysis based on Distribution Separability Criterion and
Intrinsic Dimensionality Estimation substantiates the effectiveness of our
approach in learning better representations for satellite imagery.
| [
{
"version": "v1",
"created": "Fri, 11 Sep 2015 18:32:51 GMT"
}
] | 2015-09-14T00:00:00 | [
[
"Basu",
"Saikat",
""
],
[
"Ganguly",
"Sangram",
""
],
[
"Mukhopadhyay",
"Supratik",
""
],
[
"DiBiano",
"Robert",
""
],
[
"Karki",
"Manohar",
""
],
[
"Nemani",
"Ramakrishna",
""
]
] | TITLE: DeepSat - A Learning framework for Satellite Imagery
ABSTRACT: Satellite image classification is a challenging problem that lies at the
crossroads of remote sensing, computer vision, and machine learning. Due to the
high variability inherent in satellite data, most of the current object
classification approaches are not suitable for handling satellite datasets. The
progress of satellite image analytics has also been inhibited by the lack of a
single labeled high-resolution dataset with multiple class labels. The
contributions of this paper are twofold - (1) first, we present two new
satellite datasets called SAT-4 and SAT-6, and (2) then, we propose a
classification framework that extracts features from an input image, normalizes
them and feeds the normalized feature vectors to a Deep Belief Network for
classification. On the SAT-4 dataset, our best network produces a
classification accuracy of 97.95% and outperforms three state-of-the-art object
recognition algorithms, namely - Deep Belief Networks, Convolutional Neural
Networks and Stacked Denoising Autoencoders by ~11%. On SAT-6, it produces a
classification accuracy of 93.9% and outperforms the other algorithms by ~15%.
Comparative studies with a Random Forest classifier show the advantage of an
unsupervised learning approach over traditional supervised learning techniques.
A statistical analysis based on Distribution Separability Criterion and
Intrinsic Dimensionality Estimation substantiates the effectiveness of our
approach in learning better representations for satellite imagery.
| new_dataset | 0.950549 |
1407.5599 | Bo Dai | Bo Dai, Bo Xie, Niao He, Yingyu Liang, Anant Raj, Maria-Florina
Balcan, Le Song | Scalable Kernel Methods via Doubly Stochastic Gradients | 32 pages, 22 figures | null | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The general perception is that kernel methods are not scalable, and neural
nets are the methods of choice for nonlinear learning problems. Or have we
simply not tried hard enough for kernel methods? Here we propose an approach
that scales up kernel methods using a novel concept called "doubly stochastic
functional gradients". Our approach relies on the fact that many kernel methods
can be expressed as convex optimization problems, and we solve the problems by
making two unbiased stochastic approximations to the functional gradient, one
using random training points and another using random functions associated with
the kernel, and then descending using this noisy functional gradient. We show
that a function produced by this procedure after $t$ iterations converges to
the optimal function in the reproducing kernel Hilbert space in rate $O(1/t)$,
and achieves a generalization performance of $O(1/\sqrt{t})$. This doubly
stochasticity also allows us to avoid keeping the support vectors and to
implement the algorithm in a small memory footprint, which is linear in number
of iterations and independent of data dimension. Our approach can readily scale
kernel methods up to the regimes which are dominated by neural nets. We show
that our method can achieve competitive performance to neural nets in datasets
such as 8 million handwritten digits from MNIST, 2.3 million energy materials
from MolecularSpace, and 1 million photos from ImageNet.
| [
{
"version": "v1",
"created": "Mon, 21 Jul 2014 19:05:47 GMT"
},
{
"version": "v2",
"created": "Tue, 5 Aug 2014 17:58:57 GMT"
},
{
"version": "v3",
"created": "Tue, 23 Sep 2014 15:39:03 GMT"
},
{
"version": "v4",
"created": "Thu, 10 Sep 2015 16:40:45 GMT"
}
] | 2015-09-11T00:00:00 | [
[
"Dai",
"Bo",
""
],
[
"Xie",
"Bo",
""
],
[
"He",
"Niao",
""
],
[
"Liang",
"Yingyu",
""
],
[
"Raj",
"Anant",
""
],
[
"Balcan",
"Maria-Florina",
""
],
[
"Song",
"Le",
""
]
] | TITLE: Scalable Kernel Methods via Doubly Stochastic Gradients
ABSTRACT: The general perception is that kernel methods are not scalable, and neural
nets are the methods of choice for nonlinear learning problems. Or have we
simply not tried hard enough for kernel methods? Here we propose an approach
that scales up kernel methods using a novel concept called "doubly stochastic
functional gradients". Our approach relies on the fact that many kernel methods
can be expressed as convex optimization problems, and we solve the problems by
making two unbiased stochastic approximations to the functional gradient, one
using random training points and another using random functions associated with
the kernel, and then descending using this noisy functional gradient. We show
that a function produced by this procedure after $t$ iterations converges to
the optimal function in the reproducing kernel Hilbert space in rate $O(1/t)$,
and achieves a generalization performance of $O(1/\sqrt{t})$. This doubly
stochasticity also allows us to avoid keeping the support vectors and to
implement the algorithm in a small memory footprint, which is linear in number
of iterations and independent of data dimension. Our approach can readily scale
kernel methods up to the regimes which are dominated by neural nets. We show
that our method can achieve competitive performance to neural nets in datasets
such as 8 million handwritten digits from MNIST, 2.3 million energy materials
from MolecularSpace, and 1 million photos from ImageNet.
| no_new_dataset | 0.946498 |
1505.05253 | Jun Feng | Jun Feng, Mantong Zhou, Yu Hao, Minlie Huang and Xiaoyan Zhu | Knowlege Graph Embedding by Flexible Translation | This paper has been withdraw by the author due to an error in sec3.1 | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Knowledge graph embedding refers to projecting entities and relations in
knowledge graph into continuous vector spaces. State-of-the-art methods, such
as TransE, TransH, and TransR build embeddings by treating relation as
translation from head entity to tail entity. However, previous models can not
deal with reflexive/one-to-many/many-to-one/many-to-many relations properly, or
lack of scalability and efficiency. Thus, we propose a novel method, flexible
translation, named TransF, to address the above issues. TransF regards relation
as translation between head entity vector and tail entity vector with flexible
magnitude. To evaluate the proposed model, we conduct link prediction and
triple classification on benchmark datasets. Experimental results show that our
method remarkably improve the performance compared with several
state-of-the-art baselines.
| [
{
"version": "v1",
"created": "Wed, 20 May 2015 05:57:32 GMT"
},
{
"version": "v2",
"created": "Thu, 10 Sep 2015 03:48:55 GMT"
}
] | 2015-09-11T00:00:00 | [
[
"Feng",
"Jun",
""
],
[
"Zhou",
"Mantong",
""
],
[
"Hao",
"Yu",
""
],
[
"Huang",
"Minlie",
""
],
[
"Zhu",
"Xiaoyan",
""
]
] | TITLE: Knowlege Graph Embedding by Flexible Translation
ABSTRACT: Knowledge graph embedding refers to projecting entities and relations in
knowledge graph into continuous vector spaces. State-of-the-art methods, such
as TransE, TransH, and TransR build embeddings by treating relation as
translation from head entity to tail entity. However, previous models can not
deal with reflexive/one-to-many/many-to-one/many-to-many relations properly, or
lack of scalability and efficiency. Thus, we propose a novel method, flexible
translation, named TransF, to address the above issues. TransF regards relation
as translation between head entity vector and tail entity vector with flexible
magnitude. To evaluate the proposed model, we conduct link prediction and
triple classification on benchmark datasets. Experimental results show that our
method remarkably improve the performance compared with several
state-of-the-art baselines.
| no_new_dataset | 0.947672 |
1509.02954 | Joseph Wang | Joseph Wang, Kirill Trapeznikov, Venkatesh Saligrama | Sensor Selection by Linear Programming | null | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We learn sensor trees from training data to minimize sensor acquisition costs
during test time. Our system adaptively selects sensors at each stage if
necessary to make a confident classification. We pose the problem as empirical
risk minimization over the choice of trees and node decision rules. We
decompose the problem, which is known to be intractable, into combinatorial
(tree structures) and continuous parts (node decision rules) and propose to
solve them separately. Using training data we greedily solve for the
combinatorial tree structures and for the continuous part, which is a
non-convex multilinear objective function, we derive convex surrogate loss
functions that are piecewise linear. The resulting problem can be cast as a
linear program and has the advantage of guaranteed convergence, global
optimality, repeatability and computational efficiency. We show that our
proposed approach outperforms the state-of-art on a number of benchmark
datasets.
| [
{
"version": "v1",
"created": "Wed, 9 Sep 2015 21:15:32 GMT"
}
] | 2015-09-11T00:00:00 | [
[
"Wang",
"Joseph",
""
],
[
"Trapeznikov",
"Kirill",
""
],
[
"Saligrama",
"Venkatesh",
""
]
] | TITLE: Sensor Selection by Linear Programming
ABSTRACT: We learn sensor trees from training data to minimize sensor acquisition costs
during test time. Our system adaptively selects sensors at each stage if
necessary to make a confident classification. We pose the problem as empirical
risk minimization over the choice of trees and node decision rules. We
decompose the problem, which is known to be intractable, into combinatorial
(tree structures) and continuous parts (node decision rules) and propose to
solve them separately. Using training data we greedily solve for the
combinatorial tree structures and for the continuous part, which is a
non-convex multilinear objective function, we derive convex surrogate loss
functions that are piecewise linear. The resulting problem can be cast as a
linear program and has the advantage of guaranteed convergence, global
optimality, repeatability and computational efficiency. We show that our
proposed approach outperforms the state-of-art on a number of benchmark
datasets.
| no_new_dataset | 0.948251 |
1509.03005 | David Balduzzi | David Balduzzi, Muhammad Ghifary | Compatible Value Gradients for Reinforcement Learning of Continuous Deep
Policies | 27 pages | null | null | null | cs.LG cs.AI cs.NE stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper proposes GProp, a deep reinforcement learning algorithm for
continuous policies with compatible function approximation. The algorithm is
based on two innovations. Firstly, we present a temporal-difference based
method for learning the gradient of the value-function. Secondly, we present
the deviator-actor-critic (DAC) model, which comprises three neural networks
that estimate the value function, its gradient, and determine the actor's
policy respectively. We evaluate GProp on two challenging tasks: a contextual
bandit problem constructed from nonparametric regression datasets that is
designed to probe the ability of reinforcement learning algorithms to
accurately estimate gradients; and the octopus arm, a challenging reinforcement
learning benchmark. GProp is competitive with fully supervised methods on the
bandit task and achieves the best performance to date on the octopus arm.
| [
{
"version": "v1",
"created": "Thu, 10 Sep 2015 04:14:54 GMT"
}
] | 2015-09-11T00:00:00 | [
[
"Balduzzi",
"David",
""
],
[
"Ghifary",
"Muhammad",
""
]
] | TITLE: Compatible Value Gradients for Reinforcement Learning of Continuous Deep
Policies
ABSTRACT: This paper proposes GProp, a deep reinforcement learning algorithm for
continuous policies with compatible function approximation. The algorithm is
based on two innovations. Firstly, we present a temporal-difference based
method for learning the gradient of the value-function. Secondly, we present
the deviator-actor-critic (DAC) model, which comprises three neural networks
that estimate the value function, its gradient, and determine the actor's
policy respectively. We evaluate GProp on two challenging tasks: a contextual
bandit problem constructed from nonparametric regression datasets that is
designed to probe the ability of reinforcement learning algorithms to
accurately estimate gradients; and the octopus arm, a challenging reinforcement
learning benchmark. GProp is competitive with fully supervised methods on the
bandit task and achieves the best performance to date on the octopus arm.
| no_new_dataset | 0.944074 |
1509.03247 | Arindam Chaudhuri AC | Arindam Chaudhuri | An Epsilon Hierarchical Fuzzy Twin Support Vector Regression | Research work at Samsung Research and Development Institute Delhi | null | null | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The research presents epsilon hierarchical fuzzy twin support vector
regression based on epsilon fuzzy twin support vector regression and epsilon
twin support vector regression. Epsilon FTSVR is achieved by incorporating
trapezoidal fuzzy numbers to epsilon TSVR which takes care of uncertainty
existing in forecasting problems. Epsilon FTSVR determines a pair of epsilon
insensitive proximal functions by solving two related quadratic programming
problems. The structural risk minimization principle is implemented by
introducing regularization term in primal problems of epsilon FTSVR. This
yields dual stable positive definite problems which improves regression
performance. Epsilon FTSVR is then reformulated as epsilon HFTSVR consisting of
a set of hierarchical layers each containing epsilon FTSVR. Experimental
results on both synthetic and real datasets reveal that epsilon HFTSVR has
remarkable generalization performance with minimum training time.
| [
{
"version": "v1",
"created": "Thu, 10 Sep 2015 17:37:20 GMT"
}
] | 2015-09-11T00:00:00 | [
[
"Chaudhuri",
"Arindam",
""
]
] | TITLE: An Epsilon Hierarchical Fuzzy Twin Support Vector Regression
ABSTRACT: The research presents epsilon hierarchical fuzzy twin support vector
regression based on epsilon fuzzy twin support vector regression and epsilon
twin support vector regression. Epsilon FTSVR is achieved by incorporating
trapezoidal fuzzy numbers to epsilon TSVR which takes care of uncertainty
existing in forecasting problems. Epsilon FTSVR determines a pair of epsilon
insensitive proximal functions by solving two related quadratic programming
problems. The structural risk minimization principle is implemented by
introducing regularization term in primal problems of epsilon FTSVR. This
yields dual stable positive definite problems which improves regression
performance. Epsilon FTSVR is then reformulated as epsilon HFTSVR consisting of
a set of hierarchical layers each containing epsilon FTSVR. Experimental
results on both synthetic and real datasets reveal that epsilon HFTSVR has
remarkable generalization performance with minimum training time.
| no_new_dataset | 0.948346 |
1509.03248 | George Trigeorgis | George Trigeorgis, Konstantinos Bousmalis, Stefanos Zafeiriou, Bjoern
W.Schuller | A deep matrix factorization method for learning attribute
representations | Submitted to TPAMI (16-Mar-2015) | null | null | null | cs.CV cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Semi-Non-negative Matrix Factorization is a technique that learns a
low-dimensional representation of a dataset that lends itself to a clustering
interpretation. It is possible that the mapping between this new representation
and our original data matrix contains rather complex hierarchical information
with implicit lower-level hidden attributes, that classical one level
clustering methodologies can not interpret. In this work we propose a novel
model, Deep Semi-NMF, that is able to learn such hidden representations that
allow themselves to an interpretation of clustering according to different,
unknown attributes of a given dataset. We also present a semi-supervised
version of the algorithm, named Deep WSF, that allows the use of (partial)
prior information for each of the known attributes of a dataset, that allows
the model to be used on datasets with mixed attribute knowledge. Finally, we
show that our models are able to learn low-dimensional representations that are
better suited for clustering, but also classification, outperforming
Semi-Non-negative Matrix Factorization, but also other state-of-the-art
methodologies variants.
| [
{
"version": "v1",
"created": "Thu, 10 Sep 2015 17:57:03 GMT"
}
] | 2015-09-11T00:00:00 | [
[
"Trigeorgis",
"George",
""
],
[
"Bousmalis",
"Konstantinos",
""
],
[
"Zafeiriou",
"Stefanos",
""
],
[
"Schuller",
"Bjoern W.",
""
]
] | TITLE: A deep matrix factorization method for learning attribute
representations
ABSTRACT: Semi-Non-negative Matrix Factorization is a technique that learns a
low-dimensional representation of a dataset that lends itself to a clustering
interpretation. It is possible that the mapping between this new representation
and our original data matrix contains rather complex hierarchical information
with implicit lower-level hidden attributes, that classical one level
clustering methodologies can not interpret. In this work we propose a novel
model, Deep Semi-NMF, that is able to learn such hidden representations that
allow themselves to an interpretation of clustering according to different,
unknown attributes of a given dataset. We also present a semi-supervised
version of the algorithm, named Deep WSF, that allows the use of (partial)
prior information for each of the known attributes of a dataset, that allows
the model to be used on datasets with mixed attribute knowledge. Finally, we
show that our models are able to learn low-dimensional representations that are
better suited for clustering, but also classification, outperforming
Semi-Non-negative Matrix Factorization, but also other state-of-the-art
methodologies variants.
| no_new_dataset | 0.947235 |
1509.02533 | Michael Mathioudakis | Charalampos Mavroforakis, Michael Mathioudakis and Aristides Gionis | Absorbing random-walk centrality: Theory and algorithms | 11 pages, 11 figures, short paper to appear at ICDM 2015 | null | null | null | cs.SI cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study a new notion of graph centrality based on absorbing random walks.
Given a graph $G=(V,E)$ and a set of query nodes $Q\subseteq V$, we aim to
identify the $k$ most central nodes in $G$ with respect to $Q$. Specifically,
we consider central nodes to be absorbing for random walks that start at the
query nodes $Q$. The goal is to find the set of $k$ central nodes that
minimizes the expected length of a random walk until absorption. The proposed
measure, which we call $k$ absorbing random-walk centrality, favors diverse
sets, as it is beneficial to place the $k$ absorbing nodes in different parts
of the graph so as to "intercept" random walks that start from different query
nodes.
Although similar problem definitions have been considered in the literature,
e.g., in information-retrieval settings where the goal is to diversify
web-search results, in this paper we study the problem formally and prove some
of its properties. We show that the problem is NP-hard, while the objective
function is monotone and supermodular, implying that a greedy algorithm
provides solutions with an approximation guarantee. On the other hand, the
greedy algorithm involves expensive matrix operations that make it prohibitive
to employ on large datasets. To confront this challenge, we develop more
efficient algorithms based on spectral clustering and on personalized PageRank.
| [
{
"version": "v1",
"created": "Tue, 8 Sep 2015 20:10:04 GMT"
}
] | 2015-09-10T00:00:00 | [
[
"Mavroforakis",
"Charalampos",
""
],
[
"Mathioudakis",
"Michael",
""
],
[
"Gionis",
"Aristides",
""
]
] | TITLE: Absorbing random-walk centrality: Theory and algorithms
ABSTRACT: We study a new notion of graph centrality based on absorbing random walks.
Given a graph $G=(V,E)$ and a set of query nodes $Q\subseteq V$, we aim to
identify the $k$ most central nodes in $G$ with respect to $Q$. Specifically,
we consider central nodes to be absorbing for random walks that start at the
query nodes $Q$. The goal is to find the set of $k$ central nodes that
minimizes the expected length of a random walk until absorption. The proposed
measure, which we call $k$ absorbing random-walk centrality, favors diverse
sets, as it is beneficial to place the $k$ absorbing nodes in different parts
of the graph so as to "intercept" random walks that start from different query
nodes.
Although similar problem definitions have been considered in the literature,
e.g., in information-retrieval settings where the goal is to diversify
web-search results, in this paper we study the problem formally and prove some
of its properties. We show that the problem is NP-hard, while the objective
function is monotone and supermodular, implying that a greedy algorithm
provides solutions with an approximation guarantee. On the other hand, the
greedy algorithm involves expensive matrix operations that make it prohibitive
to employ on large datasets. To confront this challenge, we develop more
efficient algorithms based on spectral clustering and on personalized PageRank.
| no_new_dataset | 0.945096 |
1407.0623 | Lamberto Ballan | Lamberto Ballan, Marco Bertini, Giuseppe Serra, Alberto Del Bimbo | A Data-Driven Approach for Tag Refinement and Localization in Web Videos | Preprint submitted to Computer Vision and Image Understanding (CVIU) | null | 10.1016/j.cviu.2015.05.009 | null | cs.CV cs.IR cs.MM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Tagging of visual content is becoming more and more widespread as web-based
services and social networks have popularized tagging functionalities among
their users. These user-generated tags are used to ease browsing and
exploration of media collections, e.g. using tag clouds, or to retrieve
multimedia content. However, not all media are equally tagged by users. Using
the current systems is easy to tag a single photo, and even tagging a part of a
photo, like a face, has become common in sites like Flickr and Facebook. On the
other hand, tagging a video sequence is more complicated and time consuming, so
that users just tag the overall content of a video. In this paper we present a
method for automatic video annotation that increases the number of tags
originally provided by users, and localizes them temporally, associating tags
to keyframes. Our approach exploits collective knowledge embedded in
user-generated tags and web sources, and visual similarity of keyframes and
images uploaded to social sites like YouTube and Flickr, as well as web sources
like Google and Bing. Given a keyframe, our method is able to select on the fly
from these visual sources the training exemplars that should be the most
relevant for this test sample, and proceeds to transfer labels across similar
images. Compared to existing video tagging approaches that require training
classifiers for each tag, our system has few parameters, is easy to implement
and can deal with an open vocabulary scenario. We demonstrate the approach on
tag refinement and localization on DUT-WEBV, a large dataset of web videos, and
show state-of-the-art results.
| [
{
"version": "v1",
"created": "Wed, 2 Jul 2014 15:48:37 GMT"
},
{
"version": "v2",
"created": "Sat, 11 Apr 2015 18:12:36 GMT"
},
{
"version": "v3",
"created": "Thu, 28 May 2015 17:12:54 GMT"
}
] | 2015-09-09T00:00:00 | [
[
"Ballan",
"Lamberto",
""
],
[
"Bertini",
"Marco",
""
],
[
"Serra",
"Giuseppe",
""
],
[
"Del Bimbo",
"Alberto",
""
]
] | TITLE: A Data-Driven Approach for Tag Refinement and Localization in Web Videos
ABSTRACT: Tagging of visual content is becoming more and more widespread as web-based
services and social networks have popularized tagging functionalities among
their users. These user-generated tags are used to ease browsing and
exploration of media collections, e.g. using tag clouds, or to retrieve
multimedia content. However, not all media are equally tagged by users. Using
the current systems is easy to tag a single photo, and even tagging a part of a
photo, like a face, has become common in sites like Flickr and Facebook. On the
other hand, tagging a video sequence is more complicated and time consuming, so
that users just tag the overall content of a video. In this paper we present a
method for automatic video annotation that increases the number of tags
originally provided by users, and localizes them temporally, associating tags
to keyframes. Our approach exploits collective knowledge embedded in
user-generated tags and web sources, and visual similarity of keyframes and
images uploaded to social sites like YouTube and Flickr, as well as web sources
like Google and Bing. Given a keyframe, our method is able to select on the fly
from these visual sources the training exemplars that should be the most
relevant for this test sample, and proceeds to transfer labels across similar
images. Compared to existing video tagging approaches that require training
classifiers for each tag, our system has few parameters, is easy to implement
and can deal with an open vocabulary scenario. We demonstrate the approach on
tag refinement and localization on DUT-WEBV, a large dataset of web videos, and
show state-of-the-art results.
| no_new_dataset | 0.948728 |
1412.5732 | Changsheng Li | Changsheng Li and Fan Wei and Weishan Dong and Qingshan Liu and
Xiangfeng Wang and Xin Zhang | Dynamic Structure Embedded Online Multiple-Output Regression for Stream
Data | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Online multiple-output regression is an important machine learning technique
for modeling, predicting, and compressing multi-dimensional correlated data
streams. In this paper, we propose a novel online multiple-output regression
method, called MORES, for stream data. MORES can \emph{dynamically} learn the
structure of the coefficients change in each update step to facilitate the
model's continuous refinement. We observe that limited expressive ability of
the regression model, especially in the preliminary stage of online update,
often leads to the variables in the residual errors being dependent. In light
of this point, MORES intends to \emph{dynamically} learn and leverage the
structure of the residual errors to improve the prediction accuracy. Moreover,
we define three statistical variables to \emph{exactly} represent all the seen
samples for \emph{incrementally} calculating prediction loss in each online
update round, which can avoid loading all the training data into memory for
updating model, and also effectively prevent drastic fluctuation of the model
in the presence of noise. Furthermore, we introduce a forgetting factor to set
different weights on samples so as to track the data streams' evolving
characteristics quickly from the latest samples. Experiments on one synthetic
dataset and three real-world datasets validate the effectiveness of the
proposed method. In addition, the update speed of MORES is at least 2000
samples processed per second on the three real-world datasets, more than 15
times faster than the state-of-the-art online learning algorithm.
| [
{
"version": "v1",
"created": "Thu, 18 Dec 2014 06:37:50 GMT"
},
{
"version": "v2",
"created": "Tue, 8 Sep 2015 03:00:55 GMT"
}
] | 2015-09-09T00:00:00 | [
[
"Li",
"Changsheng",
""
],
[
"Wei",
"Fan",
""
],
[
"Dong",
"Weishan",
""
],
[
"Liu",
"Qingshan",
""
],
[
"Wang",
"Xiangfeng",
""
],
[
"Zhang",
"Xin",
""
]
] | TITLE: Dynamic Structure Embedded Online Multiple-Output Regression for Stream
Data
ABSTRACT: Online multiple-output regression is an important machine learning technique
for modeling, predicting, and compressing multi-dimensional correlated data
streams. In this paper, we propose a novel online multiple-output regression
method, called MORES, for stream data. MORES can \emph{dynamically} learn the
structure of the coefficients change in each update step to facilitate the
model's continuous refinement. We observe that limited expressive ability of
the regression model, especially in the preliminary stage of online update,
often leads to the variables in the residual errors being dependent. In light
of this point, MORES intends to \emph{dynamically} learn and leverage the
structure of the residual errors to improve the prediction accuracy. Moreover,
we define three statistical variables to \emph{exactly} represent all the seen
samples for \emph{incrementally} calculating prediction loss in each online
update round, which can avoid loading all the training data into memory for
updating model, and also effectively prevent drastic fluctuation of the model
in the presence of noise. Furthermore, we introduce a forgetting factor to set
different weights on samples so as to track the data streams' evolving
characteristics quickly from the latest samples. Experiments on one synthetic
dataset and three real-world datasets validate the effectiveness of the
proposed method. In addition, the update speed of MORES is at least 2000
samples processed per second on the three real-world datasets, more than 15
times faster than the state-of-the-art online learning algorithm.
| no_new_dataset | 0.949902 |
1501.07492 | Huaizu Jiang | Huaizu Jiang | Weakly Supervised Learning for Salient Object Detection | technical report | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advances in supervised salient object detection has resulted in
significant performance on benchmark datasets. Training such models, however,
requires expensive pixel-wise annotations of salient objects. Moreover, many
existing salient object detection models assume that at least one salient
object exists in the input image. Such an assumption often leads to less
appealing saliency maps on the background images, which contain no salient
object at all. To avoid the requirement of expensive pixel-wise salient region
annotations, in this paper, we study weakly supervised learning approaches for
salient object detection. Given a set of background images and salient object
images, we propose a solution toward jointly addressing the salient object
existence and detection tasks. We adopt the latent SVM framework and formulate
the two problems together in a single integrated objective function: saliency
labels of superpixels are modeled as hidden variables and involved in a
classification term conditioned to the salient object existence variable, which
in turn depends on both global image and regional saliency features and
saliency label assignment. Experimental results on benchmark datasets validate
the effectiveness of our proposed approach.
