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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1702.03684 | Sebastian Bodenstedt | Sebastian Bodenstedt (1), Martin Wagner (2), Darko Kati\'c (1),
Patrick Mietkowski (2), Benjamin Mayer (2), Hannes Kenngott (2), Beat
M\"uller-Stich (2), R\"udiger Dillmann (1), Stefanie Speidel (1) ((1)
Institute for Anthropomatics and Robotics, Karlsruhe Institute of Technology,
Karlsruhe, (2) Department of General, Visceral and Transplant Surgery,
University of Heidelberg, Heidelberg) | Unsupervised temporal context learning using convolutional neural
networks for laparoscopic workflow analysis | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Computer-assisted surgery (CAS) aims to provide the surgeon with the right
type of assistance at the right moment. Such assistance systems are especially
relevant in laparoscopic surgery, where CAS can alleviate some of the drawbacks
that surgeons incur. For many assistance functions, e.g. displaying the
location of a tumor at the appropriate time or suggesting what instruments to
prepare next, analyzing the surgical workflow is a prerequisite. Since
laparoscopic interventions are performed via endoscope, the video signal is an
obvious sensor modality to rely on for workflow analysis.
Image-based workflow analysis tasks in laparoscopy, such as phase
recognition, skill assessment, video indexing or automatic annotation, require
a temporal distinction between video frames. Generally computer vision based
methods that generalize from previously seen data are used. For training such
methods, large amounts of annotated data are necessary. Annotating surgical
data requires expert knowledge, therefore collecting a sufficient amount of
data is difficult, time-consuming and not always feasible.
In this paper, we address this problem by presenting an unsupervised method
for training a convolutional neural network (CNN) to differentiate between
laparoscopic video frames on a temporal basis. We extract video frames at
regular intervals from 324 unlabeled laparoscopic interventions, resulting in a
dataset of approximately 2.2 million images. From this dataset, we extract
image pairs from the same video and train a CNN to determine their temporal
order. To solve this problem, the CNN has to extract features that are relevant
for comprehending laparoscopic workflow.
Furthermore, we demonstrate that such a CNN can be adapted for surgical
workflow segmentation. We performed image-based workflow segmentation on a
publicly available dataset of 7 cholecystectomies and 9 colorectal
interventions.
| [
{
"version": "v1",
"created": "Mon, 13 Feb 2017 09:29:50 GMT"
}
] | 2017-02-14T00:00:00 | [
[
"Bodenstedt",
"Sebastian",
""
],
[
"Wagner",
"Martin",
""
],
[
"Katić",
"Darko",
""
],
[
"Mietkowski",
"Patrick",
""
],
[
"Mayer",
"Benjamin",
""
],
[
"Kenngott",
"Hannes",
""
],
[
"Müller-Stich",
"Beat",
""
],
[
"Dillmann",
"Rüdiger",
""
],
[
"Speidel",
"Stefanie",
""
]
] | TITLE: Unsupervised temporal context learning using convolutional neural
networks for laparoscopic workflow analysis
ABSTRACT: Computer-assisted surgery (CAS) aims to provide the surgeon with the right
type of assistance at the right moment. Such assistance systems are especially
relevant in laparoscopic surgery, where CAS can alleviate some of the drawbacks
that surgeons incur. For many assistance functions, e.g. displaying the
location of a tumor at the appropriate time or suggesting what instruments to
prepare next, analyzing the surgical workflow is a prerequisite. Since
laparoscopic interventions are performed via endoscope, the video signal is an
obvious sensor modality to rely on for workflow analysis.
Image-based workflow analysis tasks in laparoscopy, such as phase
recognition, skill assessment, video indexing or automatic annotation, require
a temporal distinction between video frames. Generally computer vision based
methods that generalize from previously seen data are used. For training such
methods, large amounts of annotated data are necessary. Annotating surgical
data requires expert knowledge, therefore collecting a sufficient amount of
data is difficult, time-consuming and not always feasible.
In this paper, we address this problem by presenting an unsupervised method
for training a convolutional neural network (CNN) to differentiate between
laparoscopic video frames on a temporal basis. We extract video frames at
regular intervals from 324 unlabeled laparoscopic interventions, resulting in a
dataset of approximately 2.2 million images. From this dataset, we extract
image pairs from the same video and train a CNN to determine their temporal
order. To solve this problem, the CNN has to extract features that are relevant
for comprehending laparoscopic workflow.
Furthermore, we demonstrate that such a CNN can be adapted for surgical
workflow segmentation. We performed image-based workflow segmentation on a
publicly available dataset of 7 cholecystectomies and 9 colorectal
interventions.
| no_new_dataset | 0.615926 |
1702.03825 | Yang Zhang | Yang Zhang, Yusu Wang, Srinivasan Parthasarathy | Analyzing and Visualizing Scalar Fields on Graphs | null | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The value proposition of a dataset often resides in the implicit
interconnections or explicit relationships (patterns) among individual
entities, and is often modeled as a graph. Effective visualization of such
graphs can lead to key insights uncovering such value. In this article we
propose a visualization method to explore graphs with numerical attributes
associated with nodes (or edges) -- referred to as scalar graphs. Such
numerical attributes can represent raw content information, similarities, or
derived information reflecting important network measures such as triangle
density and centrality. The proposed visualization strategy seeks to
simultaneously uncover the relationship between attribute values and graph
topology, and relies on transforming the network to generate a terrain map. A
key objective here is to ensure that the terrain map reveals the overall
distribution of components-of-interest (e.g. dense subgraphs, k-cores) and the
relationships among them while being sensitive to the attribute values over the
graph. We also design extensions that can capture the relationship across
multiple numerical attributes (scalars). We demonstrate the efficacy of our
method on several real-world data science tasks while scaling to large graphs
with millions of nodes.
| [
{
"version": "v1",
"created": "Fri, 10 Feb 2017 07:47:48 GMT"
}
] | 2017-02-14T00:00:00 | [
[
"Zhang",
"Yang",
""
],
[
"Wang",
"Yusu",
""
],
[
"Parthasarathy",
"Srinivasan",
""
]
] | TITLE: Analyzing and Visualizing Scalar Fields on Graphs
ABSTRACT: The value proposition of a dataset often resides in the implicit
interconnections or explicit relationships (patterns) among individual
entities, and is often modeled as a graph. Effective visualization of such
graphs can lead to key insights uncovering such value. In this article we
propose a visualization method to explore graphs with numerical attributes
associated with nodes (or edges) -- referred to as scalar graphs. Such
numerical attributes can represent raw content information, similarities, or
derived information reflecting important network measures such as triangle
density and centrality. The proposed visualization strategy seeks to
simultaneously uncover the relationship between attribute values and graph
topology, and relies on transforming the network to generate a terrain map. A
key objective here is to ensure that the terrain map reveals the overall
distribution of components-of-interest (e.g. dense subgraphs, k-cores) and the
relationships among them while being sensitive to the attribute values over the
graph. We also design extensions that can capture the relationship across
multiple numerical attributes (scalars). We demonstrate the efficacy of our
method on several real-world data science tasks while scaling to large graphs
with millions of nodes.
| no_new_dataset | 0.949623 |
1702.03856 | Sameer Bansal | Sameer Bansal, Herman Kamper, Adam Lopez and Sharon Goldwater | Towards speech-to-text translation without speech recognition | To appear in EACL 2017 (short papers) | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We explore the problem of translating speech to text in low-resource
scenarios where neither automatic speech recognition (ASR) nor machine
translation (MT) are available, but we have training data in the form of audio
paired with text translations. We present the first system for this problem
applied to a realistic multi-speaker dataset, the CALLHOME Spanish-English
speech translation corpus. Our approach uses unsupervised term discovery (UTD)
to cluster repeated patterns in the audio, creating a pseudotext, which we pair
with translations to create a parallel text and train a simple bag-of-words MT
model. We identify the challenges faced by the system, finding that the
difficulty of cross-speaker UTD results in low recall, but that our system is
still able to correctly translate some content words in test data.
| [
{
"version": "v1",
"created": "Mon, 13 Feb 2017 16:30:23 GMT"
}
] | 2017-02-14T00:00:00 | [
[
"Bansal",
"Sameer",
""
],
[
"Kamper",
"Herman",
""
],
[
"Lopez",
"Adam",
""
],
[
"Goldwater",
"Sharon",
""
]
] | TITLE: Towards speech-to-text translation without speech recognition
ABSTRACT: We explore the problem of translating speech to text in low-resource
scenarios where neither automatic speech recognition (ASR) nor machine
translation (MT) are available, but we have training data in the form of audio
paired with text translations. We present the first system for this problem
applied to a realistic multi-speaker dataset, the CALLHOME Spanish-English
speech translation corpus. Our approach uses unsupervised term discovery (UTD)
to cluster repeated patterns in the audio, creating a pseudotext, which we pair
with translations to create a parallel text and train a simple bag-of-words MT
model. We identify the challenges faced by the system, finding that the
difficulty of cross-speaker UTD results in low recall, but that our system is
still able to correctly translate some content words in test data.
| no_new_dataset | 0.848972 |
1603.04037 | Umar Iqbal | Umar Iqbal, Martin Garbade, Juergen Gall | Pose for Action - Action for Pose | Accepted to FG-2017 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work we propose to utilize information about human actions to improve
pose estimation in monocular videos. To this end, we present a pictorial
structure model that exploits high-level information about activities to
incorporate higher-order part dependencies by modeling action specific
appearance models and pose priors. However, instead of using an additional
expensive action recognition framework, the action priors are efficiently
estimated by our pose estimation framework. This is achieved by starting with a
uniform action prior and updating the action prior during pose estimation. We
also show that learning the right amount of appearance sharing among action
classes improves the pose estimation. We demonstrate the effectiveness of the
proposed method on two challenging datasets for pose estimation and action
recognition with over 80,000 test images.
| [
{
"version": "v1",
"created": "Sun, 13 Mar 2016 15:09:35 GMT"
},
{
"version": "v2",
"created": "Fri, 10 Feb 2017 14:01:09 GMT"
}
] | 2017-02-13T00:00:00 | [
[
"Iqbal",
"Umar",
""
],
[
"Garbade",
"Martin",
""
],
[
"Gall",
"Juergen",
""
]
] | TITLE: Pose for Action - Action for Pose
ABSTRACT: In this work we propose to utilize information about human actions to improve
pose estimation in monocular videos. To this end, we present a pictorial
structure model that exploits high-level information about activities to
incorporate higher-order part dependencies by modeling action specific
appearance models and pose priors. However, instead of using an additional
expensive action recognition framework, the action priors are efficiently
estimated by our pose estimation framework. This is achieved by starting with a
uniform action prior and updating the action prior during pose estimation. We
also show that learning the right amount of appearance sharing among action
classes improves the pose estimation. We demonstrate the effectiveness of the
proposed method on two challenging datasets for pose estimation and action
recognition with over 80,000 test images.
| no_new_dataset | 0.947381 |
1609.03056 | Yemin Shi Shi | Yemin Shi, Yonghong Tian, Yaowei Wang, Tiejun Huang | Sequential Deep Trajectory Descriptor for Action Recognition with
Three-stream CNN | 10 pages, 29 figures, T-MM | null | 10.1109/TMM.2017.2666540 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Learning the spatial-temporal representation of motion information is crucial
to human action recognition. Nevertheless, most of the existing features or
descriptors cannot capture motion information effectively, especially for
long-term motion. To address this problem, this paper proposes a long-term
motion descriptor called sequential Deep Trajectory Descriptor (sDTD).
Specifically, we project dense trajectories into two-dimensional planes, and
subsequently a CNN-RNN network is employed to learn an effective representation
for long-term motion. Unlike the popular two-stream ConvNets, the sDTD stream
is introduced into a three-stream framework so as to identify actions from a
video sequence. Consequently, this three-stream framework can simultaneously
capture static spatial features, short-term motion and long-term motion in the
video. Extensive experiments were conducted on three challenging datasets: KTH,
HMDB51 and UCF101. Experimental results show that our method achieves
state-of-the-art performance on the KTH and UCF101 datasets, and is comparable
to the state-of-the-art methods on the HMDB51 dataset.
| [
{
"version": "v1",
"created": "Sat, 10 Sep 2016 14:24:38 GMT"
},
{
"version": "v2",
"created": "Fri, 10 Feb 2017 02:49:10 GMT"
}
] | 2017-02-13T00:00:00 | [
[
"Shi",
"Yemin",
""
],
[
"Tian",
"Yonghong",
""
],
[
"Wang",
"Yaowei",
""
],
[
"Huang",
"Tiejun",
""
]
] | TITLE: Sequential Deep Trajectory Descriptor for Action Recognition with
Three-stream CNN
ABSTRACT: Learning the spatial-temporal representation of motion information is crucial
to human action recognition. Nevertheless, most of the existing features or
descriptors cannot capture motion information effectively, especially for
long-term motion. To address this problem, this paper proposes a long-term
motion descriptor called sequential Deep Trajectory Descriptor (sDTD).
Specifically, we project dense trajectories into two-dimensional planes, and
subsequently a CNN-RNN network is employed to learn an effective representation
for long-term motion. Unlike the popular two-stream ConvNets, the sDTD stream
is introduced into a three-stream framework so as to identify actions from a
video sequence. Consequently, this three-stream framework can simultaneously
capture static spatial features, short-term motion and long-term motion in the
video. Extensive experiments were conducted on three challenging datasets: KTH,
HMDB51 and UCF101. Experimental results show that our method achieves
state-of-the-art performance on the KTH and UCF101 datasets, and is comparable
to the state-of-the-art methods on the HMDB51 dataset.
| no_new_dataset | 0.952486 |
1610.08738 | Nicolas Keriven | Nicolas Keriven (PANAMA), Nicolas Tremblay (GIPSA-CICS), Yann
Traonmilin (PANAMA), R\'emi Gribonval (PANAMA) | Compressive K-means | null | null | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The Lloyd-Max algorithm is a classical approach to perform K-means
clustering. Unfortunately, its cost becomes prohibitive as the training dataset
grows large. We propose a compressive version of K-means (CKM), that estimates
cluster centers from a sketch, i.e. from a drastically compressed
representation of the training dataset. We demonstrate empirically that CKM
performs similarly to Lloyd-Max, for a sketch size proportional to the number
of cen-troids times the ambient dimension, and independent of the size of the
original dataset. Given the sketch, the computational complexity of CKM is also
independent of the size of the dataset. Unlike Lloyd-Max which requires several
replicates, we further demonstrate that CKM is almost insensitive to
initialization. For a large dataset of 10^7 data points, we show that CKM can
run two orders of magnitude faster than five replicates of Lloyd-Max, with
similar clustering performance on artificial data. Finally, CKM achieves lower
classification errors on handwritten digits classification.
| [
{
"version": "v1",
"created": "Thu, 27 Oct 2016 12:13:05 GMT"
},
{
"version": "v2",
"created": "Wed, 30 Nov 2016 07:58:05 GMT"
},
{
"version": "v3",
"created": "Mon, 9 Jan 2017 10:40:53 GMT"
},
{
"version": "v4",
"created": "Fri, 10 Feb 2017 15:22:24 GMT"
}
] | 2017-02-13T00:00:00 | [
[
"Keriven",
"Nicolas",
"",
"PANAMA"
],
[
"Tremblay",
"Nicolas",
"",
"GIPSA-CICS"
],
[
"Traonmilin",
"Yann",
"",
"PANAMA"
],
[
"Gribonval",
"Rémi",
"",
"PANAMA"
]
] | TITLE: Compressive K-means
ABSTRACT: The Lloyd-Max algorithm is a classical approach to perform K-means
clustering. Unfortunately, its cost becomes prohibitive as the training dataset
grows large. We propose a compressive version of K-means (CKM), that estimates
cluster centers from a sketch, i.e. from a drastically compressed
representation of the training dataset. We demonstrate empirically that CKM
performs similarly to Lloyd-Max, for a sketch size proportional to the number
of cen-troids times the ambient dimension, and independent of the size of the
original dataset. Given the sketch, the computational complexity of CKM is also
independent of the size of the dataset. Unlike Lloyd-Max which requires several
replicates, we further demonstrate that CKM is almost insensitive to
initialization. For a large dataset of 10^7 data points, we show that CKM can
run two orders of magnitude faster than five replicates of Lloyd-Max, with
similar clustering performance on artificial data. Finally, CKM achieves lower
classification errors on handwritten digits classification.
| no_new_dataset | 0.941868 |
1611.00128 | David Rosen | David M. Rosen, Luca Carlone, Afonso S. Bandeira, and John J. Leonard | A Certifiably Correct Algorithm for Synchronization over the Special
Euclidean Group | 16 pages, 8 figures, to appear in the International Workshop on the
Algorithmic Foundations of Robotics (WAFR), Dec 2016 | null | null | null | cs.RO math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Many geometric estimation problems take the form of synchronization over the
special Euclidean group: estimate the values of a set of poses given noisy
measurements of a subset of their pairwise relative transforms. This problem is
typically formulated as a maximum-likelihood estimation that requires solving a
nonconvex nonlinear program, which is computationally intractable in general.
Nevertheless, in this paper we present an algorithm that is able to efficiently
recover certifiably globally optimal solutions of this estimation problem in a
non-adversarial noise regime. The crux of our approach is the development of a
semidefinite relaxation of the maximum-likelihood estimation whose minimizer
provides the exact MLE so long as the magnitude of the noise corrupting the
available measurements falls below a certain critical threshold; furthermore,
whenever exactness obtains, it is possible to verify this fact a posteriori,
thereby certifying the optimality of the recovered estimate. We develop a
specialized optimization scheme for solving large-scale instances of this
semidefinite relaxation by exploiting its low-rank, geometric, and
graph-theoretic structure to reduce it to an equivalent optimization problem on
a low-dimensional Riemannian manifold, and then design a Riemannian
truncated-Newton trust-region method to solve this reduction efficiently. We
combine this fast optimization approach with a simple rounding procedure to
produce our algorithm, SE-Sync. Experimental evaluation on a variety of
simulated and real-world pose-graph SLAM datasets shows that SE-Sync is capable
of recovering globally optimal solutions when the available measurements are
corrupted by noise up to an order of magnitude greater than that typically
encountered in robotics applications, and does so at a computational cost that
scales comparably with that of direct Newton-type local search techniques.
| [
{
"version": "v1",
"created": "Tue, 1 Nov 2016 04:54:35 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Nov 2016 22:37:03 GMT"
},
{
"version": "v3",
"created": "Fri, 10 Feb 2017 02:04:32 GMT"
}
] | 2017-02-13T00:00:00 | [
[
"Rosen",
"David M.",
""
],
[
"Carlone",
"Luca",
""
],
[
"Bandeira",
"Afonso S.",
""
],
[
"Leonard",
"John J.",
""
]
] | TITLE: A Certifiably Correct Algorithm for Synchronization over the Special
Euclidean Group
ABSTRACT: Many geometric estimation problems take the form of synchronization over the
special Euclidean group: estimate the values of a set of poses given noisy
measurements of a subset of their pairwise relative transforms. This problem is
typically formulated as a maximum-likelihood estimation that requires solving a
nonconvex nonlinear program, which is computationally intractable in general.
Nevertheless, in this paper we present an algorithm that is able to efficiently
recover certifiably globally optimal solutions of this estimation problem in a
non-adversarial noise regime. The crux of our approach is the development of a
semidefinite relaxation of the maximum-likelihood estimation whose minimizer
provides the exact MLE so long as the magnitude of the noise corrupting the
available measurements falls below a certain critical threshold; furthermore,
whenever exactness obtains, it is possible to verify this fact a posteriori,
thereby certifying the optimality of the recovered estimate. We develop a
specialized optimization scheme for solving large-scale instances of this
semidefinite relaxation by exploiting its low-rank, geometric, and
graph-theoretic structure to reduce it to an equivalent optimization problem on
a low-dimensional Riemannian manifold, and then design a Riemannian
truncated-Newton trust-region method to solve this reduction efficiently. We
combine this fast optimization approach with a simple rounding procedure to
produce our algorithm, SE-Sync. Experimental evaluation on a variety of
simulated and real-world pose-graph SLAM datasets shows that SE-Sync is capable
of recovering globally optimal solutions when the available measurements are
corrupted by noise up to an order of magnitude greater than that typically
encountered in robotics applications, and does so at a computational cost that
scales comparably with that of direct Newton-type local search techniques.
| no_new_dataset | 0.941868 |
1702.02817 | Immanuel Bayer | Immanuel Bayer, Uwe Nagel, Steffen Rendle | Graph Based Relational Features for Collective Classification | Pacific-Asia Conference on Knowledge Discovery and Data Mining | null | 10.1007/978-3-319-18032-8_35 | null | cs.IR cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Statistical Relational Learning (SRL) methods have shown that classification
accuracy can be improved by integrating relations between samples. Techniques
such as iterative classification or relaxation labeling achieve this by
propagating information between related samples during the inference process.
When only a few samples are labeled and connections between samples are sparse,
collective inference methods have shown large improvements over standard
feature-based ML methods. However, in contrast to feature based ML, collective
inference methods require complex inference procedures and often depend on the
strong assumption of label consistency among related samples. In this paper, we
introduce new relational features for standard ML methods by extracting
information from direct and indirect relations. We show empirically on three
standard benchmark datasets that our relational features yield results
comparable to collective inference methods. Finally we show that our proposal
outperforms these methods when additional information is available.
| [
{
"version": "v1",
"created": "Thu, 9 Feb 2017 12:58:23 GMT"
}
] | 2017-02-13T00:00:00 | [
[
"Bayer",
"Immanuel",
""
],
[
"Nagel",
"Uwe",
""
],
[
"Rendle",
"Steffen",
""
]
] | TITLE: Graph Based Relational Features for Collective Classification
ABSTRACT: Statistical Relational Learning (SRL) methods have shown that classification
accuracy can be improved by integrating relations between samples. Techniques
such as iterative classification or relaxation labeling achieve this by
propagating information between related samples during the inference process.
When only a few samples are labeled and connections between samples are sparse,
collective inference methods have shown large improvements over standard
feature-based ML methods. However, in contrast to feature based ML, collective
inference methods require complex inference procedures and often depend on the
strong assumption of label consistency among related samples. In this paper, we
introduce new relational features for standard ML methods by extracting
information from direct and indirect relations. We show empirically on three
standard benchmark datasets that our relational features yield results
comparable to collective inference methods. Finally we show that our proposal
outperforms these methods when additional information is available.
| no_new_dataset | 0.945651 |
1702.02970 | Jonathan Ullman | Mitali Bafna and Jonathan Ullman | The Price of Selection in Differential Privacy | null | null | null | null | cs.DS cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In the differentially private top-$k$ selection problem, we are given a
dataset $X \in \{\pm 1\}^{n \times d}$, in which each row belongs to an
individual and each column corresponds to some binary attribute, and our goal
is to find a set of $k \ll d$ columns whose means are approximately as large as
possible. Differential privacy requires that our choice of these $k$ columns
does not depend too much on any on individual's dataset. This problem can be
solved using the well known exponential mechanism and composition properties of
differential privacy. In the high-accuracy regime, where we require the error
of the selection procedure to be to be smaller than the so-called sampling
error $\alpha \approx \sqrt{\ln(d)/n}$, this procedure succeeds given a dataset
of size $n \gtrsim k \ln(d)$.
We prove a matching lower bound, showing that a dataset of size $n \gtrsim k
\ln(d)$ is necessary for private top-$k$ selection in this high-accuracy
regime. Our lower bound is the first to show that selecting the $k$ largest
columns requires more data than simply estimating the value of those $k$
columns, which can be done using a dataset of size just $n \gtrsim k$.
| [
{
"version": "v1",
"created": "Thu, 9 Feb 2017 20:11:49 GMT"
}
] | 2017-02-13T00:00:00 | [
[
"Bafna",
"Mitali",
""
],
[
"Ullman",
"Jonathan",
""
]
] | TITLE: The Price of Selection in Differential Privacy
ABSTRACT: In the differentially private top-$k$ selection problem, we are given a
dataset $X \in \{\pm 1\}^{n \times d}$, in which each row belongs to an
individual and each column corresponds to some binary attribute, and our goal
is to find a set of $k \ll d$ columns whose means are approximately as large as
possible. Differential privacy requires that our choice of these $k$ columns
does not depend too much on any on individual's dataset. This problem can be
solved using the well known exponential mechanism and composition properties of
differential privacy. In the high-accuracy regime, where we require the error
of the selection procedure to be to be smaller than the so-called sampling
error $\alpha \approx \sqrt{\ln(d)/n}$, this procedure succeeds given a dataset
of size $n \gtrsim k \ln(d)$.