| [
{
"version": "v1",
"created": "Thu, 29 Jan 2015 15:57:52 GMT"
},
{
"version": "v2",
"created": "Tue, 8 Sep 2015 13:34:24 GMT"
}
] | 2015-09-09T00:00:00 | [
[
"Jiang",
"Huaizu",
""
]
] | TITLE: Weakly Supervised Learning for Salient Object Detection
ABSTRACT: Recent advances in supervised salient object detection has resulted in
significant performance on benchmark datasets. Training such models, however,
requires expensive pixel-wise annotations of salient objects. Moreover, many
existing salient object detection models assume that at least one salient
object exists in the input image. Such an assumption often leads to less
appealing saliency maps on the background images, which contain no salient
object at all. To avoid the requirement of expensive pixel-wise salient region
annotations, in this paper, we study weakly supervised learning approaches for
salient object detection. Given a set of background images and salient object
images, we propose a solution toward jointly addressing the salient object
existence and detection tasks. We adopt the latent SVM framework and formulate
the two problems together in a single integrated objective function: saliency
labels of superpixels are modeled as hidden variables and involved in a
classification term conditioned to the salient object existence variable, which
in turn depends on both global image and regional saliency features and
saliency label assignment. Experimental results on benchmark datasets validate
the effectiveness of our proposed approach.
| no_new_dataset | 0.951051 |
1506.02108 | Chunhua Shen | Guosheng Lin, Chunhua Shen, Ian Reid, Anton van den Hengel | Deeply Learning the Messages in Message Passing Inference | 11 pages. Appearing in Proc. The Twenty-ninth Annual Conference on
Neural Information Processing Systems (NIPS), 2015, Montreal, Canada | null | null | null | cs.CV cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep structured output learning shows great promise in tasks like semantic
image segmentation. We proffer a new, efficient deep structured model learning
scheme, in which we show how deep Convolutional Neural Networks (CNNs) can be
used to estimate the messages in message passing inference for structured
prediction with Conditional Random Fields (CRFs). With such CNN message
estimators, we obviate the need to learn or evaluate potential functions for
message calculation. This confers significant efficiency for learning, since
otherwise when performing structured learning for a CRF with CNN potentials it
is necessary to undertake expensive inference for every stochastic gradient
iteration. The network output dimension for message estimation is the same as
the number of classes, in contrast to the network output for general CNN
potential functions in CRFs, which is exponential in the order of the
potentials. Hence CNN message learning has fewer network parameters and is more
scalable for cases that a large number of classes are involved. We apply our
method to semantic image segmentation on the PASCAL VOC 2012 dataset. We
achieve an intersection-over-union score of 73.4 on its test set, which is the
best reported result for methods using the VOC training images alone. This
impressive performance demonstrates the effectiveness and usefulness of our CNN
message learning method.
| [
{
"version": "v1",
"created": "Sat, 6 Jun 2015 02:52:38 GMT"
},
{
"version": "v2",
"created": "Wed, 10 Jun 2015 06:49:06 GMT"
},
{
"version": "v3",
"created": "Tue, 8 Sep 2015 04:29:45 GMT"
}
] | 2015-09-09T00:00:00 | [
[
"Lin",
"Guosheng",
""
],
[
"Shen",
"Chunhua",
""
],
[
"Reid",
"Ian",
""
],
[
"Hengel",
"Anton van den",
""
]
] | TITLE: Deeply Learning the Messages in Message Passing Inference
ABSTRACT: Deep structured output learning shows great promise in tasks like semantic
image segmentation. We proffer a new, efficient deep structured model learning
scheme, in which we show how deep Convolutional Neural Networks (CNNs) can be
used to estimate the messages in message passing inference for structured
prediction with Conditional Random Fields (CRFs). With such CNN message
estimators, we obviate the need to learn or evaluate potential functions for
message calculation. This confers significant efficiency for learning, since
otherwise when performing structured learning for a CRF with CNN potentials it
is necessary to undertake expensive inference for every stochastic gradient
iteration. The network output dimension for message estimation is the same as
the number of classes, in contrast to the network output for general CNN
potential functions in CRFs, which is exponential in the order of the
potentials. Hence CNN message learning has fewer network parameters and is more
scalable for cases that a large number of classes are involved. We apply our
method to semantic image segmentation on the PASCAL VOC 2012 dataset. We
achieve an intersection-over-union score of 73.4 on its test set, which is the
best reported result for methods using the VOC training images alone. This
impressive performance demonstrates the effectiveness and usefulness of our CNN
message learning method.
| no_new_dataset | 0.950041 |
1509.02441 | Subarna Tripathi | Subarna Tripathi, Serge Belongie, Youngbae Hwang, Truong Nguyen | Semantic Video Segmentation : Exploring Inference Efficiency | To appear in proc of ISOCC 2015 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We explore the efficiency of the CRF inference beyond image level semantic
segmentation and perform joint inference in video frames. The key idea is to
combine best of two worlds: semantic co-labeling and more expressive models.
Our formulation enables us to perform inference over ten thousand images within
seconds and makes the system amenable to perform video semantic segmentation
most effectively. On CamVid dataset, with TextonBoost unaries, our proposed
method achieves up to 8% improvement in accuracy over individual semantic image
segmentation without additional time overhead. The source code is available at
https://github.com/subtri/video_inference
| [
{
"version": "v1",
"created": "Fri, 4 Sep 2015 22:03:40 GMT"
}
] | 2015-09-09T00:00:00 | [
[
"Tripathi",
"Subarna",
""
],
[
"Belongie",
"Serge",
""
],
[
"Hwang",
"Youngbae",
""
],
[
"Nguyen",
"Truong",
""
]
] | TITLE: Semantic Video Segmentation : Exploring Inference Efficiency
ABSTRACT: We explore the efficiency of the CRF inference beyond image level semantic
segmentation and perform joint inference in video frames. The key idea is to
combine best of two worlds: semantic co-labeling and more expressive models.
Our formulation enables us to perform inference over ten thousand images within
seconds and makes the system amenable to perform video semantic segmentation
most effectively. On CamVid dataset, with TextonBoost unaries, our proposed
method achieves up to 8% improvement in accuracy over individual semantic image
segmentation without additional time overhead. The source code is available at
https://github.com/subtri/video_inference
| no_new_dataset | 0.950503 |
1408.1656 | Shengcai Liao | Shengcai Liao, Anil K. Jain, and Stan Z. Li | A Fast and Accurate Unconstrained Face Detector | This paper has been accepted by TPAMI. The source code is available
on the project page
http://www.cbsr.ia.ac.cn/users/scliao/projects/npdface/index.html | null | 10.1109/TPAMI.2015.2448075 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a method to address challenges in unconstrained face detection,
such as arbitrary pose variations and occlusions. First, a new image feature
called Normalized Pixel Difference (NPD) is proposed. NPD feature is computed
as the difference to sum ratio between two pixel values, inspired by the Weber
Fraction in experimental psychology. The new feature is scale invariant,
bounded, and is able to reconstruct the original image. Second, we propose a
deep quadratic tree to learn the optimal subset of NPD features and their
combinations, so that complex face manifolds can be partitioned by the learned
rules. This way, only a single soft-cascade classifier is needed to handle
unconstrained face detection. Furthermore, we show that the NPD features can be
efficiently obtained from a look up table, and the detection template can be
easily scaled, making the proposed face detector very fast. Experimental
results on three public face datasets (FDDB, GENKI, and CMU-MIT) show that the
proposed method achieves state-of-the-art performance in detecting
unconstrained faces with arbitrary pose variations and occlusions in cluttered
scenes.
| [
{
"version": "v1",
"created": "Wed, 6 Aug 2014 15:17:33 GMT"
},
{
"version": "v2",
"created": "Tue, 12 Aug 2014 14:24:52 GMT"
},
{
"version": "v3",
"created": "Mon, 7 Sep 2015 08:17:34 GMT"
}
] | 2015-09-08T00:00:00 | [
[
"Liao",
"Shengcai",
""
],
[
"Jain",
"Anil K.",
""
],
[
"Li",
"Stan Z.",
""
]
] | TITLE: A Fast and Accurate Unconstrained Face Detector
ABSTRACT: We propose a method to address challenges in unconstrained face detection,
such as arbitrary pose variations and occlusions. First, a new image feature
called Normalized Pixel Difference (NPD) is proposed. NPD feature is computed
as the difference to sum ratio between two pixel values, inspired by the Weber
Fraction in experimental psychology. The new feature is scale invariant,
bounded, and is able to reconstruct the original image. Second, we propose a
deep quadratic tree to learn the optimal subset of NPD features and their
combinations, so that complex face manifolds can be partitioned by the learned
rules. This way, only a single soft-cascade classifier is needed to handle
unconstrained face detection. Furthermore, we show that the NPD features can be
efficiently obtained from a look up table, and the detection template can be
easily scaled, making the proposed face detector very fast. Experimental
results on three public face datasets (FDDB, GENKI, and CMU-MIT) show that the
proposed method achieves state-of-the-art performance in detecting
unconstrained faces with arbitrary pose variations and occlusions in cluttered
scenes.
| no_new_dataset | 0.946547 |
1410.4449 | Alina S\^irbu | Alina S\^irbu, Ozalp Babaoglu | A Holistic Approach to Log Data Analysis in High-Performance Computing
Systems: The Case of IBM Blue Gene/Q | 12 pages, 7 Figures | null | null | null | cs.DC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The complexity and cost of managing high-performance computing
infrastructures are on the rise. Automating management and repair through
predictive models to minimize human interventions is an attempt to increase
system availability and contain these costs. Building predictive models that
are accurate enough to be useful in automatic management cannot be based on
restricted log data from subsystems but requires a holistic approach to data
analysis from disparate sources. Here we provide a detailed multi-scale
characterization study based on four datasets reporting power consumption,
temperature, workload, and hardware/software events for an IBM Blue Gene/Q
installation. We show that the system runs a rich parallel workload, with low
correlation among its components in terms of temperature and power, but higher
correlation in terms of events. As expected, power and temperature correlate
strongly, while events display negative correlations with load and power. Power
and workload show moderate correlations, and only at the scale of components.
The aim of the study is a systematic, integrated characterization of the
computing infrastructure and discovery of correlation sources and levels to
serve as basis for future predictive modeling efforts.
| [
{
"version": "v1",
"created": "Thu, 16 Oct 2014 14:40:00 GMT"
},
{
"version": "v2",
"created": "Tue, 10 Feb 2015 10:41:57 GMT"
},
{
"version": "v3",
"created": "Mon, 7 Sep 2015 11:08:50 GMT"
}
] | 2015-09-08T00:00:00 | [
[
"Sîrbu",
"Alina",
""
],
[
"Babaoglu",
"Ozalp",
""
]
] | TITLE: A Holistic Approach to Log Data Analysis in High-Performance Computing
Systems: The Case of IBM Blue Gene/Q
ABSTRACT: The complexity and cost of managing high-performance computing
infrastructures are on the rise. Automating management and repair through
predictive models to minimize human interventions is an attempt to increase
system availability and contain these costs. Building predictive models that
are accurate enough to be useful in automatic management cannot be based on
restricted log data from subsystems but requires a holistic approach to data
analysis from disparate sources. Here we provide a detailed multi-scale
characterization study based on four datasets reporting power consumption,
temperature, workload, and hardware/software events for an IBM Blue Gene/Q
installation. We show that the system runs a rich parallel workload, with low
correlation among its components in terms of temperature and power, but higher
correlation in terms of events. As expected, power and temperature correlate
strongly, while events display negative correlations with load and power. Power
and workload show moderate correlations, and only at the scale of components.
The aim of the study is a systematic, integrated characterization of the
computing infrastructure and discovery of correlation sources and levels to
serve as basis for future predictive modeling efforts.
| no_new_dataset | 0.934574 |
1505.02108 | Ira Kemelmacher-Shlizerman | D. Miller, E. Brossard, S. Seitz, I. Kemelmacher-Shlizerman | MegaFace: A Million Faces for Recognition at Scale | Please see http://megaface.cs.washington.edu/ for code and data | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent face recognition experiments on the LFW benchmark show that face
recognition is performing stunningly well, surpassing human recognition rates.
In this paper, we study face recognition at scale. Specifically, we have
collected from Flickr a \textbf{Million} faces and evaluated state of the art
face recognition algorithms on this dataset. We found that the performance of
algorithms varies--while all perform great on LFW, once evaluated at scale
recognition rates drop drastically for most algorithms. Interestingly, deep
learning based approach by \cite{schroff2015facenet} performs much better, but
still gets less robust at scale. We consider both verification and
identification problems, and evaluate how pose affects recognition at scale.
Moreover, we ran an extensive human study on Mechanical Turk to evaluate human
recognition at scale, and report results. All the photos are creative commons
photos and is released at \small{\url{http://megaface.cs.washington.edu/}} for
research and further experiments.
| [
{
"version": "v1",
"created": "Fri, 8 May 2015 17:39:23 GMT"
},
{
"version": "v2",
"created": "Mon, 7 Sep 2015 19:45:47 GMT"
}
] | 2015-09-08T00:00:00 | [
[
"Miller",
"D.",
""
],
[
"Brossard",
"E.",
""
],
[
"Seitz",
"S.",
""
],
[
"Kemelmacher-Shlizerman",
"I.",
""
]
] | TITLE: MegaFace: A Million Faces for Recognition at Scale
ABSTRACT: Recent face recognition experiments on the LFW benchmark show that face
recognition is performing stunningly well, surpassing human recognition rates.
In this paper, we study face recognition at scale. Specifically, we have
collected from Flickr a \textbf{Million} faces and evaluated state of the art
face recognition algorithms on this dataset. We found that the performance of
algorithms varies--while all perform great on LFW, once evaluated at scale
recognition rates drop drastically for most algorithms. Interestingly, deep
learning based approach by \cite{schroff2015facenet} performs much better, but
still gets less robust at scale. We consider both verification and
identification problems, and evaluate how pose affects recognition at scale.
Moreover, we ran an extensive human study on Mechanical Turk to evaluate human
recognition at scale, and report results. All the photos are creative commons
photos and is released at \small{\url{http://megaface.cs.washington.edu/}} for
research and further experiments.
| no_new_dataset | 0.926901 |
1509.01602 | Ivan Bogun | Ivan Bogun, Anelia Angelova and Navdeep Jaitly | Object Recognition from Short Videos for Robotic Perception | 7 pages, 6 figures, 3 tables | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep neural networks have become the primary learning technique for object
recognition. Videos, unlike still images, are temporally coherent which makes
the application of deep networks non-trivial. Here, we investigate how motion
can aid object recognition in short videos. Our approach is based on Long
Short-Term Memory (LSTM) deep networks. Unlike previous applications of LSTMs,
we implement each gate as a convolution. We show that convolutional-based LSTM
models are capable of learning motion dependencies and are able to improve the
recognition accuracy when more frames in a sequence are available. We evaluate
our approach on the Washington RGBD Object dataset and on the Washington RGBD
Scenes dataset. Our approach outperforms deep nets applied to still images and
sets a new state-of-the-art in this domain.
| [
{
"version": "v1",
"created": "Fri, 4 Sep 2015 20:48:23 GMT"
}
] | 2015-09-08T00:00:00 | [
[
"Bogun",
"Ivan",
""
],
[
"Angelova",
"Anelia",
""
],
[
"Jaitly",
"Navdeep",
""
]
] | TITLE: Object Recognition from Short Videos for Robotic Perception
ABSTRACT: Deep neural networks have become the primary learning technique for object
recognition. Videos, unlike still images, are temporally coherent which makes
the application of deep networks non-trivial. Here, we investigate how motion
can aid object recognition in short videos. Our approach is based on Long
Short-Term Memory (LSTM) deep networks. Unlike previous applications of LSTMs,
we implement each gate as a convolution. We show that convolutional-based LSTM
models are capable of learning motion dependencies and are able to improve the
recognition accuracy when more frames in a sequence are available. We evaluate
our approach on the Washington RGBD Object dataset and on the Washington RGBD
Scenes dataset. Our approach outperforms deep nets applied to still images and
sets a new state-of-the-art in this domain.
| no_new_dataset | 0.948346 |
1509.01659 | Armen Aghajanyan | Armen Aghajanyan | Gravitational Clustering | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The downfall of many supervised learning algorithms, such as neural networks,
is the inherent need for a large amount of training data. Although there is a
lot of buzz about big data, there is still the problem of doing classification
from a small dataset. Other methods such as support vector machines, although
capable of dealing with few samples, are inherently binary classifiers, and are
in need of learning strategies such as One vs All in the case of
multi-classification. In the presence of a large number of classes this can
become problematic. In this paper we present, a novel approach to supervised
learning through the method of clustering. Unlike traditional methods such as
K-Means, Gravitational Clustering does not require the initial number of
clusters, and automatically builds the clusters, individual samples can be
arbitrarily weighted and it requires only few samples while staying resilient
to over-fitting.
| [
{
"version": "v1",
"created": "Sat, 5 Sep 2015 03:37:50 GMT"
}
] | 2015-09-08T00:00:00 | [
[
"Aghajanyan",
"Armen",
""
]
] | TITLE: Gravitational Clustering
ABSTRACT: The downfall of many supervised learning algorithms, such as neural networks,
is the inherent need for a large amount of training data. Although there is a
lot of buzz about big data, there is still the problem of doing classification
from a small dataset. Other methods such as support vector machines, although
capable of dealing with few samples, are inherently binary classifiers, and are
in need of learning strategies such as One vs All in the case of
multi-classification. In the presence of a large number of classes this can
become problematic. In this paper we present, a novel approach to supervised
learning through the method of clustering. Unlike traditional methods such as
K-Means, Gravitational Clustering does not require the initial number of
clusters, and automatically builds the clusters, individual samples can be
arbitrarily weighted and it requires only few samples while staying resilient
to over-fitting.
| no_new_dataset | 0.949856 |
1509.01719 | Yuewei Lin | Yuewei Lin, Jing Chen, Yu Cao, Youjie Zhou, Lingfeng Zhang, Yuan Yan
Tang, Song Wang | Unsupervised Cross-Domain Recognition by Identifying Compact Joint
Subspaces | ICIP 2015 Top 10% paper | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper introduces a new method to solve the cross-domain recognition
problem. Different from the traditional domain adaption methods which rely on a
global domain shift for all classes between source and target domain, the
proposed method is more flexible to capture individual class variations across
domains. By adopting a natural and widely used assumption -- "the data samples
from the same class should lay on a low-dimensional subspace, even if they come
from different domains", the proposed method circumvents the limitation of the
global domain shift, and solves the cross-domain recognition by finding the
compact joint subspaces of source and target domain. Specifically, given
labeled samples in source domain, we construct subspaces for each of the
classes. Then we construct subspaces in the target domain, called anchor
subspaces, by collecting unlabeled samples that are close to each other and
highly likely all fall into the same class. The corresponding class label is
then assigned by minimizing a cost function which reflects the overlap and
topological structure consistency between subspaces across source and target
domains, and within anchor subspaces, respectively.We further combine the
anchor subspaces to corresponding source subspaces to construct the compact
joint subspaces. Subsequently, one-vs-rest SVM classifiers are trained in the
compact joint subspaces and applied to unlabeled data in the target domain. We
evaluate the proposed method on two widely used datasets: object recognition
dataset for computer vision tasks, and sentiment classification dataset for
natural language processing tasks. Comparison results demonstrate that the
proposed method outperforms the comparison methods on both datasets.
| [
{
"version": "v1",
"created": "Sat, 5 Sep 2015 17:12:21 GMT"
}
] | 2015-09-08T00:00:00 | [
[
"Lin",
"Yuewei",
""
],
[
"Chen",
"Jing",
""
],
[
"Cao",
"Yu",
""
],
[
"Zhou",
"Youjie",
""
],
[
"Zhang",
"Lingfeng",
""
],
[
"Tang",
"Yuan Yan",
""
],
[
"Wang",
"Song",
""
]
] | TITLE: Unsupervised Cross-Domain Recognition by Identifying Compact Joint
Subspaces
ABSTRACT: This paper introduces a new method to solve the cross-domain recognition
problem. Different from the traditional domain adaption methods which rely on a
global domain shift for all classes between source and target domain, the
proposed method is more flexible to capture individual class variations across
domains. By adopting a natural and widely used assumption -- "the data samples
from the same class should lay on a low-dimensional subspace, even if they come
from different domains", the proposed method circumvents the limitation of the
global domain shift, and solves the cross-domain recognition by finding the
compact joint subspaces of source and target domain. Specifically, given
labeled samples in source domain, we construct subspaces for each of the
classes. Then we construct subspaces in the target domain, called anchor
subspaces, by collecting unlabeled samples that are close to each other and
highly likely all fall into the same class. The corresponding class label is
then assigned by minimizing a cost function which reflects the overlap and
topological structure consistency between subspaces across source and target
domains, and within anchor subspaces, respectively.We further combine the
anchor subspaces to corresponding source subspaces to construct the compact
joint subspaces. Subsequently, one-vs-rest SVM classifiers are trained in the
compact joint subspaces and applied to unlabeled data in the target domain. We
evaluate the proposed method on two widely used datasets: object recognition
dataset for computer vision tasks, and sentiment classification dataset for
natural language processing tasks. Comparison results demonstrate that the
proposed method outperforms the comparison methods on both datasets.
| no_new_dataset | 0.949482 |
1509.02094 | Hyun Soo Park | Hyun Soo Park, Yedong Niu, Jianbo Shi | Future Localization from an Egocentric Depth Image | 9 pages | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents a method for future localization: to predict a set of
plausible trajectories of ego-motion given a depth image. We predict paths
avoiding obstacles, between objects, even paths turning around a corner into
space behind objects. As a byproduct of the predicted trajectories of
ego-motion, we discover in the image the empty space occluded by foreground
objects. We use no image based features such as semantic labeling/segmentation
or object detection/recognition for this algorithm. Inspired by proxemics, we
represent the space around a person using an EgoSpace map, akin to an
illustrated tourist map, that measures a likelihood of occlusion at the
egocentric coordinate system. A future trajectory of ego-motion is modeled by a
linear combination of compact trajectory bases allowing us to constrain the
predicted trajectory. We learn the relationship between the EgoSpace map and
trajectory from the EgoMotion dataset providing in-situ measurements of the
future trajectory. A cost function that takes into account partial occlusion
due to foreground objects is minimized to predict a trajectory. This cost
function generates a trajectory that passes through the occluded space, which
allows us to discover the empty space behind the foreground objects. We
quantitatively evaluate our method to show predictive validity and apply to
various real world scenes including walking, shopping, and social interactions.
| [
{
"version": "v1",
"created": "Mon, 7 Sep 2015 15:51:11 GMT"
}
] | 2015-09-08T00:00:00 | [
[
"Park",
"Hyun Soo",
""
],
[
"Niu",
"Yedong",
""
],
[
"Shi",
"Jianbo",
""
]
] | TITLE: Future Localization from an Egocentric Depth Image
ABSTRACT: This paper presents a method for future localization: to predict a set of
plausible trajectories of ego-motion given a depth image. We predict paths
avoiding obstacles, between objects, even paths turning around a corner into
space behind objects. As a byproduct of the predicted trajectories of
ego-motion, we discover in the image the empty space occluded by foreground
objects. We use no image based features such as semantic labeling/segmentation
or object detection/recognition for this algorithm. Inspired by proxemics, we
represent the space around a person using an EgoSpace map, akin to an
illustrated tourist map, that measures a likelihood of occlusion at the
egocentric coordinate system. A future trajectory of ego-motion is modeled by a
linear combination of compact trajectory bases allowing us to constrain the
predicted trajectory. We learn the relationship between the EgoSpace map and
trajectory from the EgoMotion dataset providing in-situ measurements of the
future trajectory. A cost function that takes into account partial occlusion
due to foreground objects is minimized to predict a trajectory. This cost
function generates a trajectory that passes through the occluded space, which
allows us to discover the empty space behind the foreground objects. We
quantitatively evaluate our method to show predictive validity and apply to
various real world scenes including walking, shopping, and social interactions.
| no_new_dataset | 0.919643 |
1509.01354 | Jinma Guo | Jinma Guo and Jianmin Li | CNN Based Hashing for Image Retrieval | 16 pages, 6 figures | null | null | null | cs.CV cs.LG | http://creativecommons.org/publicdomain/zero/1.0/ | Along with data on the web increasing dramatically, hashing is becoming more
and more popular as a method of approximate nearest neighbor search. Previous
supervised hashing methods utilized similarity/dissimilarity matrix to get
semantic information. But the matrix is not easy to construct for a new
dataset. Rather than to reconstruct the matrix, we proposed a straightforward
CNN-based hashing method, i.e. binarilizing the activations of a fully
connected layer with threshold 0 and taking the binary result as hash codes.
This method achieved the best performance on CIFAR-10 and was comparable with
the state-of-the-art on MNIST. And our experiments on CIFAR-10 suggested that
the signs of activations may carry more information than the relative values of
activations between samples, and that the co-adaption between feature extractor
and hash functions is important for hashing.
| [
{
"version": "v1",
"created": "Fri, 4 Sep 2015 07:08:44 GMT"
}
] | 2015-09-07T00:00:00 | [
[
"Guo",
"Jinma",
""
],
[
"Li",
"Jianmin",
""
]
] | TITLE: CNN Based Hashing for Image Retrieval
ABSTRACT: Along with data on the web increasing dramatically, hashing is becoming more
and more popular as a method of approximate nearest neighbor search. Previous
supervised hashing methods utilized similarity/dissimilarity matrix to get
semantic information. But the matrix is not easy to construct for a new
dataset. Rather than to reconstruct the matrix, we proposed a straightforward
CNN-based hashing method, i.e. binarilizing the activations of a fully
connected layer with threshold 0 and taking the binary result as hash codes.
This method achieved the best performance on CIFAR-10 and was comparable with
the state-of-the-art on MNIST. And our experiments on CIFAR-10 suggested that
the signs of activations may carry more information than the relative values of
activations between samples, and that the co-adaption between feature extractor
and hash functions is important for hashing.
| no_new_dataset | 0.94868 |
1509.01379 | Balubaid Mohammed | Mohammed A. Balubaid and Umar Manzoor | Ontology Based SMS Controller for Smart Phones | null | null | null | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Text analysis includes lexical analysis of the text and has been widely
studied and used in diverse applications. In the last decade, researchers have
proposed many efficient solutions to analyze / classify large text dataset,
however, analysis / classification of short text is still a challenge because
1) the data is very sparse 2) It contains noise words and 3) It is difficult to
understand the syntactical structure of the text. Short Messaging Service (SMS)
is a text messaging service for mobile/smart phone and this service is
frequently used by all mobile users. Because of the popularity of SMS service,
marketing companies nowadays are also using this service for direct marketing
also known as SMS marketing.In this paper, we have proposed Ontology based SMS
Controller which analyze the text message and classify it using ontology
aslegitimate or spam. The proposed system has been tested on different
scenarios and experimental results shows that the proposed solution is
effective both in terms of efficiency and time.
| [
{
"version": "v1",
"created": "Fri, 4 Sep 2015 09:29:47 GMT"
}
] | 2015-09-07T00:00:00 | [
[
"Balubaid",
"Mohammed A.",
""
],
[
"Manzoor",
"Umar",
""
]
] | TITLE: Ontology Based SMS Controller for Smart Phones
ABSTRACT: Text analysis includes lexical analysis of the text and has been widely
studied and used in diverse applications. In the last decade, researchers have
proposed many efficient solutions to analyze / classify large text dataset,
however, analysis / classification of short text is still a challenge because
1) the data is very sparse 2) It contains noise words and 3) It is difficult to
understand the syntactical structure of the text. Short Messaging Service (SMS)
is a text messaging service for mobile/smart phone and this service is
frequently used by all mobile users. Because of the popularity of SMS service,
marketing companies nowadays are also using this service for direct marketing
also known as SMS marketing.In this paper, we have proposed Ontology based SMS
Controller which analyze the text message and classify it using ontology
aslegitimate or spam. The proposed system has been tested on different
scenarios and experimental results shows that the proposed solution is
effective both in terms of efficiency and time.
| no_new_dataset | 0.944074 |
1509.01469 | Ruiqi Guo | Ruiqi Guo, Sanjiv Kumar, Krzysztof Choromanski and David Simcha | Quantization based Fast Inner Product Search | null | null | null | null | cs.AI cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a quantization based approach for fast approximate Maximum Inner
Product Search (MIPS). Each database vector is quantized in multiple subspaces
via a set of codebooks, learned directly by minimizing the inner product
quantization error. Then, the inner product of a query to a database vector is
approximated as the sum of inner products with the subspace quantizers.
Different from recently proposed LSH approaches to MIPS, the database vectors
and queries do not need to be augmented in a higher dimensional feature space.
We also provide a theoretical analysis of the proposed approach, consisting of
the concentration results under mild assumptions. Furthermore, if a small
sample of example queries is given at the training time, we propose a modified
codebook learning procedure which further improves the accuracy. Experimental
results on a variety of datasets including those arising from deep neural
networks show that the proposed approach significantly outperforms the existing
state-of-the-art.
| [
{
"version": "v1",
"created": "Fri, 4 Sep 2015 14:43:11 GMT"
}
] | 2015-09-07T00:00:00 | [
[
"Guo",
"Ruiqi",
""
],
[
"Kumar",
"Sanjiv",
""
],
[
"Choromanski",
"Krzysztof",
""
],
[
"Simcha",
"David",
""
]
] | TITLE: Quantization based Fast Inner Product Search
ABSTRACT: We propose a quantization based approach for fast approximate Maximum Inner
Product Search (MIPS). Each database vector is quantized in multiple subspaces
via a set of codebooks, learned directly by minimizing the inner product
quantization error. Then, the inner product of a query to a database vector is
approximated as the sum of inner products with the subspace quantizers.
Different from recently proposed LSH approaches to MIPS, the database vectors
and queries do not need to be augmented in a higher dimensional feature space.
We also provide a theoretical analysis of the proposed approach, consisting of
the concentration results under mild assumptions. Furthermore, if a small
sample of example queries is given at the training time, we propose a modified
codebook learning procedure which further improves the accuracy. Experimental
results on a variety of datasets including those arising from deep neural
networks show that the proposed approach significantly outperforms the existing
state-of-the-art.
| no_new_dataset | 0.941007 |
1409.6813 | Hossein Rahmani | Hossein Rahmani, Arif Mahmood, Du Huynh, Ajmal Mian | Histogram of Oriented Principal Components for Cross-View Action
Recognition | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Existing techniques for 3D action recognition are sensitive to viewpoint
variations because they extract features from depth images which are viewpoint
dependent. In contrast, we directly process pointclouds for cross-view action
recognition from unknown and unseen views. We propose the Histogram of Oriented
Principal Components (HOPC) descriptor that is robust to noise, viewpoint,
scale and action speed variations. At a 3D point, HOPC is computed by
projecting the three scaled eigenvectors of the pointcloud within its local
spatio-temporal support volume onto the vertices of a regular dodecahedron.
HOPC is also used for the detection of Spatio-Temporal Keypoints (STK) in 3D
pointcloud sequences so that view-invariant STK descriptors (or Local HOPC
descriptors) at these key locations only are used for action recognition. We
also propose a global descriptor computed from the normalized spatio-temporal
distribution of STKs in 4-D, which we refer to as STK-D. We have evaluated the
performance of our proposed descriptors against nine existing techniques on two
cross-view and three single-view human action recognition datasets. The
Experimental results show that our techniques provide significant improvement
over state-of-the-art methods.
| [
{
"version": "v1",
"created": "Wed, 24 Sep 2014 03:57:49 GMT"
},
{
"version": "v2",
"created": "Thu, 3 Sep 2015 05:12:27 GMT"
}
] | 2015-09-04T00:00:00 | [
[
"Rahmani",
"Hossein",
""
],
[
"Mahmood",
"Arif",
""
],
[
"Huynh",
"Du",
""
],
[
"Mian",
"Ajmal",
""
]
] | TITLE: Histogram of Oriented Principal Components for Cross-View Action
Recognition
ABSTRACT: Existing techniques for 3D action recognition are sensitive to viewpoint
variations because they extract features from depth images which are viewpoint
dependent. In contrast, we directly process pointclouds for cross-view action
recognition from unknown and unseen views. We propose the Histogram of Oriented
Principal Components (HOPC) descriptor that is robust to noise, viewpoint,
scale and action speed variations. At a 3D point, HOPC is computed by
projecting the three scaled eigenvectors of the pointcloud within its local
spatio-temporal support volume onto the vertices of a regular dodecahedron.
HOPC is also used for the detection of Spatio-Temporal Keypoints (STK) in 3D
pointcloud sequences so that view-invariant STK descriptors (or Local HOPC
descriptors) at these key locations only are used for action recognition. We
also propose a global descriptor computed from the normalized spatio-temporal
distribution of STKs in 4-D, which we refer to as STK-D. We have evaluated the
performance of our proposed descriptors against nine existing techniques on two
cross-view and three single-view human action recognition datasets. The
Experimental results show that our techniques provide significant improvement
over state-of-the-art methods.
| no_new_dataset | 0.94428 |
1506.00976 | Gautier Marti | Gautier Marti, Philippe Very and Philippe Donnat | Toward a generic representation of random variables for machine learning | submitted to Pattern Recognition Letters | null | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents a pre-processing and a distance which improve the
performance of machine learning algorithms working on independent and
identically distributed stochastic processes. We introduce a novel
non-parametric approach to represent random variables which splits apart
dependency and distribution without losing any information. We also propound an
associated metric leveraging this representation and its statistical estimate.
Besides experiments on synthetic datasets, the benefits of our contribution is
illustrated through the example of clustering financial time series, for
instance prices from the credit default swaps market. Results are available on
the website www.datagrapple.com and an IPython Notebook tutorial is available
at www.datagrapple.com/Tech for reproducible research.
| [
{
"version": "v1",
"created": "Tue, 2 Jun 2015 17:58:48 GMT"
},
{
"version": "v2",
"created": "Thu, 3 Sep 2015 19:23:30 GMT"
}
] | 2015-09-04T00:00:00 | [
[
"Marti",
"Gautier",
""
],
[
"Very",
"Philippe",
""
],
[
"Donnat",
"Philippe",
""
]
] | TITLE: Toward a generic representation of random variables for machine learning
ABSTRACT: This paper presents a pre-processing and a distance which improve the
performance of machine learning algorithms working on independent and
identically distributed stochastic processes. We introduce a novel
non-parametric approach to represent random variables which splits apart
dependency and distribution without losing any information. We also propound an
associated metric leveraging this representation and its statistical estimate.
Besides experiments on synthetic datasets, the benefits of our contribution is
illustrated through the example of clustering financial time series, for
instance prices from the credit default swaps market. Results are available on
the website www.datagrapple.com and an IPython Notebook tutorial is available
at www.datagrapple.com/Tech for reproducible research.
| no_new_dataset | 0.944022 |
1509.01074 | Ahmed Mohamed | Ahmed Nabil Mohamed | A Novice Guide towards Human Motion Analysis and Understanding | 35 Pages | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Human motion analysis and understanding has been, and is still, the focus of
attention of many disciplines which is considered an obvious indicator of the
wide and massive importance of the subject. The purpose of this article is to
shed some light on this very important subject, so it can be a good insight for
a novice computer vision researcher in this field by providing him/her with a
wealth of knowledge about the subject covering many directions. There are two
main contributions of this article. The first one investigates various aspects
of some disciplines (e.g., arts, philosophy, psychology, and neuroscience) that
are interested in the subject and review some of their contributions stressing
on those that can be useful for computer vision researchers. Moreover, many
examples are illustrated to indicate the benefits of integrating concepts and
results among different disciplines. The second contribution is concerned with
the subject from the computer vision aspect where we discuss the following
issues. First, we explore many demanding and promising applications to reveal
the wide and massive importance of the field. Second, we list various types of
sensors that may be used for acquiring various data. Third, we review different
taxonomies used for classifying motions. Fourth, we review various processes
involved in motion analysis. Fifth, we exhibit how different surveys are
structured. Sixth, we examine many of the most cited and recent reviews in the
field that have been published during the past two decades to reveal various
approaches used for implementing different stages of the problem and refer to
various algorithms and their suitability for different situations. Moreover, we
provide a long list of public datasets and discuss briefly some examples of
these datasets. Finally, we provide a general discussion of the subject from
the aspect of computer vision.
| [
{
"version": "v1",
"created": "Thu, 3 Sep 2015 13:25:37 GMT"
}
] | 2015-09-04T00:00:00 | [
[
"Mohamed",
"Ahmed Nabil",
""
]
] | TITLE: A Novice Guide towards Human Motion Analysis and Understanding
ABSTRACT: Human motion analysis and understanding has been, and is still, the focus of
attention of many disciplines which is considered an obvious indicator of the
wide and massive importance of the subject. The purpose of this article is to
shed some light on this very important subject, so it can be a good insight for
a novice computer vision researcher in this field by providing him/her with a
wealth of knowledge about the subject covering many directions. There are two
main contributions of this article. The first one investigates various aspects
of some disciplines (e.g., arts, philosophy, psychology, and neuroscience) that
are interested in the subject and review some of their contributions stressing
on those that can be useful for computer vision researchers. Moreover, many
examples are illustrated to indicate the benefits of integrating concepts and
results among different disciplines. The second contribution is concerned with
the subject from the computer vision aspect where we discuss the following
issues. First, we explore many demanding and promising applications to reveal
the wide and massive importance of the field. Second, we list various types of
sensors that may be used for acquiring various data. Third, we review different
taxonomies used for classifying motions. Fourth, we review various processes
involved in motion analysis. Fifth, we exhibit how different surveys are
structured. Sixth, we examine many of the most cited and recent reviews in the
field that have been published during the past two decades to reveal various
approaches used for implementing different stages of the problem and refer to
various algorithms and their suitability for different situations. Moreover, we
provide a long list of public datasets and discuss briefly some examples of
these datasets. Finally, we provide a general discussion of the subject from
the aspect of computer vision.
| no_new_dataset | 0.924756 |
1412.5129 | David Weyburne | David Weyburne | The Prevalence of Similarity of the Turbulent Wall-bounded Velocity
Profile | 17 pages, 10 figures, 1 appendix. This update adds a paragraph that
shows that the similarity equivalency argument is valid not only for whole
profile similarity but is also valid even for the case where only the outer
region similarity is considered | null | null | null | physics.flu-dyn | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Castillo and George (Castillo, L. and George, W., AIAA J. 39, 41(2001))
developed a flow governing equation approach for describing the turbulent outer
boundary layer region. The approach was used to develop similarity criteria for
the mean velocity and Reynolds shear stress profiles. Using the criteria as a
guide, Castillo, George, and coworkers examined an extensive set of
experimental datasets and claim that most of these turbulent velocity boundary
layers appear to be similar boundary layers when scaled with the Zagarola and
Smits (Zagarola, M. and Smits, A., J. Fluid Mech. 373, 33(1998)) velocity
parameter. In the work herein it is shown that their success at showing scaled
profile similarity in many of those datasets is flawed due to a similarity
problem that occurs when one combines the defect profile and the Zagarola and
Smits type of velocity scaling parameter. The same problem has been identified
in other papers in the literature and may in fact be widespread. We conclude
that similarity of the turbulent velocity profile is not as prevalent as was
claimed by Castillo, George, and coworkers. The result has implications as to
the accepted paradigm of the scaling of the turbulent boundary layer.
| [
{
"version": "v1",
"created": "Tue, 16 Dec 2014 19:15:03 GMT"
},
{
"version": "v2",
"created": "Fri, 6 Mar 2015 15:50:30 GMT"
},
{
"version": "v3",
"created": "Fri, 3 Apr 2015 16:29:42 GMT"
},
{
"version": "v4",
"created": "Wed, 3 Jun 2015 18:56:47 GMT"
},
{
"version": "v5",
"created": "Wed, 2 Sep 2015 13:46:11 GMT"
}
] | 2015-09-03T00:00:00 | [
[
"Weyburne",
"David",
""
]
] | TITLE: The Prevalence of Similarity of the Turbulent Wall-bounded Velocity
Profile
ABSTRACT: Castillo and George (Castillo, L. and George, W., AIAA J. 39, 41(2001))
developed a flow governing equation approach for describing the turbulent outer
boundary layer region. The approach was used to develop similarity criteria for
the mean velocity and Reynolds shear stress profiles. Using the criteria as a
guide, Castillo, George, and coworkers examined an extensive set of
experimental datasets and claim that most of these turbulent velocity boundary
layers appear to be similar boundary layers when scaled with the Zagarola and
Smits (Zagarola, M. and Smits, A., J. Fluid Mech. 373, 33(1998)) velocity
parameter. In the work herein it is shown that their success at showing scaled
profile similarity in many of those datasets is flawed due to a similarity
problem that occurs when one combines the defect profile and the Zagarola and
Smits type of velocity scaling parameter. The same problem has been identified
in other papers in the literature and may in fact be widespread. We conclude
that similarity of the turbulent velocity profile is not as prevalent as was
claimed by Castillo, George, and coworkers. The result has implications as to
the accepted paradigm of the scaling of the turbulent boundary layer.
| no_new_dataset | 0.954223 |
1509.00511 | Xitong Yang | Xitong Yang, Yuncheng Li, Jiebo Luo | Pinterest Board Recommendation for Twitter Users | null | null | 10.1145/2733373.2806375 | null | cs.SI cs.MM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Pinboard on Pinterest is an emerging media to engage online social media
users, on which users post online images for specific topics. Regardless of its
significance, there is little previous work specifically to facilitate
information discovery based on pinboards. This paper proposes a novel pinboard
recommendation system for Twitter users. In order to associate contents from
the two social media platforms, we propose to use MultiLabel classification to
map Twitter user followees to pinboard topics and visual diversification to
recommend pinboards given user interested topics. A preliminary experiment on a
dataset with 2000 users validated our proposed system.
| [
{
"version": "v1",
"created": "Tue, 1 Sep 2015 21:42:27 GMT"
}
] | 2015-09-03T00:00:00 | [
[
"Yang",
"Xitong",
""
],
[
"Li",
"Yuncheng",
""
],
[
"Luo",
"Jiebo",
""
]
] | TITLE: Pinterest Board Recommendation for Twitter Users
ABSTRACT: Pinboard on Pinterest is an emerging media to engage online social media
users, on which users post online images for specific topics. Regardless of its
significance, there is little previous work specifically to facilitate
information discovery based on pinboards. This paper proposes a novel pinboard
recommendation system for Twitter users. In order to associate contents from
the two social media platforms, we propose to use MultiLabel classification to
map Twitter user followees to pinboard topics and visual diversification to
recommend pinboards given user interested topics. A preliminary experiment on a
dataset with 2000 users validated our proposed system.
| no_new_dataset | 0.937211 |
1509.00533 | Scott Wisdom | Scott Wisdom, Thomas Powers, Les Atlas, and James Pitton | Enhancement and Recognition of Reverberant and Noisy Speech by Extending
Its Coherence | 22 pages | null | null | null | cs.SD cs.CL stat.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Most speech enhancement algorithms make use of the short-time Fourier
transform (STFT), which is a simple and flexible time-frequency decomposition
that estimates the short-time spectrum of a signal. However, the duration of
short STFT frames are inherently limited by the nonstationarity of speech
signals. The main contribution of this paper is a demonstration of speech
enhancement and automatic speech recognition in the presence of reverberation
and noise by extending the length of analysis windows. We accomplish this
extension by performing enhancement in the short-time fan-chirp transform
(STFChT) domain, an overcomplete time-frequency representation that is coherent
with speech signals over longer analysis window durations than the STFT. This
extended coherence is gained by using a linear model of fundamental frequency
variation of voiced speech signals. Our approach centers around using a
single-channel minimum mean-square error log-spectral amplitude (MMSE-LSA)
estimator proposed by Habets, which scales coefficients in a time-frequency
domain to suppress noise and reverberation. In the case of multiple
microphones, we preprocess the data with either a minimum variance
distortionless response (MVDR) beamformer, or a delay-and-sum beamformer (DSB).
We evaluate our algorithm on both speech enhancement and recognition tasks for
the REVERB challenge dataset. Compared to the same processing done in the STFT
domain, our approach achieves significant improvement in terms of objective
enhancement metrics (including PESQ---the ITU-T standard measurement for speech
quality). In terms of automatic speech recognition (ASR) performance as
measured by word error rate (WER), our experiments indicate that the STFT with
a long window is more effective for ASR.
| [
{
"version": "v1",
"created": "Wed, 2 Sep 2015 00:31:40 GMT"
}
] | 2015-09-03T00:00:00 | [
[
"Wisdom",
"Scott",
""
],
[
"Powers",
"Thomas",
""
],
[
"Atlas",
"Les",
""
],
[
"Pitton",
"James",
""
]
] | TITLE: Enhancement and Recognition of Reverberant and Noisy Speech by Extending
Its Coherence
ABSTRACT: Most speech enhancement algorithms make use of the short-time Fourier
transform (STFT), which is a simple and flexible time-frequency decomposition
that estimates the short-time spectrum of a signal. However, the duration of
short STFT frames are inherently limited by the nonstationarity of speech
signals. The main contribution of this paper is a demonstration of speech
enhancement and automatic speech recognition in the presence of reverberation
and noise by extending the length of analysis windows. We accomplish this
extension by performing enhancement in the short-time fan-chirp transform
(STFChT) domain, an overcomplete time-frequency representation that is coherent
with speech signals over longer analysis window durations than the STFT. This
extended coherence is gained by using a linear model of fundamental frequency
variation of voiced speech signals. Our approach centers around using a
single-channel minimum mean-square error log-spectral amplitude (MMSE-LSA)
estimator proposed by Habets, which scales coefficients in a time-frequency
domain to suppress noise and reverberation. In the case of multiple
microphones, we preprocess the data with either a minimum variance
distortionless response (MVDR) beamformer, or a delay-and-sum beamformer (DSB).
We evaluate our algorithm on both speech enhancement and recognition tasks for
the REVERB challenge dataset. Compared to the same processing done in the STFT
domain, our approach achieves significant improvement in terms of objective
enhancement metrics (including PESQ---the ITU-T standard measurement for speech
quality). In terms of automatic speech recognition (ASR) performance as
measured by word error rate (WER), our experiments indicate that the STFT with
a long window is more effective for ASR.
| no_new_dataset | 0.949949 |
1509.00568 | Michael (Micky) Fire | Michael Fire and Jonathan Schler | Exploring Online Ad Images Using a Deep Convolutional Neural Network
Approach | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Online advertising is a huge, rapidly growing advertising market in today's
world. One common form of online advertising is using image ads. A decision is
made (often in real time) every time a user sees an ad, and the advertiser is
eager to determine the best ad to display. Consequently, many algorithms have
been developed that calculate the optimal ad to show to the current user at the
present time. Typically, these algorithms focus on variations of the ad,
optimizing among different properties such as background color, image size, or
set of images. However, there is a more fundamental layer. Our study looks at
new qualities of ads that can be determined before an ad is shown (rather than
online optimization) and defines which ads are most likely to be successful.
We present a set of novel algorithms that utilize deep-learning image
processing, machine learning, and graph theory to investigate online
advertising and to construct prediction models which can foresee an image ad's
success. We evaluated our algorithms on a dataset with over 260,000 ad images,
as well as a smaller dataset specifically related to the automotive industry,
and we succeeded in constructing regression models for ad image click rate
prediction. The obtained results emphasize the great potential of using
deep-learning algorithms to effectively and efficiently analyze image ads and
to create better and more innovative online ads. Moreover, the algorithms
presented in this paper can help predict ad success and can be applied to
analyze other large-scale image corpora.
| [
{
"version": "v1",
"created": "Wed, 2 Sep 2015 06:18:27 GMT"
}
] | 2015-09-03T00:00:00 | [
[
"Fire",
"Michael",
""
],
[
"Schler",
"Jonathan",
""
]
] | TITLE: Exploring Online Ad Images Using a Deep Convolutional Neural Network
Approach
ABSTRACT: Online advertising is a huge, rapidly growing advertising market in today's
world. One common form of online advertising is using image ads. A decision is
made (often in real time) every time a user sees an ad, and the advertiser is
eager to determine the best ad to display. Consequently, many algorithms have
been developed that calculate the optimal ad to show to the current user at the
present time. Typically, these algorithms focus on variations of the ad,
optimizing among different properties such as background color, image size, or
set of images. However, there is a more fundamental layer. Our study looks at
new qualities of ads that can be determined before an ad is shown (rather than
online optimization) and defines which ads are most likely to be successful.
We present a set of novel algorithms that utilize deep-learning image
processing, machine learning, and graph theory to investigate online
advertising and to construct prediction models which can foresee an image ad's
success. We evaluated our algorithms on a dataset with over 260,000 ad images,
as well as a smaller dataset specifically related to the automotive industry,
and we succeeded in constructing regression models for ad image click rate
prediction. The obtained results emphasize the great potential of using
deep-learning algorithms to effectively and efficiently analyze image ads and
to create better and more innovative online ads. Moreover, the algorithms
presented in this paper can help predict ad success and can be applied to
analyze other large-scale image corpora.
| no_new_dataset | 0.938463 |
1502.05680 | Andrea Montanari | Andrea Montanari | Finding One Community in a Sparse Graph | 30 pages, 8 pdf figures | null | 10.1007/s10955-015-1338-2 | null | stat.ML cond-mat.stat-mech cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider a random sparse graph with bounded average degree, in which a
subset of vertices has higher connectivity than the background. In particular,
the average degree inside this subset of vertices is larger than outside (but
still bounded). Given a realization of such graph, we aim at identifying the
hidden subset of vertices. This can be regarded as a model for the problem of
finding a tightly knitted community in a social network, or a cluster in a
relational dataset.
In this paper we present two sets of contributions: $(i)$ We use the cavity
method from spin glass theory to derive an exact phase diagram for the
reconstruction problem. In particular, as the difference in edge probability
increases, the problem undergoes two phase transitions, a static phase
transition and a dynamic one. $(ii)$ We establish rigorous bounds on the
dynamic phase transition and prove that, above a certain threshold, a local
algorithm (belief propagation) correctly identify most of the hidden set. Below
the same threshold \emph{no local algorithm} can achieve this goal. However, in
this regime the subset can be identified by exhaustive search.
For small hidden sets and large average degree, the phase transition for
local algorithms takes an intriguingly simple form. Local algorithms succeed
with high probability for ${\rm deg}_{\rm in} - {\rm deg}_{\rm out} >
\sqrt{{\rm deg}_{\rm out}/e}$ and fail for ${\rm deg}_{\rm in} - {\rm deg}_{\rm
out} < \sqrt{{\rm deg}_{\rm out}/e}$ (with ${\rm deg}_{\rm in}$, ${\rm
deg}_{\rm out}$ the average degrees inside and outside the community). We argue
that spectral algorithms are also ineffective in the latter regime.
It is an open problem whether any polynomial time algorithms might succeed
for ${\rm deg}_{\rm in} - {\rm deg}_{\rm out} < \sqrt{{\rm deg}_{\rm out}/e}$.
| [
{
"version": "v1",
"created": "Thu, 19 Feb 2015 19:50:09 GMT"
},
{
"version": "v2",
"created": "Thu, 30 Jul 2015 19:46:13 GMT"
}
] | 2015-09-02T00:00:00 | [
[
"Montanari",
"Andrea",
""
]
] | TITLE: Finding One Community in a Sparse Graph
ABSTRACT: We consider a random sparse graph with bounded average degree, in which a
subset of vertices has higher connectivity than the background. In particular,
the average degree inside this subset of vertices is larger than outside (but
still bounded). Given a realization of such graph, we aim at identifying the
hidden subset of vertices. This can be regarded as a model for the problem of
finding a tightly knitted community in a social network, or a cluster in a
relational dataset.
In this paper we present two sets of contributions: $(i)$ We use the cavity
method from spin glass theory to derive an exact phase diagram for the
reconstruction problem. In particular, as the difference in edge probability
increases, the problem undergoes two phase transitions, a static phase
transition and a dynamic one. $(ii)$ We establish rigorous bounds on the
dynamic phase transition and prove that, above a certain threshold, a local
algorithm (belief propagation) correctly identify most of the hidden set. Below
the same threshold \emph{no local algorithm} can achieve this goal. However, in
this regime the subset can be identified by exhaustive search.
For small hidden sets and large average degree, the phase transition for
local algorithms takes an intriguingly simple form. Local algorithms succeed
with high probability for ${\rm deg}_{\rm in} - {\rm deg}_{\rm out} >
\sqrt{{\rm deg}_{\rm out}/e}$ and fail for ${\rm deg}_{\rm in} - {\rm deg}_{\rm
out} < \sqrt{{\rm deg}_{\rm out}/e}$ (with ${\rm deg}_{\rm in}$, ${\rm
deg}_{\rm out}$ the average degrees inside and outside the community). We argue
that spectral algorithms are also ineffective in the latter regime.
It is an open problem whether any polynomial time algorithms might succeed
for ${\rm deg}_{\rm in} - {\rm deg}_{\rm out} < \sqrt{{\rm deg}_{\rm out}/e}$.
| no_new_dataset | 0.947624 |
1506.01798 | Shyeh Tjing Loi | Shyeh Tjing Loi, Cathryn M. Trott, Tara Murphy, Iver H. Cairns, Martin
Bell, Natasha Hurley-Walker, John Morgan, Emil Lenc, A. R. Offringa, L. Feng,
P. J. Hancock, D. L. Kaplan, N. Kudryavtseva, G. Bernardi, J. D. Bowman, F.
Briggs, R. J. Cappallo, B. E. Corey, A. A. Deshpande, D. Emrich, B. M.