We prove a matching lower bound, showing that a dataset of size $n \gtrsim k
\ln(d)$ is necessary for private top-$k$ selection in this high-accuracy
regime. Our lower bound is the first to show that selecting the $k$ largest
columns requires more data than simply estimating the value of those $k$
columns, which can be done using a dataset of size just $n \gtrsim k$.
| no_new_dataset | 0.912553 |
1702.03267 | Amarjot Singh | Amarjot Singh and Nick Kingsbury | Dual-Tree Wavelet Scattering Network with Parametric Log Transformation
for Object Classification | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We introduce a ScatterNet that uses a parametric log transformation with
Dual-Tree complex wavelets to extract translation invariant representations
from a multi-resolution image. The parametric transformation aids the OLS
pruning algorithm by converting the skewed distributions into relatively
mean-symmetric distributions while the Dual-Tree wavelets improve the
computational efficiency of the network. The proposed network is shown to
outperform Mallat's ScatterNet on two image datasets, both for classification
accuracy and computational efficiency. The advantages of the proposed network
over other supervised and some unsupervised methods are also presented using
experiments performed on different training dataset sizes.
| [
{
"version": "v1",
"created": "Fri, 10 Feb 2017 18:02:05 GMT"
}
] | 2017-02-13T00:00:00 | [
[
"Singh",
"Amarjot",
""
],
[
"Kingsbury",
"Nick",
""
]
] | TITLE: Dual-Tree Wavelet Scattering Network with Parametric Log Transformation
for Object Classification
ABSTRACT: We introduce a ScatterNet that uses a parametric log transformation with
Dual-Tree complex wavelets to extract translation invariant representations
from a multi-resolution image. The parametric transformation aids the OLS
pruning algorithm by converting the skewed distributions into relatively
mean-symmetric distributions while the Dual-Tree wavelets improve the
computational efficiency of the network. The proposed network is shown to
outperform Mallat's ScatterNet on two image datasets, both for classification
accuracy and computational efficiency. The advantages of the proposed network
over other supervised and some unsupervised methods are also presented using
experiments performed on different training dataset sizes.
| no_new_dataset | 0.951594 |
1701.03918 | Yongqing Wang | Yongqing Wang, Shenghua Liu, Huawei Shen, Xueqi Cheng | Marked Temporal Dynamics Modeling based on Recurrent Neural Network | null | null | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We are now witnessing the increasing availability of event stream data, i.e.,
a sequence of events with each event typically being denoted by the time it
occurs and its mark information (e.g., event type). A fundamental problem is to
model and predict such kind of marked temporal dynamics, i.e., when the next
event will take place and what its mark will be. Existing methods either
predict only the mark or the time of the next event, or predict both of them,
yet separately. Indeed, in marked temporal dynamics, the time and the mark of
the next event are highly dependent on each other, requiring a method that
could simultaneously predict both of them. To tackle this problem, in this
paper, we propose to model marked temporal dynamics by using a mark-specific
intensity function to explicitly capture the dependency between the mark and
the time of the next event. Extensive experiments on two datasets demonstrate
that the proposed method outperforms state-of-the-art methods at predicting
marked temporal dynamics.
| [
{
"version": "v1",
"created": "Sat, 14 Jan 2017 13:26:39 GMT"
}
] | 2017-02-12T00:00:00 | [
[
"Wang",
"Yongqing",
""
],
[
"Liu",
"Shenghua",
""
],
[
"Shen",
"Huawei",
""
],
[
"Cheng",
"Xueqi",
""
]
] | TITLE: Marked Temporal Dynamics Modeling based on Recurrent Neural Network
ABSTRACT: We are now witnessing the increasing availability of event stream data, i.e.,
a sequence of events with each event typically being denoted by the time it
occurs and its mark information (e.g., event type). A fundamental problem is to
model and predict such kind of marked temporal dynamics, i.e., when the next
event will take place and what its mark will be. Existing methods either
predict only the mark or the time of the next event, or predict both of them,
yet separately. Indeed, in marked temporal dynamics, the time and the mark of
the next event are highly dependent on each other, requiring a method that
could simultaneously predict both of them. To tackle this problem, in this
paper, we propose to model marked temporal dynamics by using a mark-specific
intensity function to explicitly capture the dependency between the mark and
the time of the next event. Extensive experiments on two datasets demonstrate
that the proposed method outperforms state-of-the-art methods at predicting
marked temporal dynamics.
| no_new_dataset | 0.950549 |
1701.03947 | Tuan Tran | Tuan Tran, Claudia Nieder\'ee, Nattiya Kanhabua, Ujwal Gadiraju,
Avishek Anand | Balancing Novelty and Salience: Adaptive Learning to Rank Entities for
Timeline Summarization of High-impact Events | Published via ACM to CIKM 2015 | null | 10.1145/2806416.2806486 | null | cs.IR cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Long-running, high-impact events such as the Boston Marathon bombing often
develop through many stages and involve a large number of entities in their
unfolding. Timeline summarization of an event by key sentences eases story
digestion, but does not distinguish between what a user remembers and what she
might want to re-check. In this work, we present a novel approach for timeline
summarization of high-impact events, which uses entities instead of sentences
for summarizing the event at each individual point in time. Such entity
summaries can serve as both (1) important memory cues in a retrospective event
consideration and (2) pointers for personalized event exploration. In order to
automatically create such summaries, it is crucial to identify the "right"
entities for inclusion. We propose to learn a ranking function for entities,
with a dynamically adapted trade-off between the in-document salience of
entities and the informativeness of entities across documents, i.e., the level
of new information associated with an entity for a time point under
consideration. Furthermore, for capturing collective attention for an entity we
use an innovative soft labeling approach based on Wikipedia. Our experiments on
a real large news datasets confirm the effectiveness of the proposed methods.
| [
{
"version": "v1",
"created": "Sat, 14 Jan 2017 16:47:51 GMT"
}
] | 2017-02-12T00:00:00 | [
[
"Tran",
"Tuan",
""
],
[
"Niederée",
"Claudia",
""
],
[
"Kanhabua",
"Nattiya",
""
],
[
"Gadiraju",
"Ujwal",
""
],
[
"Anand",
"Avishek",
""
]
] | TITLE: Balancing Novelty and Salience: Adaptive Learning to Rank Entities for
Timeline Summarization of High-impact Events
ABSTRACT: Long-running, high-impact events such as the Boston Marathon bombing often
develop through many stages and involve a large number of entities in their
unfolding. Timeline summarization of an event by key sentences eases story
digestion, but does not distinguish between what a user remembers and what she
might want to re-check. In this work, we present a novel approach for timeline
summarization of high-impact events, which uses entities instead of sentences
for summarizing the event at each individual point in time. Such entity
summaries can serve as both (1) important memory cues in a retrospective event
consideration and (2) pointers for personalized event exploration. In order to
automatically create such summaries, it is crucial to identify the "right"
entities for inclusion. We propose to learn a ranking function for entities,
with a dynamically adapted trade-off between the in-document salience of
entities and the informativeness of entities across documents, i.e., the level
of new information associated with an entity for a time point under
consideration. Furthermore, for capturing collective attention for an entity we
use an innovative soft labeling approach based on Wikipedia. Our experiments on
a real large news datasets confirm the effectiveness of the proposed methods.
| no_new_dataset | 0.954858 |
1607.03392 | Christian Dansereau | Christian Dansereau, Yassine Benhajali, Celine Risterucci, Emilio
Merlo Pich, Pierre Orban, Douglas Arnold, Pierre Bellec | Statistical power and prediction accuracy in multisite resting-state
fMRI connectivity | null | NeuroImage.Vol 149, p. 220-232 (2017) | 10.1016/j.neuroimage.2017.01.072 | null | q-bio.QM cs.CE stat.ML | http://creativecommons.org/licenses/by/4.0/ | Connectivity studies using resting-state functional magnetic resonance
imaging are increasingly pooling data acquired at multiple sites. While this
may allow investigators to speed up recruitment or increase sample size,
multisite studies also potentially introduce systematic biases in connectivity
measures across sites. In this work, we measure the inter-site effect in
connectivity and its impact on our ability to detect individual and group
differences. Our study was based on real, as opposed to simulated, multisite
fMRI datasets collected in N=345 young, healthy subjects across 8 scanning
sites with 3T scanners and heterogeneous scanning protocols, drawn from the
1000 functional connectome project. We first empirically show that typical
functional networks were reliably found at the group level in all sites, and
that the amplitude of the inter-site effects was small to moderate, with a
Cohen's effect size below 0.5 on average across brain connections. We then
implemented a series of Monte-Carlo simulations, based on real data, to
evaluate the impact of the multisite effects on detection power in statistical
tests comparing two groups (with and without the effect) using a general linear
model, as well as on the prediction of group labels with a support-vector
machine. As a reference, we also implemented the same simulations with fMRI
data collected at a single site using an identical sample size. Simulations
revealed that using data from heterogeneous sites only slightly decreased our
ability to detect changes compared to a monosite study with the GLM, and had a
greater impact on prediction accuracy. Taken together, our results support the
feasibility of multisite studies in rs-fMRI provided the sample size is large
enough.
| [
{
"version": "v1",
"created": "Tue, 12 Jul 2016 15:22:52 GMT"
},
{
"version": "v2",
"created": "Thu, 1 Dec 2016 19:47:13 GMT"
},
{
"version": "v3",
"created": "Fri, 27 Jan 2017 17:57:06 GMT"
}
] | 2017-02-10T00:00:00 | [
[
"Dansereau",
"Christian",
""
],
[
"Benhajali",
"Yassine",
""
],
[
"Risterucci",
"Celine",
""
],
[
"Pich",
"Emilio Merlo",
""
],
[
"Orban",
"Pierre",
""
],
[
"Arnold",
"Douglas",
""
],
[
"Bellec",
"Pierre",
""
]
] | TITLE: Statistical power and prediction accuracy in multisite resting-state
fMRI connectivity
ABSTRACT: Connectivity studies using resting-state functional magnetic resonance
imaging are increasingly pooling data acquired at multiple sites. While this
may allow investigators to speed up recruitment or increase sample size,
multisite studies also potentially introduce systematic biases in connectivity
measures across sites. In this work, we measure the inter-site effect in
connectivity and its impact on our ability to detect individual and group
differences. Our study was based on real, as opposed to simulated, multisite
fMRI datasets collected in N=345 young, healthy subjects across 8 scanning
sites with 3T scanners and heterogeneous scanning protocols, drawn from the
1000 functional connectome project. We first empirically show that typical
functional networks were reliably found at the group level in all sites, and
that the amplitude of the inter-site effects was small to moderate, with a
Cohen's effect size below 0.5 on average across brain connections. We then
implemented a series of Monte-Carlo simulations, based on real data, to
evaluate the impact of the multisite effects on detection power in statistical
tests comparing two groups (with and without the effect) using a general linear
model, as well as on the prediction of group labels with a support-vector
machine. As a reference, we also implemented the same simulations with fMRI
data collected at a single site using an identical sample size. Simulations
revealed that using data from heterogeneous sites only slightly decreased our
ability to detect changes compared to a monosite study with the GLM, and had a
greater impact on prediction accuracy. Taken together, our results support the
feasibility of multisite studies in rs-fMRI provided the sample size is large
enough.
| no_new_dataset | 0.945197 |
1702.01446 | Nirman Kumar | Pankaj K. Agarwal and Nirman Kumar and Stavros Sintos and Subhash Suri | Efficient Algorithms for k-Regret Minimizing Sets | null | null | null | null | cs.DS cs.CG cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A regret minimizing set Q is a small size representation of a much larger
database P so that user queries executed on Q return answers whose scores are
not much worse than those on the full dataset. In particular, a k-regret
minimizing set has the property that the regret ratio between the score of the
top-1 item in Q and the score of the top-k item in P is minimized, where the
score of an item is the inner product of the item's attributes with a user's
weight (preference) vector. The problem is challenging because we want to find
a single representative set Q whose regret ratio is small with respect to all
possible user weight vectors.
We show that k-regret minimization is NP-Complete for all dimensions d >= 3.
This settles an open problem from Chester et al. [VLDB 2014], and resolves the
complexity status of the problem for all d: the problem is known to have
polynomial-time solution for d <= 2. In addition, we propose two new
approximation schemes for regret minimization, both with provable guarantees,
one based on coresets and another based on hitting sets. We also carry out
extensive experimental evaluation, and show that our schemes compute
regret-minimizing sets comparable in size to the greedy algorithm proposed in
[VLDB 14] but our schemes are significantly faster and scalable to large data
sets.
| [
{
"version": "v1",
"created": "Sun, 5 Feb 2017 19:30:44 GMT"
},
{
"version": "v2",
"created": "Thu, 9 Feb 2017 01:46:20 GMT"
}
] | 2017-02-10T00:00:00 | [
[
"Agarwal",
"Pankaj K.",
""
],
[
"Kumar",
"Nirman",
""
],
[
"Sintos",
"Stavros",
""
],
[
"Suri",
"Subhash",
""
]
] | TITLE: Efficient Algorithms for k-Regret Minimizing Sets
ABSTRACT: A regret minimizing set Q is a small size representation of a much larger
database P so that user queries executed on Q return answers whose scores are
not much worse than those on the full dataset. In particular, a k-regret
minimizing set has the property that the regret ratio between the score of the
top-1 item in Q and the score of the top-k item in P is minimized, where the
score of an item is the inner product of the item's attributes with a user's
weight (preference) vector. The problem is challenging because we want to find
a single representative set Q whose regret ratio is small with respect to all
possible user weight vectors.
We show that k-regret minimization is NP-Complete for all dimensions d >= 3.
This settles an open problem from Chester et al. [VLDB 2014], and resolves the
complexity status of the problem for all d: the problem is known to have
polynomial-time solution for d <= 2. In addition, we propose two new
approximation schemes for regret minimization, both with provable guarantees,
one based on coresets and another based on hitting sets. We also carry out
extensive experimental evaluation, and show that our schemes compute
regret-minimizing sets comparable in size to the greedy algorithm proposed in
[VLDB 14] but our schemes are significantly faster and scalable to large data
sets.
| no_new_dataset | 0.941385 |
1702.02363 | Bahadir Sahin | H. Bahadir Sahin, Caglar Tirkaz, Eray Yildiz, Mustafa Tolga Eren, Ozan
Sonmez | Automatically Annotated Turkish Corpus for Named Entity Recognition and
Text Categorization using Large-Scale Gazetteers | 10 page, 1 figure, white paper, update: added correct download link
for dataset | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Turkish Wikipedia Named-Entity Recognition and Text Categorization (TWNERTC)
dataset is a collection of automatically categorized and annotated sentences
obtained from Wikipedia. We constructed large-scale gazetteers by using a graph
crawler algorithm to extract relevant entity and domain information from a
semantic knowledge base, Freebase. The constructed gazetteers contains
approximately 300K entities with thousands of fine-grained entity types under
77 different domains. Since automated processes are prone to ambiguity, we also
introduce two new content specific noise reduction methodologies. Moreover, we
map fine-grained entity types to the equivalent four coarse-grained types:
person, loc, org, misc. Eventually, we construct six different dataset versions
and evaluate the quality of annotations by comparing ground truths from human
annotators. We make these datasets publicly available to support studies on
Turkish named-entity recognition (NER) and text categorization (TC).
| [
{
"version": "v1",
"created": "Wed, 8 Feb 2017 10:45:23 GMT"
},
{
"version": "v2",
"created": "Thu, 9 Feb 2017 08:35:12 GMT"
}
] | 2017-02-10T00:00:00 | [
[
"Sahin",
"H. Bahadir",
""
],
[
"Tirkaz",
"Caglar",
""
],
[
"Yildiz",
"Eray",
""
],
[
"Eren",
"Mustafa Tolga",
""
],
[
"Sonmez",
"Ozan",
""
]
] | TITLE: Automatically Annotated Turkish Corpus for Named Entity Recognition and
Text Categorization using Large-Scale Gazetteers
ABSTRACT: Turkish Wikipedia Named-Entity Recognition and Text Categorization (TWNERTC)
dataset is a collection of automatically categorized and annotated sentences
obtained from Wikipedia. We constructed large-scale gazetteers by using a graph
crawler algorithm to extract relevant entity and domain information from a
semantic knowledge base, Freebase. The constructed gazetteers contains
approximately 300K entities with thousands of fine-grained entity types under
77 different domains. Since automated processes are prone to ambiguity, we also
introduce two new content specific noise reduction methodologies. Moreover, we
map fine-grained entity types to the equivalent four coarse-grained types:
person, loc, org, misc. Eventually, we construct six different dataset versions
and evaluate the quality of annotations by comparing ground truths from human
annotators. We make these datasets publicly available to support studies on
Turkish named-entity recognition (NER) and text categorization (TC).
| new_dataset | 0.946892 |
1702.02640 | Zhe Gan | Zhe Gan, P. D. Singh, Ameet Joshi, Xiaodong He, Jianshu Chen, Jianfeng
Gao, Li Deng | Character-level Deep Conflation for Business Data Analytics | Accepted for publication, at ICASSP 2017 | null | null | null | cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Connecting different text attributes associated with the same entity
(conflation) is important in business data analytics since it could help merge
two different tables in a database to provide a more comprehensive profile of
an entity. However, the conflation task is challenging because two text strings
that describe the same entity could be quite different from each other for
reasons such as misspelling. It is therefore critical to develop a conflation
model that is able to truly understand the semantic meaning of the strings and
match them at the semantic level. To this end, we develop a character-level
deep conflation model that encodes the input text strings from character level
into finite dimension feature vectors, which are then used to compute the
cosine similarity between the text strings. The model is trained in an
end-to-end manner using back propagation and stochastic gradient descent to
maximize the likelihood of the correct association. Specifically, we propose
two variants of the deep conflation model, based on long-short-term memory
(LSTM) recurrent neural network (RNN) and convolutional neural network (CNN),
respectively. Both models perform well on a real-world business analytics
dataset and significantly outperform the baseline bag-of-character (BoC) model.
| [
{
"version": "v1",
"created": "Wed, 8 Feb 2017 22:24:14 GMT"
}
] | 2017-02-10T00:00:00 | [
[
"Gan",
"Zhe",
""
],
[
"Singh",
"P. D.",
""
],
[
"Joshi",
"Ameet",
""
],
[
"He",
"Xiaodong",
""
],
[
"Chen",
"Jianshu",
""
],
[
"Gao",
"Jianfeng",
""
],
[
"Deng",
"Li",
""
]
] | TITLE: Character-level Deep Conflation for Business Data Analytics
ABSTRACT: Connecting different text attributes associated with the same entity
(conflation) is important in business data analytics since it could help merge
two different tables in a database to provide a more comprehensive profile of
an entity. However, the conflation task is challenging because two text strings
that describe the same entity could be quite different from each other for
reasons such as misspelling. It is therefore critical to develop a conflation
model that is able to truly understand the semantic meaning of the strings and
match them at the semantic level. To this end, we develop a character-level
deep conflation model that encodes the input text strings from character level
into finite dimension feature vectors, which are then used to compute the
cosine similarity between the text strings. The model is trained in an
end-to-end manner using back propagation and stochastic gradient descent to
maximize the likelihood of the correct association. Specifically, we propose
two variants of the deep conflation model, based on long-short-term memory
(LSTM) recurrent neural network (RNN) and convolutional neural network (CNN),
respectively. Both models perform well on a real-world business analytics
dataset and significantly outperform the baseline bag-of-character (BoC) model.
| no_new_dataset | 0.952309 |
1702.02661 | U. N. Niranjan | U.N. Niranjan, Arun Rajkumar | Inductive Pairwise Ranking: Going Beyond the n log(n) Barrier | null | null | null | null | cs.LG cs.IT math.IT stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study the problem of ranking a set of items from nonactively chosen
pairwise preferences where each item has feature information with it. We
propose and characterize a very broad class of preference matrices giving rise
to the Feature Low Rank (FLR) model, which subsumes several models ranging from
the classic Bradley-Terry-Luce (BTL) (Bradley and Terry 1952) and Thurstone
(Thurstone 1927) models to the recently proposed blade-chest (Chen and Joachims
2016) and generic low-rank preference (Rajkumar and Agarwal 2016) models. We
use the technique of matrix completion in the presence of side information to
develop the Inductive Pairwise Ranking (IPR) algorithm that provably learns a
good ranking under the FLR model, in a sample-efficient manner. In practice,
through systematic synthetic simulations, we confirm our theoretical findings
regarding improvements in the sample complexity due to the use of feature
information. Moreover, on popular real-world preference learning datasets, with
as less as 10% sampling of the pairwise comparisons, our method recovers a good
ranking.
| [
{
"version": "v1",
"created": "Thu, 9 Feb 2017 00:17:39 GMT"
}
] | 2017-02-10T00:00:00 | [
[
"Niranjan",
"U. N.",
""
],
[
"Rajkumar",
"Arun",
""
]
] | TITLE: Inductive Pairwise Ranking: Going Beyond the n log(n) Barrier
ABSTRACT: We study the problem of ranking a set of items from nonactively chosen
pairwise preferences where each item has feature information with it. We
propose and characterize a very broad class of preference matrices giving rise
to the Feature Low Rank (FLR) model, which subsumes several models ranging from
the classic Bradley-Terry-Luce (BTL) (Bradley and Terry 1952) and Thurstone
(Thurstone 1927) models to the recently proposed blade-chest (Chen and Joachims
2016) and generic low-rank preference (Rajkumar and Agarwal 2016) models. We
use the technique of matrix completion in the presence of side information to
develop the Inductive Pairwise Ranking (IPR) algorithm that provably learns a
good ranking under the FLR model, in a sample-efficient manner. In practice,
through systematic synthetic simulations, we confirm our theoretical findings
regarding improvements in the sample complexity due to the use of feature
information. Moreover, on popular real-world preference learning datasets, with
as less as 10% sampling of the pairwise comparisons, our method recovers a good
ranking.
| no_new_dataset | 0.94699 |
1702.02676 | Arman Afrasiyabi | Arman Afrasiyabi, Ozan Yildiz, Baris Nasir, Fatos T. Yarman Vural and
A. Enis Cetin | Energy Saving Additive Neural Network | 8 pages (double column), 2 figures, 1 table | null | null | null | cs.NE cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | In recent years, machine learning techniques based on neural networks for
mobile computing become increasingly popular. Classical multi-layer neural
networks require matrix multiplications at each stage. Multiplication operation
is not an energy efficient operation and consequently it drains the battery of
the mobile device. In this paper, we propose a new energy efficient neural
network with the universal approximation property over space of Lebesgue
integrable functions. This network, called, additive neural network, is very
suitable for mobile computing. The neural structure is based on a novel vector
product definition, called ef-operator, that permits a multiplier-free
implementation. In ef-operation, the "product" of two real numbers is defined
as the sum of their absolute values, with the sign determined by the sign of
the product of the numbers. This "product" is used to construct a vector
product in $R^N$. The vector product induces the $l_1$ norm. The proposed
additive neural network successfully solves the XOR problem. The experiments on
MNIST dataset show that the classification performances of the proposed
additive neural networks are very similar to the corresponding multi-layer
perceptron and convolutional neural networks (LeNet).
| [
{
"version": "v1",
"created": "Thu, 9 Feb 2017 02:02:27 GMT"
}
] | 2017-02-10T00:00:00 | [
[
"Afrasiyabi",
"Arman",
""
],
[
"Yildiz",
"Ozan",
""
],
[
"Nasir",
"Baris",
""
],
[
"Vural",
"Fatos T. Yarman",
""
],
[
"Cetin",
"A. Enis",
""
]
] | TITLE: Energy Saving Additive Neural Network
ABSTRACT: In recent years, machine learning techniques based on neural networks for
mobile computing become increasingly popular. Classical multi-layer neural
networks require matrix multiplications at each stage. Multiplication operation
is not an energy efficient operation and consequently it drains the battery of
the mobile device. In this paper, we propose a new energy efficient neural
network with the universal approximation property over space of Lebesgue
integrable functions. This network, called, additive neural network, is very
suitable for mobile computing. The neural structure is based on a novel vector
product definition, called ef-operator, that permits a multiplier-free
implementation. In ef-operation, the "product" of two real numbers is defined
as the sum of their absolute values, with the sign determined by the sign of
the product of the numbers. This "product" is used to construct a vector
product in $R^N$. The vector product induces the $l_1$ norm. The proposed
additive neural network successfully solves the XOR problem. The experiments on
MNIST dataset show that the classification performances of the proposed
additive neural networks are very similar to the corresponding multi-layer
perceptron and convolutional neural networks (LeNet).
| no_new_dataset | 0.949576 |
1702.02743 | Juan Felipe Perez-Juste Abascal | Juan F P J Abascal (CREATIS), Manuel Desco, Juan Parra-Robles | Incorporation of prior knowledge of the signal behavior into the
reconstruction to accelerate the acquisition of MR diffusion data | null | null | null | null | physics.med-ph cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Diffusion MRI measurements using hyperpolarized gases are generally acquired
during patient breath hold, which yields a compromise between achievable image
resolution, lung coverage and number of b-values. In this work, we propose a
novel method that accelerates the acquisition of MR diffusion data by
undersampling in both spatial and b-value dimensions, thanks to incorporating
knowledge about the signal decay into the reconstruction (SIDER). SIDER is
compared to total variation (TV) reconstruction by assessing their effect on
both the recovery of ventilation images and estimated mean alveolar dimensions
(MAD). Both methods are assessed by retrospectively undersampling diffusion
datasets of normal volunteers and COPD patients (n=8) for acceleration factors
between x2 and x10. TV led to large errors and artefacts for acceleration
factors equal or larger than x5. SIDER improved TV, presenting lower errors and
histograms of MAD closer to those obtained from fully sampled data for
accelerations factors up to x10. SIDER preserved image quality at all
acceleration factors but images were slightly smoothed and some details were
lost at x10. In conclusion, we have developed and validated a novel compressed
sensing method for lung MRI imaging and achieved high acceleration factors,
which can be used to increase the amount of data acquired during a breath-hold.