Gaensler, R. Goeke, L. J. Greenhill, B. J. Hazelton, M. Johnston-Hollitt, J.
C. Kasper, E. Kratzenberg, C. J. Lonsdale, M. J. Lynch, S. R. McWhirter, D.
A. Mitchell, M. F. Morales, E. Morgan, D. Oberoi, S. M. Ord, T. Prabu, A. E.
E. Rogers, A. Roshi, N. Udaya Shankar, K. S. Srivani, R. Subrahmanyan, S. J.
Tingay, M. Waterson, R. B. Wayth, R. L. Webster, A. R. Whitney, A. Williams
and C. L. Williams | Power spectrum analysis of ionospheric fluctuations with the Murchison
Widefield Array | Accepted for publication in Radio Science | null | 10.1002/2015RS005711 | null | astro-ph.IM physics.space-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Low-frequency, wide field-of-view (FoV) radio telescopes such as the
Murchison Widefield Array (MWA) enable the ionosphere to be sampled at high
spatial completeness. We present the results of the first power spectrum
analysis of ionospheric fluctuations in MWA data, where we examined the
position offsets of radio sources appearing in two datasets. The refractive
shifts in the positions of celestial sources are proportional to spatial
gradients in the electron column density transverse to the line of sight. These
can be used to probe plasma structures and waves in the ionosphere. The
regional (10-100 km) scales probed by the MWA, determined by the size of its
FoV and the spatial density of radio sources (typically thousands in a single
FoV), complement the global (100-1000 km) scales of GPS studies and local
(0.01-1 km) scales of radar scattering measurements. Our data exhibit a range
of complex structures and waves. Some fluctuations have the characteristics of
travelling ionospheric disturbances (TIDs), while others take the form of
narrow, slowly-drifting bands aligned along the Earth's magnetic field.
| [
{
"version": "v1",
"created": "Fri, 5 Jun 2015 07:31:32 GMT"
}
] | 2015-09-02T00:00:00 | [
[
"Loi",
"Shyeh Tjing",
""
],
[
"Trott",
"Cathryn M.",
""
],
[
"Murphy",
"Tara",
""
],
[
"Cairns",
"Iver H.",
""
],
[
"Bell",
"Martin",
""
],
[
"Hurley-Walker",
"Natasha",
""
],
[
"Morgan",
"John",
""
],
[
"Lenc",
"Emil",
""
],
[
"Offringa",
"A. R.",
""
],
[
"Feng",
"L.",
""
],
[
"Hancock",
"P. J.",
""
],
[
"Kaplan",
"D. L.",
""
],
[
"Kudryavtseva",
"N.",
""
],
[
"Bernardi",
"G.",
""
],
[
"Bowman",
"J. D.",
""
],
[
"Briggs",
"F.",
""
],
[
"Cappallo",
"R. J.",
""
],
[
"Corey",
"B. E.",
""
],
[
"Deshpande",
"A. A.",
""
],
[
"Emrich",
"D.",
""
],
[
"Gaensler",
"B. M.",
""
],
[
"Goeke",
"R.",
""
],
[
"Greenhill",
"L. J.",
""
],
[
"Hazelton",
"B. J.",
""
],
[
"Johnston-Hollitt",
"M.",
""
],
[
"Kasper",
"J. C.",
""
],
[
"Kratzenberg",
"E.",
""
],
[
"Lonsdale",
"C. J.",
""
],
[
"Lynch",
"M. J.",
""
],
[
"McWhirter",
"S. R.",
""
],
[
"Mitchell",
"D. A.",
""
],
[
"Morales",
"M. F.",
""
],
[
"Morgan",
"E.",
""
],
[
"Oberoi",
"D.",
""
],
[
"Ord",
"S. M.",
""
],
[
"Prabu",
"T.",
""
],
[
"Rogers",
"A. E. E.",
""
],
[
"Roshi",
"A.",
""
],
[
"Shankar",
"N. Udaya",
""
],
[
"Srivani",
"K. S.",
""
],
[
"Subrahmanyan",
"R.",
""
],
[
"Tingay",
"S. J.",
""
],
[
"Waterson",
"M.",
""
],
[
"Wayth",
"R. B.",
""
],
[
"Webster",
"R. L.",
""
],
[
"Whitney",
"A. R.",
""
],
[
"Williams",
"A.",
""
],
[
"Williams",
"C. L.",
""
]
] | TITLE: Power spectrum analysis of ionospheric fluctuations with the Murchison
Widefield Array
ABSTRACT: Low-frequency, wide field-of-view (FoV) radio telescopes such as the
Murchison Widefield Array (MWA) enable the ionosphere to be sampled at high
spatial completeness. We present the results of the first power spectrum
analysis of ionospheric fluctuations in MWA data, where we examined the
position offsets of radio sources appearing in two datasets. The refractive
shifts in the positions of celestial sources are proportional to spatial
gradients in the electron column density transverse to the line of sight. These
can be used to probe plasma structures and waves in the ionosphere. The
regional (10-100 km) scales probed by the MWA, determined by the size of its
FoV and the spatial density of radio sources (typically thousands in a single
FoV), complement the global (100-1000 km) scales of GPS studies and local
(0.01-1 km) scales of radar scattering measurements. Our data exhibit a range
of complex structures and waves. Some fluctuations have the characteristics of
travelling ionospheric disturbances (TIDs), while others take the form of
narrow, slowly-drifting bands aligned along the Earth's magnetic field.
| no_new_dataset | 0.940626 |
1506.07324 | Mariusz Tarnopolski | Mariusz Tarnopolski | Analysis of Fermi gamma-ray burst duration distribution | 6 pages, 3 figures; matches the version to be published | A&A 581, A29 (2015) | 10.1051/0004-6361/201526415 | null | astro-ph.HE astro-ph.CO hep-ph physics.space-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Two classes of GRBs, short and long, have been determined without any doubts,
and are usually prescribed to different physical scenarios. A third class,
intermediate in $T_{90}$ durations, has been reported to be present in the
datasets of BATSE, Swift, RHESSI and possibly BeppoSAX. The latest release of
$>1500$ GRBs observed by Fermi gives an opportunity to further investigate the
duration distribution. The aim of this paper is to investigate whether a third
class is present in the $\log T_{90}$ distribution, or is it described by a
bimodal distribution. A standard $\chi^2$ fitting of a mixture of Gaussians is
applied to 25 histograms with different binnings. Different binnings give
various values of the fitting parameters, as well as the shape of the fitted
curve. Among five statistically significant fits none is trimodal. Locations of
the Gaussian components are in agreement with previous works. However, a
trimodal distribution, understood in the sense of having three separated peaks,
is not found for any binning. It is concluded that the duration distribution in
Fermi data is well described by a mixture of three log-normal distributions,
but it is intrinsically bimodal, hence no third class is present in the
$T_{90}$ data of Fermi. It is suggested that the log-normal fit may not be an
adequate model.
| [
{
"version": "v1",
"created": "Wed, 24 Jun 2015 11:28:41 GMT"
},
{
"version": "v2",
"created": "Tue, 7 Jul 2015 15:39:44 GMT"
}
] | 2015-09-02T00:00:00 | [
[
"Tarnopolski",
"Mariusz",
""
]
] | TITLE: Analysis of Fermi gamma-ray burst duration distribution
ABSTRACT: Two classes of GRBs, short and long, have been determined without any doubts,
and are usually prescribed to different physical scenarios. A third class,
intermediate in $T_{90}$ durations, has been reported to be present in the
datasets of BATSE, Swift, RHESSI and possibly BeppoSAX. The latest release of
$>1500$ GRBs observed by Fermi gives an opportunity to further investigate the
duration distribution. The aim of this paper is to investigate whether a third
class is present in the $\log T_{90}$ distribution, or is it described by a
bimodal distribution. A standard $\chi^2$ fitting of a mixture of Gaussians is
applied to 25 histograms with different binnings. Different binnings give
various values of the fitting parameters, as well as the shape of the fitted
curve. Among five statistically significant fits none is trimodal. Locations of
the Gaussian components are in agreement with previous works. However, a
trimodal distribution, understood in the sense of having three separated peaks,
is not found for any binning. It is concluded that the duration distribution in
Fermi data is well described by a mixture of three log-normal distributions,
but it is intrinsically bimodal, hence no third class is present in the
$T_{90}$ data of Fermi. It is suggested that the log-normal fit may not be an
adequate model.
| no_new_dataset | 0.947381 |
1509.00083 | Samuel Kadoury | Samuel Kadoury, Eugene Vorontsov, An Tang | Metastatic liver tumour segmentation from discriminant Grassmannian
manifolds | null | Physics in Medicine and Biology 60 (2015) | 10.1088/0031-9155/60/16/6459 | null | cs.LG cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The early detection, diagnosis and monitoring of liver cancer progression can
be achieved with the precise delineation of metastatic tumours. However,
accurate automated segmentation remains challenging due to the presence of
noise, inhomogeneity and the high appearance variability of malignant tissue.
In this paper, we propose an unsupervised metastatic liver tumour segmentation
framework using a machine learning approach based on discriminant Grassmannian
manifolds which learns the appearance of tumours with respect to normal tissue.
First, the framework learns within-class and between-class similarity
distributions from a training set of images to discover the optimal manifold
discrimination between normal and pathological tissue in the liver. Second, a
conditional optimisation scheme computes nonlocal pairwise as well as
pattern-based clique potentials from the manifold subspace to recognise regions
with similar labelings and to incorporate global consistency in the
segmentation process. The proposed framework was validated on a clinical
database of 43 CT images from patients with metastatic liver cancer. Compared
to state-of-the-art methods, our method achieves a better performance on two
separate datasets of metastatic liver tumours from different clinical sites,
yielding an overall mean Dice similarity coefficient of 90.7 +/- 2.4 in over 50
tumours with an average volume of 27.3 mm3.
| [
{
"version": "v1",
"created": "Mon, 31 Aug 2015 21:45:40 GMT"
}
] | 2015-09-02T00:00:00 | [
[
"Kadoury",
"Samuel",
""
],
[
"Vorontsov",
"Eugene",
""
],
[
"Tang",
"An",
""
]
] | TITLE: Metastatic liver tumour segmentation from discriminant Grassmannian
manifolds
ABSTRACT: The early detection, diagnosis and monitoring of liver cancer progression can
be achieved with the precise delineation of metastatic tumours. However,
accurate automated segmentation remains challenging due to the presence of
noise, inhomogeneity and the high appearance variability of malignant tissue.
In this paper, we propose an unsupervised metastatic liver tumour segmentation
framework using a machine learning approach based on discriminant Grassmannian
manifolds which learns the appearance of tumours with respect to normal tissue.
First, the framework learns within-class and between-class similarity
distributions from a training set of images to discover the optimal manifold
discrimination between normal and pathological tissue in the liver. Second, a
conditional optimisation scheme computes nonlocal pairwise as well as
pattern-based clique potentials from the manifold subspace to recognise regions
with similar labelings and to incorporate global consistency in the
segmentation process. The proposed framework was validated on a clinical
database of 43 CT images from patients with metastatic liver cancer. Compared
to state-of-the-art methods, our method achieves a better performance on two
separate datasets of metastatic liver tumours from different clinical sites,
yielding an overall mean Dice similarity coefficient of 90.7 +/- 2.4 in over 50
tumours with an average volume of 27.3 mm3.
| no_new_dataset | 0.948155 |
1509.00313 | Amit K.C. | Amit Kumar K.C., Damien Delannay and Christophe De Vleeschouwer | Iterative hypothesis testing for multi-object tracking in presence of
features with variable reliability | 21 pages, 8 figures, submitted to CVIU: Special Issue on Visual
Tracking | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper assumes prior detections of multiple targets at each time instant,
and uses a graph-based approach to connect those detections across time, based
on their position and appearance estimates. In contrast to most earlier works
in the field, our framework has been designed to exploit the appearance
features, even when they are only sporadically available, or affected by a
non-stationary noise, along the sequence of detections. This is done by
implementing an iterative hypothesis testing strategy to progressively
aggregate the detections into short trajectories, named tracklets.
Specifically, each iteration considers a node, named key-node, and investigates
how to link this key-node with other nodes in its neighborhood, under the
assumption that the target appearance is defined by the key-node appearance
estimate. This is done through shortest path computation in a temporal
neighborhood of the key-node. The approach is conservative in that it only
aggregates the shortest paths that are sufficiently better compared to
alternative paths. It is also multi-scale in that the size of the investigated
neighborhood is increased proportionally to the number of detections already
aggregated into the key-node. The multi-scale nature of the process and the
progressive relaxation of its conservativeness makes it both computationally
efficient and effective.
Experimental validations are performed extensively on a toy example, a 15
minutes long multi-view basketball dataset, and other monocular pedestrian
datasets.
| [
{
"version": "v1",
"created": "Tue, 1 Sep 2015 14:27:50 GMT"
}
] | 2015-09-02T00:00:00 | [
[
"C.",
"Amit Kumar K.",
""
],
[
"Delannay",
"Damien",
""
],
[
"De Vleeschouwer",
"Christophe",
""
]
] | TITLE: Iterative hypothesis testing for multi-object tracking in presence of
features with variable reliability
ABSTRACT: This paper assumes prior detections of multiple targets at each time instant,
and uses a graph-based approach to connect those detections across time, based
on their position and appearance estimates. In contrast to most earlier works
in the field, our framework has been designed to exploit the appearance
features, even when they are only sporadically available, or affected by a
non-stationary noise, along the sequence of detections. This is done by
implementing an iterative hypothesis testing strategy to progressively
aggregate the detections into short trajectories, named tracklets.
Specifically, each iteration considers a node, named key-node, and investigates
how to link this key-node with other nodes in its neighborhood, under the
assumption that the target appearance is defined by the key-node appearance
estimate. This is done through shortest path computation in a temporal
neighborhood of the key-node. The approach is conservative in that it only
aggregates the shortest paths that are sufficiently better compared to
alternative paths. It is also multi-scale in that the size of the investigated
neighborhood is increased proportionally to the number of detections already
aggregated into the key-node. The multi-scale nature of the process and the
progressive relaxation of its conservativeness makes it both computationally
efficient and effective.
Experimental validations are performed extensively on a toy example, a 15
minutes long multi-view basketball dataset, and other monocular pedestrian
datasets.
| no_new_dataset | 0.941547 |
1509.00386 | Antonio Spanu | A. Spanu (1,2,3), M. de' Michieli Vitturi (1) and S. Barsotti (1,4)
((1) Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Pisa, Italy,
(2) Scuola Normale Superiore di Pisa, Italy, (3) Now at Deutschen Zentrums
fur Luft- und Raumfahrt, Germany, (4) Now at Icelandic Meteorological Office,
Iceland) | Reconstructing eruptive source parameters from tephra deposit: a
numerical approach for medium-sized explosive eruptions | Article:24 pages, 9 figures, 1 table, Auxiliary Material:xi pages, 6
figures, 1 table | null | null | null | physics.geo-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Since the seventies, several reconstruction techniques have been proposed,
and are currently used, to extrapolate and quantify eruptive parameters from
sampled deposit datasets. Discrete numbers of tephra ground loadings or
stratigraphic records are usually processed to estimate source eruptive values.
Reconstruction techniques like Pyle, Power law and Weibull are adopted as
standard to quantify the erupted mass (or volume) whereas Voronoi for
reconstructing the granulometry. Reconstructed values can be affected by large
uncertainty due to complexities occurring within the atmospheric dispersion and
deposition of volcanic particles. Here we want to quantify the sensitivity of
reconstruction techniques, and to quantify how much estimated values of mass
and grain size differ from emitted and deposited ones. We adopted a numerical
approach simulating with a dispersal code a mild explosive event occurring at
Mt. Etna, with eruptive parameters similar to those estimated for eruptions
occurred in the last decade. Then we created a synthetic deposit by integrating
the mass on the ground computed by the model over the computational domain
(>50000 km2). Multiple samplings of the simulated deposit are used for
generating a large dataset of sampling tests afterwards processed with standard
reconstruction techniques. Results are then compared and evaluated through a
statistical analysis, based on 2000 sampling tests of 100 samplings points. On
average, all the used techniques underestimate deposited and emitted mass. A
similar analysis, carried on Voronoi results, shows that information on the
total grain size distribution is strongly deteriorated. Here we present a new
method allowing an estimate of the deficiency in deposited mass for each
simulated class. Finally a sensitivity study on eruptive parameters is
presented in order to generalize our results to a wider range of eruptive
conditions.
| [
{
"version": "v1",
"created": "Tue, 1 Sep 2015 16:44:29 GMT"
}
] | 2015-09-02T00:00:00 | [
[
"Spanu",
"A.",
""
],
[
"Vitturi",
"M. de' Michieli",
""
],
[
"Barsotti",
"S.",
""
]
] | TITLE: Reconstructing eruptive source parameters from tephra deposit: a
numerical approach for medium-sized explosive eruptions
ABSTRACT: Since the seventies, several reconstruction techniques have been proposed,
and are currently used, to extrapolate and quantify eruptive parameters from
sampled deposit datasets. Discrete numbers of tephra ground loadings or
stratigraphic records are usually processed to estimate source eruptive values.
Reconstruction techniques like Pyle, Power law and Weibull are adopted as
standard to quantify the erupted mass (or volume) whereas Voronoi for
reconstructing the granulometry. Reconstructed values can be affected by large
uncertainty due to complexities occurring within the atmospheric dispersion and
deposition of volcanic particles. Here we want to quantify the sensitivity of
reconstruction techniques, and to quantify how much estimated values of mass
and grain size differ from emitted and deposited ones. We adopted a numerical
approach simulating with a dispersal code a mild explosive event occurring at
Mt. Etna, with eruptive parameters similar to those estimated for eruptions
occurred in the last decade. Then we created a synthetic deposit by integrating
the mass on the ground computed by the model over the computational domain
(>50000 km2). Multiple samplings of the simulated deposit are used for
generating a large dataset of sampling tests afterwards processed with standard
reconstruction techniques. Results are then compared and evaluated through a
statistical analysis, based on 2000 sampling tests of 100 samplings points. On
average, all the used techniques underestimate deposited and emitted mass. A
similar analysis, carried on Voronoi results, shows that information on the
total grain size distribution is strongly deteriorated. Here we present a new
method allowing an estimate of the deficiency in deposited mass for each
simulated class. Finally a sensitivity study on eruptive parameters is
presented in order to generalize our results to a wider range of eruptive
conditions.
| no_new_dataset | 0.950273 |
1502.05243 | Shanmuganathan Raman | Aalok Gangopadhyay, Shivam Mani Tripathi, Ishan Jindal, Shanmuganathan
Raman | SA-CNN: Dynamic Scene Classification using Convolutional Neural Networks | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The task of classifying videos of natural dynamic scenes into appropriate
classes has gained lot of attention in recent years. The problem especially
becomes challenging when the camera used to capture the video is dynamic. In
this paper, we analyse the performance of statistical aggregation (SA)
techniques on various pre-trained convolutional neural network(CNN) models to
address this problem. The proposed approach works by extracting CNN activation
features for a number of frames in a video and then uses an aggregation scheme
in order to obtain a robust feature descriptor for the video. We show through
results that the proposed approach performs better than the-state-of-the arts
for the Maryland and YUPenn dataset. The final descriptor obtained is powerful
enough to distinguish among dynamic scenes and is even capable of addressing
the scenario where the camera motion is dominant and the scene dynamics are
complex. Further, this paper shows an extensive study on the performance of
various aggregation methods and their combinations. We compare the proposed
approach with other dynamic scene classification algorithms on two publicly
available datasets - Maryland and YUPenn to demonstrate the superior
performance of the proposed approach.
| [
{
"version": "v1",
"created": "Tue, 17 Feb 2015 12:25:27 GMT"
},
{
"version": "v2",
"created": "Sat, 29 Aug 2015 06:01:02 GMT"
}
] | 2015-09-01T00:00:00 | [
[
"Gangopadhyay",
"Aalok",
""
],
[
"Tripathi",
"Shivam Mani",
""
],
[
"Jindal",
"Ishan",
""
],
[
"Raman",
"Shanmuganathan",
""
]
] | TITLE: SA-CNN: Dynamic Scene Classification using Convolutional Neural Networks
ABSTRACT: The task of classifying videos of natural dynamic scenes into appropriate
classes has gained lot of attention in recent years. The problem especially
becomes challenging when the camera used to capture the video is dynamic. In
this paper, we analyse the performance of statistical aggregation (SA)
techniques on various pre-trained convolutional neural network(CNN) models to
address this problem. The proposed approach works by extracting CNN activation
features for a number of frames in a video and then uses an aggregation scheme
in order to obtain a robust feature descriptor for the video. We show through
results that the proposed approach performs better than the-state-of-the arts
for the Maryland and YUPenn dataset. The final descriptor obtained is powerful
enough to distinguish among dynamic scenes and is even capable of addressing
the scenario where the camera motion is dominant and the scene dynamics are
complex. Further, this paper shows an extensive study on the performance of
various aggregation methods and their combinations. We compare the proposed
approach with other dynamic scene classification algorithms on two publicly
available datasets - Maryland and YUPenn to demonstrate the superior
performance of the proposed approach.
| no_new_dataset | 0.951051 |
1503.06289 | Pritheega Magalingam | Pritheega Magalingam, Stephen Davis, Asha Rao | Using shortest path to discover criminal community | null | DIIN584 2015 | 10.1016/j.diin.2015.08.002 | null | cs.SI physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Extracting communities using existing community detection algorithms yields
dense sub-networks that are difficult to analyse. Extracting a smaller sample
that embodies the relationships of a list of suspects is an important part of
the beginning of an investigation. In this paper, we present the efficacy of
our shortest paths network search algorithm (SPNSA) that begins with an
"algorithm feed", a small subset of nodes of particular interest, and builds an
investigative sub-network. The algorithm feed may consist of known criminals or
suspects, or persons of influence. This sets our approach apart from existing
community detection algorithms. We apply the SPNSA on the Enron Dataset of
e-mail communications starting with those convicted of money laundering in
relation to the collapse of Enron as the algorithm feed. The algorithm produces
sparse and small sub-networks that could feasibly identify a list of persons
and relationships to be further investigated. In contrast, we show that
identifying sub-networks of interest using either community detection
algorithms or a k-Neighbourhood approach produces sub-networks of much larger
size and complexity. When the 18 top managers of Enron were used as the
algorithm feed, the resulting sub-network identified 4 convicted criminals that
were not managers and so not part of the algorithm feed. We also directly
tested the SPNSA by removing one of the convicted criminals from the algorithm
feed and re-running the algorithm; in 5 out of 9 cases the left out criminal
occurred in the resulting sub-network.
| [
{
"version": "v1",
"created": "Sat, 21 Mar 2015 12:27:49 GMT"
}
] | 2015-09-01T00:00:00 | [
[
"Magalingam",
"Pritheega",
""
],
[
"Davis",
"Stephen",
""
],
[
"Rao",
"Asha",
""
]
] | TITLE: Using shortest path to discover criminal community
ABSTRACT: Extracting communities using existing community detection algorithms yields
dense sub-networks that are difficult to analyse. Extracting a smaller sample
that embodies the relationships of a list of suspects is an important part of
the beginning of an investigation. In this paper, we present the efficacy of
our shortest paths network search algorithm (SPNSA) that begins with an
"algorithm feed", a small subset of nodes of particular interest, and builds an
investigative sub-network. The algorithm feed may consist of known criminals or
suspects, or persons of influence. This sets our approach apart from existing
community detection algorithms. We apply the SPNSA on the Enron Dataset of
e-mail communications starting with those convicted of money laundering in
relation to the collapse of Enron as the algorithm feed. The algorithm produces
sparse and small sub-networks that could feasibly identify a list of persons
and relationships to be further investigated. In contrast, we show that
identifying sub-networks of interest using either community detection
algorithms or a k-Neighbourhood approach produces sub-networks of much larger
size and complexity. When the 18 top managers of Enron were used as the
algorithm feed, the resulting sub-network identified 4 convicted criminals that
were not managers and so not part of the algorithm feed. We also directly
tested the SPNSA by removing one of the convicted criminals from the algorithm
feed and re-running the algorithm; in 5 out of 9 cases the left out criminal
occurred in the resulting sub-network.
| no_new_dataset | 0.948106 |
1508.07551 | Karim Awudu | Awudu Karim and Shangbo Zhou | X-TREPAN: a multi class regression and adapted extraction of
comprehensible decision tree in artificial neural networks | 17 Pages, 8 Tables, 8 Figures, 6 Equations | null | null | null | cs.LG cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work, the TREPAN algorithm is enhanced and extended for extracting
decision trees from neural networks. We empirically evaluated the performance
of the algorithm on a set of databases from real world events. This benchmark
enhancement was achieved by adapting Single-test TREPAN and C4.5 decision tree
induction algorithms to analyze the datasets. The models are then compared with
X-TREPAN for comprehensibility and classification accuracy. Furthermore, we
validate the experimentations by applying statistical methods. Finally, the
modified algorithm is extended to work with multi-class regression problems and
the ability to comprehend generalized feed forward networks is achieved.
| [
{
"version": "v1",
"created": "Sun, 30 Aug 2015 10:14:48 GMT"
}
] | 2015-09-01T00:00:00 | [
[
"Karim",
"Awudu",
""
],
[
"Zhou",
"Shangbo",
""
]
] | TITLE: X-TREPAN: a multi class regression and adapted extraction of
comprehensible decision tree in artificial neural networks
ABSTRACT: In this work, the TREPAN algorithm is enhanced and extended for extracting
decision trees from neural networks. We empirically evaluated the performance
of the algorithm on a set of databases from real world events. This benchmark
enhancement was achieved by adapting Single-test TREPAN and C4.5 decision tree
induction algorithms to analyze the datasets. The models are then compared with
X-TREPAN for comprehensibility and classification accuracy. Furthermore, we
validate the experimentations by applying statistical methods. Finally, the
modified algorithm is extended to work with multi-class regression problems and
the ability to comprehend generalized feed forward networks is achieved.
| no_new_dataset | 0.949482 |
1106.2233 | Xiaowen Dong | Xiaowen Dong, Pascal Frossard, Pierre Vandergheynst and Nikolai
Nefedov | Clustering with Multi-Layer Graphs: A Spectral Perspective | null | IEEE Transactions on Signal Processing, vol. 60, no. 11, pp.
5820-5831, November 2012 | 10.1109/TSP.2012.2212886 | null | cs.LG cs.CV cs.SI stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Observational data usually comes with a multimodal nature, which means that
it can be naturally represented by a multi-layer graph whose layers share the
same set of vertices (users) with different edges (pairwise relationships). In
this paper, we address the problem of combining different layers of the
multi-layer graph for improved clustering of the vertices compared to using
layers independently. We propose two novel methods, which are based on joint
matrix factorization and graph regularization framework respectively, to
efficiently combine the spectrum of the multiple graph layers, namely the
eigenvectors of the graph Laplacian matrices. In each case, the resulting
combination, which we call a "joint spectrum" of multiple graphs, is used for
clustering the vertices. We evaluate our approaches by simulations with several
real world social network datasets. Results demonstrate the superior or
competitive performance of the proposed methods over state-of-the-art technique
and common baseline methods, such as co-regularization and summation of
information from individual graphs.
| [
{
"version": "v1",
"created": "Sat, 11 Jun 2011 12:43:18 GMT"
}
] | 2015-08-31T00:00:00 | [
[
"Dong",
"Xiaowen",
""
],
[
"Frossard",
"Pascal",
""
],
[
"Vandergheynst",
"Pierre",
""
],
[
"Nefedov",
"Nikolai",
""
]
] | TITLE: Clustering with Multi-Layer Graphs: A Spectral Perspective
ABSTRACT: Observational data usually comes with a multimodal nature, which means that
it can be naturally represented by a multi-layer graph whose layers share the
same set of vertices (users) with different edges (pairwise relationships). In
this paper, we address the problem of combining different layers of the
multi-layer graph for improved clustering of the vertices compared to using
layers independently. We propose two novel methods, which are based on joint
matrix factorization and graph regularization framework respectively, to
efficiently combine the spectrum of the multiple graph layers, namely the
eigenvectors of the graph Laplacian matrices. In each case, the resulting
combination, which we call a "joint spectrum" of multiple graphs, is used for
clustering the vertices. We evaluate our approaches by simulations with several
real world social network datasets. Results demonstrate the superior or
competitive performance of the proposed methods over state-of-the-art technique
and common baseline methods, such as co-regularization and summation of
information from individual graphs.
| no_new_dataset | 0.947284 |
1303.2221 | Xiaowen Dong | Xiaowen Dong, Pascal Frossard, Pierre Vandergheynst, Nikolai Nefedov | Clustering on Multi-Layer Graphs via Subspace Analysis on Grassmann
Manifolds | null | IEEE Transactions on Signal Processing, vol. 62, no. 4, pp.
905-918, February 2014 | 10.1109/TSP.2013.2295553 | null | cs.LG cs.CV cs.SI stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Relationships between entities in datasets are often of multiple nature, like
geographical distance, social relationships, or common interests among people
in a social network, for example. This information can naturally be modeled by
a set of weighted and undirected graphs that form a global multilayer graph,
where the common vertex set represents the entities and the edges on different
layers capture the similarities of the entities in term of the different
modalities. In this paper, we address the problem of analyzing multi-layer
graphs and propose methods for clustering the vertices by efficiently merging
the information provided by the multiple modalities. To this end, we propose to
combine the characteristics of individual graph layers using tools from
subspace analysis on a Grassmann manifold. The resulting combination can then
be viewed as a low dimensional representation of the original data which
preserves the most important information from diverse relationships between
entities. We use this information in new clustering methods and test our
algorithm on several synthetic and real world datasets where we demonstrate
superior or competitive performances compared to baseline and state-of-the-art
techniques. Our generic framework further extends to numerous analysis and
learning problems that involve different types of information on graphs.
| [
{
"version": "v1",
"created": "Sat, 9 Mar 2013 15:31:48 GMT"
}
] | 2015-08-31T00:00:00 | [
[
"Dong",
"Xiaowen",
""
],
[
"Frossard",
"Pascal",
""
],
[
"Vandergheynst",
"Pierre",
""
],
[
"Nefedov",
"Nikolai",
""
]
] | TITLE: Clustering on Multi-Layer Graphs via Subspace Analysis on Grassmann
Manifolds
ABSTRACT: Relationships between entities in datasets are often of multiple nature, like
geographical distance, social relationships, or common interests among people
in a social network, for example. This information can naturally be modeled by
a set of weighted and undirected graphs that form a global multilayer graph,
where the common vertex set represents the entities and the edges on different
layers capture the similarities of the entities in term of the different
modalities. In this paper, we address the problem of analyzing multi-layer
graphs and propose methods for clustering the vertices by efficiently merging
the information provided by the multiple modalities. To this end, we propose to
combine the characteristics of individual graph layers using tools from
subspace analysis on a Grassmann manifold. The resulting combination can then
be viewed as a low dimensional representation of the original data which
preserves the most important information from diverse relationships between
entities. We use this information in new clustering methods and test our
algorithm on several synthetic and real world datasets where we demonstrate
superior or competitive performances compared to baseline and state-of-the-art
techniques. Our generic framework further extends to numerous analysis and
learning problems that involve different types of information on graphs.
| no_new_dataset | 0.946349 |
1508.02593 | Denis Krompass | Denis Krompa{\ss} and Stephan Baier and Volker Tresp | Type-Constrained Representation Learning in Knowledge Graphs | null | null | null | null | cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large knowledge graphs increasingly add value to various applications that
require machines to recognize and understand queries and their semantics, as in
search or question answering systems. Latent variable models have increasingly
gained attention for the statistical modeling of knowledge graphs, showing
promising results in tasks related to knowledge graph completion and cleaning.
Besides storing facts about the world, schema-based knowledge graphs are backed
by rich semantic descriptions of entities and relation-types that allow
machines to understand the notion of things and their semantic relationships.
In this work, we study how type-constraints can generally support the
statistical modeling with latent variable models. More precisely, we integrated
prior knowledge in form of type-constraints in various state of the art latent
variable approaches. Our experimental results show that prior knowledge on
relation-types significantly improves these models up to 77% in link-prediction
tasks. The achieved improvements are especially prominent when a low model
complexity is enforced, a crucial requirement when these models are applied to
very large datasets. Unfortunately, type-constraints are neither always
available nor always complete e.g., they can become fuzzy when entities lack
proper typing. We show that in these cases, it can be beneficial to apply a
local closed-world assumption that approximates the semantics of relation-types
based on observations made in the data.
| [
{
"version": "v1",
"created": "Tue, 11 Aug 2015 13:49:07 GMT"
},
{
"version": "v2",
"created": "Fri, 28 Aug 2015 09:00:31 GMT"
}
] | 2015-08-31T00:00:00 | [
[
"Krompaß",
"Denis",
""
],
[
"Baier",
"Stephan",
""
],
[
"Tresp",
"Volker",
""
]
] | TITLE: Type-Constrained Representation Learning in Knowledge Graphs
ABSTRACT: Large knowledge graphs increasingly add value to various applications that
require machines to recognize and understand queries and their semantics, as in
search or question answering systems. Latent variable models have increasingly
gained attention for the statistical modeling of knowledge graphs, showing
promising results in tasks related to knowledge graph completion and cleaning.
Besides storing facts about the world, schema-based knowledge graphs are backed
by rich semantic descriptions of entities and relation-types that allow
machines to understand the notion of things and their semantic relationships.
In this work, we study how type-constraints can generally support the
statistical modeling with latent variable models. More precisely, we integrated
prior knowledge in form of type-constraints in various state of the art latent
variable approaches. Our experimental results show that prior knowledge on
relation-types significantly improves these models up to 77% in link-prediction
tasks. The achieved improvements are especially prominent when a low model
complexity is enforced, a crucial requirement when these models are applied to
very large datasets. Unfortunately, type-constraints are neither always
available nor always complete e.g., they can become fuzzy when entities lack
proper typing. We show that in these cases, it can be beneficial to apply a
local closed-world assumption that approximates the semantics of relation-types
based on observations made in the data.
| no_new_dataset | 0.944228 |
1508.07053 | Kenji Hata | Kenji Hata, Sherman Leung, Ranjay Krishna, Michael S. Bernstein, Li
Fei-Fei | SentenceRacer: A Game with a Purpose for Image Sentence Annotation | 2 pages, 2 figures, 2 tables, potential CSCW poster submission | null | null | null | cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recently datasets that contain sentence descriptions of images have enabled
models that can automatically generate image captions. However, collecting
these datasets are still very expensive. Here, we present SentenceRacer, an
online game that gathers and verifies descriptions of images at no cost.
Similar to the game hangman, players compete to uncover words in a sentence
that ultimately describes an image. SentenceRacer both generates and verifies
that the sentences are accurate descriptions. We show that SentenceRacer
generates annotations of higher quality than those generated on Amazon
Mechanical Turk (AMT).
| [
{
"version": "v1",
"created": "Thu, 27 Aug 2015 23:03:17 GMT"
}
] | 2015-08-31T00:00:00 | [
[
"Hata",
"Kenji",
""
],
[
"Leung",
"Sherman",
""
],
[
"Krishna",
"Ranjay",
""
],
[
"Bernstein",
"Michael S.",
""
],
[
"Fei-Fei",
"Li",
""
]
] | TITLE: SentenceRacer: A Game with a Purpose for Image Sentence Annotation
ABSTRACT: Recently datasets that contain sentence descriptions of images have enabled
models that can automatically generate image captions. However, collecting
these datasets are still very expensive. Here, we present SentenceRacer, an
online game that gathers and verifies descriptions of images at no cost.
Similar to the game hangman, players compete to uncover words in a sentence
that ultimately describes an image. SentenceRacer both generates and verifies
that the sentences are accurate descriptions. We show that SentenceRacer
generates annotations of higher quality than those generated on Amazon
Mechanical Turk (AMT).
| no_new_dataset | 0.89974 |
1508.07148 | Thanh-Toan Do | Thanh-Toan Do, Anh-Zung Doan, Ngai-Man Cheung | Discrete Hashing with Deep Neural Network | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper addresses the problem of learning binary hash codes for large
scale image search by proposing a novel hashing method based on deep neural
network. The advantage of our deep model over previous deep model used in
hashing is that our model contains necessary criteria for producing good codes
such as similarity preserving, balance and independence. Another advantage of
our method is that instead of relaxing the binary constraint of codes during
the learning process as most previous works, in this paper, by introducing the
auxiliary variable, we reformulate the optimization into two sub-optimization
steps allowing us to efficiently solve binary constraints without any
relaxation.
The proposed method is also extended to the supervised hashing by leveraging
the label information such that the learned binary codes preserve the pairwise
label of inputs.
The experimental results on three benchmark datasets show the proposed
methods outperform state-of-the-art hashing methods.
| [
{
"version": "v1",
"created": "Fri, 28 Aug 2015 09:38:05 GMT"
}
] | 2015-08-31T00:00:00 | [
[
"Do",
"Thanh-Toan",
""
],
[
"Doan",
"Anh-Zung",
""
],
[
"Cheung",
"Ngai-Man",
""
]
] | TITLE: Discrete Hashing with Deep Neural Network
ABSTRACT: This paper addresses the problem of learning binary hash codes for large
scale image search by proposing a novel hashing method based on deep neural
network. The advantage of our deep model over previous deep model used in
hashing is that our model contains necessary criteria for producing good codes
such as similarity preserving, balance and independence. Another advantage of
our method is that instead of relaxing the binary constraint of codes during
the learning process as most previous works, in this paper, by introducing the
auxiliary variable, we reformulate the optimization into two sub-optimization
steps allowing us to efficiently solve binary constraints without any
relaxation.
The proposed method is also extended to the supervised hashing by leveraging
the label information such that the learned binary codes preserve the pairwise
label of inputs.
The experimental results on three benchmark datasets show the proposed
methods outperform state-of-the-art hashing methods.
| no_new_dataset | 0.947527 |
1508.07275 | Luiz Capretz Dr. | Ali Bou Nassif, Mohammad Azzeh, Luiz Fernando Capretz, Danny Ho | A Comparison Between Decision Trees and Decision Tree Forest Models for
Software Development Effort Estimation | 3rd International Conference on Communications and Information
Technology (ICCIT), Beirut, Lebanon, pp. 220-224, 2013 | null | 10.1109/ICCITechnology.2013.6579553 | null | cs.SE cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Accurate software effort estimation has been a challenge for many software
practitioners and project managers. Underestimation leads to disruption in the
projects estimated cost and delivery. On the other hand, overestimation causes
outbidding and financial losses in business. Many software estimation models
exist; however, none have been proven to be the best in all situations. In this
paper, a decision tree forest (DTF) model is compared to a traditional decision
tree (DT) model, as well as a multiple linear regression model (MLR). The
evaluation was conducted using ISBSG and Desharnais industrial datasets.
Results show that the DTF model is competitive and can be used as an
alternative in software effort prediction.
| [
{
"version": "v1",
"created": "Fri, 28 Aug 2015 16:52:21 GMT"
}
] | 2015-08-31T00:00:00 | [
[
"Nassif",
"Ali Bou",
""
],
[
"Azzeh",
"Mohammad",
""
],
[
"Capretz",
"Luiz Fernando",
""
],
[
"Ho",
"Danny",
""
]
] | TITLE: A Comparison Between Decision Trees and Decision Tree Forest Models for
Software Development Effort Estimation
ABSTRACT: Accurate software effort estimation has been a challenge for many software
practitioners and project managers. Underestimation leads to disruption in the
projects estimated cost and delivery. On the other hand, overestimation causes
outbidding and financial losses in business. Many software estimation models
exist; however, none have been proven to be the best in all situations. In this
paper, a decision tree forest (DTF) model is compared to a traditional decision
tree (DT) model, as well as a multiple linear regression model (MLR). The
evaluation was conducted using ISBSG and Desharnais industrial datasets.
Results show that the DTF model is competitive and can be used as an
alternative in software effort prediction.
| no_new_dataset | 0.951953 |
1508.07292 | Anastasios Noulas Anastasios Noulas | Anastasios Noulas, Vsevolod Salnikov, Renaud Lambiotte, Cecilia
Mascolo | Mining Open Datasets for Transparency in Taxi Transport in Metropolitan
Environments | null | null | null | null | cs.CY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Uber has recently been introducing novel practices in urban taxi transport.
Journey prices can change dynamically in almost real time and also vary
geographically from one area to another in a city, a strategy known as surge
pricing. In this paper, we explore the power of the new generation of open
datasets towards understanding the impact of the new disruption technologies
that emerge in the area of public transport. With our primary goal being a more
transparent economic landscape for urban commuters, we provide a direct price
comparison between Uber and the Yellow Cab company in New York. We discover
that Uber, despite its lower standard pricing rates, effectively charges higher
fares on average, especially during short in length, but frequent in
occurrence, taxi journeys. Building on this insight, we develop a smartphone
application, OpenStreetCab, that offers a personalized consultation to mobile
users on which taxi provider is cheaper for their journey. Almost five months
after its launch, the app has attracted more than three thousand users in a
single city. Their journey queries have provided additional insights on the
potential savings similar technologies can have for urban commuters, with a
highlight being that on average, a user in New York saves 6 U.S. Dollars per
taxi journey if they pick the cheapest taxi provider. We run extensive
experiments to show how Uber's surge pricing is the driving factor of higher
journey prices and therefore higher potential savings for our application's
users. Finally, motivated by the observation that Uber's surge pricing is
occurring more frequently that intuitively expected, we formulate a prediction
task where the aim becomes to predict a geographic area's tendency to surge.
Using exogenous to Uber datasets we show how it is possible to estimate
customer demand within an area, and by extension surge pricing, with high
accuracy.
| [
{
"version": "v1",
"created": "Thu, 27 Aug 2015 16:33:52 GMT"
}
] | 2015-08-31T00:00:00 | [
[
"Noulas",
"Anastasios",
""
],
[
"Salnikov",
"Vsevolod",
""
],
[
"Lambiotte",
"Renaud",
""
],
[
"Mascolo",
"Cecilia",
""
]
] | TITLE: Mining Open Datasets for Transparency in Taxi Transport in Metropolitan
Environments
ABSTRACT: Uber has recently been introducing novel practices in urban taxi transport.
Journey prices can change dynamically in almost real time and also vary
geographically from one area to another in a city, a strategy known as surge
pricing. In this paper, we explore the power of the new generation of open
datasets towards understanding the impact of the new disruption technologies
that emerge in the area of public transport. With our primary goal being a more
transparent economic landscape for urban commuters, we provide a direct price
comparison between Uber and the Yellow Cab company in New York. We discover
that Uber, despite its lower standard pricing rates, effectively charges higher
fares on average, especially during short in length, but frequent in
occurrence, taxi journeys. Building on this insight, we develop a smartphone
application, OpenStreetCab, that offers a personalized consultation to mobile
users on which taxi provider is cheaper for their journey. Almost five months
after its launch, the app has attracted more than three thousand users in a
single city. Their journey queries have provided additional insights on the
potential savings similar technologies can have for urban commuters, with a
highlight being that on average, a user in New York saves 6 U.S. Dollars per
taxi journey if they pick the cheapest taxi provider. We run extensive
experiments to show how Uber's surge pricing is the driving factor of higher
journey prices and therefore higher potential savings for our application's
users. Finally, motivated by the observation that Uber's surge pricing is
occurring more frequently that intuitively expected, we formulate a prediction
task where the aim becomes to predict a geographic area's tendency to surge.
Using exogenous to Uber datasets we show how it is possible to estimate
customer demand within an area, and by extension surge pricing, with high
accuracy.
| no_new_dataset | 0.91957 |
1508.07306 | Ashwin Machanavajjhala | Yan Chen and Ashwin Machanavajjhala | On the Privacy Properties of Variants on the Sparse Vector Technique | 8 pages | null | null | null | cs.DB cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The sparse vector technique is a powerful differentially private primitive
that allows an analyst to check whether queries in a stream are greater or
lesser than a threshold. This technique has a unique property -- the algorithm
works by adding noise with a finite variance to the queries and the threshold,
and guarantees privacy that only degrades with (a) the maximum sensitivity of
any one query in stream, and (b) the number of positive answers output by the
algorithm. Recent work has developed variants of this algorithm, which we call
{\em generalized private threshold testing}, and are claimed to have privacy
guarantees that do not depend on the number of positive or negative answers
output by the algorithm. These algorithms result in a significant improvement
in utility over the sparse vector technique for a given privacy budget, and
have found applications in frequent itemset mining, feature selection in
machine learning and generating synthetic data.
In this paper we critically analyze the privacy properties of generalized
private threshold testing. We show that generalized private threshold testing
does not satisfy \epsilon-differential privacy for any finite \epsilon. We
identify a subtle error in the privacy analysis of this technique in prior
work. Moreover, we show an adversary can use generalized private threshold
testing to recover counts from the datasets (especially small counts) exactly
with high accuracy, and thus can result in individuals being reidentified. We
demonstrate our attacks empirically on real datasets.
| [
{
"version": "v1",
"created": "Fri, 28 Aug 2015 18:42:56 GMT"
}
] | 2015-08-31T00:00:00 | [
[
"Chen",
"Yan",
""
],
[
"Machanavajjhala",
"Ashwin",
""
]
] | TITLE: On the Privacy Properties of Variants on the Sparse Vector Technique
ABSTRACT: The sparse vector technique is a powerful differentially private primitive
that allows an analyst to check whether queries in a stream are greater or
lesser than a threshold. This technique has a unique property -- the algorithm
works by adding noise with a finite variance to the queries and the threshold,
and guarantees privacy that only degrades with (a) the maximum sensitivity of
any one query in stream, and (b) the number of positive answers output by the
algorithm. Recent work has developed variants of this algorithm, which we call
{\em generalized private threshold testing}, and are claimed to have privacy
guarantees that do not depend on the number of positive or negative answers
output by the algorithm. These algorithms result in a significant improvement
in utility over the sparse vector technique for a given privacy budget, and
have found applications in frequent itemset mining, feature selection in
machine learning and generating synthetic data.
In this paper we critically analyze the privacy properties of generalized
private threshold testing. We show that generalized private threshold testing
does not satisfy \epsilon-differential privacy for any finite \epsilon. We
identify a subtle error in the privacy analysis of this technique in prior
work. Moreover, we show an adversary can use generalized private threshold
testing to recover counts from the datasets (especially small counts) exactly
with high accuracy, and thus can result in individuals being reidentified. We
demonstrate our attacks empirically on real datasets.
| no_new_dataset | 0.949809 |
1502.06719 | Anamika Chhabra | Anamika Chhabra, S. R. S. Iyengar, Poonam Saini, Rajesh Shreedhar
Bhat, Vijay Kumar | Ecosystem: A Characteristic Of Crowdsourced Environments | 21 pages, 9 figures, 7 tables | null | null | null | cs.CY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The phenomenal success of certain crowdsourced online platforms, such as
Wikipedia, is accredited to their ability to tap the crowd's potential to
collaboratively build knowledge. While it is well known that the crowd's
collective wisdom surpasses the cumulative individual expertise, little is
understood on the dynamics of knowledge building in a crowdsourced environment.
A proper understanding of the dynamics of knowledge building in a crowdsourced
environment would enable one in the better designing of such environments to
solicit knowledge from the crowd. Our experiment on crowdsourced systems based
on annotations shows that an important reason for the rapid knowledge building
in such environments is due to variance in expertise. First, we used as our
test bed, a customized Crowdsourced Annotation System (CAS) which provides a
group of users the facility to annotate a given document while trying to
understand it. Our results showed the presence of different genres of
proficiency amongst the users of an annotation system. We observed that the
ecosystem in crowdsourced annotation system comprised of mainly four categories
of contributors, namely: Probers, Solvers, Articulators and Explorers. We
inferred from our experiment that the knowledge garnering mainly happens due to
the synergetic interaction across these categories. Further, we conducted an
analysis on the dataset of Wikipedia and Stack Overflow and noticed the
ecosystem presence in these portals as well. From this study, we claim that the
ecosystem is a universal characteristic of all crowdsourced portals.
| [
{
"version": "v1",
"created": "Tue, 24 Feb 2015 09:11:19 GMT"
},
{
"version": "v2",
"created": "Sat, 16 May 2015 09:35:01 GMT"
},
{
"version": "v3",
"created": "Tue, 11 Aug 2015 06:11:37 GMT"
},
{
"version": "v4",
"created": "Thu, 27 Aug 2015 16:46:22 GMT"
}
] | 2015-08-28T00:00:00 | [
[
"Chhabra",
"Anamika",
""
],
[
"Iyengar",
"S. R. S.",
""
],
[
"Saini",
"Poonam",
""
],
[
"Bhat",
"Rajesh Shreedhar",
""
],
[
"Kumar",
"Vijay",
""
]
] | TITLE: Ecosystem: A Characteristic Of Crowdsourced Environments
ABSTRACT: The phenomenal success of certain crowdsourced online platforms, such as
Wikipedia, is accredited to their ability to tap the crowd's potential to
collaboratively build knowledge. While it is well known that the crowd's
collective wisdom surpasses the cumulative individual expertise, little is
understood on the dynamics of knowledge building in a crowdsourced environment.
A proper understanding of the dynamics of knowledge building in a crowdsourced
environment would enable one in the better designing of such environments to
solicit knowledge from the crowd. Our experiment on crowdsourced systems based
on annotations shows that an important reason for the rapid knowledge building
in such environments is due to variance in expertise. First, we used as our
test bed, a customized Crowdsourced Annotation System (CAS) which provides a
group of users the facility to annotate a given document while trying to
understand it. Our results showed the presence of different genres of
proficiency amongst the users of an annotation system. We observed that the
ecosystem in crowdsourced annotation system comprised of mainly four categories
of contributors, namely: Probers, Solvers, Articulators and Explorers. We
inferred from our experiment that the knowledge garnering mainly happens due to
the synergetic interaction across these categories. Further, we conducted an
analysis on the dataset of Wikipedia and Stack Overflow and noticed the
ecosystem presence in these portals as well. From this study, we claim that the
ecosystem is a universal characteristic of all crowdsourced portals.
| no_new_dataset | 0.940079 |
1506.03487 | John Wieting | John Wieting, Mohit Bansal, Kevin Gimpel, Karen Livescu, and Dan Roth | From Paraphrase Database to Compositional Paraphrase Model and Back | 2015 TACL paper updated with an appendix describing new 300
dimensional embeddings. Submitted 1/2015. Accepted 2/2015. Published 6/2015 | TACL Vol 3 (2015) pg 345-358 | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The Paraphrase Database (PPDB; Ganitkevitch et al., 2013) is an extensive
semantic resource, consisting of a list of phrase pairs with (heuristic)
confidence estimates. However, it is still unclear how it can best be used, due
to the heuristic nature of the confidences and its necessarily incomplete
coverage. We propose models to leverage the phrase pairs from the PPDB to build
parametric paraphrase models that score paraphrase pairs more accurately than
the PPDB's internal scores while simultaneously improving its coverage. They
allow for learning phrase embeddings as well as improved word embeddings.
Moreover, we introduce two new, manually annotated datasets to evaluate
short-phrase paraphrasing models. Using our paraphrase model trained using
PPDB, we achieve state-of-the-art results on standard word and bigram
similarity tasks and beat strong baselines on our new short phrase paraphrase
tasks.
| [
{
"version": "v1",
"created": "Wed, 10 Jun 2015 21:29:28 GMT"
},
{
"version": "v2",
"created": "Wed, 26 Aug 2015 21:18:00 GMT"
}
] | 2015-08-28T00:00:00 | [
[
"Wieting",
"John",
""
],
[
"Bansal",
"Mohit",
""
],
[
"Gimpel",
"Kevin",
""
],
[
"Livescu",
"Karen",
""
],
[
"Roth",
"Dan",
""
]
] | TITLE: From Paraphrase Database to Compositional Paraphrase Model and Back
ABSTRACT: The Paraphrase Database (PPDB; Ganitkevitch et al., 2013) is an extensive
semantic resource, consisting of a list of phrase pairs with (heuristic)
confidence estimates. However, it is still unclear how it can best be used, due
to the heuristic nature of the confidences and its necessarily incomplete
coverage. We propose models to leverage the phrase pairs from the PPDB to build
parametric paraphrase models that score paraphrase pairs more accurately than
the PPDB's internal scores while simultaneously improving its coverage. They
allow for learning phrase embeddings as well as improved word embeddings.