This methodology is expected to improve the accuracy of estimated lung
microstructure dimensions and widen the possibilities of studying lung diseases
with MRI.
| [
{
"version": "v1",
"created": "Thu, 9 Feb 2017 08:26:53 GMT"
}
] | 2017-02-10T00:00:00 | [
[
"Abascal",
"Juan F P J",
"",
"CREATIS"
],
[
"Desco",
"Manuel",
""
],
[
"Parra-Robles",
"Juan",
""
]
] | TITLE: Incorporation of prior knowledge of the signal behavior into the
reconstruction to accelerate the acquisition of MR diffusion data
ABSTRACT: Diffusion MRI measurements using hyperpolarized gases are generally acquired
during patient breath hold, which yields a compromise between achievable image
resolution, lung coverage and number of b-values. In this work, we propose a
novel method that accelerates the acquisition of MR diffusion data by
undersampling in both spatial and b-value dimensions, thanks to incorporating
knowledge about the signal decay into the reconstruction (SIDER). SIDER is
compared to total variation (TV) reconstruction by assessing their effect on
both the recovery of ventilation images and estimated mean alveolar dimensions
(MAD). Both methods are assessed by retrospectively undersampling diffusion
datasets of normal volunteers and COPD patients (n=8) for acceleration factors
between x2 and x10. TV led to large errors and artefacts for acceleration
factors equal or larger than x5. SIDER improved TV, presenting lower errors and
histograms of MAD closer to those obtained from fully sampled data for
accelerations factors up to x10. SIDER preserved image quality at all
acceleration factors but images were slightly smoothed and some details were
lost at x10. In conclusion, we have developed and validated a novel compressed
sensing method for lung MRI imaging and achieved high acceleration factors,
which can be used to increase the amount of data acquired during a breath-hold.
This methodology is expected to improve the accuracy of estimated lung
microstructure dimensions and widen the possibilities of studying lung diseases
with MRI.
| no_new_dataset | 0.95452 |
1702.02805 | Qi Guo | Qi Guo, Ce Zhu, Zhiqiang Xia, Zhengtao Wang, Yipeng Liu | Attribute-controlled face photo synthesis from simple line drawing | 5 pages, 5 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Face photo synthesis from simple line drawing is a one-to-many task as simple
line drawing merely contains the contour of human face. Previous exemplar-based
methods are over-dependent on the datasets and are hard to generalize to
complicated natural scenes. Recently, several works utilize deep neural
networks to increase the generalization, but they are still limited in the
controllability of the users. In this paper, we propose a deep generative model
to synthesize face photo from simple line drawing controlled by face attributes
such as hair color and complexion. In order to maximize the controllability of
face attributes, an attribute-disentangled variational auto-encoder (AD-VAE) is
firstly introduced to learn latent representations disentangled with respect to
specified attributes. Then we conduct photo synthesis from simple line drawing
based on AD-VAE. Experiments show that our model can well disentangle the
variations of attributes from other variations of face photos and synthesize
detailed photorealistic face images with desired attributes. Regarding
background and illumination as the style and human face as the content, we can
also synthesize face photos with the target style of a style photo.
| [
{
"version": "v1",
"created": "Thu, 9 Feb 2017 12:21:36 GMT"
}
] | 2017-02-10T00:00:00 | [
[
"Guo",
"Qi",
""
],
[
"Zhu",
"Ce",
""
],
[
"Xia",
"Zhiqiang",
""
],
[
"Wang",
"Zhengtao",
""
],
[
"Liu",
"Yipeng",
""
]
] | TITLE: Attribute-controlled face photo synthesis from simple line drawing
ABSTRACT: Face photo synthesis from simple line drawing is a one-to-many task as simple
line drawing merely contains the contour of human face. Previous exemplar-based
methods are over-dependent on the datasets and are hard to generalize to
complicated natural scenes. Recently, several works utilize deep neural
networks to increase the generalization, but they are still limited in the
controllability of the users. In this paper, we propose a deep generative model
to synthesize face photo from simple line drawing controlled by face attributes
such as hair color and complexion. In order to maximize the controllability of
face attributes, an attribute-disentangled variational auto-encoder (AD-VAE) is
firstly introduced to learn latent representations disentangled with respect to
specified attributes. Then we conduct photo synthesis from simple line drawing
based on AD-VAE. Experiments show that our model can well disentangle the
variations of attributes from other variations of face photos and synthesize
detailed photorealistic face images with desired attributes. Regarding
background and illumination as the style and human face as the content, we can
also synthesize face photos with the target style of a style photo.
| no_new_dataset | 0.946151 |
1702.02925 | Wei Li | Wei Li, Farnaz Abtahi, Zhigang Zhu, Lijun Yin | EAC-Net: A Region-based Deep Enhancing and Cropping Approach for Facial
Action Unit Detection | The paper is accepted by FG 2017 | null | null | null | cs.CV | http://creativecommons.org/publicdomain/zero/1.0/ | In this paper, we propose a deep learning based approach for facial action
unit detection by enhancing and cropping the regions of interest. The approach
is implemented by adding two novel nets (layers): the enhancing layers and the
cropping layers, to a pretrained CNN model. For the enhancing layers, we
designed an attention map based on facial landmark features and applied it to a
pretrained neural network to conduct enhanced learning (The E-Net). For the
cropping layers, we crop facial regions around the detected landmarks and
design convolutional layers to learn deeper features for each facial region
(C-Net). We then fuse the E-Net and the C-Net to obtain our Enhancing and
Cropping (EAC) Net, which can learn both feature enhancing and region cropping
functions. Our approach shows significant improvement in performance compared
to the state-of-the-art methods applied to BP4D and DISFA AU datasets.
| [
{
"version": "v1",
"created": "Thu, 9 Feb 2017 18:16:44 GMT"
}
] | 2017-02-10T00:00:00 | [
[
"Li",
"Wei",
""
],
[
"Abtahi",
"Farnaz",
""
],
[
"Zhu",
"Zhigang",
""
],
[
"Yin",
"Lijun",
""
]
] | TITLE: EAC-Net: A Region-based Deep Enhancing and Cropping Approach for Facial
Action Unit Detection
ABSTRACT: In this paper, we propose a deep learning based approach for facial action
unit detection by enhancing and cropping the regions of interest. The approach
is implemented by adding two novel nets (layers): the enhancing layers and the
cropping layers, to a pretrained CNN model. For the enhancing layers, we
designed an attention map based on facial landmark features and applied it to a
pretrained neural network to conduct enhanced learning (The E-Net). For the
cropping layers, we crop facial regions around the detected landmarks and
design convolutional layers to learn deeper features for each facial region
(C-Net). We then fuse the E-Net and the C-Net to obtain our Enhancing and
Cropping (EAC) Net, which can learn both feature enhancing and region cropping
functions. Our approach shows significant improvement in performance compared
to the state-of-the-art methods applied to BP4D and DISFA AU datasets.
| no_new_dataset | 0.951142 |
1503.00173 | Jonathan Mei | Jonathan Mei and Jos\'e M. F. Moura | Signal Processing on Graphs: Causal Modeling of Unstructured Data | null | IEEE Transactions on Signal Processing, vol. 65, no. 8, pp.
2077-2092, April 15, 2017 | 10.1109/TSP.2016.2634543 | null | cs.IT math.IT stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Many applications collect a large number of time series, for example, the
financial data of companies quoted in a stock exchange, the health care data of
all patients that visit the emergency room of a hospital, or the temperature
sequences continuously measured by weather stations across the US. These data
are often referred to as unstructured. A first task in its analytics is to
derive a low dimensional representation, a graph or discrete manifold, that
describes well the interrelations among the time series and their
intrarelations across time. This paper presents a computationally tractable
algorithm for estimating this graph that structures the data. The resulting
graph is directed and weighted, possibly capturing causal relations, not just
reciprocal correlations as in many existing approaches in the literature. A
convergence analysis is carried out. The algorithm is demonstrated on random
graph datasets and real network time series datasets, and its performance is
compared to that of related methods. The adjacency matrices estimated with the
new method are close to the true graph in the simulated data and consistent
with prior physical knowledge in the real dataset tested.
| [
{
"version": "v1",
"created": "Sat, 28 Feb 2015 20:28:05 GMT"
},
{
"version": "v2",
"created": "Thu, 14 Apr 2016 20:58:45 GMT"
},
{
"version": "v3",
"created": "Tue, 13 Sep 2016 13:19:02 GMT"
},
{
"version": "v4",
"created": "Mon, 31 Oct 2016 22:05:33 GMT"
},
{
"version": "v5",
"created": "Wed, 30 Nov 2016 19:12:41 GMT"
},
{
"version": "v6",
"created": "Wed, 8 Feb 2017 15:49:58 GMT"
}
] | 2017-02-09T00:00:00 | [
[
"Mei",
"Jonathan",
""
],
[
"Moura",
"José M. F.",
""
]
] | TITLE: Signal Processing on Graphs: Causal Modeling of Unstructured Data
ABSTRACT: Many applications collect a large number of time series, for example, the
financial data of companies quoted in a stock exchange, the health care data of
all patients that visit the emergency room of a hospital, or the temperature
sequences continuously measured by weather stations across the US. These data
are often referred to as unstructured. A first task in its analytics is to
derive a low dimensional representation, a graph or discrete manifold, that
describes well the interrelations among the time series and their
intrarelations across time. This paper presents a computationally tractable
algorithm for estimating this graph that structures the data. The resulting
graph is directed and weighted, possibly capturing causal relations, not just
reciprocal correlations as in many existing approaches in the literature. A
convergence analysis is carried out. The algorithm is demonstrated on random
graph datasets and real network time series datasets, and its performance is
compared to that of related methods. The adjacency matrices estimated with the
new method are close to the true graph in the simulated data and consistent
with prior physical knowledge in the real dataset tested.
| no_new_dataset | 0.946843 |
1611.01839 | Eunsol Choi | Eunsol Choi, Daniel Hewlett, Alexandre Lacoste, Illia Polosukhin,
Jakob Uszkoreit, Jonathan Berant | Hierarchical Question Answering for Long Documents | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a framework for question answering that can efficiently scale to
longer documents while maintaining or even improving performance of
state-of-the-art models. While most successful approaches for reading
comprehension rely on recurrent neural networks (RNNs), running them over long
documents is prohibitively slow because it is difficult to parallelize over
sequences. Inspired by how people first skim the document, identify relevant
parts, and carefully read these parts to produce an answer, we combine a
coarse, fast model for selecting relevant sentences and a more expensive RNN
for producing the answer from those sentences. We treat sentence selection as a
latent variable trained jointly from the answer only using reinforcement
learning. Experiments demonstrate the state of the art performance on a
challenging subset of the Wikireading and on a new dataset, while speeding up
the model by 3.5x-6.7x.
| [
{
"version": "v1",
"created": "Sun, 6 Nov 2016 20:24:40 GMT"
},
{
"version": "v2",
"created": "Wed, 8 Feb 2017 07:42:34 GMT"
}
] | 2017-02-09T00:00:00 | [
[
"Choi",
"Eunsol",
""
],
[
"Hewlett",
"Daniel",
""
],
[
"Lacoste",
"Alexandre",
""
],
[
"Polosukhin",
"Illia",
""
],
[
"Uszkoreit",
"Jakob",
""
],
[
"Berant",
"Jonathan",
""
]
] | TITLE: Hierarchical Question Answering for Long Documents
ABSTRACT: We present a framework for question answering that can efficiently scale to
longer documents while maintaining or even improving performance of
state-of-the-art models. While most successful approaches for reading
comprehension rely on recurrent neural networks (RNNs), running them over long
documents is prohibitively slow because it is difficult to parallelize over
sequences. Inspired by how people first skim the document, identify relevant
parts, and carefully read these parts to produce an answer, we combine a
coarse, fast model for selecting relevant sentences and a more expensive RNN
for producing the answer from those sentences. We treat sentence selection as a
latent variable trained jointly from the answer only using reinforcement
learning. Experiments demonstrate the state of the art performance on a
challenging subset of the Wikireading and on a new dataset, while speeding up
the model by 3.5x-6.7x.
| new_dataset | 0.961714 |
1612.00729 | Sowmya Vajjala | Sowmya Vajjala | Automated assessment of non-native learner essays: Investigating the
role of linguistic features | Article accepted for publication at: International Journal of
Artificial Intelligence in Education (IJAIED). To appear in early 2017
(journal url: http://www.springer.com/computer/ai/journal/40593) | null | 10.1007/s40593-017-0142-3 | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Automatic essay scoring (AES) refers to the process of scoring free text
responses to given prompts, considering human grader scores as the gold
standard. Writing such essays is an essential component of many language and
aptitude exams. Hence, AES became an active and established area of research,
and there are many proprietary systems used in real life applications today.
However, not much is known about which specific linguistic features are useful
for prediction and how much of this is consistent across datasets. This article
addresses that by exploring the role of various linguistic features in
automatic essay scoring using two publicly available datasets of non-native
English essays written in test taking scenarios. The linguistic properties are
modeled by encoding lexical, syntactic, discourse and error types of learner
language in the feature set. Predictive models are then developed using these
features on both datasets and the most predictive features are compared. While
the results show that the feature set used results in good predictive models
with both datasets, the question "what are the most predictive features?" has a
different answer for each dataset.
| [
{
"version": "v1",
"created": "Fri, 2 Dec 2016 16:22:49 GMT"
}
] | 2017-02-09T00:00:00 | [
[
"Vajjala",
"Sowmya",
""
]
] | TITLE: Automated assessment of non-native learner essays: Investigating the
role of linguistic features
ABSTRACT: Automatic essay scoring (AES) refers to the process of scoring free text
responses to given prompts, considering human grader scores as the gold
standard. Writing such essays is an essential component of many language and
aptitude exams. Hence, AES became an active and established area of research,
and there are many proprietary systems used in real life applications today.
However, not much is known about which specific linguistic features are useful
for prediction and how much of this is consistent across datasets. This article
addresses that by exploring the role of various linguistic features in
automatic essay scoring using two publicly available datasets of non-native
English essays written in test taking scenarios. The linguistic properties are
modeled by encoding lexical, syntactic, discourse and error types of learner
language in the feature set. Predictive models are then developed using these
features on both datasets and the most predictive features are compared. While
the results show that the feature set used results in good predictive models
with both datasets, the question "what are the most predictive features?" has a
different answer for each dataset.
| no_new_dataset | 0.939248 |
1702.02367 | Claudio Greco | Claudio Greco, Alessandro Suglia, Pierpaolo Basile, Gaetano Rossiello,
Giovanni Semeraro | Iterative Multi-document Neural Attention for Multiple Answer Prediction | Paper accepted and presented at the Deep Understanding and Reasoning:
A challenge for Next-generation Intelligent Agents (URANIA) workshop, held in
the context of the AI*IA 2016 conference | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | People have information needs of varying complexity, which can be solved by
an intelligent agent able to answer questions formulated in a proper way,
eventually considering user context and preferences. In a scenario in which the
user profile can be considered as a question, intelligent agents able to answer
questions can be used to find the most relevant answers for a given user. In
this work we propose a novel model based on Artificial Neural Networks to
answer questions with multiple answers by exploiting multiple facts retrieved
from a knowledge base. The model is evaluated on the factoid Question Answering
and top-n recommendation tasks of the bAbI Movie Dialog dataset. After
assessing the performance of the model on both tasks, we try to define the
long-term goal of a conversational recommender system able to interact using
natural language and to support users in their information seeking processes in
a personalized way.
| [
{
"version": "v1",
"created": "Wed, 8 Feb 2017 10:58:02 GMT"
}
] | 2017-02-09T00:00:00 | [
[
"Greco",
"Claudio",
""
],
[
"Suglia",
"Alessandro",
""
],
[
"Basile",
"Pierpaolo",
""
],
[
"Rossiello",
"Gaetano",
""
],
[
"Semeraro",
"Giovanni",
""
]
] | TITLE: Iterative Multi-document Neural Attention for Multiple Answer Prediction
ABSTRACT: People have information needs of varying complexity, which can be solved by
an intelligent agent able to answer questions formulated in a proper way,
eventually considering user context and preferences. In a scenario in which the
user profile can be considered as a question, intelligent agents able to answer
questions can be used to find the most relevant answers for a given user. In
this work we propose a novel model based on Artificial Neural Networks to
answer questions with multiple answers by exploiting multiple facts retrieved
from a knowledge base. The model is evaluated on the factoid Question Answering
and top-n recommendation tasks of the bAbI Movie Dialog dataset. After
assessing the performance of the model on both tasks, we try to define the
long-term goal of a conversational recommender system able to interact using
natural language and to support users in their information seeking processes in
a personalized way.
| no_new_dataset | 0.946597 |
1702.02508 | Corneliu Arsene Dr | Corneliu Arsene, Peter Pormann, William Sellers, Siam Bhayro | Computational Techniques in Multispectral Image Processing: Application
to the Syriac Galen Palimpsest | 29 February - 2 March 2016, Second International Conference on
Natural Sciences and Technology in Manuscript Analysis, Centre for the study
of Manuscript Cultures, Hamburg, Germany | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multispectral and hyperspectral image analysis has experienced much
development in the last decade. The application of these methods to palimpsests
has produced significant results, enabling researchers to recover texts that
would be otherwise lost under the visible overtext, by improving the contrast
between the undertext and the overtext. In this paper we explore an extended
number of multispectral and hyperspectral image analysis methods, consisting of
supervised and unsupervised dimensionality reduction techniques, on a part of
the Syriac Galen Palimpsest dataset (www.digitalgalen.net). Of this extended
set of methods, eight methods gave good results: three were supervised methods
Generalized Discriminant Analysis (GDA), Linear Discriminant Analysis (LDA),
and Neighborhood Component Analysis (NCA); and the other five methods were
unsupervised methods (but still used in a supervised way) Gaussian Process
Latent Variable Model (GPLVM), Isomap, Landmark Isomap, Principal Component
Analysis (PCA), and Probabilistic Principal Component Analysis (PPCA). The
relative success of these methods was determined visually, using color
pictures, on the basis of whether the undertext was distinguishable from the
overtext, resulting in the following ranking of the methods: LDA, NCA, GDA,
Isomap, Landmark Isomap, PPCA, PCA, and GPLVM. These results were compared with
those obtained using the Canonical Variates Analysis (CVA) method on the same
dataset, which showed remarkably accuracy (LDA is a particular case of CVA
where the objects are classified to two classes).
| [
{
"version": "v1",
"created": "Tue, 31 Jan 2017 13:03:20 GMT"
}
] | 2017-02-09T00:00:00 | [
[
"Arsene",
"Corneliu",
""
],
[
"Pormann",
"Peter",
""
],
[
"Sellers",
"William",
""
],
[
"Bhayro",
"Siam",
""
]
] | TITLE: Computational Techniques in Multispectral Image Processing: Application
to the Syriac Galen Palimpsest
ABSTRACT: Multispectral and hyperspectral image analysis has experienced much
development in the last decade. The application of these methods to palimpsests
has produced significant results, enabling researchers to recover texts that
would be otherwise lost under the visible overtext, by improving the contrast
between the undertext and the overtext. In this paper we explore an extended
number of multispectral and hyperspectral image analysis methods, consisting of
supervised and unsupervised dimensionality reduction techniques, on a part of
the Syriac Galen Palimpsest dataset (www.digitalgalen.net). Of this extended
set of methods, eight methods gave good results: three were supervised methods
Generalized Discriminant Analysis (GDA), Linear Discriminant Analysis (LDA),
and Neighborhood Component Analysis (NCA); and the other five methods were
unsupervised methods (but still used in a supervised way) Gaussian Process
Latent Variable Model (GPLVM), Isomap, Landmark Isomap, Principal Component
Analysis (PCA), and Probabilistic Principal Component Analysis (PPCA). The
relative success of these methods was determined visually, using color
pictures, on the basis of whether the undertext was distinguishable from the
overtext, resulting in the following ranking of the methods: LDA, NCA, GDA,
Isomap, Landmark Isomap, PPCA, PCA, and GPLVM. These results were compared with
those obtained using the Canonical Variates Analysis (CVA) method on the same
dataset, which showed remarkably accuracy (LDA is a particular case of CVA
where the objects are classified to two classes).
| no_new_dataset | 0.9462 |
1702.02512 | Yi Zhou | Yi Zhou, Laurent Kneip and Hongdong Li | Semi-Dense Visual Odometry for RGB-D Cameras Using Approximate Nearest
Neighbour Fields | ICRA 2017 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents a robust and efficient semi-dense visual odometry
solution for RGB-D cameras. The core of our method is a 2D-3D ICP pipeline
which estimates the pose of the sensor by registering the projection of a 3D
semi-dense map of the reference frame with the 2D semi-dense region extracted
in the current frame. The processing is speeded up by efficiently implemented
approximate nearest neighbour fields under the Euclidean distance criterion,
which permits the use of compact Gauss-Newton updates in the optimization. The
registration is formulated as a maximum a posterior problem to deal with
outliers and sensor noises, and consequently the equivalent weighted least
squares problem is solved by iteratively reweighted least squares method. A
variety of robust weight functions are tested and the optimum is determined
based on the characteristics of the sensor model. Extensive evaluation on
publicly available RGB-D datasets shows that the proposed method predominantly
outperforms existing state-of-the-art methods.
| [
{
"version": "v1",
"created": "Mon, 6 Feb 2017 00:12:37 GMT"
}
] | 2017-02-09T00:00:00 | [
[
"Zhou",
"Yi",
""
],
[
"Kneip",
"Laurent",
""
],
[
"Li",
"Hongdong",
""
]
] | TITLE: Semi-Dense Visual Odometry for RGB-D Cameras Using Approximate Nearest
Neighbour Fields
ABSTRACT: This paper presents a robust and efficient semi-dense visual odometry
solution for RGB-D cameras. The core of our method is a 2D-3D ICP pipeline
which estimates the pose of the sensor by registering the projection of a 3D
semi-dense map of the reference frame with the 2D semi-dense region extracted
in the current frame. The processing is speeded up by efficiently implemented
approximate nearest neighbour fields under the Euclidean distance criterion,
which permits the use of compact Gauss-Newton updates in the optimization. The
registration is formulated as a maximum a posterior problem to deal with
outliers and sensor noises, and consequently the equivalent weighted least
squares problem is solved by iteratively reweighted least squares method. A
variety of robust weight functions are tested and the optimum is determined
based on the characteristics of the sensor model. Extensive evaluation on
publicly available RGB-D datasets shows that the proposed method predominantly
outperforms existing state-of-the-art methods.