Moreover, we introduce two new, manually annotated datasets to evaluate
short-phrase paraphrasing models. Using our paraphrase model trained using
PPDB, we achieve state-of-the-art results on standard word and bigram
similarity tasks and beat strong baselines on our new short phrase paraphrase
tasks.
| new_dataset | 0.95877 |
1508.06708 | Sijin Li | Sijin Li, Weichen Zhang, Antoni B. Chan | Maximum-Margin Structured Learning with Deep Networks for 3D Human Pose
Estimation | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper focuses on structured-output learning using deep neural networks
for 3D human pose estimation from monocular images. Our network takes an image
and 3D pose as inputs and outputs a score value, which is high when the
image-pose pair matches and low otherwise. The network structure consists of a
convolutional neural network for image feature extraction, followed by two
sub-networks for transforming the image features and pose into a joint
embedding. The score function is then the dot-product between the image and
pose embeddings. The image-pose embedding and score function are jointly
trained using a maximum-margin cost function. Our proposed framework can be
interpreted as a special form of structured support vector machines where the
joint feature space is discriminatively learned using deep neural networks. We
test our framework on the Human3.6m dataset and obtain state-of-the-art results
compared to other recent methods. Finally, we present visualizations of the
image-pose embedding space, demonstrating the network has learned a high-level
embedding of body-orientation and pose-configuration.
| [
{
"version": "v1",
"created": "Thu, 27 Aug 2015 03:21:15 GMT"
}
] | 2015-08-28T00:00:00 | [
[
"Li",
"Sijin",
""
],
[
"Zhang",
"Weichen",
""
],
[
"Chan",
"Antoni B.",
""
]
] | TITLE: Maximum-Margin Structured Learning with Deep Networks for 3D Human Pose
Estimation
ABSTRACT: This paper focuses on structured-output learning using deep neural networks
for 3D human pose estimation from monocular images. Our network takes an image
and 3D pose as inputs and outputs a score value, which is high when the
image-pose pair matches and low otherwise. The network structure consists of a
convolutional neural network for image feature extraction, followed by two
sub-networks for transforming the image features and pose into a joint
embedding. The score function is then the dot-product between the image and
pose embeddings. The image-pose embedding and score function are jointly
trained using a maximum-margin cost function. Our proposed framework can be
interpreted as a special form of structured support vector machines where the
joint feature space is discriminatively learned using deep neural networks. We
test our framework on the Human3.6m dataset and obtain state-of-the-art results
compared to other recent methods. Finally, we present visualizations of the
image-pose embedding space, demonstrating the network has learned a high-level
embedding of body-orientation and pose-configuration.
| no_new_dataset | 0.946745 |
1508.06878 | Kunal Bhattacharya | Kunal Bhattacharya, Asim Ghosh, Daniel Monsivais, Robin I. M. Dunbar
and Kimmo Kaski | Sex differences in social focus across the lifecycle in humans | 11 pages, 6 figures | null | null | null | physics.soc-ph cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Age and gender are two important factors that play crucial roles in the way
organisms allocate their social effort. In this study, we analyse a large
mobile phone dataset to explore the way lifehistory influences human sociality
and the way social networks are structured. Our results indicate that these
aspects of human behaviour are strongly related to the age and gender such that
younger individuals have more contacts and, among them, males more than
females. However, the rate of decrease in the number of contacts with age
differs between males and females, such that there is a reversal in the number
of contacts around the late 30s. We suggest that this pattern can be attributed
to the difference in reproductive investments that are made by the two sexes.
We analyse the inequality in social investment patterns and suggest that the
age and gender-related differences that we find reflect the constraints imposed
by reproduction in a context where time (a form of social capital) is limited.
| [
{
"version": "v1",
"created": "Thu, 27 Aug 2015 14:40:38 GMT"
}
] | 2015-08-28T00:00:00 | [
[
"Bhattacharya",
"Kunal",
""
],
[
"Ghosh",
"Asim",
""
],
[
"Monsivais",
"Daniel",
""
],
[
"Dunbar",
"Robin I. M.",
""
],
[
"Kaski",
"Kimmo",
""
]
] | TITLE: Sex differences in social focus across the lifecycle in humans
ABSTRACT: Age and gender are two important factors that play crucial roles in the way
organisms allocate their social effort. In this study, we analyse a large
mobile phone dataset to explore the way lifehistory influences human sociality
and the way social networks are structured. Our results indicate that these
aspects of human behaviour are strongly related to the age and gender such that
younger individuals have more contacts and, among them, males more than
females. However, the rate of decrease in the number of contacts with age
differs between males and females, such that there is a reversal in the number
of contacts around the late 30s. We suggest that this pattern can be attributed
to the difference in reproductive investments that are made by the two sexes.
We analyse the inequality in social investment patterns and suggest that the
age and gender-related differences that we find reflect the constraints imposed
by reproduction in a context where time (a form of social capital) is limited.
| no_new_dataset | 0.928668 |
1508.06976 | Byung Suk Lee | Saurav Acharya, Byung Suk Lee and Paul Hines | Real-time Top-K Predictive Query Processing over Event Streams | null | null | null | null | cs.DB cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper addresses the problem of predicting the k events that are most
likely to occur next, over historical real-time event streams. Existing
approaches to causal prediction queries have a number of limitations. First,
they exhaustively search over an acyclic causal network to find the most likely
k effect events; however, data from real event streams frequently reflect
cyclic causality. Second, they contain conservative assumptions intended to
exclude all possible non-causal links in the causal network; it leads to the
omission of many less-frequent but important causal links. We overcome these
limitations by proposing a novel event precedence model and a run-time causal
inference mechanism. The event precedence model constructs a first order
absorbing Markov chain incrementally over event streams, where an edge between
two events signifies a temporal precedence relationship between them, which is
a necessary condition for causality. Then, the run-time causal inference
mechanism learns causal relationships dynamically during query processing. This
is done by removing some of the temporal precedence relationships that do not
exhibit causality in the presence of other events in the event precedence
model. This paper presents two query processing algorithms -- one performs
exhaustive search on the model and the other performs a more efficient reduced
search with early termination. Experiments using two real datasets (cascading
blackouts in power systems and web page views) verify the effectiveness of the
probabilistic top-k prediction queries and the efficiency of the algorithms.
Specifically, the reduced search algorithm reduced runtime, relative to
exhaustive search, by 25-80% (depending on the application) with only a small
reduction in accuracy.
| [
{
"version": "v1",
"created": "Wed, 26 Aug 2015 15:02:09 GMT"
}
] | 2015-08-28T00:00:00 | [
[
"Acharya",
"Saurav",
""
],
[
"Lee",
"Byung Suk",
""
],
[
"Hines",
"Paul",
""
]
] | TITLE: Real-time Top-K Predictive Query Processing over Event Streams
ABSTRACT: This paper addresses the problem of predicting the k events that are most
likely to occur next, over historical real-time event streams. Existing
approaches to causal prediction queries have a number of limitations. First,
they exhaustively search over an acyclic causal network to find the most likely
k effect events; however, data from real event streams frequently reflect
cyclic causality. Second, they contain conservative assumptions intended to
exclude all possible non-causal links in the causal network; it leads to the
omission of many less-frequent but important causal links. We overcome these
limitations by proposing a novel event precedence model and a run-time causal
inference mechanism. The event precedence model constructs a first order
absorbing Markov chain incrementally over event streams, where an edge between
two events signifies a temporal precedence relationship between them, which is
a necessary condition for causality. Then, the run-time causal inference
mechanism learns causal relationships dynamically during query processing. This
is done by removing some of the temporal precedence relationships that do not
exhibit causality in the presence of other events in the event precedence
model. This paper presents two query processing algorithms -- one performs
exhaustive search on the model and the other performs a more efficient reduced
search with early termination. Experiments using two real datasets (cascading
blackouts in power systems and web page views) verify the effectiveness of the
probabilistic top-k prediction queries and the efficiency of the algorithms.
Specifically, the reduced search algorithm reduced runtime, relative to
exhaustive search, by 25-80% (depending on the application) with only a small
reduction in accuracy.
| no_new_dataset | 0.957517 |
1502.01199 | Reza Farrahi Moghaddam | Reza Farrahi Moghaddam and Mohamed Cheriet | A Multiple-Expert Binarization Framework for Multispectral Images | 12 pages, 8 figures, 6 tables. Presented at ICDAR'15 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work, a multiple-expert binarization framework for multispectral
images is proposed. The framework is based on a constrained subspace selection
limited to the spectral bands combined with state-of-the-art gray-level
binarization methods. The framework uses a binarization wrapper to enhance the
performance of the gray-level binarization. Nonlinear preprocessing of the
individual spectral bands is used to enhance the textual information. An
evolutionary optimizer is considered to obtain the optimal and some suboptimal
3-band subspaces from which an ensemble of experts is then formed. The
framework is applied to a ground truth multispectral dataset with promising
results. In addition, a generalization to the cross-validation approach is
developed that not only evaluates generalizability of the framework, it also
provides a practical instance of the selected experts that could be then
applied to unseen inputs despite the small size of the given ground truth
dataset.
| [
{
"version": "v1",
"created": "Wed, 4 Feb 2015 14:01:38 GMT"
},
{
"version": "v2",
"created": "Thu, 5 Feb 2015 18:56:40 GMT"
},
{
"version": "v3",
"created": "Mon, 9 Feb 2015 17:42:37 GMT"
},
{
"version": "v4",
"created": "Wed, 11 Feb 2015 18:04:21 GMT"
},
{
"version": "v5",
"created": "Mon, 13 Apr 2015 14:49:16 GMT"
},
{
"version": "v6",
"created": "Wed, 26 Aug 2015 13:27:54 GMT"
}
] | 2015-08-27T00:00:00 | [
[
"Moghaddam",
"Reza Farrahi",
""
],
[
"Cheriet",
"Mohamed",
""
]
] | TITLE: A Multiple-Expert Binarization Framework for Multispectral Images
ABSTRACT: In this work, a multiple-expert binarization framework for multispectral
images is proposed. The framework is based on a constrained subspace selection
limited to the spectral bands combined with state-of-the-art gray-level
binarization methods. The framework uses a binarization wrapper to enhance the
performance of the gray-level binarization. Nonlinear preprocessing of the
individual spectral bands is used to enhance the textual information. An
evolutionary optimizer is considered to obtain the optimal and some suboptimal
3-band subspaces from which an ensemble of experts is then formed. The
framework is applied to a ground truth multispectral dataset with promising
results. In addition, a generalization to the cross-validation approach is
developed that not only evaluates generalizability of the framework, it also
provides a practical instance of the selected experts that could be then
applied to unseen inputs despite the small size of the given ground truth
dataset.
| no_new_dataset | 0.944587 |
1508.06314 | Maher Salloum | Maher Salloum, Nathan Fabian, David M. Hensinger, Jeremy A. Templeton | Compressed Sensing and Reconstruction of Unstructured Mesh Datasets | 18 pages, 7 figures | null | null | SAND2015-4995C | cs.IT cs.DC cs.SY math.IT math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Exascale computing promises quantities of data too large to efficiently store
and transfer across networks in order to be able to analyze and visualize the
results. We investigate Compressive Sensing (CS) as a way to reduce the size of
the data as it is being stored. CS works by sampling the data on the
computational cluster within an alternative function space such as wavelet
bases, and then reconstructing back to the original space on visualization
platforms. While much work has gone into exploring CS on structured data sets,
such as image data, we investigate its usefulness for point clouds such as
unstructured mesh datasets found in many finite element simulations. We sample
using second generation wavelets (SGW) and reconstruct using the Stagewise
Orthogonal Matching Pursuit (StOMP) algorithm. We analyze the compression
ratios achievable and quality of reconstructed results at each compression
rate. We are able to achieve compression ratios between 10 and 30 on moderate
size datasets with minimal visual deterioration as a result of the lossy
compression.
| [
{
"version": "v1",
"created": "Tue, 25 Aug 2015 21:46:30 GMT"
}
] | 2015-08-27T00:00:00 | [
[
"Salloum",
"Maher",
""
],
[
"Fabian",
"Nathan",
""
],
[
"Hensinger",
"David M.",
""
],
[
"Templeton",
"Jeremy A.",
""
]
] | TITLE: Compressed Sensing and Reconstruction of Unstructured Mesh Datasets
ABSTRACT: Exascale computing promises quantities of data too large to efficiently store
and transfer across networks in order to be able to analyze and visualize the
results. We investigate Compressive Sensing (CS) as a way to reduce the size of
the data as it is being stored. CS works by sampling the data on the
computational cluster within an alternative function space such as wavelet
bases, and then reconstructing back to the original space on visualization
platforms. While much work has gone into exploring CS on structured data sets,
such as image data, we investigate its usefulness for point clouds such as
unstructured mesh datasets found in many finite element simulations. We sample
using second generation wavelets (SGW) and reconstruct using the Stagewise
Orthogonal Matching Pursuit (StOMP) algorithm. We analyze the compression
ratios achievable and quality of reconstructed results at each compression
rate. We are able to achieve compression ratios between 10 and 30 on moderate
size datasets with minimal visual deterioration as a result of the lossy
compression.
| no_new_dataset | 0.9463 |
1508.06380 | Suman Saha | Suman Saha and Satya P. Ghrera | Network Community Detection on Metric Space | null | Algorithms 2015, 8(3), 680-696 | 10.3390/a8030680 | null | cs.SI physics.soc-ph | http://creativecommons.org/licenses/by/4.0/ | Community detection in a complex network is an important problem of much
interest in recent years. In general, a community detection algorithm chooses
an objective function and captures the communities of the network by optimizing
the objective function, and then, one uses various heuristics to solve the
optimization problem to extract the interesting communities for the user. In
this article, we demonstrate the procedure to transform a graph into points of
a metric space and develop the methods of community detection with the help of
a metric defined for a pair of points. We have also studied and analyzed the
community structure of the network therein. The results obtained with our
approach are very competitive with most of the well-known algorithms in the
literature, and this is justified over the large collection of datasets. On the
other hand, it can be observed that time taken by our algorithm is quite less
compared to other methods and justifies the theoretical findings.
| [
{
"version": "v1",
"created": "Wed, 26 Aug 2015 06:55:20 GMT"
}
] | 2015-08-27T00:00:00 | [
[
"Saha",
"Suman",
""
],
[
"Ghrera",
"Satya P.",
""
]
] | TITLE: Network Community Detection on Metric Space
ABSTRACT: Community detection in a complex network is an important problem of much
interest in recent years. In general, a community detection algorithm chooses
an objective function and captures the communities of the network by optimizing
the objective function, and then, one uses various heuristics to solve the
optimization problem to extract the interesting communities for the user. In
this article, we demonstrate the procedure to transform a graph into points of
a metric space and develop the methods of community detection with the help of
a metric defined for a pair of points. We have also studied and analyzed the
community structure of the network therein. The results obtained with our
approach are very competitive with most of the well-known algorithms in the
literature, and this is justified over the large collection of datasets. On the
other hand, it can be observed that time taken by our algorithm is quite less
compared to other methods and justifies the theoretical findings.
| no_new_dataset | 0.951323 |
1409.0080 | Wei Lu | Wei Lu, Shanshan Chen, Keqian Li, Laks V.S. Lakshmanan | Show Me the Money: Dynamic Recommendations for Revenue Maximization | Conference version published in PVLDB 7(14). To be presented in the
VLDB Conference 2015, in Hawaii. This version gives a detailed submodularity
proof | null | null | null | cs.DB cs.GT cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recommender Systems (RS) play a vital role in applications such as e-commerce
and on-demand content streaming. Research on RS has mainly focused on the
customer perspective, i.e., accurate prediction of user preferences and
maximization of user utilities. As a result, most existing techniques are not
explicitly built for revenue maximization, the primary business goal of
enterprises. In this work, we explore and exploit a novel connection between RS
and the profitability of a business. As recommendations can be seen as an
information channel between a business and its customers, it is interesting and
important to investigate how to make strategic dynamic recommendations leading
to maximum possible revenue. To this end, we propose a novel \model that takes
into account a variety of factors including prices, valuations, saturation
effects, and competition amongst products. Under this model, we study the
problem of finding revenue-maximizing recommendation strategies over a finite
time horizon. We show that this problem is NP-hard, but approximation
guarantees can be obtained for a slightly relaxed version, by establishing an
elegant connection to matroid theory. Given the prohibitively high complexity
of the approximation algorithm, we also design intelligent heuristics for the
original problem. Finally, we conduct extensive experiments on two real and
synthetic datasets and demonstrate the efficiency, scalability, and
effectiveness our algorithms, and that they significantly outperform several
intuitive baselines.
| [
{
"version": "v1",
"created": "Sat, 30 Aug 2014 04:15:15 GMT"
},
{
"version": "v2",
"created": "Sat, 6 Sep 2014 01:37:15 GMT"
},
{
"version": "v3",
"created": "Tue, 25 Aug 2015 18:21:48 GMT"
}
] | 2015-08-26T00:00:00 | [
[
"Lu",
"Wei",
""
],
[
"Chen",
"Shanshan",
""
],
[
"Li",
"Keqian",
""
],
[
"Lakshmanan",
"Laks V. S.",
""
]
] | TITLE: Show Me the Money: Dynamic Recommendations for Revenue Maximization
ABSTRACT: Recommender Systems (RS) play a vital role in applications such as e-commerce
and on-demand content streaming. Research on RS has mainly focused on the
customer perspective, i.e., accurate prediction of user preferences and
maximization of user utilities. As a result, most existing techniques are not
explicitly built for revenue maximization, the primary business goal of
enterprises. In this work, we explore and exploit a novel connection between RS
and the profitability of a business. As recommendations can be seen as an
information channel between a business and its customers, it is interesting and
important to investigate how to make strategic dynamic recommendations leading
to maximum possible revenue. To this end, we propose a novel \model that takes
into account a variety of factors including prices, valuations, saturation
effects, and competition amongst products. Under this model, we study the
problem of finding revenue-maximizing recommendation strategies over a finite
time horizon. We show that this problem is NP-hard, but approximation
guarantees can be obtained for a slightly relaxed version, by establishing an
elegant connection to matroid theory. Given the prohibitively high complexity
of the approximation algorithm, we also design intelligent heuristics for the
original problem. Finally, we conduct extensive experiments on two real and
synthetic datasets and demonstrate the efficiency, scalability, and
effectiveness our algorithms, and that they significantly outperform several
intuitive baselines.
| no_new_dataset | 0.939081 |
1505.02445 | Tomaso Aste | Guido Previde Massara, T. Di Matteo, Tomaso Aste | Network Filtering for Big Data: Triangulated Maximally Filtered Graph | 16 pages, 7 Figures, 2 Tables | null | null | null | cs.DS cond-mat.stat-mech cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a network-filtering method, the Triangulated Maximally Filtered
Graph (TMFG), that provides an approximate solution to the Weighted Maximal
Planar Graph problem. The underlying idea of TMFG consists in building a
triangulation that maximizes a score function associated with the amount of
information retained by the network. TMFG uses as weights any arbitrary
similarity measure to arrange data into a meaningful network structure that can
be used for clustering, community detection and modeling. The method is fast,
adaptable and scalable to very large datasets, it allows online updating and
learning as new data can be inserted and deleted with combinations of local and
non-local moves. TMFG permits readjustments of the network in consequence of
changes in the strength of the similarity measure. The method is based on local
topological moves and can therefore take advantage of parallel and GPUs
computing. We discuss how this network-filtering method can be used intuitively
and efficiently for big data studies and its significance from an
information-theoretic perspective.
| [
{
"version": "v1",
"created": "Sun, 10 May 2015 21:47:38 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Aug 2015 16:02:37 GMT"
}
] | 2015-08-26T00:00:00 | [
[
"Massara",
"Guido Previde",
""
],
[
"Di Matteo",
"T.",
""
],
[
"Aste",
"Tomaso",
""
]
] | TITLE: Network Filtering for Big Data: Triangulated Maximally Filtered Graph
ABSTRACT: We propose a network-filtering method, the Triangulated Maximally Filtered
Graph (TMFG), that provides an approximate solution to the Weighted Maximal
Planar Graph problem. The underlying idea of TMFG consists in building a
triangulation that maximizes a score function associated with the amount of
information retained by the network. TMFG uses as weights any arbitrary
similarity measure to arrange data into a meaningful network structure that can
be used for clustering, community detection and modeling. The method is fast,
adaptable and scalable to very large datasets, it allows online updating and
learning as new data can be inserted and deleted with combinations of local and
non-local moves. TMFG permits readjustments of the network in consequence of
changes in the strength of the similarity measure. The method is based on local
topological moves and can therefore take advantage of parallel and GPUs
computing. We discuss how this network-filtering method can be used intuitively
and efficiently for big data studies and its significance from an
information-theoretic perspective.
| no_new_dataset | 0.947866 |
1508.06206 | Sahar Vahdati | Angelo Di Iorio, Christoph Lange, Anastasia Dimou, Sahar Vahdati | Semantic Publishing Challenge - Assessing the Quality of Scientific
Output by Information Extraction and Interlinking | To appear in: E. Cabrio and M. Stankovic and M. Dragoni and A.
Gangemi and R. Navigli and V. Presutti and D. Garigliotti and A. L. Gentile
and A. Nuzzolese and A. Di Iorio and A. Dimou and C. Lange and S. Vahdati and
A. Freitas and C. Unger and D. Reforgiato Recupero (eds.). Semantic Web
Evaluation Challenges 2015. Communications in Computer and Information
Science, Springer, 2015. arXiv admin note: text overlap with arXiv:1408.3863 | null | null | null | cs.DL | http://creativecommons.org/licenses/by/4.0/ | The Semantic Publishing Challenge series aims at investigating novel
approaches for improving scholarly publishing using Linked Data technology. In
2014 we had bootstrapped this effort with a focus on extracting information
from non-semantic publications - computer science workshop proceedings volumes
and their papers - to assess their quality. The objective of this second
edition was to improve information extraction but also to interlink the 2014
dataset with related ones in the LOD Cloud, thus paving the way for
sophisticated end-user services.
| [
{
"version": "v1",
"created": "Tue, 25 Aug 2015 16:17:24 GMT"
}
] | 2015-08-26T00:00:00 | [
[
"Di Iorio",
"Angelo",
""
],
[
"Lange",
"Christoph",
""
],
[
"Dimou",
"Anastasia",
""
],
[
"Vahdati",
"Sahar",
""
]
] | TITLE: Semantic Publishing Challenge - Assessing the Quality of Scientific
Output by Information Extraction and Interlinking
ABSTRACT: The Semantic Publishing Challenge series aims at investigating novel
approaches for improving scholarly publishing using Linked Data technology. In
2014 we had bootstrapped this effort with a focus on extracting information
from non-semantic publications - computer science workshop proceedings volumes
and their papers - to assess their quality. The objective of this second
edition was to improve information extraction but also to interlink the 2014
dataset with related ones in the LOD Cloud, thus paving the way for
sophisticated end-user services.
| no_new_dataset | 0.941708 |
1501.04505 | Kaihua Zhang | Kaihua Zhang, Qingshan Liu, Yi Wu, Ming-Hsuan Yang | Robust Visual Tracking via Convolutional Networks | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Deep networks have been successfully applied to visual tracking by learning a
generic representation offline from numerous training images. However the
offline training is time-consuming and the learned generic representation may
be less discriminative for tracking specific objects. In this paper we present
that, even without offline training with a large amount of auxiliary data,
simple two-layer convolutional networks can be powerful enough to develop a
robust representation for visual tracking. In the first frame, we employ the
k-means algorithm to extract a set of normalized patches from the target region
as fixed filters, which integrate a series of adaptive contextual filters
surrounding the target to define a set of feature maps in the subsequent
frames. These maps measure similarities between each filter and the useful
local intensity patterns across the target, thereby encoding its local
structural information. Furthermore, all the maps form together a global
representation, which is built on mid-level features, thereby remaining close
to image-level information, and hence the inner geometric layout of the target
is also well preserved. A simple soft shrinkage method with an adaptive
threshold is employed to de-noise the global representation, resulting in a
robust sparse representation. The representation is updated via a simple and
effective online strategy, allowing it to robustly adapt to target appearance
variations. Our convolution networks have surprisingly lightweight structure,
yet perform favorably against several state-of-the-art methods on the CVPR2013
tracking benchmark dataset with 50 challenging videos.
| [
{
"version": "v1",
"created": "Mon, 19 Jan 2015 14:39:51 GMT"
},
{
"version": "v2",
"created": "Mon, 24 Aug 2015 06:07:22 GMT"
}
] | 2015-08-25T00:00:00 | [
[
"Zhang",
"Kaihua",
""
],
[
"Liu",
"Qingshan",
""
],
[
"Wu",
"Yi",
""
],
[
"Yang",
"Ming-Hsuan",
""
]
] | TITLE: Robust Visual Tracking via Convolutional Networks
ABSTRACT: Deep networks have been successfully applied to visual tracking by learning a
generic representation offline from numerous training images. However the
offline training is time-consuming and the learned generic representation may
be less discriminative for tracking specific objects. In this paper we present
that, even without offline training with a large amount of auxiliary data,
simple two-layer convolutional networks can be powerful enough to develop a
robust representation for visual tracking. In the first frame, we employ the
k-means algorithm to extract a set of normalized patches from the target region
as fixed filters, which integrate a series of adaptive contextual filters
surrounding the target to define a set of feature maps in the subsequent
frames. These maps measure similarities between each filter and the useful
local intensity patterns across the target, thereby encoding its local
structural information. Furthermore, all the maps form together a global
representation, which is built on mid-level features, thereby remaining close
to image-level information, and hence the inner geometric layout of the target
is also well preserved. A simple soft shrinkage method with an adaptive
threshold is employed to de-noise the global representation, resulting in a
robust sparse representation. The representation is updated via a simple and
effective online strategy, allowing it to robustly adapt to target appearance
variations. Our convolution networks have surprisingly lightweight structure,
yet perform favorably against several state-of-the-art methods on the CVPR2013
tracking benchmark dataset with 50 challenging videos.
| no_new_dataset | 0.946646 |
1506.02275 | Jacob Eisenstein | Umashanthi Pavalanathan and Jacob Eisenstein | Confounds and Consequences in Geotagged Twitter Data | final version for EMNLP 2015 | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Twitter is often used in quantitative studies that identify
geographically-preferred topics, writing styles, and entities. These studies
rely on either GPS coordinates attached to individual messages, or on the
user-supplied location field in each profile. In this paper, we compare these
data acquisition techniques and quantify the biases that they introduce; we
also measure their effects on linguistic analysis and text-based geolocation.
GPS-tagging and self-reported locations yield measurably different corpora, and
these linguistic differences are partially attributable to differences in
dataset composition by age and gender. Using a latent variable model to induce
age and gender, we show how these demographic variables interact with geography
to affect language use. We also show that the accuracy of text-based
geolocation varies with population demographics, giving the best results for
men above the age of 40.
| [
{
"version": "v1",
"created": "Sun, 7 Jun 2015 15:29:26 GMT"
},
{
"version": "v2",
"created": "Sat, 22 Aug 2015 15:25:59 GMT"
}
] | 2015-08-25T00:00:00 | [
[
"Pavalanathan",
"Umashanthi",
""
],
[
"Eisenstein",
"Jacob",
""
]
] | TITLE: Confounds and Consequences in Geotagged Twitter Data
ABSTRACT: Twitter is often used in quantitative studies that identify
geographically-preferred topics, writing styles, and entities. These studies
rely on either GPS coordinates attached to individual messages, or on the
user-supplied location field in each profile. In this paper, we compare these
data acquisition techniques and quantify the biases that they introduce; we
also measure their effects on linguistic analysis and text-based geolocation.