| no_new_dataset | 0.944228 |
1702.02537 | Olasimbo Ayodeji Arigbabu | Olasimbo Ayodeji Arigbabu, Sharifah Mumtazah Syed Ahmad, Wan Azizun
Wan Adnan, Salman Yussof, Saif Mahmood | Soft Biometrics: Gender Recognition from Unconstrained Face Images using
Local Feature Descriptor | null | Journal of Information and Communication Technology (JICT), 2015 | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Gender recognition from unconstrained face images is a challenging task due
to the high degree of misalignment, pose, expression, and illumination
variation. In previous works, the recognition of gender from unconstrained face
images is approached by utilizing image alignment, exploiting multiple samples
per individual to improve the learning ability of the classifier, or learning
gender based on prior knowledge about pose and demographic distributions of the
dataset. However, image alignment increases the complexity and time of
computation, while the use of multiple samples or having prior knowledge about
data distribution is unrealistic in practical applications. This paper presents
an approach for gender recognition from unconstrained face images. Our
technique exploits the robustness of local feature descriptor to photometric
variations to extract the shape description of the 2D face image using a single
sample image per individual. The results obtained from experiments on Labeled
Faces in the Wild (LFW) dataset describe the effectiveness of the proposed
method. The essence of this study is to investigate the most suitable functions
and parameter settings for recognizing gender from unconstrained face images.
| [
{
"version": "v1",
"created": "Wed, 8 Feb 2017 17:34:53 GMT"
}
] | 2017-02-09T00:00:00 | [
[
"Arigbabu",
"Olasimbo Ayodeji",
""
],
[
"Ahmad",
"Sharifah Mumtazah Syed",
""
],
[
"Adnan",
"Wan Azizun Wan",
""
],
[
"Yussof",
"Salman",
""
],
[
"Mahmood",
"Saif",
""
]
] | TITLE: Soft Biometrics: Gender Recognition from Unconstrained Face Images using
Local Feature Descriptor
ABSTRACT: Gender recognition from unconstrained face images is a challenging task due
to the high degree of misalignment, pose, expression, and illumination
variation. In previous works, the recognition of gender from unconstrained face
images is approached by utilizing image alignment, exploiting multiple samples
per individual to improve the learning ability of the classifier, or learning
gender based on prior knowledge about pose and demographic distributions of the
dataset. However, image alignment increases the complexity and time of
computation, while the use of multiple samples or having prior knowledge about
data distribution is unrealistic in practical applications. This paper presents
an approach for gender recognition from unconstrained face images. Our
technique exploits the robustness of local feature descriptor to photometric
variations to extract the shape description of the 2D face image using a single
sample image per individual. The results obtained from experiments on Labeled
Faces in the Wild (LFW) dataset describe the effectiveness of the proposed
method. The essence of this study is to investigate the most suitable functions
and parameter settings for recognizing gender from unconstrained face images.
| no_new_dataset | 0.95469 |
1405.6623 | Michael May | Andrew F. Magee and Michael R. May and Brian R. Moore | The Dawn of Open Access to Phylogenetic Data | null | null | 10.1371/journal.pone.0110268 | null | q-bio.PE cs.DL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The scientific enterprise depends critically on the preservation of and open
access to published data. This basic tenet applies acutely to phylogenies
(estimates of evolutionary relationships among species). Increasingly,
phylogenies are estimated from increasingly large, genome-scale datasets using
increasingly complex statistical methods that require increasing levels of
expertise and computational investment. Moreover, the resulting phylogenetic
data provide an explicit historical perspective that critically informs
research in a vast and growing number of scientific disciplines. One such use
is the study of changes in rates of lineage diversification (speciation -
extinction) through time. As part of a meta-analysis in this area, we sought to
collect phylogenetic data (comprising nucleotide sequence alignment and tree
files) from 217 studies published in 46 journals over a 13-year period. We
document our attempts to procure those data (from online archives and by direct
request to corresponding authors), and report results of analyses (using
Bayesian logistic regression) to assess the impact of various factors on the
success of our efforts. Overall, complete phylogenetic data for ~60% of these
studies are effectively lost to science. Our study indicates that phylogenetic
data are more likely to be deposited in online archives and/or shared upon
request when: (1) the publishing journal has a strong data-sharing policy; (2)
the publishing journal has a higher impact factor, and; (3) the data are
requested from faculty rather than students. Although the situation appears
dire, our analyses suggest that it is far from hopeless: recent initiatives by
the scientific community -- including policy changes by journals and funding
agencies -- are improving the state of affairs.
| [
{
"version": "v1",
"created": "Fri, 23 May 2014 00:20:42 GMT"
}
] | 2017-02-08T00:00:00 | [
[
"Magee",
"Andrew F.",
""
],
[
"May",
"Michael R.",
""
],
[
"Moore",
"Brian R.",
""
]
] | TITLE: The Dawn of Open Access to Phylogenetic Data
ABSTRACT: The scientific enterprise depends critically on the preservation of and open
access to published data. This basic tenet applies acutely to phylogenies
(estimates of evolutionary relationships among species). Increasingly,
phylogenies are estimated from increasingly large, genome-scale datasets using
increasingly complex statistical methods that require increasing levels of
expertise and computational investment. Moreover, the resulting phylogenetic
data provide an explicit historical perspective that critically informs
research in a vast and growing number of scientific disciplines. One such use
is the study of changes in rates of lineage diversification (speciation -
extinction) through time. As part of a meta-analysis in this area, we sought to
collect phylogenetic data (comprising nucleotide sequence alignment and tree
files) from 217 studies published in 46 journals over a 13-year period. We
document our attempts to procure those data (from online archives and by direct
request to corresponding authors), and report results of analyses (using
Bayesian logistic regression) to assess the impact of various factors on the
success of our efforts. Overall, complete phylogenetic data for ~60% of these
studies are effectively lost to science. Our study indicates that phylogenetic
data are more likely to be deposited in online archives and/or shared upon
request when: (1) the publishing journal has a strong data-sharing policy; (2)
the publishing journal has a higher impact factor, and; (3) the data are
requested from faculty rather than students. Although the situation appears
dire, our analyses suggest that it is far from hopeless: recent initiatives by
the scientific community -- including policy changes by journals and funding
agencies -- are improving the state of affairs.
| no_new_dataset | 0.947088 |
1506.04422 | Rafael Pinto | Rafael Pinto and Paulo Engel | A Fast Incremental Gaussian Mixture Model | 10 pages, no figures, draft submission to Plos One | null | 10.1371/journal.pone.0139931 | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This work builds upon previous efforts in online incremental learning, namely
the Incremental Gaussian Mixture Network (IGMN). The IGMN is capable of
learning from data streams in a single-pass by improving its model after
analyzing each data point and discarding it thereafter. Nevertheless, it
suffers from the scalability point-of-view, due to its asymptotic time
complexity of $\operatorname{O}\bigl(NKD^3\bigr)$ for $N$ data points, $K$
Gaussian components and $D$ dimensions, rendering it inadequate for
high-dimensional data. In this paper, we manage to reduce this complexity to
$\operatorname{O}\bigl(NKD^2\bigr)$ by deriving formulas for working directly
with precision matrices instead of covariance matrices. The final result is a
much faster and scalable algorithm which can be applied to high dimensional
tasks. This is confirmed by applying the modified algorithm to high-dimensional
classification datasets.
| [
{
"version": "v1",
"created": "Sun, 14 Jun 2015 17:02:49 GMT"
},
{
"version": "v2",
"created": "Thu, 18 Jun 2015 17:04:01 GMT"
}
] | 2017-02-08T00:00:00 | [
[
"Pinto",
"Rafael",
""
],
[
"Engel",
"Paulo",
""
]
] | TITLE: A Fast Incremental Gaussian Mixture Model
ABSTRACT: This work builds upon previous efforts in online incremental learning, namely
the Incremental Gaussian Mixture Network (IGMN). The IGMN is capable of
learning from data streams in a single-pass by improving its model after
analyzing each data point and discarding it thereafter. Nevertheless, it
suffers from the scalability point-of-view, due to its asymptotic time
complexity of $\operatorname{O}\bigl(NKD^3\bigr)$ for $N$ data points, $K$
Gaussian components and $D$ dimensions, rendering it inadequate for
high-dimensional data. In this paper, we manage to reduce this complexity to
$\operatorname{O}\bigl(NKD^2\bigr)$ by deriving formulas for working directly
with precision matrices instead of covariance matrices. The final result is a
much faster and scalable algorithm which can be applied to high dimensional
tasks. This is confirmed by applying the modified algorithm to high-dimensional
classification datasets.
| no_new_dataset | 0.94366 |
1604.01277 | Ramon Fraga Pereira | Ramon Fraga Pereira and Felipe Meneguzzi | Landmark-Based Plan Recognition | Accepted as short paper in the 22nd European Conference on Artificial
Intelligence, ECAI 2016 | null | null | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recognition of goals and plans using incomplete evidence from action
execution can be done efficiently by using planning techniques. In many
applications it is important to recognize goals and plans not only accurately,
but also quickly. In this paper, we develop a heuristic approach for
recognizing plans based on planning techniques that rely on ordering
constraints to filter candidate goals from observations. These ordering
constraints are called landmarks in the planning literature, which are facts or
actions that cannot be avoided to achieve a goal. We show the applicability of
planning landmarks in two settings: first, we use it directly to develop a
heuristic-based plan recognition approach; second, we refine an existing
planning-based plan recognition approach by pre-filtering its candidate goals.
Our empirical evaluation shows that our approach is not only substantially more
accurate than the state-of-the-art in all available datasets, it is also an
order of magnitude faster.
| [
{
"version": "v1",
"created": "Tue, 5 Apr 2016 14:44:03 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Jun 2016 17:56:47 GMT"
},
{
"version": "v3",
"created": "Tue, 7 Feb 2017 01:15:59 GMT"
}
] | 2017-02-08T00:00:00 | [
[
"Pereira",
"Ramon Fraga",
""
],
[
"Meneguzzi",
"Felipe",
""
]
] | TITLE: Landmark-Based Plan Recognition
ABSTRACT: Recognition of goals and plans using incomplete evidence from action
execution can be done efficiently by using planning techniques. In many
applications it is important to recognize goals and plans not only accurately,
but also quickly. In this paper, we develop a heuristic approach for
recognizing plans based on planning techniques that rely on ordering
constraints to filter candidate goals from observations. These ordering
constraints are called landmarks in the planning literature, which are facts or
actions that cannot be avoided to achieve a goal. We show the applicability of
planning landmarks in two settings: first, we use it directly to develop a
heuristic-based plan recognition approach; second, we refine an existing
planning-based plan recognition approach by pre-filtering its candidate goals.
Our empirical evaluation shows that our approach is not only substantially more
accurate than the state-of-the-art in all available datasets, it is also an
order of magnitude faster.
| no_new_dataset | 0.954858 |
1608.06108 | Andrea Cuttone | Andrea Cuttone, Per B{\ae}kgaard, Vedran Sekara, H{\aa}kan Jonsson,
Jakob Eg Larsen, Sune Lehmann | SensibleSleep: A Bayesian Model for Learning Sleep Patterns from
Smartphone Events | null | null | 10.1371/journal.pone.0169901 | null | cs.CY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a Bayesian model for extracting sleep patterns from smartphone
events. Our method is able to identify individuals' daily sleep periods and
their evolution over time, and provides an estimation of the probability of
sleep and wake transitions. The model is fitted to more than 400 participants
from two different datasets, and we verify the results against ground truth
from dedicated armband sleep trackers. We show that the model is able to
produce reliable sleep estimates with an accuracy of 0.89, both at the
individual and at the collective level. Moreover the Bayesian model is able to
quantify uncertainty and encode prior knowledge about sleep patterns. Compared
with existing smartphone-based systems, our method requires only screen on/off
events, and is therefore much less intrusive in terms of privacy and more
battery-efficient.
| [
{
"version": "v1",
"created": "Mon, 22 Aug 2016 10:18:56 GMT"
}
] | 2017-02-08T00:00:00 | [
[
"Cuttone",
"Andrea",
""
],
[
"Bækgaard",
"Per",
""
],
[
"Sekara",
"Vedran",
""
],
[
"Jonsson",
"Håkan",
""
],
[
"Larsen",
"Jakob Eg",
""
],
[
"Lehmann",
"Sune",
""
]
] | TITLE: SensibleSleep: A Bayesian Model for Learning Sleep Patterns from
Smartphone Events
ABSTRACT: We propose a Bayesian model for extracting sleep patterns from smartphone
events. Our method is able to identify individuals' daily sleep periods and
their evolution over time, and provides an estimation of the probability of
sleep and wake transitions. The model is fitted to more than 400 participants
from two different datasets, and we verify the results against ground truth
from dedicated armband sleep trackers. We show that the model is able to
produce reliable sleep estimates with an accuracy of 0.89, both at the
individual and at the collective level. Moreover the Bayesian model is able to
quantify uncertainty and encode prior knowledge about sleep patterns. Compared
with existing smartphone-based systems, our method requires only screen on/off
events, and is therefore much less intrusive in terms of privacy and more
battery-efficient.
| no_new_dataset | 0.946597 |
1611.02770 | Xinyun Chen | Yanpei Liu, Xinyun Chen, Chang Liu, Dawn Song | Delving into Transferable Adversarial Examples and Black-box Attacks | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | An intriguing property of deep neural networks is the existence of
adversarial examples, which can transfer among different architectures. These
transferable adversarial examples may severely hinder deep neural network-based
applications. Previous works mostly study the transferability using small scale
datasets. In this work, we are the first to conduct an extensive study of the
transferability over large models and a large scale dataset, and we are also
the first to study the transferability of targeted adversarial examples with
their target labels. We study both non-targeted and targeted adversarial
examples, and show that while transferable non-targeted adversarial examples
are easy to find, targeted adversarial examples generated using existing
approaches almost never transfer with their target labels. Therefore, we
propose novel ensemble-based approaches to generating transferable adversarial
examples. Using such approaches, we observe a large proportion of targeted
adversarial examples that are able to transfer with their target labels for the
first time. We also present some geometric studies to help understanding the
transferable adversarial examples. Finally, we show that the adversarial
examples generated using ensemble-based approaches can successfully attack
Clarifai.com, which is a black-box image classification system.
| [
{
"version": "v1",
"created": "Tue, 8 Nov 2016 23:25:00 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Nov 2016 22:28:51 GMT"
},
{
"version": "v3",
"created": "Tue, 7 Feb 2017 14:24:44 GMT"
}
] | 2017-02-08T00:00:00 | [
[
"Liu",
"Yanpei",
""
],
[
"Chen",
"Xinyun",
""
],
[
"Liu",
"Chang",
""
],
[
"Song",
"Dawn",
""
]
] | TITLE: Delving into Transferable Adversarial Examples and Black-box Attacks
ABSTRACT: An intriguing property of deep neural networks is the existence of
adversarial examples, which can transfer among different architectures. These
transferable adversarial examples may severely hinder deep neural network-based
applications. Previous works mostly study the transferability using small scale
datasets. In this work, we are the first to conduct an extensive study of the
transferability over large models and a large scale dataset, and we are also
the first to study the transferability of targeted adversarial examples with
their target labels. We study both non-targeted and targeted adversarial
examples, and show that while transferable non-targeted adversarial examples
are easy to find, targeted adversarial examples generated using existing
approaches almost never transfer with their target labels. Therefore, we
propose novel ensemble-based approaches to generating transferable adversarial
examples. Using such approaches, we observe a large proportion of targeted
adversarial examples that are able to transfer with their target labels for the
first time. We also present some geometric studies to help understanding the
transferable adversarial examples. Finally, we show that the adversarial
examples generated using ensemble-based approaches can successfully attack
Clarifai.com, which is a black-box image classification system.
| no_new_dataset | 0.940298 |
1611.08481 | Harm de Vries | Harm de Vries, Florian Strub, Sarath Chandar, Olivier Pietquin, Hugo
Larochelle, Aaron Courville | GuessWhat?! Visual object discovery through multi-modal dialogue | 23 pages; CVPR 2017 submission; see https://guesswhat.ai | null | null | null | cs.AI cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We introduce GuessWhat?!, a two-player guessing game as a testbed for
research on the interplay of computer vision and dialogue systems. The goal of
the game is to locate an unknown object in a rich image scene by asking a
sequence of questions. Higher-level image understanding, like spatial reasoning
and language grounding, is required to solve the proposed task. Our key
contribution is the collection of a large-scale dataset consisting of 150K
human-played games with a total of 800K visual question-answer pairs on 66K
images. We explain our design decisions in collecting the dataset and introduce
the oracle and questioner tasks that are associated with the two players of the
game. We prototyped deep learning models to establish initial baselines of the
introduced tasks.
| [
{
"version": "v1",
"created": "Wed, 23 Nov 2016 20:56:13 GMT"
},
{
"version": "v2",
"created": "Mon, 6 Feb 2017 12:52:53 GMT"
}
] | 2017-02-08T00:00:00 | [
[
"de Vries",
"Harm",
""
],
[
"Strub",
"Florian",
""
],
[
"Chandar",
"Sarath",
""
],
[
"Pietquin",
"Olivier",
""
],
[
"Larochelle",
"Hugo",
""
],
[
"Courville",
"Aaron",
""
]
] | TITLE: GuessWhat?! Visual object discovery through multi-modal dialogue
ABSTRACT: We introduce GuessWhat?!, a two-player guessing game as a testbed for
research on the interplay of computer vision and dialogue systems. The goal of
the game is to locate an unknown object in a rich image scene by asking a
sequence of questions. Higher-level image understanding, like spatial reasoning
and language grounding, is required to solve the proposed task. Our key
contribution is the collection of a large-scale dataset consisting of 150K
human-played games with a total of 800K visual question-answer pairs on 66K
images. We explain our design decisions in collecting the dataset and introduce
the oracle and questioner tasks that are associated with the two players of the
game. We prototyped deep learning models to establish initial baselines of the
introduced tasks.
| new_dataset | 0.956634 |
1611.09830 | Tong Wang | Adam Trischler, Tong Wang, Xingdi Yuan, Justin Harris, Alessandro
Sordoni, Philip Bachman, Kaheer Suleman | NewsQA: A Machine Comprehension Dataset | null | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present NewsQA, a challenging machine comprehension dataset of over
100,000 human-generated question-answer pairs. Crowdworkers supply questions
and answers based on a set of over 10,000 news articles from CNN, with answers
consisting of spans of text from the corresponding articles. We collect this
dataset through a four-stage process designed to solicit exploratory questions
that require reasoning. A thorough analysis confirms that NewsQA demands
abilities beyond simple word matching and recognizing textual entailment. We
measure human performance on the dataset and compare it to several strong
neural models. The performance gap between humans and machines (0.198 in F1)
indicates that significant progress can be made on NewsQA through future
research. The dataset is freely available at
https://datasets.maluuba.com/NewsQA.
| [
{
"version": "v1",
"created": "Tue, 29 Nov 2016 20:38:07 GMT"
},
{
"version": "v2",
"created": "Thu, 22 Dec 2016 18:12:57 GMT"
},
{
"version": "v3",
"created": "Tue, 7 Feb 2017 16:27:59 GMT"
}
] | 2017-02-08T00:00:00 | [
[
"Trischler",
"Adam",
""
],
[
"Wang",
"Tong",
""
],
[
"Yuan",
"Xingdi",
""
],
[
"Harris",
"Justin",
""
],
[
"Sordoni",
"Alessandro",
""
],
[
"Bachman",
"Philip",
""
],
[
"Suleman",
"Kaheer",
""
]
] | TITLE: NewsQA: A Machine Comprehension Dataset
ABSTRACT: We present NewsQA, a challenging machine comprehension dataset of over
100,000 human-generated question-answer pairs. Crowdworkers supply questions
and answers based on a set of over 10,000 news articles from CNN, with answers
consisting of spans of text from the corresponding articles. We collect this
dataset through a four-stage process designed to solicit exploratory questions
that require reasoning. A thorough analysis confirms that NewsQA demands
abilities beyond simple word matching and recognizing textual entailment. We
measure human performance on the dataset and compare it to several strong
neural models. The performance gap between humans and machines (0.198 in F1)
indicates that significant progress can be made on NewsQA through future
research. The dataset is freely available at
https://datasets.maluuba.com/NewsQA.
| new_dataset | 0.962321 |
1702.01992 | John Arevalo | John Arevalo, Thamar Solorio, Manuel Montes-y-G\'omez, Fabio A.
Gonz\'alez | Gated Multimodal Units for Information Fusion | null | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents a novel model for multimodal learning based on gated
neural networks. The Gated Multimodal Unit (GMU) model is intended to be used
as an internal unit in a neural network architecture whose purpose is to find
an intermediate representation based on a combination of data from different
modalities. The GMU learns to decide how modalities influence the activation of
the unit using multiplicative gates. It was evaluated on a multilabel scenario
for genre classification of movies using the plot and the poster. The GMU
improved the macro f-score performance of single-modality approaches and
outperformed other fusion strategies, including mixture of experts models.
Along with this work, the MM-IMDb dataset is released which, to the best of our
knowledge, is the largest publicly available multimodal dataset for genre
prediction on movies.
| [
{
"version": "v1",
"created": "Tue, 7 Feb 2017 13:05:19 GMT"
}
] | 2017-02-08T00:00:00 | [
[
"Arevalo",
"John",
""
],
[
"Solorio",
"Thamar",
""
],
[
"Montes-y-Gómez",
"Manuel",
""
],
[
"González",
"Fabio A.",
""
]
] | TITLE: Gated Multimodal Units for Information Fusion
ABSTRACT: This paper presents a novel model for multimodal learning based on gated
neural networks. The Gated Multimodal Unit (GMU) model is intended to be used
as an internal unit in a neural network architecture whose purpose is to find
an intermediate representation based on a combination of data from different
modalities. The GMU learns to decide how modalities influence the activation of
the unit using multiplicative gates. It was evaluated on a multilabel scenario
for genre classification of movies using the plot and the poster. The GMU
improved the macro f-score performance of single-modality approaches and
outperformed other fusion strategies, including mixture of experts models.
Along with this work, the MM-IMDb dataset is released which, to the best of our
knowledge, is the largest publicly available multimodal dataset for genre
prediction on movies.
| new_dataset | 0.95594 |
1702.02125 | Eug\'enio Rodrigues | Eug\'enio Rodrigues and Lu\'isa Dias Pereira and Ad\'elio Rodrigues
Gaspar and \'Alvaro Gomes and Manuel Carlos Gameiro da Silva | Estimation of classrooms occupancy using a multi-layer perceptron | 6 pages, 2 figures, conference article | null | null | null | cs.NE cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents a multi-layer perceptron model for the estimation of
classrooms number of occupants from sensed indoor environmental data-relative
humidity, air temperature, and carbon dioxide concentration. The modelling
datasets were collected from two classrooms in the Secondary School of Pombal,
Portugal. The number of occupants and occupation periods were obtained from
class attendance reports. However, post-class occupancy was unknown and the
developed model is used to reconstruct the classrooms occupancy by filling the
unreported periods. Different model structure and environment variables
combination were tested. The model with best accuracy had as input vector 10
variables of five averaged time intervals of relative humidity and carbon
dioxide concentration. The model presented a mean square error of 1.99,
coefficient of determination of 0.96 with a significance of p-value < 0.001,
and a mean absolute error of 1 occupant. These results show promising
estimation capabilities in uncertain indoor environment conditions.
| [
{
"version": "v1",
"created": "Tue, 7 Feb 2017 18:17:25 GMT"
}
] | 2017-02-08T00:00:00 | [
[
"Rodrigues",
"Eugénio",
""
],
[
"Pereira",
"Luísa Dias",
""
],
[
"Gaspar",
"Adélio Rodrigues",
""
],
[
"Gomes",
"Álvaro",
""
],
[
"da Silva",
"Manuel Carlos Gameiro",
""
]
] | TITLE: Estimation of classrooms occupancy using a multi-layer perceptron
ABSTRACT: This paper presents a multi-layer perceptron model for the estimation of
classrooms number of occupants from sensed indoor environmental data-relative
humidity, air temperature, and carbon dioxide concentration. The modelling
datasets were collected from two classrooms in the Secondary School of Pombal,
Portugal. The number of occupants and occupation periods were obtained from
class attendance reports. However, post-class occupancy was unknown and the
developed model is used to reconstruct the classrooms occupancy by filling the
unreported periods. Different model structure and environment variables
combination were tested. The model with best accuracy had as input vector 10
variables of five averaged time intervals of relative humidity and carbon
dioxide concentration. The model presented a mean square error of 1.99,
coefficient of determination of 0.96 with a significance of p-value < 0.001,
and a mean absolute error of 1 occupant. These results show promising
estimation capabilities in uncertain indoor environment conditions.
| no_new_dataset | 0.951729 |
1310.2880 | Adrian Barbu | Adrian Barbu, Yiyuan She, Liangjing Ding, Gary Gramajo | Feature Selection with Annealing for Computer Vision and Big Data
Learning | 18 pages, 9 figures | IEEE Transactions on Pattern Analysis and Machine Intelligence,
vol. 39, no 2, pp 272 - 286, 2017 | 10.1109/TPAMI.2016.2544315 | null | stat.ML cs.CV cs.LG math.ST stat.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Many computer vision and medical imaging problems are faced with learning
from large-scale datasets, with millions of observations and features. In this
paper we propose a novel efficient learning scheme that tightens a sparsity
constraint by gradually removing variables based on a criterion and a schedule.