GPS-tagging and self-reported locations yield measurably different corpora, and
these linguistic differences are partially attributable to differences in
dataset composition by age and gender. Using a latent variable model to induce
age and gender, we show how these demographic variables interact with geography
to affect language use. We also show that the accuracy of text-based
geolocation varies with population demographics, giving the best results for
men above the age of 40.
| no_new_dataset | 0.951188 |
1508.05710 | Zixuan Zhuang | Zixuan Zhuang | An Experimental Study of Distributed Quantile Estimation | M.S. Thesis | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Quantiles are very important statistics information used to describe the
distribution of datasets. Given the quantiles of a dataset, we can easily know
the distribution of the dataset, which is a fundamental problem in data
analysis. However, quite often, computing quantiles directly is inappropriate
due to the memory limitations. Further, in many settings such as data streaming
and sensor network model, even the data size is unpredictable. Although the
quantiles computation has been widely studied, it was mostly in the sequential
setting. In this paper, we study several quantile computation algorithms in the
distributed setting and compare them in terms of space usage, running time, and
accuracy. Moreover, we provide detailed experimental comparisons between
several popular algorithms. Our work focuses on the approximate quantile
algorithms which provide error bounds. Approximate quantiles have received more
attentions than exact ones since they are often faster, can be more easily
adapted to the distributed setting while giving sufficiently good statistical
information on the data sets.
| [
{
"version": "v1",
"created": "Mon, 24 Aug 2015 07:49:38 GMT"
}
] | 2015-08-25T00:00:00 | [
[
"Zhuang",
"Zixuan",
""
]
] | TITLE: An Experimental Study of Distributed Quantile Estimation
ABSTRACT: Quantiles are very important statistics information used to describe the
distribution of datasets. Given the quantiles of a dataset, we can easily know
the distribution of the dataset, which is a fundamental problem in data
analysis. However, quite often, computing quantiles directly is inappropriate
due to the memory limitations. Further, in many settings such as data streaming
and sensor network model, even the data size is unpredictable. Although the
quantiles computation has been widely studied, it was mostly in the sequential
setting. In this paper, we study several quantile computation algorithms in the
distributed setting and compare them in terms of space usage, running time, and
accuracy. Moreover, we provide detailed experimental comparisons between
several popular algorithms. Our work focuses on the approximate quantile
algorithms which provide error bounds. Approximate quantiles have received more
attentions than exact ones since they are often faster, can be more easily
adapted to the distributed setting while giving sufficiently good statistical
information on the data sets.
| no_new_dataset | 0.9455 |
1508.05817 | Marco Guerini | Marco Guerini, G\"ozde \"Ozbal, Carlo Strapparava | Echoes of Persuasion: The Effect of Euphony in Persuasive Communication | null | null | null | null | cs.CL cs.CY cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | While the effect of various lexical, syntactic, semantic and stylistic
features have been addressed in persuasive language from a computational point
of view, the persuasive effect of phonetics has received little attention. By
modeling a notion of euphony and analyzing four datasets comprising persuasive
and non-persuasive sentences in different domains (political speeches, movie
quotes, slogans and tweets), we explore the impact of sounds on different forms
of persuasiveness. We conduct a series of analyses and prediction experiments
within and across datasets. Our results highlight the positive role of phonetic
devices on persuasion.
| [
{
"version": "v1",
"created": "Mon, 24 Aug 2015 14:15:39 GMT"
}
] | 2015-08-25T00:00:00 | [
[
"Guerini",
"Marco",
""
],
[
"Özbal",
"Gözde",
""
],
[
"Strapparava",
"Carlo",
""
]
] | TITLE: Echoes of Persuasion: The Effect of Euphony in Persuasive Communication
ABSTRACT: While the effect of various lexical, syntactic, semantic and stylistic
features have been addressed in persuasive language from a computational point
of view, the persuasive effect of phonetics has received little attention. By
modeling a notion of euphony and analyzing four datasets comprising persuasive
and non-persuasive sentences in different domains (political speeches, movie
quotes, slogans and tweets), we explore the impact of sounds on different forms
of persuasiveness. We conduct a series of analyses and prediction experiments
within and across datasets. Our results highlight the positive role of phonetic
devices on persuasion.
| no_new_dataset | 0.940024 |
1406.5975 | Yogesh Simmhan | Yogesh Simmhan, Charith Wickramaarachchi, Alok Kumbhare, Marc Frincu,
Soonil Nagarkar, Santosh Ravi, Cauligi Raghavendra, Viktor Prasanna | Scalable Analytics over Distributed Time-series Graphs using GoFFish | null | Proceedings of the IEEE International Parallel and Distributed
Processing Symposium (IPDPS) (2015) pp. 809-818 | 10.1109/IPDPS.2015.66 | null | cs.DC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Graphs are a key form of Big Data, and performing scalable analytics over
them is invaluable to many domains. As our ability to collect data grows, there
is an emerging class of inter-connected data which accumulates or varies over
time, and on which novel analytics - both over the network structure and across
the time-variant attribute values - is necessary. We introduce the notion of
time-series graph analytics and propose Gopher, a scalable programming
abstraction to develop algorithms and analytics on such datasets. Our
abstraction leverages a sub-graph centric programming model and extends it to
the temporal dimension using an iterative BSP (Bulk Synchronous Parallel)
approach. Gopher is co-designed with GoFS, a distributed storage specialized
for time-series graphs, as part of the GoFFish distributed analytics platform.
We examine storage optimizations for GoFS, design patterns in Gopher to
leverage the distributed data layout, and evaluate the GoFFish platform using
time-series graph data and applications on a commodity cluster.
| [
{
"version": "v1",
"created": "Mon, 23 Jun 2014 16:48:03 GMT"
}
] | 2015-08-21T00:00:00 | [
[
"Simmhan",
"Yogesh",
""
],
[
"Wickramaarachchi",
"Charith",
""
],
[
"Kumbhare",
"Alok",
""
],
[
"Frincu",
"Marc",
""
],
[
"Nagarkar",
"Soonil",
""
],
[
"Ravi",
"Santosh",
""
],
[
"Raghavendra",
"Cauligi",
""
],
[
"Prasanna",
"Viktor",
""
]
] | TITLE: Scalable Analytics over Distributed Time-series Graphs using GoFFish
ABSTRACT: Graphs are a key form of Big Data, and performing scalable analytics over
them is invaluable to many domains. As our ability to collect data grows, there
is an emerging class of inter-connected data which accumulates or varies over
time, and on which novel analytics - both over the network structure and across
the time-variant attribute values - is necessary. We introduce the notion of
time-series graph analytics and propose Gopher, a scalable programming
abstraction to develop algorithms and analytics on such datasets. Our
abstraction leverages a sub-graph centric programming model and extends it to
the temporal dimension using an iterative BSP (Bulk Synchronous Parallel)
approach. Gopher is co-designed with GoFS, a distributed storage specialized
for time-series graphs, as part of the GoFFish distributed analytics platform.
We examine storage optimizations for GoFS, design patterns in Gopher to
leverage the distributed data layout, and evaluate the GoFFish platform using
time-series graph data and applications on a commodity cluster.
| no_new_dataset | 0.942454 |
1411.4046 | Mohammad Ali Keyvanrad | Mohammad Ali Keyvanrad, Mohammad Mehdi Homayounpour | Deep Belief Network Training Improvement Using Elite Samples Minimizing
Free Energy | 18 pages. arXiv admin note: substantial text overlap with
arXiv:1408.3264 | Int. J. Patt. Recogn. Artif. Intell. 29, 1551006 (2015) | 10.1142/S0218001415510064 | null | cs.LG cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Nowadays this is very popular to use deep architectures in machine learning.
Deep Belief Networks (DBNs) are deep architectures that use stack of Restricted
Boltzmann Machines (RBM) to create a powerful generative model using training
data. In this paper we present an improvement in a common method that is
usually used in training of RBMs. The new method uses free energy as a
criterion to obtain elite samples from generative model. We argue that these
samples can more accurately compute gradient of log probability of training
data. According to the results, an error rate of 0.99% was achieved on MNIST
test set. This result shows that the proposed method outperforms the method
presented in the first paper introducing DBN (1.25% error rate) and general
classification methods such as SVM (1.4% error rate) and KNN (with 1.6% error
rate). In another test using ISOLET dataset, letter classification error
dropped to 3.59% compared to 5.59% error rate achieved in those papers using
this dataset. The implemented method is available online at
"http://ceit.aut.ac.ir/~keyvanrad/DeeBNet Toolbox.html".
| [
{
"version": "v1",
"created": "Fri, 14 Nov 2014 16:57:48 GMT"
}
] | 2015-08-21T00:00:00 | [
[
"Keyvanrad",
"Mohammad Ali",
""
],
[
"Homayounpour",
"Mohammad Mehdi",
""
]
] | TITLE: Deep Belief Network Training Improvement Using Elite Samples Minimizing
Free Energy
ABSTRACT: Nowadays this is very popular to use deep architectures in machine learning.
Deep Belief Networks (DBNs) are deep architectures that use stack of Restricted
Boltzmann Machines (RBM) to create a powerful generative model using training
data. In this paper we present an improvement in a common method that is
usually used in training of RBMs. The new method uses free energy as a
criterion to obtain elite samples from generative model. We argue that these
samples can more accurately compute gradient of log probability of training
data. According to the results, an error rate of 0.99% was achieved on MNIST
test set. This result shows that the proposed method outperforms the method
presented in the first paper introducing DBN (1.25% error rate) and general
classification methods such as SVM (1.4% error rate) and KNN (with 1.6% error
rate). In another test using ISOLET dataset, letter classification error
dropped to 3.59% compared to 5.59% error rate achieved in those papers using
this dataset. The implemented method is available online at
"http://ceit.aut.ac.ir/~keyvanrad/DeeBNet Toolbox.html".
| no_new_dataset | 0.949576 |
1508.04785 | KuanTing Chen | KuanTing Chen, Kezhen Chen, Peizhong Cong, Winston H. Hsu, Jiebo Luo | Who are the Devils Wearing Prada in New York City? | null | null | 10.1145/2733373.2809930 | null | cs.CV cs.CY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Fashion is a perpetual topic in human social life, and the mass has the
penchant to emulate what large city residents and celebrities wear. Undeniably,
New York City is such a bellwether large city with all kinds of fashion
leadership. Consequently, to study what the fashion trends are during this
year, it is very helpful to learn the fashion trends of New York City.
Discovering fashion trends in New York City could boost many applications such
as clothing recommendation and advertising. Does the fashion trend in the New
York Fashion Show actually influence the clothing styles on the public? To
answer this question, we design a novel system that consists of three major
components: (1) constructing a large dataset from the New York Fashion Shows
and New York street chic in order to understand the likely clothing fashion
trends in New York, (2) utilizing a learning-based approach to discover fashion
attributes as the representative characteristics of fashion trends, and (3)
comparing the analysis results from the New York Fashion Shows and street-chic
images to verify whether the fashion shows have actual influence on the people
in New York City. Through the preliminary experiments over a large clothing
dataset, we demonstrate the effectiveness of our proposed system, and obtain
useful insights on fashion trends and fashion influence.
| [
{
"version": "v1",
"created": "Wed, 19 Aug 2015 20:28:31 GMT"
}
] | 2015-08-21T00:00:00 | [
[
"Chen",
"KuanTing",
""
],
[
"Chen",
"Kezhen",
""
],
[
"Cong",
"Peizhong",
""
],
[
"Hsu",
"Winston H.",
""
],
[
"Luo",
"Jiebo",
""
]
] | TITLE: Who are the Devils Wearing Prada in New York City?
ABSTRACT: Fashion is a perpetual topic in human social life, and the mass has the
penchant to emulate what large city residents and celebrities wear. Undeniably,
New York City is such a bellwether large city with all kinds of fashion
leadership. Consequently, to study what the fashion trends are during this
year, it is very helpful to learn the fashion trends of New York City.
Discovering fashion trends in New York City could boost many applications such
as clothing recommendation and advertising. Does the fashion trend in the New
York Fashion Show actually influence the clothing styles on the public? To
answer this question, we design a novel system that consists of three major
components: (1) constructing a large dataset from the New York Fashion Shows
and New York street chic in order to understand the likely clothing fashion
trends in New York, (2) utilizing a learning-based approach to discover fashion
attributes as the representative characteristics of fashion trends, and (3)
comparing the analysis results from the New York Fashion Shows and street-chic
images to verify whether the fashion shows have actual influence on the people
in New York City. Through the preliminary experiments over a large clothing
dataset, we demonstrate the effectiveness of our proposed system, and obtain
useful insights on fashion trends and fashion influence.
| no_new_dataset | 0.66861 |
1508.04870 | Natasha Holmes | N.G. Holmes, Carl E. Wieman, and D.A. Bonn | Teaching Critical Thinking | Proceedings of the National Academy of Sciences (2015) | null | 10.1073/pnas.1505329112 | null | physics.ed-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The ability to make decisions based on data, with its inherent uncertainties
and variability, is a complex and vital skill in the modern world. The need for
such quantitative critical thinking occurs in many different contexts, and
while it is an important goal of education, that goal is seldom being achieved.
We argue that the key element for developing this ability is repeated practice
in making decisions based on data, with feedback on those decisions. We
demonstrate a structure for providing suitable practice that can be applied in
any instructional setting that involves the acquisition of data and relating
that data to scientific models. This study reports the results of applying that
structure in an introductory physics lab course. Students in an experimental
condition were repeatedly instructed to make and act on quantitative
comparisons between datasets, and between data and models, an approach that is
common to all science disciplines. These instructions were slowly faded across
the course. After the instructions had been removed, students in the
experimental condition were 12 times more likely to spontaneously propose or
make changes to improve their experimental methods than a control group, who
performed traditional experimental activities. They were also four times more
likely to identify and explain a limitation of a physical model using their
data. Students in the experimental condition also showed much more
sophisticated reasoning about their data. These differences between the groups
were seen to persist into a subsequent course taken the following year.
| [
{
"version": "v1",
"created": "Thu, 20 Aug 2015 03:51:42 GMT"
}
] | 2015-08-21T00:00:00 | [
[
"Holmes",
"N. G.",
""
],
[
"Wieman",
"Carl E.",
""
],
[
"Bonn",
"D. A.",
""
]
] | TITLE: Teaching Critical Thinking
ABSTRACT: The ability to make decisions based on data, with its inherent uncertainties
and variability, is a complex and vital skill in the modern world. The need for
such quantitative critical thinking occurs in many different contexts, and
while it is an important goal of education, that goal is seldom being achieved.
We argue that the key element for developing this ability is repeated practice
in making decisions based on data, with feedback on those decisions. We
demonstrate a structure for providing suitable practice that can be applied in
any instructional setting that involves the acquisition of data and relating
that data to scientific models. This study reports the results of applying that
structure in an introductory physics lab course. Students in an experimental
condition were repeatedly instructed to make and act on quantitative
comparisons between datasets, and between data and models, an approach that is
common to all science disciplines. These instructions were slowly faded across
the course. After the instructions had been removed, students in the
experimental condition were 12 times more likely to spontaneously propose or
make changes to improve their experimental methods than a control group, who
performed traditional experimental activities. They were also four times more
likely to identify and explain a limitation of a physical model using their
data. Students in the experimental condition also showed much more
sophisticated reasoning about their data. These differences between the groups
were seen to persist into a subsequent course taken the following year.
| no_new_dataset | 0.942082 |
1508.04909 | Alain Rakotomamonjy | Alain Rakotomamonjy (LITIS), Gilles Gasso (LITIS) | Histogram of gradients of Time-Frequency Representations for Audio scene
detection | null | null | null | null | cs.SD cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper addresses the problem of audio scenes classification and
contributes to the state of the art by proposing a novel feature. We build this
feature by considering histogram of gradients (HOG) of time-frequency
representation of an audio scene. Contrarily to classical audio features like
MFCC, we make the hypothesis that histogram of gradients are able to encode
some relevant informations in a time-frequency {representation:} namely, the
local direction of variation (in time and frequency) of the signal spectral
power. In addition, in order to gain more invariance and robustness, histogram
of gradients are locally pooled. We have evaluated the relevance of {the novel
feature} by comparing its performances with state-of-the-art competitors, on
several datasets, including a novel one that we provide, as part of our
contribution. This dataset, that we make publicly available, involves $19$
classes and contains about $900$ minutes of audio scene recording. We thus
believe that it may be the next standard dataset for evaluating audio scene
classification algorithms. Our comparison results clearly show that our
HOG-based features outperform its competitors
| [
{
"version": "v1",
"created": "Thu, 20 Aug 2015 08:07:10 GMT"
}
] | 2015-08-21T00:00:00 | [
[
"Rakotomamonjy",
"Alain",
"",
"LITIS"
],
[
"Gasso",
"Gilles",
"",
"LITIS"
]
] | TITLE: Histogram of gradients of Time-Frequency Representations for Audio scene
detection
ABSTRACT: This paper addresses the problem of audio scenes classification and
contributes to the state of the art by proposing a novel feature. We build this
feature by considering histogram of gradients (HOG) of time-frequency
representation of an audio scene. Contrarily to classical audio features like
MFCC, we make the hypothesis that histogram of gradients are able to encode
some relevant informations in a time-frequency {representation:} namely, the
local direction of variation (in time and frequency) of the signal spectral
power. In addition, in order to gain more invariance and robustness, histogram
of gradients are locally pooled. We have evaluated the relevance of {the novel
feature} by comparing its performances with state-of-the-art competitors, on
several datasets, including a novel one that we provide, as part of our
contribution. This dataset, that we make publicly available, involves $19$
classes and contains about $900$ minutes of audio scene recording. We thus
believe that it may be the next standard dataset for evaluating audio scene
classification algorithms. Our comparison results clearly show that our
HOG-based features outperform its competitors
| new_dataset | 0.964119 |
1508.04957 | Aggeliki Dimitriou | Aggeliki Dimitriou, Ananya Dass, Dimitri Theodoratos | Cohesiveness Relationships to Empower Keyword Search on Tree Data on the
Web | null | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Keyword search is the most popular querying technique on semistructured data.
Keyword queries are simple and con- venient. However, as a consequence of their
imprecision, the quality of their answers is poor and the existing algorithms
do not scale satisfactorily.
In this paper, we introduce the novel concept of cohesive keyword queries for
tree data. Intuitively, a cohesiveness relationship on keywords indicates that
they should form a cohesive whole in a query result. Cohesive keyword queries
allow term nesting and keyword repetition. Although more expressive, they are
as simple as flat keyword queries. We provide formal semantics for cohesive
keyword queries rank- ing query results on the proximity of the keyword
instances. We design a stack based algorithm which efficiently evaluates
cohesive keyword queries. Our experiments demonstrate that our approach
outperforms in quality previous filtering semantics and our algorithm scales
smoothly on queries of even 20 keywords on large datasets.
| [
{
"version": "v1",
"created": "Thu, 20 Aug 2015 11:17:04 GMT"
}
] | 2015-08-21T00:00:00 | [
[
"Dimitriou",
"Aggeliki",
""
],
[
"Dass",
"Ananya",
""
],
[
"Theodoratos",
"Dimitri",
""
]
] | TITLE: Cohesiveness Relationships to Empower Keyword Search on Tree Data on the
Web
ABSTRACT: Keyword search is the most popular querying technique on semistructured data.
Keyword queries are simple and con- venient. However, as a consequence of their
imprecision, the quality of their answers is poor and the existing algorithms
do not scale satisfactorily.
In this paper, we introduce the novel concept of cohesive keyword queries for
tree data. Intuitively, a cohesiveness relationship on keywords indicates that
they should form a cohesive whole in a query result. Cohesive keyword queries
allow term nesting and keyword repetition. Although more expressive, they are
as simple as flat keyword queries. We provide formal semantics for cohesive
keyword queries rank- ing query results on the proximity of the keyword
instances. We design a stack based algorithm which efficiently evaluates
cohesive keyword queries. Our experiments demonstrate that our approach
outperforms in quality previous filtering semantics and our algorithm scales
smoothly on queries of even 20 keywords on large datasets.
| no_new_dataset | 0.951278 |
1508.05003 | Suvrit Sra | Suvrit Sra, Adams Wei Yu, Mu Li, Alexander J. Smola | AdaDelay: Delay Adaptive Distributed Stochastic Convex Optimization | 19 pages | null | null | null | stat.ML cs.LG math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study distributed stochastic convex optimization under the delayed
gradient model where the server nodes perform parameter updates, while the
worker nodes compute stochastic gradients. We discuss, analyze, and experiment
with a setup motivated by the behavior of real-world distributed computation
networks, where the machines are differently slow at different time. Therefore,
we allow the parameter updates to be sensitive to the actual delays
experienced, rather than to worst-case bounds on the maximum delay. This
sensitivity leads to larger stepsizes, that can help gain rapid initial
convergence without having to wait too long for slower machines, while
maintaining the same asymptotic complexity. We obtain encouraging improvements
to overall convergence for distributed experiments on real datasets with up to
billions of examples and features.
| [
{
"version": "v1",
"created": "Thu, 20 Aug 2015 15:11:11 GMT"
}
] | 2015-08-21T00:00:00 | [
[
"Sra",
"Suvrit",
""
],
[
"Yu",
"Adams Wei",
""
],
[
"Li",
"Mu",
""
],
[
"Smola",
"Alexander J.",
""
]
] | TITLE: AdaDelay: Delay Adaptive Distributed Stochastic Convex Optimization
ABSTRACT: We study distributed stochastic convex optimization under the delayed
gradient model where the server nodes perform parameter updates, while the
worker nodes compute stochastic gradients. We discuss, analyze, and experiment
with a setup motivated by the behavior of real-world distributed computation
networks, where the machines are differently slow at different time. Therefore,
we allow the parameter updates to be sensitive to the actual delays
experienced, rather than to worst-case bounds on the maximum delay. This
sensitivity leads to larger stepsizes, that can help gain rapid initial
convergence without having to wait too long for slower machines, while
maintaining the same asymptotic complexity. We obtain encouraging improvements
to overall convergence for distributed experiments on real datasets with up to
billions of examples and features.
| no_new_dataset | 0.946051 |
1508.04525 | Wei Zhang | Wei Zhang, Yang Yu, Osho Gupta, Judith Gelernter | Recognizing Extended Spatiotemporal Expressions by Actively Trained
Average Perceptron Ensembles | 10 pages | null | null | null | cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Precise geocoding and time normalization for text requires that location and
time phrases be identified. Many state-of-the-art geoparsers and temporal
parsers suffer from low recall. Categories commonly missed by parsers are:
nouns used in a non- spatiotemporal sense, adjectival and adverbial phrases,
prepositional phrases, and numerical phrases. We collected and annotated data
set by querying commercial web searches API with such spatiotemporal
expressions as were missed by state-of-the- art parsers. Due to the high cost
of sentence annotation, active learning was used to label training data, and a
new strategy was designed to better select training examples to reduce labeling
cost. For the learning algorithm, we applied an average perceptron trained
Featurized Hidden Markov Model (FHMM). Five FHMM instances were used to create
an ensemble, with the output phrase selected by voting. Our ensemble model was
tested on a range of sequential labeling tasks, and has shown competitive
performance. Our contributions include (1) an new dataset annotated with named
entities and expanded spatiotemporal expressions; (2) a comparison of inference
algorithms for ensemble models showing the superior accuracy of Belief
Propagation over Viterbi Decoding; (3) a new example re-weighting method for
active ensemble learning that 'memorizes' the latest examples trained; (4) a
spatiotemporal parser that jointly recognizes expanded spatiotemporal
expressions as well as named entities.
| [
{
"version": "v1",
"created": "Wed, 19 Aug 2015 04:17:47 GMT"
}
] | 2015-08-20T00:00:00 | [
[
"Zhang",
"Wei",
""
],
[
"Yu",
"Yang",
""
],
[
"Gupta",
"Osho",
""
],
[
"Gelernter",
"Judith",
""
]
] | TITLE: Recognizing Extended Spatiotemporal Expressions by Actively Trained
Average Perceptron Ensembles
ABSTRACT: Precise geocoding and time normalization for text requires that location and
time phrases be identified. Many state-of-the-art geoparsers and temporal
parsers suffer from low recall. Categories commonly missed by parsers are:
nouns used in a non- spatiotemporal sense, adjectival and adverbial phrases,
prepositional phrases, and numerical phrases. We collected and annotated data
set by querying commercial web searches API with such spatiotemporal
expressions as were missed by state-of-the- art parsers. Due to the high cost
of sentence annotation, active learning was used to label training data, and a
new strategy was designed to better select training examples to reduce labeling
cost. For the learning algorithm, we applied an average perceptron trained
Featurized Hidden Markov Model (FHMM). Five FHMM instances were used to create
an ensemble, with the output phrase selected by voting. Our ensemble model was
tested on a range of sequential labeling tasks, and has shown competitive
performance. Our contributions include (1) an new dataset annotated with named
entities and expanded spatiotemporal expressions; (2) a comparison of inference
algorithms for ensemble models showing the superior accuracy of Belief
Propagation over Viterbi Decoding; (3) a new example re-weighting method for
active ensemble learning that 'memorizes' the latest examples trained; (4) a
spatiotemporal parser that jointly recognizes expanded spatiotemporal
expressions as well as named entities.
| no_new_dataset | 0.559079 |
1508.04537 | Jun He | Hao Wu, Jun He, Bo Li, Yijian Pei | Personalized QoS Prediction of Cloud Services via Learning
Neighborhood-based Model | null | null | null | null | cs.DC cs.PF | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The explosion of cloud services on the Internet brings new challenges in
service discovery and selection. Particularly, the demand for efficient
quality-of-service (QoS) evaluation is becoming urgently strong. To address
this issue, this paper proposes neighborhood-based approach for QoS prediction
of cloud services by taking advantages of collaborative intelligence. Different
from heuristic collaborative filtering and matrix factorization, we define a
formal neighborhood-based prediction framework which allows an efficient global
optimization scheme, and then exploit different baseline estimate component to
improve predictive performance. To validate the proposed methods, a large-scale
QoS-specific dataset which consists of invocation records from 339 service
users on 5,825 web services on a world-scale distributed network is used.
Experimental results demonstrate that the learned neighborhood-based models can
overcome existing difficulties of heuristic collaborative filtering methods and
achieve superior performance than state-of-the-art prediction methods.
| [
{
"version": "v1",
"created": "Wed, 19 Aug 2015 06:32:54 GMT"
}
] | 2015-08-20T00:00:00 | [
[
"Wu",
"Hao",
""
],
[
"He",
"Jun",
""
],
[
"Li",
"Bo",
""
],
[
"Pei",
"Yijian",
""
]
] | TITLE: Personalized QoS Prediction of Cloud Services via Learning
Neighborhood-based Model
ABSTRACT: The explosion of cloud services on the Internet brings new challenges in
service discovery and selection. Particularly, the demand for efficient
quality-of-service (QoS) evaluation is becoming urgently strong. To address
this issue, this paper proposes neighborhood-based approach for QoS prediction
of cloud services by taking advantages of collaborative intelligence. Different
from heuristic collaborative filtering and matrix factorization, we define a
formal neighborhood-based prediction framework which allows an efficient global
optimization scheme, and then exploit different baseline estimate component to
improve predictive performance. To validate the proposed methods, a large-scale
QoS-specific dataset which consists of invocation records from 339 service
users on 5,825 web services on a world-scale distributed network is used.