The attractive fact that the problem size keeps dropping throughout the
iterations makes it particularly suitable for big data learning. Our approach
applies generically to the optimization of any differentiable loss function,
and finds applications in regression, classification and ranking. The resultant
algorithms build variable screening into estimation and are extremely simple to
implement. We provide theoretical guarantees of convergence and selection
consistency. In addition, one dimensional piecewise linear response functions
are used to account for nonlinearity and a second order prior is imposed on
these functions to avoid overfitting. Experiments on real and synthetic data
show that the proposed method compares very well with other state of the art
methods in regression, classification and ranking while being computationally
very efficient and scalable.
| [
{
"version": "v1",
"created": "Thu, 10 Oct 2013 16:47:22 GMT"
},
{
"version": "v2",
"created": "Wed, 4 Jun 2014 22:42:51 GMT"
},
{
"version": "v3",
"created": "Tue, 30 Sep 2014 00:33:36 GMT"
},
{
"version": "v4",
"created": "Wed, 1 Oct 2014 20:03:42 GMT"
},
{
"version": "v5",
"created": "Thu, 3 Sep 2015 13:20:26 GMT"
},
{
"version": "v6",
"created": "Wed, 24 Feb 2016 02:02:20 GMT"
},
{
"version": "v7",
"created": "Thu, 17 Mar 2016 14:55:09 GMT"
}
] | 2017-02-07T00:00:00 | [
[
"Barbu",
"Adrian",
""
],
[
"She",
"Yiyuan",
""
],
[
"Ding",
"Liangjing",
""
],
[
"Gramajo",
"Gary",
""
]
] | TITLE: Feature Selection with Annealing for Computer Vision and Big Data
Learning
ABSTRACT: Many computer vision and medical imaging problems are faced with learning
from large-scale datasets, with millions of observations and features. In this
paper we propose a novel efficient learning scheme that tightens a sparsity
constraint by gradually removing variables based on a criterion and a schedule.
The attractive fact that the problem size keeps dropping throughout the
iterations makes it particularly suitable for big data learning. Our approach
applies generically to the optimization of any differentiable loss function,
and finds applications in regression, classification and ranking. The resultant
algorithms build variable screening into estimation and are extremely simple to
implement. We provide theoretical guarantees of convergence and selection
consistency. In addition, one dimensional piecewise linear response functions
are used to account for nonlinearity and a second order prior is imposed on
these functions to avoid overfitting. Experiments on real and synthetic data
show that the proposed method compares very well with other state of the art
methods in regression, classification and ranking while being computationally
very efficient and scalable.
| no_new_dataset | 0.94801 |
1502.05137 | Andreas Bulling | Hosnieh Sattar, Sabine M\"uller, Mario Fritz, Andreas Bulling | Prediction of Search Targets From Fixations in Open-World Settings | null | null | 10.1109/CVPR.2015.7298700 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Previous work on predicting the target of visual search from human fixations
only considered closed-world settings in which training labels are available
and predictions are performed for a known set of potential targets. In this
work we go beyond the state of the art by studying search target prediction in
an open-world setting in which we no longer assume that we have fixation data
to train for the search targets. We present a dataset containing fixation data
of 18 users searching for natural images from three image categories within
synthesised image collages of about 80 images. In a closed-world baseline
experiment we show that we can predict the correct target image out of a
candidate set of five images. We then present a new problem formulation for
search target prediction in the open-world setting that is based on learning
compatibilities between fixations and potential targets.
| [
{
"version": "v1",
"created": "Wed, 18 Feb 2015 07:04:04 GMT"
},
{
"version": "v2",
"created": "Wed, 4 Mar 2015 09:26:03 GMT"
},
{
"version": "v3",
"created": "Sat, 11 Apr 2015 14:56:51 GMT"
}
] | 2017-02-07T00:00:00 | [
[
"Sattar",
"Hosnieh",
""
],
[
"Müller",
"Sabine",
""
],
[
"Fritz",
"Mario",
""
],
[
"Bulling",
"Andreas",
""
]
] | TITLE: Prediction of Search Targets From Fixations in Open-World Settings
ABSTRACT: Previous work on predicting the target of visual search from human fixations
only considered closed-world settings in which training labels are available
and predictions are performed for a known set of potential targets. In this
work we go beyond the state of the art by studying search target prediction in
an open-world setting in which we no longer assume that we have fixation data
to train for the search targets. We present a dataset containing fixation data
of 18 users searching for natural images from three image categories within
synthesised image collages of about 80 images. In a closed-world baseline
experiment we show that we can predict the correct target image out of a
candidate set of five images. We then present a new problem formulation for
search target prediction in the open-world setting that is based on learning
compatibilities between fixations and potential targets.
| new_dataset | 0.961534 |
1504.02863 | Andreas Bulling | Xucong Zhang, Yusuke Sugano, Mario Fritz, Andreas Bulling | Appearance-Based Gaze Estimation in the Wild | null | null | 10.1109/CVPR.2015.7299081 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Appearance-based gaze estimation is believed to work well in real-world
settings, but existing datasets have been collected under controlled laboratory
conditions and methods have been not evaluated across multiple datasets. In
this work we study appearance-based gaze estimation in the wild. We present the
MPIIGaze dataset that contains 213,659 images we collected from 15 participants
during natural everyday laptop use over more than three months. Our dataset is
significantly more variable than existing ones with respect to appearance and
illumination. We also present a method for in-the-wild appearance-based gaze
estimation using multimodal convolutional neural networks that significantly
outperforms state-of-the art methods in the most challenging cross-dataset
evaluation. We present an extensive evaluation of several state-of-the-art
image-based gaze estimation algorithms on three current datasets, including our
own. This evaluation provides clear insights and allows us to identify key
research challenges of gaze estimation in the wild.
| [
{
"version": "v1",
"created": "Sat, 11 Apr 2015 11:52:33 GMT"
}
] | 2017-02-07T00:00:00 | [
[
"Zhang",
"Xucong",
""
],
[
"Sugano",
"Yusuke",
""
],
[
"Fritz",
"Mario",
""
],
[
"Bulling",
"Andreas",
""
]
] | TITLE: Appearance-Based Gaze Estimation in the Wild
ABSTRACT: Appearance-based gaze estimation is believed to work well in real-world
settings, but existing datasets have been collected under controlled laboratory
conditions and methods have been not evaluated across multiple datasets. In
this work we study appearance-based gaze estimation in the wild. We present the
MPIIGaze dataset that contains 213,659 images we collected from 15 participants
during natural everyday laptop use over more than three months. Our dataset is
significantly more variable than existing ones with respect to appearance and
illumination. We also present a method for in-the-wild appearance-based gaze
estimation using multimodal convolutional neural networks that significantly
outperforms state-of-the art methods in the most challenging cross-dataset
evaluation. We present an extensive evaluation of several state-of-the-art
image-based gaze estimation algorithms on three current datasets, including our
own. This evaluation provides clear insights and allows us to identify key
research challenges of gaze estimation in the wild.
| new_dataset | 0.962532 |
1505.05916 | Erroll Wood | Erroll Wood, Tadas Baltrusaitis, Xucong Zhang, Yusuke Sugano, Peter
Robinson, and Andreas Bulling | Rendering of Eyes for Eye-Shape Registration and Gaze Estimation | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Images of the eye are key in several computer vision problems, such as shape
registration and gaze estimation. Recent large-scale supervised methods for
these problems require time-consuming data collection and manual annotation,
which can be unreliable. We propose synthesizing perfectly labelled
photo-realistic training data in a fraction of the time. We used computer
graphics techniques to build a collection of dynamic eye-region models from
head scan geometry. These were randomly posed to synthesize close-up eye images
for a wide range of head poses, gaze directions, and illumination conditions.
We used our model's controllability to verify the importance of realistic
illumination and shape variations in eye-region training data. Finally, we
demonstrate the benefits of our synthesized training data (SynthesEyes) by
out-performing state-of-the-art methods for eye-shape registration as well as
cross-dataset appearance-based gaze estimation in the wild.
| [
{
"version": "v1",
"created": "Thu, 21 May 2015 22:12:31 GMT"
}
] | 2017-02-07T00:00:00 | [
[
"Wood",
"Erroll",
""
],
[
"Baltrusaitis",
"Tadas",
""
],
[
"Zhang",
"Xucong",
""
],
[
"Sugano",
"Yusuke",
""
],
[
"Robinson",
"Peter",
""
],
[
"Bulling",
"Andreas",
""
]
] | TITLE: Rendering of Eyes for Eye-Shape Registration and Gaze Estimation
ABSTRACT: Images of the eye are key in several computer vision problems, such as shape
registration and gaze estimation. Recent large-scale supervised methods for
these problems require time-consuming data collection and manual annotation,
which can be unreliable. We propose synthesizing perfectly labelled
photo-realistic training data in a fraction of the time. We used computer
graphics techniques to build a collection of dynamic eye-region models from
head scan geometry. These were randomly posed to synthesize close-up eye images
for a wide range of head poses, gaze directions, and illumination conditions.
We used our model's controllability to verify the importance of realistic
illumination and shape variations in eye-region training data. Finally, we
demonstrate the benefits of our synthesized training data (SynthesEyes) by
out-performing state-of-the-art methods for eye-shape registration as well as
cross-dataset appearance-based gaze estimation in the wild.
| no_new_dataset | 0.948537 |
1511.05768 | Andreas Bulling | Marc Tonsen, Xucong Zhang, Yusuke Sugano, Andreas Bulling | Labeled pupils in the wild: A dataset for studying pupil detection in
unconstrained environments | null | null | 10.1145/2857491.2857520 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present labelled pupils in the wild (LPW), a novel dataset of 66
high-quality, high-speed eye region videos for the development and evaluation
of pupil detection algorithms. The videos in our dataset were recorded from 22
participants in everyday locations at about 95 FPS using a state-of-the-art
dark-pupil head-mounted eye tracker. They cover people with different
ethnicities, a diverse set of everyday indoor and outdoor illumination
environments, as well as natural gaze direction distributions. The dataset also
includes participants wearing glasses, contact lenses, as well as make-up. We
benchmark five state-of-the-art pupil detection algorithms on our dataset with
respect to robustness and accuracy. We further study the influence of image
resolution, vision aids, as well as recording location (indoor, outdoor) on
pupil detection performance. Our evaluations provide valuable insights into the
general pupil detection problem and allow us to identify key challenges for
robust pupil detection on head-mounted eye trackers.
| [
{
"version": "v1",
"created": "Wed, 18 Nov 2015 13:30:55 GMT"
}
] | 2017-02-07T00:00:00 | [
[
"Tonsen",
"Marc",
""
],
[
"Zhang",
"Xucong",
""
],
[
"Sugano",
"Yusuke",
""
],
[
"Bulling",
"Andreas",
""
]
] | TITLE: Labeled pupils in the wild: A dataset for studying pupil detection in
unconstrained environments
ABSTRACT: We present labelled pupils in the wild (LPW), a novel dataset of 66
high-quality, high-speed eye region videos for the development and evaluation
of pupil detection algorithms. The videos in our dataset were recorded from 22
participants in everyday locations at about 95 FPS using a state-of-the-art
dark-pupil head-mounted eye tracker. They cover people with different
ethnicities, a diverse set of everyday indoor and outdoor illumination
environments, as well as natural gaze direction distributions. The dataset also
includes participants wearing glasses, contact lenses, as well as make-up. We
benchmark five state-of-the-art pupil detection algorithms on our dataset with
respect to robustness and accuracy. We further study the influence of image
resolution, vision aids, as well as recording location (indoor, outdoor) on
pupil detection performance. Our evaluations provide valuable insights into the
general pupil detection problem and allow us to identify key challenges for
robust pupil detection on head-mounted eye trackers.
| new_dataset | 0.964288 |
1512.07158 | Baichuan Zhang | Baichuan Zhang, Noman Mohammed, Vachik Dave, Mohammad Al Hasan | Feature Selection for Classification under Anonymity Constraint | Transactions on Data Privacy 2017 | null | null | null | cs.LG cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Over the last decade, proliferation of various online platforms and their
increasing adoption by billions of users have heightened the privacy risk of a
user enormously. In fact, security researchers have shown that sparse microdata
containing information about online activities of a user although anonymous,
can still be used to disclose the identity of the user by cross-referencing the
data with other data sources. To preserve the privacy of a user, in existing
works several methods (k-anonymity, l-diversity, differential privacy) are
proposed that ensure a dataset which is meant to share or publish bears small
identity disclosure risk. However, the majority of these methods modify the
data in isolation, without considering their utility in subsequent knowledge
discovery tasks, which makes these datasets less informative. In this work, we
consider labeled data that are generally used for classification, and propose
two methods for feature selection considering two goals: first, on the reduced
feature set the data has small disclosure risk, and second, the utility of the
data is preserved for performing a classification task. Experimental results on
various real-world datasets show that the method is effective and useful in
practice.
| [
{
"version": "v1",
"created": "Tue, 22 Dec 2015 17:06:01 GMT"
},
{
"version": "v2",
"created": "Sat, 13 Feb 2016 03:05:36 GMT"
},
{
"version": "v3",
"created": "Fri, 19 Feb 2016 02:01:57 GMT"
},
{
"version": "v4",
"created": "Thu, 17 Mar 2016 02:30:33 GMT"
},
{
"version": "v5",
"created": "Thu, 1 Dec 2016 01:05:59 GMT"
},
{
"version": "v6",
"created": "Tue, 31 Jan 2017 15:47:47 GMT"
},
{
"version": "v7",
"created": "Mon, 6 Feb 2017 01:14:37 GMT"
}
] | 2017-02-07T00:00:00 | [
[
"Zhang",
"Baichuan",
""
],
[
"Mohammed",
"Noman",
""
],
[
"Dave",
"Vachik",
""
],
[
"Hasan",
"Mohammad Al",
""
]
] | TITLE: Feature Selection for Classification under Anonymity Constraint
ABSTRACT: Over the last decade, proliferation of various online platforms and their
increasing adoption by billions of users have heightened the privacy risk of a
user enormously. In fact, security researchers have shown that sparse microdata
containing information about online activities of a user although anonymous,
can still be used to disclose the identity of the user by cross-referencing the
data with other data sources. To preserve the privacy of a user, in existing
works several methods (k-anonymity, l-diversity, differential privacy) are
proposed that ensure a dataset which is meant to share or publish bears small
identity disclosure risk. However, the majority of these methods modify the
data in isolation, without considering their utility in subsequent knowledge
discovery tasks, which makes these datasets less informative. In this work, we
consider labeled data that are generally used for classification, and propose
two methods for feature selection considering two goals: first, on the reduced
feature set the data has small disclosure risk, and second, the utility of the
data is preserved for performing a classification task. Experimental results on
various real-world datasets show that the method is effective and useful in
practice.
| no_new_dataset | 0.952175 |
1601.01006 | Fei Han | Fei Han, Brian Reily, William Hoff, Hao Zhang | Space-Time Representation of People Based on 3D Skeletal Data: A Review | Our paper has been accepted by the journal Computer Vision and Image
Understanding, see
http://www.sciencedirect.com/science/article/pii/S1077314217300279, Computer
Vision and Image Understanding, 2017 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Spatiotemporal human representation based on 3D visual perception data is a
rapidly growing research area. Based on the information sources, these
representations can be broadly categorized into two groups based on RGB-D
information or 3D skeleton data. Recently, skeleton-based human representations
have been intensively studied and kept attracting an increasing attention, due
to their robustness to variations of viewpoint, human body scale and motion
speed as well as the realtime, online performance. This paper presents a
comprehensive survey of existing space-time representations of people based on
3D skeletal data, and provides an informative categorization and analysis of
these methods from the perspectives, including information modality,
representation encoding, structure and transition, and feature engineering. We
also provide a brief overview of skeleton acquisition devices and construction
methods, enlist a number of public benchmark datasets with skeleton data, and
discuss potential future research directions.
| [
{
"version": "v1",
"created": "Tue, 5 Jan 2016 22:38:36 GMT"
},
{
"version": "v2",
"created": "Thu, 21 Jan 2016 06:00:39 GMT"
},
{
"version": "v3",
"created": "Sat, 4 Feb 2017 01:08:55 GMT"
}
] | 2017-02-07T00:00:00 | [
[
"Han",
"Fei",
""
],
[
"Reily",
"Brian",
""
],
[
"Hoff",
"William",
""
],
[
"Zhang",
"Hao",
""
]
] | TITLE: Space-Time Representation of People Based on 3D Skeletal Data: A Review
ABSTRACT: Spatiotemporal human representation based on 3D visual perception data is a
rapidly growing research area. Based on the information sources, these
representations can be broadly categorized into two groups based on RGB-D
information or 3D skeleton data. Recently, skeleton-based human representations
have been intensively studied and kept attracting an increasing attention, due
to their robustness to variations of viewpoint, human body scale and motion
speed as well as the realtime, online performance. This paper presents a
comprehensive survey of existing space-time representations of people based on
3D skeletal data, and provides an informative categorization and analysis of
these methods from the perspectives, including information modality,
representation encoding, structure and transition, and feature engineering. We
also provide a brief overview of skeleton acquisition devices and construction
methods, enlist a number of public benchmark datasets with skeleton data, and
discuss potential future research directions.
| no_new_dataset | 0.947624 |
1605.05110 | Zhen Xu | Zhen Xu, Bingquan Liu, Baoxun Wang, Chengjie Sun, Xiaolong Wang | Incorporating Loose-Structured Knowledge into Conversation Modeling via
Recall-Gate LSTM | under review of IJCNN 2017; 10 pages, 5 figures | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Modeling human conversations is the essence for building satisfying chat-bots
with multi-turn dialog ability. Conversation modeling will notably benefit from
domain knowledge since the relationships between sentences can be clarified due
to semantic hints introduced by knowledge. In this paper, a deep neural network
is proposed to incorporate background knowledge for conversation modeling.
Through a specially designed Recall gate, domain knowledge can be transformed
into the extra global memory of Long Short-Term Memory (LSTM), so as to enhance
LSTM by cooperating with its local memory to capture the implicit semantic
relevance between sentences within conversations. In addition, this paper
introduces the loose structured domain knowledge base, which can be built with
slight amount of manual work and easily adopted by the Recall gate. Our model
is evaluated on the context-oriented response selecting task, and experimental
results on both two datasets have shown that our approach is promising for
modeling human conversations and building key components of automatic chatting
systems.
| [
{
"version": "v1",
"created": "Tue, 17 May 2016 11:03:25 GMT"
},
{
"version": "v2",
"created": "Mon, 6 Feb 2017 03:43:17 GMT"
}
] | 2017-02-07T00:00:00 | [
[
"Xu",
"Zhen",
""
],
[
"Liu",
"Bingquan",
""
],
[
"Wang",
"Baoxun",
""
],
[
"Sun",
"Chengjie",
""
],
[
"Wang",
"Xiaolong",
""
]
] | TITLE: Incorporating Loose-Structured Knowledge into Conversation Modeling via
Recall-Gate LSTM
ABSTRACT: Modeling human conversations is the essence for building satisfying chat-bots
with multi-turn dialog ability. Conversation modeling will notably benefit from
domain knowledge since the relationships between sentences can be clarified due
to semantic hints introduced by knowledge. In this paper, a deep neural network
is proposed to incorporate background knowledge for conversation modeling.
Through a specially designed Recall gate, domain knowledge can be transformed
into the extra global memory of Long Short-Term Memory (LSTM), so as to enhance
LSTM by cooperating with its local memory to capture the implicit semantic
relevance between sentences within conversations. In addition, this paper
introduces the loose structured domain knowledge base, which can be built with
slight amount of manual work and easily adopted by the Recall gate. Our model
is evaluated on the context-oriented response selecting task, and experimental
results on both two datasets have shown that our approach is promising for
modeling human conversations and building key components of automatic chatting
systems.
| no_new_dataset | 0.942135 |
1605.06423 | Jonathan Huggins | Jonathan H. Huggins, Trevor Campbell, Tamara Broderick | Coresets for Scalable Bayesian Logistic Regression | In Proceedings of Advances in Neural Information Processing Systems
(NIPS 2016) | null | null | null | stat.CO cs.DS stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The use of Bayesian methods in large-scale data settings is attractive
because of the rich hierarchical models, uncertainty quantification, and prior
specification they provide. Standard Bayesian inference algorithms are
computationally expensive, however, making their direct application to large
datasets difficult or infeasible. Recent work on scaling Bayesian inference has
focused on modifying the underlying algorithms to, for example, use only a
random data subsample at each iteration. We leverage the insight that data is
often redundant to instead obtain a weighted subset of the data (called a
coreset) that is much smaller than the original dataset. We can then use this
small coreset in any number of existing posterior inference algorithms without
modification. In this paper, we develop an efficient coreset construction
algorithm for Bayesian logistic regression models. We provide theoretical
guarantees on the size and approximation quality of the coreset -- both for
fixed, known datasets, and in expectation for a wide class of data generative
models. Crucially, the proposed approach also permits efficient construction of
the coreset in both streaming and parallel settings, with minimal additional
effort. We demonstrate the efficacy of our approach on a number of synthetic
and real-world datasets, and find that, in practice, the size of the coreset is
independent of the original dataset size. Furthermore, constructing the coreset
takes a negligible amount of time compared to that required to run MCMC on it.
| [
{
"version": "v1",
"created": "Fri, 20 May 2016 16:26:45 GMT"
},
{
"version": "v2",
"created": "Thu, 27 Oct 2016 14:12:19 GMT"
},
{
"version": "v3",
"created": "Mon, 6 Feb 2017 15:11:30 GMT"
}
] | 2017-02-07T00:00:00 | [
[
"Huggins",
"Jonathan H.",
""
],
[
"Campbell",
"Trevor",
""
],
[
"Broderick",
"Tamara",
""
]
] | TITLE: Coresets for Scalable Bayesian Logistic Regression
ABSTRACT: The use of Bayesian methods in large-scale data settings is attractive
because of the rich hierarchical models, uncertainty quantification, and prior
specification they provide. Standard Bayesian inference algorithms are
computationally expensive, however, making their direct application to large
datasets difficult or infeasible. Recent work on scaling Bayesian inference has
focused on modifying the underlying algorithms to, for example, use only a
random data subsample at each iteration. We leverage the insight that data is
often redundant to instead obtain a weighted subset of the data (called a
coreset) that is much smaller than the original dataset. We can then use this
small coreset in any number of existing posterior inference algorithms without
modification. In this paper, we develop an efficient coreset construction
algorithm for Bayesian logistic regression models. We provide theoretical
guarantees on the size and approximation quality of the coreset -- both for
fixed, known datasets, and in expectation for a wide class of data generative
models. Crucially, the proposed approach also permits efficient construction of
the coreset in both streaming and parallel settings, with minimal additional
effort. We demonstrate the efficacy of our approach on a number of synthetic
and real-world datasets, and find that, in practice, the size of the coreset is
independent of the original dataset size. Furthermore, constructing the coreset
takes a negligible amount of time compared to that required to run MCMC on it.
| no_new_dataset | 0.949809 |
1610.01101 | Damek Davis | Aleksandr Aravkin and Damek Davis | A SMART Stochastic Algorithm for Nonconvex Optimization with
Applications to Robust Machine Learning | 33 pages, 5 figures | null | null | null | stat.ML cs.LG math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we show how to transform any optimization problem that arises
from fitting a machine learning model into one that (1) detects and removes
contaminated data from the training set while (2) simultaneously fitting the
trimmed model on the uncontaminated data that remains. To solve the resulting
nonconvex optimization problem, we introduce a fast stochastic
proximal-gradient algorithm that incorporates prior knowledge through nonsmooth
regularization. For datasets of size $n$, our approach requires
$O(n^{2/3}/\varepsilon)$ gradient evaluations to reach $\varepsilon$-accuracy
and, when a certain error bound holds, the complexity improves to $O(\kappa
n^{2/3}\log(1/\varepsilon))$. These rates are $n^{1/3}$ times better than those
achieved by typical, full gradient methods.
| [
{
"version": "v1",
"created": "Tue, 4 Oct 2016 17:24:43 GMT"
},
{
"version": "v2",
"created": "Sun, 5 Feb 2017 15:24:39 GMT"
}
] | 2017-02-07T00:00:00 | [
[
"Aravkin",
"Aleksandr",
""
],
[
"Davis",
"Damek",
""
]
] | TITLE: A SMART Stochastic Algorithm for Nonconvex Optimization with
Applications to Robust Machine Learning
ABSTRACT: In this paper, we show how to transform any optimization problem that arises
from fitting a machine learning model into one that (1) detects and removes
contaminated data from the training set while (2) simultaneously fitting the
trimmed model on the uncontaminated data that remains. To solve the resulting
nonconvex optimization problem, we introduce a fast stochastic
proximal-gradient algorithm that incorporates prior knowledge through nonsmooth
regularization. For datasets of size $n$, our approach requires
$O(n^{2/3}/\varepsilon)$ gradient evaluations to reach $\varepsilon$-accuracy
and, when a certain error bound holds, the complexity improves to $O(\kappa
n^{2/3}\log(1/\varepsilon))$. These rates are $n^{1/3}$ times better than those
achieved by typical, full gradient methods.
| no_new_dataset | 0.946597 |
1610.06227 | Mohammad Sadegh Rasooli | Mohammad Sadegh Rasooli, Michael Collins | Cross-Lingual Syntactic Transfer with Limited Resources | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We describe a simple but effective method for cross-lingual syntactic
transfer of dependency parsers, in the scenario where a large amount of
translation data is not available. The method makes use of three steps: 1) a
method for deriving cross-lingual word clusters, which can then be used in a
multilingual parser; 2) a method for transferring lexical information from a
target language to source language treebanks; 3) a method for integrating these
steps with the density-driven annotation projection method of Rasooli and
Collins (2015). Experiments show improvements over the state-of-the-art in
several languages used in previous work, in a setting where the only source of
translation data is the Bible, a considerably smaller corpus than the Europarl
corpus used in previous work. Results using the Europarl corpus as a source of
translation data show additional improvements over the results of Rasooli and
Collins (2015). We conclude with results on 38 datasets from the Universal
Dependencies corpora.
| [
{
"version": "v1",
"created": "Wed, 19 Oct 2016 21:25:39 GMT"
},
{
"version": "v2",
"created": "Sat, 4 Feb 2017 04:05:00 GMT"
}
] | 2017-02-07T00:00:00 | [
[
"Rasooli",
"Mohammad Sadegh",
""
],
[
"Collins",
"Michael",
""
]
] | TITLE: Cross-Lingual Syntactic Transfer with Limited Resources
ABSTRACT: We describe a simple but effective method for cross-lingual syntactic
transfer of dependency parsers, in the scenario where a large amount of
translation data is not available. The method makes use of three steps: 1) a
method for deriving cross-lingual word clusters, which can then be used in a
multilingual parser; 2) a method for transferring lexical information from a
target language to source language treebanks; 3) a method for integrating these
steps with the density-driven annotation projection method of Rasooli and
Collins (2015). Experiments show improvements over the state-of-the-art in
several languages used in previous work, in a setting where the only source of
translation data is the Bible, a considerably smaller corpus than the Europarl
corpus used in previous work. Results using the Europarl corpus as a source of
translation data show additional improvements over the results of Rasooli and
Collins (2015). We conclude with results on 38 datasets from the Universal
Dependencies corpora.
| no_new_dataset | 0.9462 |
1611.07810 | Tegan Maharaj | Tegan Maharaj and Nicolas Ballas and Anna Rohrbach and Aaron Courville
and Christopher Pal | A dataset and exploration of models for understanding video data through
fill-in-the-blank question-answering | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | While deep convolutional neural networks frequently approach or exceed
human-level performance at benchmark tasks involving static images, extending
this success to moving images is not straightforward. Having models which can
learn to understand video is of interest for many applications, including
content recommendation, prediction, summarization, event/object detection and
understanding human visual perception, but many domains lack sufficient data to
explore and perfect video models. In order to address the need for a simple,
quantitative benchmark for developing and understanding video, we present
MovieFIB, a fill-in-the-blank question-answering dataset with over 300,000
examples, based on descriptive video annotations for the visually impaired. In
addition to presenting statistics and a description of the dataset, we perform
a detailed analysis of 5 different models' predictions, and compare these with
human performance. We investigate the relative importance of language, static
(2D) visual features, and moving (3D) visual features; the effects of
increasing dataset size, the number of frames sampled; and of vocabulary size.
We illustrate that: this task is not solvable by a language model alone; our
model combining 2D and 3D visual information indeed provides the best result;
all models perform significantly worse than human-level. We provide human
evaluations for responses given by different models and find that accuracy on
the MovieFIB evaluation corresponds well with human judgement. We suggest
avenues for improving video models, and hope that the proposed dataset can be
useful for measuring and encouraging progress in this very interesting field.
| [
{
"version": "v1",
"created": "Wed, 23 Nov 2016 14:22:51 GMT"
},
{
"version": "v2",
"created": "Sun, 5 Feb 2017 17:51:19 GMT"
}
] | 2017-02-07T00:00:00 | [
[
"Maharaj",
"Tegan",
""
],
[
"Ballas",
"Nicolas",
""
],
[
"Rohrbach",
"Anna",
""
],
[
"Courville",
"Aaron",
""
],
[
"Pal",
"Christopher",
""
]
] | TITLE: A dataset and exploration of models for understanding video data through
fill-in-the-blank question-answering
ABSTRACT: While deep convolutional neural networks frequently approach or exceed
human-level performance at benchmark tasks involving static images, extending
this success to moving images is not straightforward. Having models which can
learn to understand video is of interest for many applications, including
content recommendation, prediction, summarization, event/object detection and
understanding human visual perception, but many domains lack sufficient data to
explore and perfect video models. In order to address the need for a simple,
quantitative benchmark for developing and understanding video, we present
MovieFIB, a fill-in-the-blank question-answering dataset with over 300,000
examples, based on descriptive video annotations for the visually impaired. In
addition to presenting statistics and a description of the dataset, we perform
a detailed analysis of 5 different models' predictions, and compare these with
human performance. We investigate the relative importance of language, static
(2D) visual features, and moving (3D) visual features; the effects of
increasing dataset size, the number of frames sampled; and of vocabulary size.
We illustrate that: this task is not solvable by a language model alone; our
model combining 2D and 3D visual information indeed provides the best result;
all models perform significantly worse than human-level. We provide human
evaluations for responses given by different models and find that accuracy on
the MovieFIB evaluation corresponds well with human judgement. We suggest
avenues for improving video models, and hope that the proposed dataset can be
useful for measuring and encouraging progress in this very interesting field.
| no_new_dataset | 0.912124 |
1612.00157 | Yi\u{g}it Baran Can | Yi\u{g}it Baran Can, Efe Il{\i}cak, Tolga \c{C}ukur | Fast 3D Variable-FOV Reconstruction for Parallel Imaging with Localized
Sensitivities | Accepted, to be presented at ISMRM 25th Annual Meeting 2017 | null | null | null | physics.med-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Several successful iterative approaches have recently been proposed for
parallel-imaging reconstructions of variable-density (VD) acquisitions, but
they often induce substantial computational burden for non-Cartesian data. Here
we propose a generalized variable-FOV PILS reconstruction 3D VD Cartesian and
non-Cartesian data. The proposed method separates k-space into non-intersecting
annuli based on sampling density, and sets the 3D reconstruction FOV for each
annulus based on the respective sampling density. The variable-FOV method is
compared against conventional gridding, PILS, and ESPIRiT reconstructions.
Results indicate that the proposed method yields better artifact suppression
compared to gridding and PILS, and improves noise conditioning relative to
ESPIRiT, enabling fast and high-quality reconstructions of 3D datasets.
| [
{
"version": "v1",
"created": "Thu, 1 Dec 2016 06:14:34 GMT"
},
{
"version": "v2",
"created": "Mon, 6 Feb 2017 13:06:28 GMT"
}
] | 2017-02-07T00:00:00 | [
[
"Can",
"Yiğit Baran",
""
],
[
"Ilıcak",
"Efe",
""
],
[
"Çukur",
"Tolga",
""
]
] | TITLE: Fast 3D Variable-FOV Reconstruction for Parallel Imaging with Localized
Sensitivities
ABSTRACT: Several successful iterative approaches have recently been proposed for
parallel-imaging reconstructions of variable-density (VD) acquisitions, but
they often induce substantial computational burden for non-Cartesian data. Here
we propose a generalized variable-FOV PILS reconstruction 3D VD Cartesian and
non-Cartesian data. The proposed method separates k-space into non-intersecting
annuli based on sampling density, and sets the 3D reconstruction FOV for each
annulus based on the respective sampling density. The variable-FOV method is
compared against conventional gridding, PILS, and ESPIRiT reconstructions.
Results indicate that the proposed method yields better artifact suppression
compared to gridding and PILS, and improves noise conditioning relative to
ESPIRiT, enabling fast and high-quality reconstructions of 3D datasets.
| no_new_dataset | 0.951142 |
1612.07386 | David Rosen | David M. Rosen, Luca Carlone, Afonso S. Bandeira, and John J. Leonard | SE-Sync: A Certifiably Correct Algorithm for Synchronization over the
Special Euclidean Group | 49 Pages, 20 figures | null | null | null | cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Many important geometric estimation problems take the form of synchronization
over the special Euclidean group: estimate the values of a set of poses given a
set of relative measurements between them. This problem is typically formulated
as a nonconvex maximum-likelihood estimation that is computationally hard to
solve in general. Nevertheless, in this paper we present an algorithm that is
able to efficiently recover certifiably globally optimal solutions of the
special Euclidean synchronization problem in a non-adversarial noise regime.
The crux of our approach is the development of a semidefinite relaxation of the
maximum-likelihood estimation whose minimizer provides an exact MLE so long as
the magnitude of the noise corrupting the available measurements falls below a
certain critical threshold; furthermore, whenever exactness obtains, it is
possible to verify this fact a posteriori, thereby certifying the optimality of
the recovered estimate. We develop a specialized optimization scheme for
solving large-scale instances of this relaxation by exploiting its low-rank,
geometric, and graph-theoretic structure to reduce it to an equivalent
optimization problem on a low-dimensional Riemannian manifold, and design a
truncated-Newton trust-region method to solve this reduction efficiently.
Finally, we combine this fast optimization approach with a simple rounding
procedure to produce our algorithm, SE-Sync. Experimental evaluation on a
variety of simulated and real-world pose-graph SLAM datasets shows that SE-Sync
is able to recover certifiably globally optimal solutions when the available
measurements are corrupted by noise up to an order of magnitude greater than
that typically encountered in robotics and computer vision applications, and
does so more than an order of magnitude faster than the Gauss-Newton-based
approach that forms the basis of current state-of-the-art techniques.
| [
{
"version": "v1",
"created": "Wed, 21 Dec 2016 23:21:29 GMT"
},
{
"version": "v2",
"created": "Sun, 5 Feb 2017 03:49:42 GMT"
}
] | 2017-02-07T00:00:00 | [
[
"Rosen",
"David M.",
""
],
[
"Carlone",
"Luca",
""
],
[
"Bandeira",
"Afonso S.",
""
],
[
"Leonard",
"John J.",
""
]
] | TITLE: SE-Sync: A Certifiably Correct Algorithm for Synchronization over the
Special Euclidean Group
ABSTRACT: Many important geometric estimation problems take the form of synchronization
over the special Euclidean group: estimate the values of a set of poses given a
set of relative measurements between them. This problem is typically formulated
as a nonconvex maximum-likelihood estimation that is computationally hard to
solve in general. Nevertheless, in this paper we present an algorithm that is
able to efficiently recover certifiably globally optimal solutions of the
special Euclidean synchronization problem in a non-adversarial noise regime.
The crux of our approach is the development of a semidefinite relaxation of the
maximum-likelihood estimation whose minimizer provides an exact MLE so long as
the magnitude of the noise corrupting the available measurements falls below a
certain critical threshold; furthermore, whenever exactness obtains, it is
possible to verify this fact a posteriori, thereby certifying the optimality of
the recovered estimate. We develop a specialized optimization scheme for
solving large-scale instances of this relaxation by exploiting its low-rank,
geometric, and graph-theoretic structure to reduce it to an equivalent
optimization problem on a low-dimensional Riemannian manifold, and design a
truncated-Newton trust-region method to solve this reduction efficiently.
Finally, we combine this fast optimization approach with a simple rounding
procedure to produce our algorithm, SE-Sync. Experimental evaluation on a
variety of simulated and real-world pose-graph SLAM datasets shows that SE-Sync
is able to recover certifiably globally optimal solutions when the available
measurements are corrupted by noise up to an order of magnitude greater than
that typically encountered in robotics and computer vision applications, and
does so more than an order of magnitude faster than the Gauss-Newton-based
approach that forms the basis of current state-of-the-art techniques.
| no_new_dataset | 0.944893 |
1702.00956 | Suwon Shon | Suwon Shon, Hanseok Ko | KU-ISPL Speaker Recognition Systems under Language mismatch condition
for NIST 2016 Speaker Recognition Evaluation | SRE16, NIST SRE 2016 system description | null | null | null | cs.SD cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Korea University Intelligent Signal Processing Lab. (KU-ISPL) developed
speaker recognition system for SRE16 fixed training condition. Data for
evaluation trials are collected from outside North America, spoken in Tagalog
and Cantonese while training data only is spoken English. Thus, main issue for
SRE16 is compensating the discrepancy between different languages. As
development dataset which is spoken in Cebuano and Mandarin, we could prepare
the evaluation trials through preliminary experiments to compensate the
language mismatched condition. Our team developed 4 different approaches to
extract i-vectors and applied state-of-the-art techniques as backend. To
compensate language mismatch, we investigated and endeavored unique method such
as unsupervised language clustering, inter language variability compensation
and gender/language dependent score normalization.
| [
{
"version": "v1",
"created": "Fri, 3 Feb 2017 10:15:29 GMT"
},
{
"version": "v2",
"created": "Mon, 6 Feb 2017 03:37:28 GMT"
}
] | 2017-02-07T00:00:00 | [
[
"Shon",
"Suwon",
""
],
[
"Ko",
"Hanseok",
""
]
] | TITLE: KU-ISPL Speaker Recognition Systems under Language mismatch condition
for NIST 2016 Speaker Recognition Evaluation
ABSTRACT: Korea University Intelligent Signal Processing Lab. (KU-ISPL) developed
speaker recognition system for SRE16 fixed training condition. Data for
evaluation trials are collected from outside North America, spoken in Tagalog
and Cantonese while training data only is spoken English. Thus, main issue for
SRE16 is compensating the discrepancy between different languages. As
development dataset which is spoken in Cebuano and Mandarin, we could prepare
the evaluation trials through preliminary experiments to compensate the
language mismatched condition. Our team developed 4 different approaches to
extract i-vectors and applied state-of-the-art techniques as backend. To
compensate language mismatch, we investigated and endeavored unique method such
as unsupervised language clustering, inter language variability compensation
and gender/language dependent score normalization.
| no_new_dataset | 0.934813 |
1702.01151 | Helge Holzmann | Helge Holzmann, Wolfgang Nejdl, Avishek Anand | The Dawn of Today's Popular Domains: A Study of the Archived German Web
over 18 Years | JCDL 2016, Newark, NJ, USA | null | 10.1145/2910896.2910901 | null | cs.DL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The Web has been around and maturing for 25 years. The popular websites of
today have undergone vast changes during this period, with a few being there
almost since the beginning and many new ones becoming popular over the years.
This makes it worthwhile to take a look at how these sites have evolved and
what they might tell us about the future of the Web. We therefore embarked on a
longitudinal study spanning almost the whole period of the Web, based on data
collected by the Internet Archive starting in 1996, to retrospectively analyze
how the popular Web as of now has evolved over the past 18 years.
For our study we focused on the German Web, specifically on the top 100 most
popular websites in 17 categories. This paper presents a selection of the most
interesting findings in terms of volume, size as well as age of the Web. While
related work in the field of Web Dynamics has mainly focused on change rates
and analyzed datasets spanning less than a year, we looked at the evolution of
websites over 18 years. We found that around 70% of the pages we investigated
are younger than a year, with an observed exponential growth in age as well as
in size up to now. If this growth rate continues, the number of pages from the
popular domains will almost double in the next two years. In addition, we give
insights into our data set, provided by the Internet Archive, which hosts the
largest and most complete Web archive as of today.
| [
{
"version": "v1",
"created": "Fri, 3 Feb 2017 20:45:56 GMT"
}
] | 2017-02-07T00:00:00 | [
[
"Holzmann",
"Helge",
""
],
[
"Nejdl",
"Wolfgang",
""
],
[
"Anand",
"Avishek",
""
]
] | TITLE: The Dawn of Today's Popular Domains: A Study of the Archived German Web
over 18 Years
ABSTRACT: The Web has been around and maturing for 25 years. The popular websites of
today have undergone vast changes during this period, with a few being there
almost since the beginning and many new ones becoming popular over the years.
This makes it worthwhile to take a look at how these sites have evolved and
what they might tell us about the future of the Web. We therefore embarked on a
longitudinal study spanning almost the whole period of the Web, based on data
collected by the Internet Archive starting in 1996, to retrospectively analyze
how the popular Web as of now has evolved over the past 18 years.
For our study we focused on the German Web, specifically on the top 100 most
popular websites in 17 categories. This paper presents a selection of the most
interesting findings in terms of volume, size as well as age of the Web. While
related work in the field of Web Dynamics has mainly focused on change rates
and analyzed datasets spanning less than a year, we looked at the evolution of
websites over 18 years. We found that around 70% of the pages we investigated
are younger than a year, with an observed exponential growth in age as well as
in size up to now. If this growth rate continues, the number of pages from the
popular domains will almost double in the next two years. In addition, we give
insights into our data set, provided by the Internet Archive, which hosts the
largest and most complete Web archive as of today.
| no_new_dataset | 0.886174 |
1702.01159 | Helge Holzmann | Helge Holzmann, Wolfgang Nejdl, Avishek Anand | On the Applicability of Delicious for Temporal Search on Web Archives | SIGIR 2016, Pisa, Italy | null | 10.1145/2911451.2914724 | null | cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Web archives are large longitudinal collections that store webpages from the
past, which might be missing on the current live Web. Consequently, temporal
search over such collections is essential for finding prominent missing
webpages and tasks like historical analysis. However, this has been challenging
due to the lack of popularity information and proper ground truth to evaluate
temporal retrieval models. In this paper we investigate the applicability of
external longitudinal resources to identify important and popular websites in
the past and analyze the social bookmarking service Delicious for this purpose.
The timestamped bookmarks on Delicious provide explicit cues about popular
time periods in the past along with relevant descriptors. These are valuable to
identify important documents in the past for a given temporal query. Focusing
purely on recall, we analyzed more than 12,000 queries and find that using
Delicious yields average recall values from 46% up to 100%, when limiting
ourselves to the best represented queries in the considered dataset. This
constitutes an attractive and low-overhead approach for quick access into Web
archives by not dealing with the actual contents.
| [
{
"version": "v1",
"created": "Fri, 3 Feb 2017 21:06:47 GMT"
}
] | 2017-02-07T00:00:00 | [
[
"Holzmann",
"Helge",
""
],
[
"Nejdl",
"Wolfgang",
""
],
[
"Anand",
"Avishek",
""
]
] | TITLE: On the Applicability of Delicious for Temporal Search on Web Archives
ABSTRACT: Web archives are large longitudinal collections that store webpages from the
past, which might be missing on the current live Web. Consequently, temporal
search over such collections is essential for finding prominent missing
webpages and tasks like historical analysis. However, this has been challenging
due to the lack of popularity information and proper ground truth to evaluate
temporal retrieval models. In this paper we investigate the applicability of
external longitudinal resources to identify important and popular websites in
the past and analyze the social bookmarking service Delicious for this purpose.
The timestamped bookmarks on Delicious provide explicit cues about popular
time periods in the past along with relevant descriptors. These are valuable to
identify important documents in the past for a given temporal query. Focusing
purely on recall, we analyzed more than 12,000 queries and find that using
Delicious yields average recall values from 46% up to 100%, when limiting
ourselves to the best represented queries in the considered dataset. This
constitutes an attractive and low-overhead approach for quick access into Web
archives by not dealing with the actual contents.
| no_new_dataset | 0.947769 |
1702.01167 | Andrey Kuehlkamp | Andrey Kuehlkamp and Kevin W. Bowyer | An Analysis of 1-to-First Matching in Iris Recognition | 2016 IEEE Winter Conference on Applications of Computer Vision (WACV) | null | 10.1109/WACV.2016.7477687 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Iris recognition systems are a mature technology that is widely used
throughout the world. In identification (as opposed to verification) mode, an
iris to be recognized is typically matched against all N enrolled irises. This
is the classic "1-to-N search". In order to improve the speed of large-scale
identification, a modified "1-to-First" search has been used in some
operational systems. A 1-to-First search terminates with the first
below-threshold match that is found, whereas a 1-to-N search always finds the
best match across all enrollments. We know of no previous studies that evaluate
how the accuracy of 1-to-First search differs from that of 1-to-N search. Using
a dataset of over 50,000 iris images from 2,800 different irises, we perform
experiments to evaluate the relative accuracy of 1-to-First and 1-to-N search.
We evaluate how the accuracy difference changes with larger numbers of enrolled
irises, and with larger ranges of rotational difference allowed between iris
images. We find that False Match error rate for 1-to-First is higher than for
1-to-N, and the the difference grows with larger number of enrolled irises and
with larger range of rotation.
| [
{
"version": "v1",
"created": "Fri, 3 Feb 2017 21:24:10 GMT"
}
] | 2017-02-07T00:00:00 | [
[
"Kuehlkamp",
"Andrey",
""
],
[
"Bowyer",
"Kevin W.",
""
]
] | TITLE: An Analysis of 1-to-First Matching in Iris Recognition
ABSTRACT: Iris recognition systems are a mature technology that is widely used
throughout the world. In identification (as opposed to verification) mode, an
iris to be recognized is typically matched against all N enrolled irises. This
is the classic "1-to-N search". In order to improve the speed of large-scale
identification, a modified "1-to-First" search has been used in some
operational systems. A 1-to-First search terminates with the first
below-threshold match that is found, whereas a 1-to-N search always finds the
best match across all enrollments. We know of no previous studies that evaluate
how the accuracy of 1-to-First search differs from that of 1-to-N search. Using
a dataset of over 50,000 iris images from 2,800 different irises, we perform
experiments to evaluate the relative accuracy of 1-to-First and 1-to-N search.
We evaluate how the accuracy difference changes with larger numbers of enrolled
irises, and with larger ranges of rotational difference allowed between iris
images. We find that False Match error rate for 1-to-First is higher than for
1-to-N, and the the difference grows with larger number of enrolled irises and
with larger range of rotation.
| no_new_dataset | 0.884489 |
1702.01184 | Benjamin Mako Hill | Benjamin Mako Hill, Andr\'es Monroy-Hern\'andez | A longitudinal dataset of five years of public activity in the Scratch
online community | null | Scientific Data 4, Article number: 170002, 2017 | 10.1038/sdata.2017.2 | null | cs.CY cs.HC cs.SI | http://creativecommons.org/licenses/by/4.0/ | Scratch is a programming environment and an online community where young
people can create, share, learn, and communicate. In collaboration with the
Scratch Team at MIT, we created a longitudinal dataset of public activity in
the Scratch online community during its first five years (2007-2012). The
dataset comprises 32 tables with information on more than 1 million Scratch
users, nearly 2 million Scratch projects, more than 10 million comments, more
than 30 million visits to Scratch projects, and more. To help researchers
understand this dataset, and to establish the validity of the data, we also
include the source code of every version of the software that operated the
website, as well as the software used to generate this dataset. We believe this
is the largest and most comprehensive downloadable dataset of youth programming
artifacts and communication.
| [
{
"version": "v1",
"created": "Fri, 3 Feb 2017 22:02:24 GMT"
}
] | 2017-02-07T00:00:00 | [
[
"Hill",
"Benjamin Mako",
""
],
[
"Monroy-Hernández",
"Andrés",
""
]
] | TITLE: A longitudinal dataset of five years of public activity in the Scratch
online community
ABSTRACT: Scratch is a programming environment and an online community where young
people can create, share, learn, and communicate. In collaboration with the
Scratch Team at MIT, we created a longitudinal dataset of public activity in
the Scratch online community during its first five years (2007-2012). The
dataset comprises 32 tables with information on more than 1 million Scratch
users, nearly 2 million Scratch projects, more than 10 million comments, more
than 30 million visits to Scratch projects, and more. To help researchers
understand this dataset, and to establish the validity of the data, we also
include the source code of every version of the software that operated the
website, as well as the software used to generate this dataset. We believe this
is the largest and most comprehensive downloadable dataset of youth programming
artifacts and communication.
| new_dataset | 0.958654 |
1702.01268 | Giorgio Valentini | Jessica Gliozzo | Network-based methods for outcome prediction in the "sample space" | MSc Thesis, Advisor: G. Valentini, Co-Advisors: A. Paccanaro and M.