Experimental results demonstrate that the learned neighborhood-based models can
overcome existing difficulties of heuristic collaborative filtering methods and
achieve superior performance than state-of-the-art prediction methods.
| no_new_dataset | 0.802633 |
1508.04546 | Alexander Krull | Alexander Krull, Eric Brachmann, Frank Michel, Michael Ying Yang,
Stefan Gumhold, Carsten Rother | Learning Analysis-by-Synthesis for 6D Pose Estimation in RGB-D Images | 16 pages, 8 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Analysis-by-synthesis has been a successful approach for many tasks in
computer vision, such as 6D pose estimation of an object in an RGB-D image
which is the topic of this work. The idea is to compare the observation with
the output of a forward process, such as a rendered image of the object of
interest in a particular pose. Due to occlusion or complicated sensor noise, it
can be difficult to perform this comparison in a meaningful way. We propose an
approach that "learns to compare", while taking these difficulties into
account. This is done by describing the posterior density of a particular
object pose with a convolutional neural network (CNN) that compares an observed
and rendered image. The network is trained with the maximum likelihood
paradigm. We observe empirically that the CNN does not specialize to the
geometry or appearance of specific objects, and it can be used with objects of
vastly different shapes and appearances, and in different backgrounds. Compared
to state-of-the-art, we demonstrate a significant improvement on two different
datasets which include a total of eleven objects, cluttered background, and
heavy occlusion.
| [
{
"version": "v1",
"created": "Wed, 19 Aug 2015 07:24:14 GMT"
}
] | 2015-08-20T00:00:00 | [
[
"Krull",
"Alexander",
""
],
[
"Brachmann",
"Eric",
""
],
[
"Michel",
"Frank",
""
],
[
"Yang",
"Michael Ying",
""
],
[
"Gumhold",
"Stefan",
""
],
[
"Rother",
"Carsten",
""
]
] | TITLE: Learning Analysis-by-Synthesis for 6D Pose Estimation in RGB-D Images
ABSTRACT: Analysis-by-synthesis has been a successful approach for many tasks in
computer vision, such as 6D pose estimation of an object in an RGB-D image
which is the topic of this work. The idea is to compare the observation with
the output of a forward process, such as a rendered image of the object of
interest in a particular pose. Due to occlusion or complicated sensor noise, it
can be difficult to perform this comparison in a meaningful way. We propose an
approach that "learns to compare", while taking these difficulties into
account. This is done by describing the posterior density of a particular
object pose with a convolutional neural network (CNN) that compares an observed
and rendered image. The network is trained with the maximum likelihood
paradigm. We observe empirically that the CNN does not specialize to the
geometry or appearance of specific objects, and it can be used with objects of
vastly different shapes and appearances, and in different backgrounds. Compared
to state-of-the-art, we demonstrate a significant improvement on two different
datasets which include a total of eleven objects, cluttered background, and
heavy occlusion.
| no_new_dataset | 0.949106 |
1508.04586 | Veronica Vilaplana | Ver\'onica Vilaplana | Saliency maps on image hierarchies | Accepted for publication in Signal Processing: Image Communications,
2015 | null | 10.1016/j.image.2015.07.012 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper we propose two saliency models for salient object segmentation
based on a hierarchical image segmentation, a tree-like structure that
represents regions at different scales from the details to the whole image
(e.g. gPb-UCM, BPT). The first model is based on a hierarchy of image
partitions. The saliency at each level is computed on a region basis, taking
into account the contrast between regions. The maps obtained for the different
partitions are then integrated into a final saliency map. The second model
directly works on the structure created by the segmentation algorithm,
computing saliency at each node and integrating these cues in a straightforward
manner into a single saliency map. We show that the proposed models produce
high quality saliency maps. Objective evaluation demonstrates that the two
methods achieve state-of-the-art performance in several benchmark datasets.
| [
{
"version": "v1",
"created": "Wed, 19 Aug 2015 10:07:07 GMT"
}
] | 2015-08-20T00:00:00 | [
[
"Vilaplana",
"Verónica",
""
]
] | TITLE: Saliency maps on image hierarchies
ABSTRACT: In this paper we propose two saliency models for salient object segmentation
based on a hierarchical image segmentation, a tree-like structure that
represents regions at different scales from the details to the whole image
(e.g. gPb-UCM, BPT). The first model is based on a hierarchy of image
partitions. The saliency at each level is computed on a region basis, taking
into account the contrast between regions. The maps obtained for the different
partitions are then integrated into a final saliency map. The second model
directly works on the structure created by the segmentation algorithm,
computing saliency at each node and integrating these cues in a straightforward
manner into a single saliency map. We show that the proposed models produce
high quality saliency maps. Objective evaluation demonstrates that the two
methods achieve state-of-the-art performance in several benchmark datasets.
| no_new_dataset | 0.956145 |
1410.8616 | Abhijit Chandra | Abhijit Chandra and Oliva Kar | Data Driven Prognosis: A multi-physics approach verified via balloon
burst experiment | null | null | 10.1098/rspa.2014.0525 | null | cs.CE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A multi-physics formulation for Data Driven Prognosis (DDP) is developed.
Unlike traditional predictive strategies that require controlled off-line
measurements or training for determination of constitutive parameters to derive
the transitional statistics, the proposed DDP algorithm relies solely on in
situ measurements. It utilizes a deterministic mechanics framework, but the
stochastic nature of the solution arises naturally from the underlying
assumptions regarding the order of the conservation potential as well as the
number of dimensions involved. The proposed DDP scheme is capable of predicting
onset of instabilities. Since the need for off-line testing (or training) is
obviated, it can be easily implemented for systems where such a priori testing
is difficult or even impossible to conduct. The prognosis capability is
demonstrated here via a balloon burst experiment where the instability is
predicted utilizing only on-line visual observations. The DDP scheme never
failed to predict the incipient failure, and no false positives were issued.
The DDP algorithm is applicable to others types of datasets. Time horizons of
DDP predictions can be adjusted by using memory over different time windows.
Thus, a big dataset can be parsed in time to make a range of predictions over
varying time horizons.
| [
{
"version": "v1",
"created": "Fri, 31 Oct 2014 02:05:09 GMT"
}
] | 2015-08-19T00:00:00 | [
[
"Chandra",
"Abhijit",
""
],
[
"Kar",
"Oliva",
""
]
] | TITLE: Data Driven Prognosis: A multi-physics approach verified via balloon
burst experiment
ABSTRACT: A multi-physics formulation for Data Driven Prognosis (DDP) is developed.
Unlike traditional predictive strategies that require controlled off-line
measurements or training for determination of constitutive parameters to derive
the transitional statistics, the proposed DDP algorithm relies solely on in
situ measurements. It utilizes a deterministic mechanics framework, but the
stochastic nature of the solution arises naturally from the underlying
assumptions regarding the order of the conservation potential as well as the
number of dimensions involved. The proposed DDP scheme is capable of predicting
onset of instabilities. Since the need for off-line testing (or training) is
obviated, it can be easily implemented for systems where such a priori testing
is difficult or even impossible to conduct. The prognosis capability is
demonstrated here via a balloon burst experiment where the instability is
predicted utilizing only on-line visual observations. The DDP scheme never
failed to predict the incipient failure, and no false positives were issued.
The DDP algorithm is applicable to others types of datasets. Time horizons of
DDP predictions can be adjusted by using memory over different time windows.
Thus, a big dataset can be parsed in time to make a range of predictions over
varying time horizons.
| no_new_dataset | 0.949295 |
1412.5808 | Johannes Niedermayer | Johannes Niedermayer, Peer Kr\"oger | Minimizing the Number of Matching Queries for Object Retrieval | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | To increase the computational efficiency of interest-point based object
retrieval, researchers have put remarkable research efforts into improving the
efficiency of kNN-based feature matching, pursuing to match thousands of
features against a database within fractions of a second. However, due to the
high-dimensional nature of image features that reduces the effectivity of index
structures (curse of dimensionality), due to the vast amount of features stored
in image databases (images are often represented by up to several thousand
features), this ultimate goal demanded to trade query runtimes for query
precision. In this paper we address an approach complementary to indexing in
order to improve the runtimes of retrieval by querying only the most promising
keypoint descriptors, as this affects matching runtimes linearly and can
therefore lead to increased efficiency. As this reduction of kNN queries
reduces the number of tentative correspondences, a loss of query precision is
minimized by an additional image-level correspondence generation stage with a
computational performance independent of the underlying indexing structure. We
evaluate such an adaption of the standard recognition pipeline on a variety of
datasets using both SIFT and state-of-the-art binary descriptors. Our results
suggest that decreasing the number of queried descriptors does not necessarily
imply a reduction in the result quality as long as alternative ways of
increasing query recall (by thoroughly selecting k) and MAP (using image-level
correspondence generation) are considered.
| [
{
"version": "v1",
"created": "Thu, 18 Dec 2014 11:20:39 GMT"
},
{
"version": "v2",
"created": "Mon, 30 Mar 2015 10:10:14 GMT"
},
{
"version": "v3",
"created": "Tue, 18 Aug 2015 07:08:05 GMT"
}
] | 2015-08-19T00:00:00 | [
[
"Niedermayer",
"Johannes",
""
],
[
"Kröger",
"Peer",
""
]
] | TITLE: Minimizing the Number of Matching Queries for Object Retrieval
ABSTRACT: To increase the computational efficiency of interest-point based object
retrieval, researchers have put remarkable research efforts into improving the
efficiency of kNN-based feature matching, pursuing to match thousands of
features against a database within fractions of a second. However, due to the
high-dimensional nature of image features that reduces the effectivity of index
structures (curse of dimensionality), due to the vast amount of features stored
in image databases (images are often represented by up to several thousand
features), this ultimate goal demanded to trade query runtimes for query
precision. In this paper we address an approach complementary to indexing in
order to improve the runtimes of retrieval by querying only the most promising
keypoint descriptors, as this affects matching runtimes linearly and can
therefore lead to increased efficiency. As this reduction of kNN queries
reduces the number of tentative correspondences, a loss of query precision is
minimized by an additional image-level correspondence generation stage with a
computational performance independent of the underlying indexing structure. We
evaluate such an adaption of the standard recognition pipeline on a variety of
datasets using both SIFT and state-of-the-art binary descriptors. Our results
suggest that decreasing the number of queried descriptors does not necessarily
imply a reduction in the result quality as long as alternative ways of
increasing query recall (by thoroughly selecting k) and MAP (using image-level
correspondence generation) are considered.
| no_new_dataset | 0.950686 |
1506.07862 | Mariusz Tarnopolski | Mariusz Tarnopolski | On the limit between short and long GRBs | 6 pages, 1 figure; matches the version to published in Ap&SS | null | 10.1007/s10509-015-2473-6 | null | astro-ph.HE physics.space-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Two classes of GRBs have been identified thus far without doubt and are
prescribed to different physical scenarios -- NS-NS or NS-BH mergers, and
collapse of massive stars, for short and long GRBs, respectively. The existence
of two distinct populations was inferred through a bimodal distribution of the
observed durations $T_{90}$, and the commonly applied $2\,{\rm s}$ limit
between short and long GRBs was obtained by fitting a parabola between the two
peaks in binned data from BATSE 1B. Herein, by means of a maximum likelihood
(ML) method a mixture of two Gaussians is fitted to the datasets from BATSE,
$Swift$, $BeppoSAX$, and $Fermi$ in search for a local minimum that might serve
as a new, more proper, limit for the two GRB classes. It is found that $Swift$
and $BeppoSAX$ distributions are unimodal, hence no local minimum is present,
$Fermi$ is consistent with the conventional limit, whereas BATSE gives the
limit significantly longer (equal to $3.38\pm 0.27\,{\rm s}$) than $2\,{\rm
s}$. These new values change the fractions of short and long GRBs in the
samples examined, and imply that the observed $T_{90}$ durations are detector
dependent, hence no universal limiting value may be applied to all satellites
due to their different instrument specifications. Because of this, and due to
the strong overlap of the two-Gaussian components, the straightforward
association of short GRBs to mergers and long ones to collapsars is ambiguous.
| [
{
"version": "v1",
"created": "Thu, 25 Jun 2015 19:29:43 GMT"
},
{
"version": "v2",
"created": "Thu, 30 Jul 2015 20:15:13 GMT"
}
] | 2015-08-19T00:00:00 | [
[
"Tarnopolski",
"Mariusz",
""
]
] | TITLE: On the limit between short and long GRBs
ABSTRACT: Two classes of GRBs have been identified thus far without doubt and are
prescribed to different physical scenarios -- NS-NS or NS-BH mergers, and
collapse of massive stars, for short and long GRBs, respectively. The existence
of two distinct populations was inferred through a bimodal distribution of the
observed durations $T_{90}$, and the commonly applied $2\,{\rm s}$ limit
between short and long GRBs was obtained by fitting a parabola between the two
peaks in binned data from BATSE 1B. Herein, by means of a maximum likelihood
(ML) method a mixture of two Gaussians is fitted to the datasets from BATSE,
$Swift$, $BeppoSAX$, and $Fermi$ in search for a local minimum that might serve
as a new, more proper, limit for the two GRB classes. It is found that $Swift$
and $BeppoSAX$ distributions are unimodal, hence no local minimum is present,
$Fermi$ is consistent with the conventional limit, whereas BATSE gives the
limit significantly longer (equal to $3.38\pm 0.27\,{\rm s}$) than $2\,{\rm
s}$. These new values change the fractions of short and long GRBs in the
samples examined, and imply that the observed $T_{90}$ durations are detector
dependent, hence no universal limiting value may be applied to all satellites
due to their different instrument specifications. Because of this, and due to
the strong overlap of the two-Gaussian components, the straightforward
association of short GRBs to mergers and long ones to collapsars is ambiguous.
| no_new_dataset | 0.945298 |
1507.06821 | Andreas Eitel | Andreas Eitel, Jost Tobias Springenberg, Luciano Spinello, Martin
Riedmiller, Wolfram Burgard | Multimodal Deep Learning for Robust RGB-D Object Recognition | Final version submitted to IROS'2015, results unchanged,
reformulation of some text passages in abstract and introduction | null | null | null | cs.CV cs.LG cs.NE cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Robust object recognition is a crucial ingredient of many, if not all,
real-world robotics applications. This paper leverages recent progress on
Convolutional Neural Networks (CNNs) and proposes a novel RGB-D architecture
for object recognition. Our architecture is composed of two separate CNN
processing streams - one for each modality - which are consecutively combined
with a late fusion network. We focus on learning with imperfect sensor data, a
typical problem in real-world robotics tasks. For accurate learning, we
introduce a multi-stage training methodology and two crucial ingredients for
handling depth data with CNNs. The first, an effective encoding of depth
information for CNNs that enables learning without the need for large depth
datasets. The second, a data augmentation scheme for robust learning with depth
images by corrupting them with realistic noise patterns. We present
state-of-the-art results on the RGB-D object dataset and show recognition in
challenging RGB-D real-world noisy settings.
| [
{
"version": "v1",
"created": "Fri, 24 Jul 2015 12:20:19 GMT"
},
{
"version": "v2",
"created": "Tue, 18 Aug 2015 13:04:29 GMT"
}
] | 2015-08-19T00:00:00 | [
[
"Eitel",
"Andreas",
""
],
[
"Springenberg",
"Jost Tobias",
""
],
[
"Spinello",
"Luciano",
""
],
[
"Riedmiller",
"Martin",
""
],
[
"Burgard",
"Wolfram",
""
]
] | TITLE: Multimodal Deep Learning for Robust RGB-D Object Recognition
ABSTRACT: Robust object recognition is a crucial ingredient of many, if not all,
real-world robotics applications. This paper leverages recent progress on
Convolutional Neural Networks (CNNs) and proposes a novel RGB-D architecture
for object recognition. Our architecture is composed of two separate CNN
processing streams - one for each modality - which are consecutively combined
with a late fusion network. We focus on learning with imperfect sensor data, a
typical problem in real-world robotics tasks. For accurate learning, we
introduce a multi-stage training methodology and two crucial ingredients for
handling depth data with CNNs. The first, an effective encoding of depth
information for CNNs that enables learning without the need for large depth
datasets. The second, a data augmentation scheme for robust learning with depth
images by corrupting them with realistic noise patterns. We present
state-of-the-art results on the RGB-D object dataset and show recognition in
challenging RGB-D real-world noisy settings.
| no_new_dataset | 0.94545 |
1508.04190 | Zongbo Hao | Hao Zongbo, Lu Linlin, Zhang Qianni, Wu Jie, Izquierdo Ebroul, Yang
Juanyu, Zhao Jun | Action Recognition based on Subdivision-Fusion Model | Accepted by BMVC2015 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper proposes a novel Subdivision-Fusion Model (SFM) to recognize human
actions. In most action recognition tasks, overlapping feature distribution is
a common problem leading to overfitting. In the subdivision stage of the
proposed SFM, samples in each category are clustered. Then, such samples are
grouped into multiple more concentrated subcategories. Boundaries for the
subcategories are easier to find and as consequence overfitting is avoided. In
the subsequent fusion stage, the multi-subcategories classification results are
converted back to the original category recognition problem. Two methods to
determine the number of clusters are provided. The proposed model has been
thoroughly tested with four popular datasets. In the Hollywood2 dataset, an
accuracy of 79.4% is achieved, outperforming the state-of-the-art accuracy of
64.3%. The performance on the YouTube Action dataset has been improved from
75.8% to 82.5%, while considerably improvements are also observed on the KTH
and UCF50 datasets.
| [
{
"version": "v1",
"created": "Tue, 18 Aug 2015 01:38:08 GMT"
}
] | 2015-08-19T00:00:00 | [
[
"Zongbo",
"Hao",
""
],
[
"Linlin",
"Lu",
""
],
[
"Qianni",
"Zhang",
""
],
[
"Jie",
"Wu",
""
],
[
"Ebroul",
"Izquierdo",
""
],
[
"Juanyu",
"Yang",
""
],
[
"Jun",
"Zhao",
""
]
] | TITLE: Action Recognition based on Subdivision-Fusion Model
ABSTRACT: This paper proposes a novel Subdivision-Fusion Model (SFM) to recognize human
actions. In most action recognition tasks, overlapping feature distribution is
a common problem leading to overfitting. In the subdivision stage of the
proposed SFM, samples in each category are clustered. Then, such samples are
grouped into multiple more concentrated subcategories. Boundaries for the
subcategories are easier to find and as consequence overfitting is avoided. In
the subsequent fusion stage, the multi-subcategories classification results are
converted back to the original category recognition problem. Two methods to
determine the number of clusters are provided. The proposed model has been
thoroughly tested with four popular datasets. In the Hollywood2 dataset, an
accuracy of 79.4% is achieved, outperforming the state-of-the-art accuracy of
64.3%. The performance on the YouTube Action dataset has been improved from
75.8% to 82.5%, while considerably improvements are also observed on the KTH
and UCF50 datasets.
| no_new_dataset | 0.955026 |
1508.04198 | Yifan Fu | Yifan Fu and Junbin Gao and Xia Hong and David Tien | Low Rank Representation on Riemannian Manifold of Square Root Densities | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we present a novel low rank representation (LRR) algorithm for
data lying on the manifold of square root densities. Unlike traditional LRR
methods which rely on the assumption that the data points are vectors in the
Euclidean space, our new algorithm is designed to incorporate the intrinsic
geometric structure and geodesic distance of the manifold. Experiments on
several computer vision datasets showcase its noise robustness and superior
performance on classification and subspace clustering compared to other
state-of-the-art approaches.
| [
{
"version": "v1",
"created": "Tue, 18 Aug 2015 02:33:30 GMT"
}
] | 2015-08-19T00:00:00 | [
[
"Fu",
"Yifan",
""
],
[
"Gao",
"Junbin",
""
],
[
"Hong",
"Xia",
""
],
[
"Tien",
"David",
""
]
] | TITLE: Low Rank Representation on Riemannian Manifold of Square Root Densities
ABSTRACT: In this paper, we present a novel low rank representation (LRR) algorithm for
data lying on the manifold of square root densities. Unlike traditional LRR
methods which rely on the assumption that the data points are vectors in the
Euclidean space, our new algorithm is designed to incorporate the intrinsic
geometric structure and geodesic distance of the manifold. Experiments on
several computer vision datasets showcase its noise robustness and superior
performance on classification and subspace clustering compared to other
state-of-the-art approaches.
| no_new_dataset | 0.956553 |
1508.04333 | Shouvick Mondal | Shouvick Mondal and Arko Banerjee | ESDF: Ensemble Selection using Diversity and Frequency | Conference: National Conference on Research Trends in Computer
Science and Application (NCRTCSA-2014) Date: 8th February 2014 Organized by:
Dept. of Computer Application, Siliguri Institute of Technology, India In
Association With: Computer Society of India, Siliguri Chapter Technically
Sponsored By: IEEE, Kolkata Section Paper Id: NCRTCSA118. Shouvick Mondal et
al.; ESDF: Ensemble Selection using Diversity and Frequency; Proceedings of
NCRTCSA 2014; pp. 28-33, 2014 | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recently ensemble selection for consensus clustering has emerged as a
research problem in Machine Intelligence. Normally consensus clustering
algorithms take into account the entire ensemble of clustering, where there is
a tendency of generating a very large size ensemble before computing its
consensus. One can avoid considering the entire ensemble and can judiciously
select few partitions in the ensemble without compromising on the quality of
the consensus. This may result in an efficient consensus computation technique
and may save unnecessary computational overheads. The ensemble selection
problem addresses this issue of consensus clustering. In this paper, we propose
an efficient method of ensemble selection for a large ensemble. We prioritize
the partitions in the ensemble based on diversity and frequency. Our method
selects top K of the partitions in order of priority, where K is decided by the
user. We observe that considering jointly the diversity and frequency helps in
identifying few representative partitions whose consensus is qualitatively
better than the consensus of the entire ensemble. Experimental analysis on a
large number of datasets shows our method gives better results than earlier
ensemble selection methods.
| [
{
"version": "v1",
"created": "Tue, 18 Aug 2015 14:43:57 GMT"
}
] | 2015-08-19T00:00:00 | [
[
"Mondal",
"Shouvick",
""
],
[
"Banerjee",
"Arko",
""
]
] | TITLE: ESDF: Ensemble Selection using Diversity and Frequency
ABSTRACT: Recently ensemble selection for consensus clustering has emerged as a
research problem in Machine Intelligence. Normally consensus clustering
algorithms take into account the entire ensemble of clustering, where there is
a tendency of generating a very large size ensemble before computing its
consensus. One can avoid considering the entire ensemble and can judiciously
select few partitions in the ensemble without compromising on the quality of
the consensus. This may result in an efficient consensus computation technique
and may save unnecessary computational overheads. The ensemble selection
problem addresses this issue of consensus clustering. In this paper, we propose
an efficient method of ensemble selection for a large ensemble. We prioritize
the partitions in the ensemble based on diversity and frequency. Our method
selects top K of the partitions in order of priority, where K is decided by the
user. We observe that considering jointly the diversity and frequency helps in
identifying few representative partitions whose consensus is qualitatively
better than the consensus of the entire ensemble. Experimental analysis on a
large number of datasets shows our method gives better results than earlier
ensemble selection methods.
| no_new_dataset | 0.954984 |
1508.04389 | Rajeev Ranjan | Rajeev Ranjan, Vishal M. Patel, Rama Chellappa | A Deep Pyramid Deformable Part Model for Face Detection | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a face detection algorithm based on Deformable Part Models and
deep pyramidal features. The proposed method called DP2MFD is able to detect
faces of various sizes and poses in unconstrained conditions. It reduces the
gap in training and testing of DPM on deep features by adding a normalization
layer to the deep convolutional neural network (CNN). Extensive experiments on
four publicly available unconstrained face detection datasets show that our
method is able to capture the meaningful structure of faces and performs
significantly better than many competitive face detection algorithms.
| [
{
"version": "v1",
"created": "Tue, 18 Aug 2015 17:24:09 GMT"
}
] | 2015-08-19T00:00:00 | [
[
"Ranjan",
"Rajeev",
""
],
[
"Patel",
"Vishal M.",
""
],
[
"Chellappa",
"Rama",
""
]
] | TITLE: A Deep Pyramid Deformable Part Model for Face Detection
ABSTRACT: We present a face detection algorithm based on Deformable Part Models and
deep pyramidal features. The proposed method called DP2MFD is able to detect
faces of various sizes and poses in unconstrained conditions. It reduces the
gap in training and testing of DPM on deep features by adding a normalization
layer to the deep convolutional neural network (CNN). Extensive experiments on
four publicly available unconstrained face detection datasets show that our
method is able to capture the meaningful structure of faces and performs
significantly better than many competitive face detection algorithms.
| no_new_dataset | 0.948917 |
1506.00379 | Yankai Lin | Yankai Lin, Zhiyuan Liu, Huanbo Luan, Maosong Sun, Siwei Rao, Song Liu | Modeling Relation Paths for Representation Learning of Knowledge Bases | 10 pages | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Representation learning of knowledge bases (KBs) aims to embed both entities
and relations into a low-dimensional space. Most existing methods only consider
direct relations in representation learning. We argue that multiple-step
relation paths also contain rich inference patterns between entities, and
propose a path-based representation learning model. This model considers
relation paths as translations between entities for representation learning,
and addresses two key challenges: (1) Since not all relation paths are
reliable, we design a path-constraint resource allocation algorithm to measure
the reliability of relation paths. (2) We represent relation paths via semantic
composition of relation embeddings. Experimental results on real-world datasets
show that, as compared with baselines, our model achieves significant and
consistent improvements on knowledge base completion and relation extraction
from text.
| [
{
"version": "v1",
"created": "Mon, 1 Jun 2015 08:22:49 GMT"
},
{
"version": "v2",
"created": "Sat, 15 Aug 2015 09:28:49 GMT"
}
] | 2015-08-18T00:00:00 | [
[
"Lin",
"Yankai",
""
],
[
"Liu",
"Zhiyuan",
""
],
[
"Luan",
"Huanbo",
""
],
[
"Sun",
"Maosong",
""
],
[
"Rao",
"Siwei",
""
],
[
"Liu",
"Song",
""
]
] | TITLE: Modeling Relation Paths for Representation Learning of Knowledge Bases
ABSTRACT: Representation learning of knowledge bases (KBs) aims to embed both entities
and relations into a low-dimensional space. Most existing methods only consider
direct relations in representation learning. We argue that multiple-step
relation paths also contain rich inference patterns between entities, and
propose a path-based representation learning model. This model considers
relation paths as translations between entities for representation learning,
and addresses two key challenges: (1) Since not all relation paths are
reliable, we design a path-constraint resource allocation algorithm to measure
the reliability of relation paths. (2) We represent relation paths via semantic
composition of relation embeddings. Experimental results on real-world datasets
show that, as compared with baselines, our model achieves significant and
consistent improvements on knowledge base completion and relation extraction
from text.
| no_new_dataset | 0.948058 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.