Re, 92 pages, 36 figures, 10 tables | null | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this thesis we present the novel semi-supervised network-based algorithm
P-Net, which is able to rank and classify patients with respect to a specific
phenotype or clinical outcome under study. The peculiar and innovative
characteristic of this method is that it builds a network of samples/patients,
where the nodes represent the samples and the edges are functional or genetic
relationships between individuals (e.g. similarity of expression profiles), to
predict the phenotype under study. In other words, it constructs the network in
the "sample space" and not in the "biomarker space" (where nodes represent
biomolecules (e.g. genes, proteins) and edges represent functional or genetic
relationships between nodes), as usual in state-of-the-art methods. To assess
the performances of P-Net, we apply it on three different publicly available
datasets from patients afflicted with a specific type of tumor: pancreatic
cancer, melanoma and ovarian cancer dataset, by using the data and following
the experimental set-up proposed in two recently published papers [Barter et
al., 2014, Winter et al., 2012]. We show that network-based methods in the
"sample space" can achieve results competitive with classical supervised
inductive systems. Moreover, the graph representation of the samples can be
easily visualized through networks and can be used to gain visual clues about
the relationships between samples, taking into account the phenotype associated
or predicted for each sample. To our knowledge this is one of the first works
that proposes graph-based algorithms working in the "sample space" of the
biomolecular profiles of the patients to predict their phenotype or outcome,
thus contributing to a novel research line in the framework of the Network
Medicine.
| [
{
"version": "v1",
"created": "Sat, 4 Feb 2017 11:18:53 GMT"
}
] | 2017-02-07T00:00:00 | [
[
"Gliozzo",
"Jessica",
""
]
] | TITLE: Network-based methods for outcome prediction in the "sample space"
ABSTRACT: In this thesis we present the novel semi-supervised network-based algorithm
P-Net, which is able to rank and classify patients with respect to a specific
phenotype or clinical outcome under study. The peculiar and innovative
characteristic of this method is that it builds a network of samples/patients,
where the nodes represent the samples and the edges are functional or genetic
relationships between individuals (e.g. similarity of expression profiles), to
predict the phenotype under study. In other words, it constructs the network in
the "sample space" and not in the "biomarker space" (where nodes represent
biomolecules (e.g. genes, proteins) and edges represent functional or genetic
relationships between nodes), as usual in state-of-the-art methods. To assess
the performances of P-Net, we apply it on three different publicly available
datasets from patients afflicted with a specific type of tumor: pancreatic
cancer, melanoma and ovarian cancer dataset, by using the data and following
the experimental set-up proposed in two recently published papers [Barter et
al., 2014, Winter et al., 2012]. We show that network-based methods in the
"sample space" can achieve results competitive with classical supervised
inductive systems. Moreover, the graph representation of the samples can be
easily visualized through networks and can be used to gain visual clues about
the relationships between samples, taking into account the phenotype associated
or predicted for each sample. To our knowledge this is one of the first works
that proposes graph-based algorithms working in the "sample space" of the
biomolecular profiles of the patients to predict their phenotype or outcome,
thus contributing to a novel research line in the framework of the Network
Medicine.
| no_new_dataset | 0.942612 |
1702.01434 | Muhammad Qasim Pasta | Muhammad Qasim Pasta, Faraz Zaidi, C\'eline Rozenblat | Generating online social networks based on socio-demographic attributes | arXiv admin note: substantial text overlap with arXiv:1311.3508 | J Complex Netw 2014, 2 (4): 475-494 | null | null | cs.SI physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent years have seen tremendous growth of many online social networks such
as Facebook, LinkedIn and MySpace. People connect to each other through these
networks forming large social communities providing researchers rich datasets
to understand, model and predict social interactions and behaviors. New
contacts in these networks can be formed due to an individual's demographic
attributes such as age group, gender, geographic location, or due to a
network's structural dynamics such as triadic closure and preferential
attachment, or a combination of both demographic and structural
characteristics.
A number of network generation models have been proposed in the last decade
to explain the structure, evolution and processes taking place in different
types of networks, and notably social networks. Network generation models
studied in the literature primarily consider structural properties, and in some
cases an individual's demographic profile in the formation of new social
contacts. These models do not present a mechanism to combine both structural
and demographic characteristics for the formation of new links. In this paper,
we propose a new network generation algorithm which incorporates both these
characteristics to model network formation. We use different publicly available
Facebook datasets as benchmarks to demonstrate the correctness of the proposed
network generation model. The proposed model is flexible and thus can generate
networks with varying demographic and structural properties.
| [
{
"version": "v1",
"created": "Sun, 5 Feb 2017 18:04:29 GMT"
}
] | 2017-02-07T00:00:00 | [
[
"Pasta",
"Muhammad Qasim",
""
],
[
"Zaidi",
"Faraz",
""
],
[
"Rozenblat",
"Céline",
""
]
] | TITLE: Generating online social networks based on socio-demographic attributes
ABSTRACT: Recent years have seen tremendous growth of many online social networks such
as Facebook, LinkedIn and MySpace. People connect to each other through these
networks forming large social communities providing researchers rich datasets
to understand, model and predict social interactions and behaviors. New
contacts in these networks can be formed due to an individual's demographic
attributes such as age group, gender, geographic location, or due to a
network's structural dynamics such as triadic closure and preferential
attachment, or a combination of both demographic and structural
characteristics.
A number of network generation models have been proposed in the last decade
to explain the structure, evolution and processes taking place in different
types of networks, and notably social networks. Network generation models
studied in the literature primarily consider structural properties, and in some
cases an individual's demographic profile in the formation of new social
contacts. These models do not present a mechanism to combine both structural
and demographic characteristics for the formation of new links. In this paper,
we propose a new network generation algorithm which incorporates both these
characteristics to model network formation. We use different publicly available
Facebook datasets as benchmarks to demonstrate the correctness of the proposed
network generation model. The proposed model is flexible and thus can generate
networks with varying demographic and structural properties.
| no_new_dataset | 0.954308 |
1702.01466 | Hongyu Gong | Hongyu Gong, Jiaqi Mu, Suma Bhat, Pramod Viswanath | Prepositions in Context | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Prepositions are highly polysemous, and their variegated senses encode
significant semantic information. In this paper we match each preposition's
complement and attachment and their interplay crucially to the geometry of the
word vectors to the left and right of the preposition. Extracting such features
from the vast number of instances of each preposition and clustering them makes
for an efficient preposition sense disambigution (PSD) algorithm, which is
comparable to and better than state-of-the-art on two benchmark datasets. Our
reliance on no external linguistic resource allows us to scale the PSD
algorithm to a large WikiCorpus and learn sense-specific preposition
representations -- which we show to encode semantic relations and paraphrasing
of verb particle compounds, via simple vector operations.
| [
{
"version": "v1",
"created": "Sun, 5 Feb 2017 23:16:01 GMT"
}
] | 2017-02-07T00:00:00 | [
[
"Gong",
"Hongyu",
""
],
[
"Mu",
"Jiaqi",
""
],
[
"Bhat",
"Suma",
""
],
[
"Viswanath",
"Pramod",
""
]
] | TITLE: Prepositions in Context
ABSTRACT: Prepositions are highly polysemous, and their variegated senses encode
significant semantic information. In this paper we match each preposition's
complement and attachment and their interplay crucially to the geometry of the
word vectors to the left and right of the preposition. Extracting such features
from the vast number of instances of each preposition and clustering them makes
for an efficient preposition sense disambigution (PSD) algorithm, which is
comparable to and better than state-of-the-art on two benchmark datasets. Our
reliance on no external linguistic resource allows us to scale the PSD
algorithm to a large WikiCorpus and learn sense-specific preposition
representations -- which we show to encode semantic relations and paraphrasing
of verb particle compounds, via simple vector operations.
| no_new_dataset | 0.952618 |
1702.01638 | Xinyu Li | Xinyu Li, Yanyi Zhang, Jianyu Zhang, Shuhong Chen, Ivan Marsic,
Richard A. Farneth, Randall S. Burd | Concurrent Activity Recognition with Multimodal CNN-LSTM Structure | 14 pages, 12 figures, under review | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We introduce a system that recognizes concurrent activities from real-world
data captured by multiple sensors of different types. The recognition is
achieved in two steps. First, we extract spatial and temporal features from the
multimodal data. We feed each datatype into a convolutional neural network that
extracts spatial features, followed by a long-short term memory network that
extracts temporal information in the sensory data. The extracted features are
then fused for decision making in the second step. Second, we achieve
concurrent activity recognition with a single classifier that encodes a binary
output vector in which elements indicate whether the corresponding activity
types are currently in progress. We tested our system with three datasets from
different domains recorded using different sensors and achieved performance
comparable to existing systems designed specifically for those domains. Our
system is the first to address the concurrent activity recognition with
multisensory data using a single model, which is scalable, simple to train and
easy to deploy.
| [
{
"version": "v1",
"created": "Mon, 6 Feb 2017 15:01:45 GMT"
}
] | 2017-02-07T00:00:00 | [
[
"Li",
"Xinyu",
""
],
[
"Zhang",
"Yanyi",
""
],
[
"Zhang",
"Jianyu",
""
],
[
"Chen",
"Shuhong",
""
],
[
"Marsic",
"Ivan",
""
],
[
"Farneth",
"Richard A.",
""
],
[
"Burd",
"Randall S.",
""
]
] | TITLE: Concurrent Activity Recognition with Multimodal CNN-LSTM Structure
ABSTRACT: We introduce a system that recognizes concurrent activities from real-world
data captured by multiple sensors of different types. The recognition is
achieved in two steps. First, we extract spatial and temporal features from the
multimodal data. We feed each datatype into a convolutional neural network that
extracts spatial features, followed by a long-short term memory network that
extracts temporal information in the sensory data. The extracted features are
then fused for decision making in the second step. Second, we achieve
concurrent activity recognition with a single classifier that encodes a binary
output vector in which elements indicate whether the corresponding activity
types are currently in progress. We tested our system with three datasets from
different domains recorded using different sensors and achieved performance
comparable to existing systems designed specifically for those domains. Our
system is the first to address the concurrent activity recognition with
multisensory data using a single model, which is scalable, simple to train and
easy to deploy.
| no_new_dataset | 0.948537 |
1702.01711 | Rodrigo Agerri | I\~naki San Vicente, Rodrigo Agerri, German Rigau | Q-WordNet PPV: Simple, Robust and (almost) Unsupervised Generation of
Polarity Lexicons for Multiple Languages | 8 pages plus 2 pages of references | Proceedings of the 14th Conference of the European Chapter of the
Association for Computational Linguistics (EACL 2014), pages 88-97,
Gothenburg, Sweden, April 26-30 2014 | null | null | cs.CL | http://creativecommons.org/licenses/by-nc-sa/4.0/ | This paper presents a simple, robust and (almost) unsupervised
dictionary-based method, qwn-ppv (Q-WordNet as Personalized PageRanking Vector)
to automatically generate polarity lexicons. We show that qwn-ppv outperforms
other automatically generated lexicons for the four extrinsic evaluations
presented here. It also shows very competitive and robust results with respect
to manually annotated ones. Results suggest that no single lexicon is best for
every task and dataset and that the intrinsic evaluation of polarity lexicons
is not a good performance indicator on a Sentiment Analysis task. The qwn-ppv
method allows to easily create quality polarity lexicons whenever no
domain-based annotated corpora are available for a given language.
| [
{
"version": "v1",
"created": "Mon, 6 Feb 2017 17:14:29 GMT"
}
] | 2017-02-07T00:00:00 | [
[
"Vicente",
"Iñaki San",
""
],
[
"Agerri",
"Rodrigo",
""
],
[
"Rigau",
"German",
""
]
] | TITLE: Q-WordNet PPV: Simple, Robust and (almost) Unsupervised Generation of
Polarity Lexicons for Multiple Languages
ABSTRACT: This paper presents a simple, robust and (almost) unsupervised
dictionary-based method, qwn-ppv (Q-WordNet as Personalized PageRanking Vector)
to automatically generate polarity lexicons. We show that qwn-ppv outperforms
other automatically generated lexicons for the four extrinsic evaluations
presented here. It also shows very competitive and robust results with respect
to manually annotated ones. Results suggest that no single lexicon is best for
every task and dataset and that the intrinsic evaluation of polarity lexicons
is not a good performance indicator on a Sentiment Analysis task. The qwn-ppv
method allows to easily create quality polarity lexicons whenever no
domain-based annotated corpora are available for a given language.
| no_new_dataset | 0.948917 |
1702.01713 | Nikolaos Polatidis Mr | Nikolaos Polatidis, Christos K. Georgiadis | A dynamic multi-level collaborative filtering method for improved
recommendations | null | Computer Standards & Interfaces, 51, 14-21 (2017) | 10.1016/j.csi.2016.10.014 | null | cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | One of the most used approaches for providing recommendations in various
online environments such as e-commerce is collaborative filtering. Although,
this is a simple method for recommending items or services, accuracy and
quality problems still exist. Thus, we propose a dynamic multi-level
collaborative filtering method that improves the quality of the
recommendations. The proposed method is based on positive and negative
adjustments and can be used in different domains that utilize collaborative
filtering to increase the quality of the user experience. Furthermore, the
effectiveness of the proposed method is shown by providing an extensive
experimental evaluation based on three real datasets and by comparisons to
alternative methods.
| [
{
"version": "v1",
"created": "Mon, 6 Feb 2017 17:19:07 GMT"
}
] | 2017-02-07T00:00:00 | [
[
"Polatidis",
"Nikolaos",
""
],
[
"Georgiadis",
"Christos K.",
""
]
] | TITLE: A dynamic multi-level collaborative filtering method for improved
recommendations
ABSTRACT: One of the most used approaches for providing recommendations in various
online environments such as e-commerce is collaborative filtering. Although,
this is a simple method for recommending items or services, accuracy and
quality problems still exist. Thus, we propose a dynamic multi-level
collaborative filtering method that improves the quality of the
recommendations. The proposed method is based on positive and negative
adjustments and can be used in different domains that utilize collaborative
filtering to increase the quality of the user experience. Furthermore, the
effectiveness of the proposed method is shown by providing an extensive
experimental evaluation based on three real datasets and by comparisons to
alternative methods.
| no_new_dataset | 0.952131 |
1702.01721 | Afshin Dehghan | Afshin Dehghan, Syed Zain Masood, Guang Shu, Enrique. G. Ortiz | View Independent Vehicle Make, Model and Color Recognition Using
Convolutional Neural Network | 7 Pages | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper describes the details of Sighthound's fully automated vehicle
make, model and color recognition system. The backbone of our system is a deep
convolutional neural network that is not only computationally inexpensive, but
also provides state-of-the-art results on several competitive benchmarks.
Additionally, our deep network is trained on a large dataset of several million
images which are labeled through a semi-automated process. Finally we test our
system on several public datasets as well as our own internal test dataset. Our
results show that we outperform other methods on all benchmarks by significant
margins. Our model is available to developers through the Sighthound Cloud API
at https://www.sighthound.com/products/cloud
| [
{
"version": "v1",
"created": "Mon, 6 Feb 2017 17:47:08 GMT"
}
] | 2017-02-07T00:00:00 | [
[
"Dehghan",
"Afshin",
""
],
[
"Masood",
"Syed Zain",
""
],
[
"Shu",
"Guang",
""
],
[
"Ortiz",
"Enrique. G.",
""
]
] | TITLE: View Independent Vehicle Make, Model and Color Recognition Using
Convolutional Neural Network
ABSTRACT: This paper describes the details of Sighthound's fully automated vehicle
make, model and color recognition system. The backbone of our system is a deep
convolutional neural network that is not only computationally inexpensive, but
also provides state-of-the-art results on several competitive benchmarks.
Additionally, our deep network is trained on a large dataset of several million
images which are labeled through a semi-automated process. Finally we test our
system on several public datasets as well as our own internal test dataset. Our
results show that we outperform other methods on all benchmarks by significant
margins. Our model is available to developers through the Sighthound Cloud API
at https://www.sighthound.com/products/cloud
| new_dataset | 0.955068 |
1604.01170 | Antonia Godoy-Lorite | Antonia Godoy-Lorite, Roger Guimera, Cristopher Moore, Marta
Sales-Pardo | Accurate and scalable social recommendation using mixed-membership
stochastic block models | 9 pages, 4 figures | Proc. Natl. Acad. Sci. USA 113 (50) , 14207 -14212 (2016) | 10.1073/pnas.1606316113 | null | cs.SI cs.IR cs.LG physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With ever-increasing amounts of online information available, modeling and
predicting individual preferences-for books or articles, for example-is
becoming more and more important. Good predictions enable us to improve advice
to users, and obtain a better understanding of the socio-psychological
processes that determine those preferences. We have developed a collaborative
filtering model, with an associated scalable algorithm, that makes accurate
predictions of individuals' preferences. Our approach is based on the explicit
assumption that there are groups of individuals and of items, and that the
preferences of an individual for an item are determined only by their group
memberships. Importantly, we allow each individual and each item to belong
simultaneously to mixtures of different groups and, unlike many popular
approaches, such as matrix factorization, we do not assume implicitly or
explicitly that individuals in each group prefer items in a single group of
items. The resulting overlapping groups and the predicted preferences can be
inferred with a expectation-maximization algorithm whose running time scales
linearly (per iteration). Our approach enables us to predict individual
preferences in large datasets, and is considerably more accurate than the
current algorithms for such large datasets.
| [
{
"version": "v1",
"created": "Tue, 5 Apr 2016 08:28:08 GMT"
},
{
"version": "v2",
"created": "Wed, 6 Apr 2016 07:55:35 GMT"
}
] | 2017-02-06T00:00:00 | [
[
"Godoy-Lorite",
"Antonia",
""
],
[
"Guimera",
"Roger",
""
],
[
"Moore",
"Cristopher",
""
],
[
"Sales-Pardo",
"Marta",
""
]
] | TITLE: Accurate and scalable social recommendation using mixed-membership
stochastic block models
ABSTRACT: With ever-increasing amounts of online information available, modeling and
predicting individual preferences-for books or articles, for example-is
becoming more and more important. Good predictions enable us to improve advice
to users, and obtain a better understanding of the socio-psychological
processes that determine those preferences. We have developed a collaborative
filtering model, with an associated scalable algorithm, that makes accurate
predictions of individuals' preferences. Our approach is based on the explicit
assumption that there are groups of individuals and of items, and that the
preferences of an individual for an item are determined only by their group
memberships. Importantly, we allow each individual and each item to belong
simultaneously to mixtures of different groups and, unlike many popular
approaches, such as matrix factorization, we do not assume implicitly or
explicitly that individuals in each group prefer items in a single group of
items. The resulting overlapping groups and the predicted preferences can be
inferred with a expectation-maximization algorithm whose running time scales
linearly (per iteration). Our approach enables us to predict individual
preferences in large datasets, and is considerably more accurate than the
current algorithms for such large datasets.
| no_new_dataset | 0.944893 |
1702.00820 | Theodoros Rekatsinas | Theodoros Rekatsinas, Xu Chu, Ihab F. Ilyas, Christopher R\'e | HoloClean: Holistic Data Repairs with Probabilistic Inference | null | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We introduce HoloClean, a framework for holistic data repairing driven by
probabilistic inference. HoloClean unifies existing qualitative data repairing
approaches, which rely on integrity constraints or external data sources, with
quantitative data repairing methods, which leverage statistical properties of
the input data. Given an inconsistent dataset as input, HoloClean automatically
generates a probabilistic program that performs data repairing. Inspired by
recent theoretical advances in probabilistic inference, we introduce a series
of optimizations which ensure that inference over HoloClean's probabilistic
model scales to instances with millions of tuples. We show that HoloClean
scales to instances with millions of tuples and find data repairs with an
average precision of ~90% and an average recall of above ~76% across a diverse
array of datasets exhibiting different types of errors. This yields an average
F1 improvement of more than 2x against state-of-the-art methods.
| [
{
"version": "v1",
"created": "Thu, 2 Feb 2017 20:25:41 GMT"
}
] | 2017-02-06T00:00:00 | [
[
"Rekatsinas",
"Theodoros",
""
],
[
"Chu",
"Xu",
""
],
[
"Ilyas",
"Ihab F.",
""
],
[
"Ré",
"Christopher",
""
]
] | TITLE: HoloClean: Holistic Data Repairs with Probabilistic Inference
ABSTRACT: We introduce HoloClean, a framework for holistic data repairing driven by
probabilistic inference. HoloClean unifies existing qualitative data repairing
approaches, which rely on integrity constraints or external data sources, with
quantitative data repairing methods, which leverage statistical properties of
the input data. Given an inconsistent dataset as input, HoloClean automatically
generates a probabilistic program that performs data repairing. Inspired by
recent theoretical advances in probabilistic inference, we introduce a series
of optimizations which ensure that inference over HoloClean's probabilistic
model scales to instances with millions of tuples. We show that HoloClean
scales to instances with millions of tuples and find data repairs with an
average precision of ~90% and an average recall of above ~76% across a diverse
array of datasets exhibiting different types of errors. This yields an average
F1 improvement of more than 2x against state-of-the-art methods.
| no_new_dataset | 0.94699 |
1702.00833 | Maciej Wielgosz | Maciej Wielgosz and Andrzej Skocze\'n and Matej Mertik | Recurrent Neural Networks for anomaly detection in the Post-Mortem time
series of LHC superconducting magnets | Related to arxiv:1611.06241 | null | null | null | physics.ins-det cs.LG physics.acc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents a model based on Deep Learning algorithms of LSTM and GRU
for facilitating an anomaly detection in Large Hadron Collider superconducting
magnets. We used high resolution data available in Post Mortem database to
train a set of models and chose the best possible set of their
hyper-parameters. Using Deep Learning approach allowed to examine a vast body
of data and extract the fragments which require further experts examination and
are regarded as anomalies. The presented method does not require tedious manual
threshold setting and operator attention at the stage of the system setup.
Instead, the automatic approach is proposed, which achieves according to our
experiments accuracy of 99%. This is reached for the largest dataset of 302 MB
and the following architecture of the network: single layer LSTM, 128 cells, 20
epochs of training, look_back=16, look_ahead=128, grid=100 and optimizer Adam.
All the experiments were run on GPU Nvidia Tesla K80
| [
{
"version": "v1",
"created": "Thu, 2 Feb 2017 21:32:32 GMT"
}
] | 2017-02-06T00:00:00 | [
[
"Wielgosz",
"Maciej",
""
],
[
"Skoczeń",
"Andrzej",
""
],
[
"Mertik",
"Matej",
""
]
] | TITLE: Recurrent Neural Networks for anomaly detection in the Post-Mortem time
series of LHC superconducting magnets
ABSTRACT: This paper presents a model based on Deep Learning algorithms of LSTM and GRU
for facilitating an anomaly detection in Large Hadron Collider superconducting
magnets. We used high resolution data available in Post Mortem database to
train a set of models and chose the best possible set of their
hyper-parameters. Using Deep Learning approach allowed to examine a vast body
of data and extract the fragments which require further experts examination and
are regarded as anomalies. The presented method does not require tedious manual
threshold setting and operator attention at the stage of the system setup.
Instead, the automatic approach is proposed, which achieves according to our
experiments accuracy of 99%. This is reached for the largest dataset of 302 MB
and the following architecture of the network: single layer LSTM, 128 cells, 20
epochs of training, look_back=16, look_ahead=128, grid=100 and optimizer Adam.
All the experiments were run on GPU Nvidia Tesla K80
| no_new_dataset | 0.953275 |
1702.00926 | Seungryong Kim | Seungryong Kim, Dongbo Min, Bumsub Ham, Sangryul Jeon, Stephen Lin,
Kwanghoon Sohn | FCSS: Fully Convolutional Self-Similarity for Dense Semantic
Correspondence | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a descriptor, called fully convolutional self-similarity (FCSS),
for dense semantic correspondence. To robustly match points among different
instances within the same object class, we formulate FCSS using local
self-similarity (LSS) within a fully convolutional network. In contrast to
existing CNN-based descriptors, FCSS is inherently insensitive to intra-class
appearance variations because of its LSS-based structure, while maintaining the
precise localization ability of deep neural networks. The sampling patterns of
local structure and the self-similarity measure are jointly learned within the
proposed network in an end-to-end and multi-scale manner. As training data for
semantic correspondence is rather limited, we propose to leverage object
candidate priors provided in existing image datasets and also correspondence
consistency between object pairs to enable weakly-supervised learning.
Experiments demonstrate that FCSS outperforms conventional handcrafted
descriptors and CNN-based descriptors on various benchmarks.
| [
{
"version": "v1",
"created": "Fri, 3 Feb 2017 07:36:44 GMT"
}
] | 2017-02-06T00:00:00 | [
[
"Kim",
"Seungryong",
""
],
[
"Min",
"Dongbo",
""
],
[
"Ham",
"Bumsub",
""
],
[
"Jeon",
"Sangryul",
""
],
[
"Lin",
"Stephen",
""
],
[
"Sohn",
"Kwanghoon",
""
]
] | TITLE: FCSS: Fully Convolutional Self-Similarity for Dense Semantic
Correspondence
ABSTRACT: We present a descriptor, called fully convolutional self-similarity (FCSS),
for dense semantic correspondence. To robustly match points among different
instances within the same object class, we formulate FCSS using local
self-similarity (LSS) within a fully convolutional network. In contrast to
existing CNN-based descriptors, FCSS is inherently insensitive to intra-class
appearance variations because of its LSS-based structure, while maintaining the
precise localization ability of deep neural networks. The sampling patterns of
local structure and the self-similarity measure are jointly learned within the
proposed network in an end-to-end and multi-scale manner. As training data for
semantic correspondence is rather limited, we propose to leverage object
candidate priors provided in existing image datasets and also correspondence
consistency between object pairs to enable weakly-supervised learning.
Experiments demonstrate that FCSS outperforms conventional handcrafted
descriptors and CNN-based descriptors on various benchmarks.
| no_new_dataset | 0.949623 |
1702.01015 | Helge Holzmann | Helge Holzmann, Vinay Goel, Avishek Anand | ArchiveSpark: Efficient Web Archive Access, Extraction and Derivation | JCDL 2016, Newark, NJ, USA | null | 10.1145/2910896.2910902 | null | cs.DL cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Web archives are a valuable resource for researchers of various disciplines.
However, to use them as a scholarly source, researchers require a tool that
provides efficient access to Web archive data for extraction and derivation of
smaller datasets. Besides efficient access we identify five other objectives
based on practical researcher needs such as ease of use, extensibility and
reusability.
Towards these objectives we propose ArchiveSpark, a framework for efficient,
distributed Web archive processing that builds a research corpus by working on
existing and standardized data formats commonly held by Web archiving
institutions. Performance optimizations in ArchiveSpark, facilitated by the use
of a widely available metadata index, result in significant speed-ups of data
processing. Our benchmarks show that ArchiveSpark is faster than alternative
approaches without depending on any additional data stores while improving
usability by seamlessly integrating queries and derivations with external
tools.
| [
{
"version": "v1",
"created": "Fri, 3 Feb 2017 14:17:02 GMT"
}
] | 2017-02-06T00:00:00 | [
[
"Holzmann",
"Helge",
""
],
[
"Goel",
"Vinay",
""
],
[
"Anand",
"Avishek",
""
]
] | TITLE: ArchiveSpark: Efficient Web Archive Access, Extraction and Derivation
ABSTRACT: Web archives are a valuable resource for researchers of various disciplines.
However, to use them as a scholarly source, researchers require a tool that
provides efficient access to Web archive data for extraction and derivation of
smaller datasets. Besides efficient access we identify five other objectives
based on practical researcher needs such as ease of use, extensibility and
reusability.
Towards these objectives we propose ArchiveSpark, a framework for efficient,
distributed Web archive processing that builds a research corpus by working on
existing and standardized data formats commonly held by Web archiving
institutions. Performance optimizations in ArchiveSpark, facilitated by the use
of a widely available metadata index, result in significant speed-ups of data
processing. Our benchmarks show that ArchiveSpark is faster than alternative
approaches without depending on any additional data stores while improving
usability by seamlessly integrating queries and derivations with external
tools.
| no_new_dataset | 0.945147 |
1609.07042 | Xiang Xiang | Xiang Xiang and Trac D. Tran | Pose-Selective Max Pooling for Measuring Similarity | The tutorial and program associated with this paper are available at
https://github.com/eglxiang/ytf yet for non-commercial use | null | null | null | cs.CV cs.AI cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we deal with two challenges for measuring the similarity of
the subject identities in practical video-based face recognition - the
variation of the head pose in uncontrolled environments and the computational
expense of processing videos. Since the frame-wise feature mean is unable to
characterize the pose diversity among frames, we define and preserve the
overall pose diversity and closeness in a video. Then, identity will be the
only source of variation across videos since the pose varies even within a
single video. Instead of simply using all the frames, we select those faces
whose pose point is closest to the centroid of the K-means cluster containing
that pose point. Then, we represent a video as a bag of frame-wise deep face
features while the number of features has been reduced from hundreds to K.
Since the video representation can well represent the identity, now we measure
the subject similarity between two videos as the max correlation among all
possible pairs in the two bags of features. On the official 5,000 video-pairs
of the YouTube Face dataset for face verification, our algorithm achieves a
comparable performance with VGG-face that averages over deep features of all
frames. Other vision tasks can also benefit from the generic idea of employing
geometric cues to improve the descriptiveness of deep features.
| [
{
"version": "v1",
"created": "Thu, 22 Sep 2016 15:59:38 GMT"
},
{
"version": "v2",
"created": "Mon, 26 Sep 2016 18:21:05 GMT"
},
{
"version": "v3",
"created": "Thu, 10 Nov 2016 04:05:29 GMT"
},
{
"version": "v4",
"created": "Mon, 14 Nov 2016 04:10:09 GMT"
}
] | 2017-02-03T00:00:00 | [
[
"Xiang",
"Xiang",
""
],
[
"Tran",
"Trac D.",
""
]
] | TITLE: Pose-Selective Max Pooling for Measuring Similarity
ABSTRACT: In this paper, we deal with two challenges for measuring the similarity of
the subject identities in practical video-based face recognition - the
variation of the head pose in uncontrolled environments and the computational
expense of processing videos. Since the frame-wise feature mean is unable to
characterize the pose diversity among frames, we define and preserve the
overall pose diversity and closeness in a video. Then, identity will be the
only source of variation across videos since the pose varies even within a
single video. Instead of simply using all the frames, we select those faces
whose pose point is closest to the centroid of the K-means cluster containing
that pose point. Then, we represent a video as a bag of frame-wise deep face
features while the number of features has been reduced from hundreds to K.
Since the video representation can well represent the identity, now we measure
the subject similarity between two videos as the max correlation among all
possible pairs in the two bags of features. On the official 5,000 video-pairs
of the YouTube Face dataset for face verification, our algorithm achieves a
comparable performance with VGG-face that averages over deep features of all
frames. Other vision tasks can also benefit from the generic idea of employing
geometric cues to improve the descriptiveness of deep features.
| no_new_dataset | 0.947962 |
1701.09123 | Rodrigo Agerri | Rodrigo Agerri and German Rigau | Robust Multilingual Named Entity Recognition with Shallow
Semi-Supervised Features | 26 pages, 19 tables (submitted for publication on September 2015),
Artificial Intelligence (2016) | Artificial Intelligence, 238, 63-82 (2016) | 10.1016/j.artint.2016.05.003 | null | cs.CL cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | We present a multilingual Named Entity Recognition approach based on a robust
and general set of features across languages and datasets. Our system combines
shallow local information with clustering semi-supervised features induced on
large amounts of unlabeled text. Understanding via empirical experimentation
how to effectively combine various types of clustering features allows us to
seamlessly export our system to other datasets and languages. The result is a
simple but highly competitive system which obtains state of the art results
across five languages and twelve datasets. The results are reported on standard
shared task evaluation data such as CoNLL for English, Spanish and Dutch.
Furthermore, and despite the lack of linguistically motivated features, we also
report best results for languages such as Basque and German. In addition, we
demonstrate that our method also obtains very competitive results even when the
amount of supervised data is cut by half, alleviating the dependency on
manually annotated data. Finally, the results show that our emphasis on
clustering features is crucial to develop robust out-of-domain models. The
system and models are freely available to facilitate its use and guarantee the
reproducibility of results.
| [
{
"version": "v1",
"created": "Tue, 31 Jan 2017 16:36:06 GMT"
}
] | 2017-02-03T00:00:00 | [
[
"Agerri",
"Rodrigo",
""
],
[
"Rigau",
"German",
""
]
] | TITLE: Robust Multilingual Named Entity Recognition with Shallow
Semi-Supervised Features
ABSTRACT: We present a multilingual Named Entity Recognition approach based on a robust
and general set of features across languages and datasets. Our system combines
shallow local information with clustering semi-supervised features induced on
large amounts of unlabeled text. Understanding via empirical experimentation
how to effectively combine various types of clustering features allows us to
seamlessly export our system to other datasets and languages. The result is a
simple but highly competitive system which obtains state of the art results
across five languages and twelve datasets. The results are reported on standard
shared task evaluation data such as CoNLL for English, Spanish and Dutch.
Furthermore, and despite the lack of linguistically motivated features, we also
report best results for languages such as Basque and German. In addition, we
demonstrate that our method also obtains very competitive results even when the
amount of supervised data is cut by half, alleviating the dependency on
manually annotated data. Finally, the results show that our emphasis on
clustering features is crucial to develop robust out-of-domain models. The
system and models are freely available to facilitate its use and guarantee the
reproducibility of results.
| no_new_dataset | 0.946151 |
1702.00552 | Aziz Mohaisen | Omar Al-Ibrahim and Aziz Mohaisen and Charles Kamhoua and Kevin Kwiat
and Laurent Njilla | Beyond Free Riding: Quality of Indicators for Assessing Participation in
Information Sharing for Threat Intelligence | null | null | null | null | cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Threat intelligence sharing has become a growing concept, whereby entities
can exchange patterns of threats with each other, in the form of indicators, to
a community of trust for threat analysis and incident response. However,
sharing threat-related information have posed various risks to an organization
that pertains to its security, privacy, and competitiveness. Given the
coinciding benefits and risks of threat information sharing, some entities have
adopted an elusive behavior of "free-riding" so that they can acquire the
benefits of sharing without contributing much to the community. So far,
understanding the effectiveness of sharing has been viewed from the perspective
of the amount of information exchanged as opposed to its quality. In this
paper, we introduce the notion of quality of indicators (\qoi) for the
assessment of the level of contribution by participants in information sharing
for threat intelligence. We exemplify this notion through various metrics,
including correctness, relevance, utility, and uniqueness of indicators. In
order to realize the notion of \qoi, we conducted an empirical study and taken
a benchmark approach to define quality metrics, then we obtained a reference
dataset and utilized tools from the machine learning literature for quality
assessment. We compared these results against a model that only considers the
volume of information as a metric for contribution, and unveiled various
interesting observations, including the ability to spot low quality
contributions that are synonym to free riding in threat information sharing.
| [
{
"version": "v1",
"created": "Thu, 2 Feb 2017 06:35:55 GMT"
}
] | 2017-02-03T00:00:00 | [
[
"Al-Ibrahim",
"Omar",
""
],
[
"Mohaisen",
"Aziz",
""
],
[
"Kamhoua",
"Charles",
""
],
[
"Kwiat",
"Kevin",
""
],
[
"Njilla",
"Laurent",
""
]
] | TITLE: Beyond Free Riding: Quality of Indicators for Assessing Participation in
Information Sharing for Threat Intelligence
ABSTRACT: Threat intelligence sharing has become a growing concept, whereby entities
can exchange patterns of threats with each other, in the form of indicators, to
a community of trust for threat analysis and incident response. However,
sharing threat-related information have posed various risks to an organization
that pertains to its security, privacy, and competitiveness. Given the
coinciding benefits and risks of threat information sharing, some entities have
adopted an elusive behavior of "free-riding" so that they can acquire the
benefits of sharing without contributing much to the community. So far,
understanding the effectiveness of sharing has been viewed from the perspective
of the amount of information exchanged as opposed to its quality. In this
paper, we introduce the notion of quality of indicators (\qoi) for the
assessment of the level of contribution by participants in information sharing
for threat intelligence. We exemplify this notion through various metrics,
including correctness, relevance, utility, and uniqueness of indicators. In
order to realize the notion of \qoi, we conducted an empirical study and taken
a benchmark approach to define quality metrics, then we obtained a reference
dataset and utilized tools from the machine learning literature for quality
assessment. We compared these results against a model that only considers the
volume of information as a metric for contribution, and unveiled various
interesting observations, including the ability to spot low quality
contributions that are synonym to free riding in threat information sharing.
| no_new_dataset | 0.950732 |
1702.00583 | Mikhail Breslav | Mikhail Breslav, Tyson L. Hedrick, Stan Sclaroff, Margrit Betke | Automating Image Analysis by Annotating Landmarks with Deep Neural
Networks | 30 pages | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Image and video analysis is often a crucial step in the study of animal
behavior and kinematics. Often these analyses require that the position of one
or more animal landmarks are annotated (marked) in numerous images. The process
of annotating landmarks can require a significant amount of time and tedious
labor, which motivates the need for algorithms that can automatically annotate
landmarks. In the community of scientists that use image and video analysis to
study the 3D flight of animals, there has been a trend of developing more
automated approaches for annotating landmarks, yet they fall short of being
generally applicable. Inspired by the success of Deep Neural Networks (DNNs) on
many problems in the field of computer vision, we investigate how suitable DNNs
are for accurate and automatic annotation of landmarks in video datasets
representative of those collected by scientists studying animals.
Our work shows, through extensive experimentation on videos of hawkmoths,
that DNNs are suitable for automatic and accurate landmark localization. In
particular, we show that one of our proposed DNNs is more accurate than the
current best algorithm for automatic localization of landmarks on hawkmoth
videos. Moreover, we demonstrate how these annotations can be used to
quantitatively analyze the 3D flight of a hawkmoth. To facilitate the use of
DNNs by scientists from many different fields, we provide a self contained
explanation of what DNNs are, how they work, and how to apply them to other
datasets using the freely available library Caffe and supplemental code that we
provide.
| [
{
"version": "v1",
"created": "Thu, 2 Feb 2017 08:53:10 GMT"
}
] | 2017-02-03T00:00:00 | [
[
"Breslav",
"Mikhail",
""
],
[
"Hedrick",
"Tyson L.",
""
],
[
"Sclaroff",
"Stan",
""
],
[
"Betke",
"Margrit",
""
]
] | TITLE: Automating Image Analysis by Annotating Landmarks with Deep Neural
Networks
ABSTRACT: Image and video analysis is often a crucial step in the study of animal
behavior and kinematics. Often these analyses require that the position of one
or more animal landmarks are annotated (marked) in numerous images. The process
of annotating landmarks can require a significant amount of time and tedious
labor, which motivates the need for algorithms that can automatically annotate
landmarks. In the community of scientists that use image and video analysis to
study the 3D flight of animals, there has been a trend of developing more
automated approaches for annotating landmarks, yet they fall short of being
generally applicable. Inspired by the success of Deep Neural Networks (DNNs) on
many problems in the field of computer vision, we investigate how suitable DNNs
are for accurate and automatic annotation of landmarks in video datasets
representative of those collected by scientists studying animals.
Our work shows, through extensive experimentation on videos of hawkmoths,
that DNNs are suitable for automatic and accurate landmark localization. In
particular, we show that one of our proposed DNNs is more accurate than the
current best algorithm for automatic localization of landmarks on hawkmoth
videos. Moreover, we demonstrate how these annotations can be used to
quantitatively analyze the 3D flight of a hawkmoth. To facilitate the use of
DNNs by scientists from many different fields, we provide a self contained
explanation of what DNNs are, how they work, and how to apply them to other
datasets using the freely available library Caffe and supplemental code that we
provide.
| no_new_dataset | 0.94256 |
1702.00585 | Massimo Franceschet | Massimo Franceschet and Enrico Bozzo | The temporalized Massey's method | arXiv admin note: text overlap with arXiv:1701.03363 | null | null | null | cs.SI physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose and throughly investigate a temporalized version of the popular
Massey's technique for rating actors in sport competitions. The method can be
described as a dynamic temporal process in which team ratings are updated at
every match according to their performance during the match and the strength of
the opponent team. Using the Italian soccer dataset, we empirically show that
the method has a good foresight prediction accuracy.
| [
{
"version": "v1",
"created": "Thu, 2 Feb 2017 08:54:32 GMT"
}
] | 2017-02-03T00:00:00 | [
[
"Franceschet",
"Massimo",
""
],
[
"Bozzo",
"Enrico",
""
]
] | TITLE: The temporalized Massey's method
ABSTRACT: We propose and throughly investigate a temporalized version of the popular
Massey's technique for rating actors in sport competitions. The method can be
described as a dynamic temporal process in which team ratings are updated at
every match according to their performance during the match and the strength of
the opponent team. Using the Italian soccer dataset, we empirically show that
the method has a good foresight prediction accuracy.
| no_new_dataset | 0.952353 |
1702.00619 | Tarcisio Souza Costa | Tarcisio Souza and Elena Demidova and Thomas Risse and Helge Holzmann
and Gerhard Gossen and Julian Szymanski | Semantic URL Analytics to Support Efficient Annotation of Large Scale
Web Archives | null | null | 10.1007/978-3-319-27932-9_14 | null | cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Long-term Web archives comprise Web documents gathered over longer time
periods and can easily reach hundreds of terabytes in size. Semantic
annotations such as named entities can facilitate intelligent access to the Web
archive data. However, the annotation of the entire archive content on this
scale is often infeasible. The most efficient way to access the documents
within Web archives is provided through their URLs, which are typically stored
in dedicated index files.The URLs of the archived Web documents can contain
semantic information and can offer an efficient way to obtain initial semantic
annotations for the archived documents. In this paper, we analyse the
applicability of semantic analysis techniques such as named entity extraction
to the URLs in a Web archive. We evaluate the precision of the named entity
extraction from the URLs in the Popular German Web dataset and analyse the
proportion of the archived URLs from 1,444 popular domains in the time interval
from 2000 to 2012 to which these techniques are applicable. Our results
demonstrate that named entity recognition can be successfully applied to a
large number of URLs in our Web archive and provide a good starting point to
efficiently annotate large scale collections of Web documents.
| [
{
"version": "v1",
"created": "Thu, 2 Feb 2017 11:09:53 GMT"
}
] | 2017-02-03T00:00:00 | [
[
"Souza",
"Tarcisio",
""
],
[
"Demidova",
"Elena",
""
],
[
"Risse",
"Thomas",
""
],
[
"Holzmann",
"Helge",
""
],
[
"Gossen",
"Gerhard",
""
],
[
"Szymanski",
"Julian",
""
]
] | TITLE: Semantic URL Analytics to Support Efficient Annotation of Large Scale
Web Archives
ABSTRACT: Long-term Web archives comprise Web documents gathered over longer time
periods and can easily reach hundreds of terabytes in size. Semantic
annotations such as named entities can facilitate intelligent access to the Web
archive data. However, the annotation of the entire archive content on this
scale is often infeasible. The most efficient way to access the documents
within Web archives is provided through their URLs, which are typically stored
in dedicated index files.The URLs of the archived Web documents can contain
semantic information and can offer an efficient way to obtain initial semantic
annotations for the archived documents. In this paper, we analyse the
applicability of semantic analysis techniques such as named entity extraction
to the URLs in a Web archive. We evaluate the precision of the named entity
extraction from the URLs in the Popular German Web dataset and analyse the
proportion of the archived URLs from 1,444 popular domains in the time interval
from 2000 to 2012 to which these techniques are applicable. Our results
demonstrate that named entity recognition can be successfully applied to a
large number of URLs in our Web archive and provide a good starting point to
efficiently annotate large scale collections of Web documents.
| no_new_dataset | 0.95418 |
1602.07480 | Lluis Gomez | Lluis Gomez, Anguelos Nicolaou, Dimosthenis Karatzas | Improving patch-based scene text script identification with ensembles of
conjoined networks | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper focuses on the problem of script identification in scene text
images. Facing this problem with state of the art CNN classifiers is not
straightforward, as they fail to address a key characteristic of scene text
instances: their extremely variable aspect ratio. Instead of resizing input
images to a fixed aspect ratio as in the typical use of holistic CNN
classifiers, we propose here a patch-based classification framework in order to
preserve discriminative parts of the image that are characteristic of its
class. We describe a novel method based on the use of ensembles of conjoined
networks to jointly learn discriminative stroke-parts representations and their
relative importance in a patch-based classification scheme. Our experiments
with this learning procedure demonstrate state-of-the-art results in two public
script identification datasets. In addition, we propose a new public benchmark
dataset for the evaluation of multi-lingual scene text end-to-end reading
systems. Experiments done in this dataset demonstrate the key role of script
identification in a complete end-to-end system that combines our script
identification method with a previously published text detector and an
off-the-shelf OCR engine.
| [
{
"version": "v1",
"created": "Wed, 24 Feb 2016 12:33:25 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Feb 2017 13:17:57 GMT"
}
] | 2017-02-02T00:00:00 | [
[
"Gomez",
"Lluis",
""
],
[
"Nicolaou",
"Anguelos",
""
],
[
"Karatzas",
"Dimosthenis",
""
]
] | TITLE: Improving patch-based scene text script identification with ensembles of
conjoined networks
ABSTRACT: This paper focuses on the problem of script identification in scene text
images. Facing this problem with state of the art CNN classifiers is not
straightforward, as they fail to address a key characteristic of scene text
instances: their extremely variable aspect ratio. Instead of resizing input
images to a fixed aspect ratio as in the typical use of holistic CNN
classifiers, we propose here a patch-based classification framework in order to
preserve discriminative parts of the image that are characteristic of its
class. We describe a novel method based on the use of ensembles of conjoined
networks to jointly learn discriminative stroke-parts representations and their
relative importance in a patch-based classification scheme. Our experiments
with this learning procedure demonstrate state-of-the-art results in two public
script identification datasets. In addition, we propose a new public benchmark
dataset for the evaluation of multi-lingual scene text end-to-end reading
systems. Experiments done in this dataset demonstrate the key role of script
identification in a complete end-to-end system that combines our script
identification method with a previously published text detector and an
off-the-shelf OCR engine.
| new_dataset | 0.969469 |
1604.02619 | Lluis Gomez | Lluis Gomez-Bigorda and Dimosthenis Karatzas | TextProposals: a Text-specific Selective Search Algorithm for Word
Spotting in the Wild | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Motivated by the success of powerful while expensive techniques to recognize
words in a holistic way, object proposals techniques emerge as an alternative
to the traditional text detectors. In this paper we introduce a novel object
proposals method that is specifically designed for text. We rely on a
similarity based region grouping algorithm that generates a hierarchy of word
hypotheses. Over the nodes of this hierarchy it is possible to apply a holistic
word recognition method in an efficient way.
Our experiments demonstrate that the presented method is superior in its
ability of producing good quality word proposals when compared with
class-independent algorithms. We show impressive recall rates with a few
thousand proposals in different standard benchmarks, including focused or
incidental text datasets, and multi-language scenarios. Moreover, the
combination of our object proposals with existing whole-word recognizers shows
competitive performance in end-to-end word spotting, and, in some benchmarks,
outperforms previously published results. Concretely, in the challenging
ICDAR2015 Incidental Text dataset, we overcome in more than 10 percent f-score
the best-performing method in the last ICDAR Robust Reading Competition. Source
code of the complete end-to-end system is available at
https://github.com/lluisgomez/TextProposals
| [
{
"version": "v1",
"created": "Sun, 10 Apr 2016 00:03:16 GMT"
},
{
"version": "v2",
"created": "Tue, 24 May 2016 17:03:13 GMT"
},
{
"version": "v3",
"created": "Wed, 1 Feb 2017 15:35:28 GMT"
}
] | 2017-02-02T00:00:00 | [
[
"Gomez-Bigorda",
"Lluis",
""
],
[
"Karatzas",
"Dimosthenis",
""
]
] | TITLE: TextProposals: a Text-specific Selective Search Algorithm for Word
Spotting in the Wild
ABSTRACT: Motivated by the success of powerful while expensive techniques to recognize
words in a holistic way, object proposals techniques emerge as an alternative
to the traditional text detectors. In this paper we introduce a novel object
proposals method that is specifically designed for text. We rely on a
similarity based region grouping algorithm that generates a hierarchy of word
hypotheses. Over the nodes of this hierarchy it is possible to apply a holistic
word recognition method in an efficient way.
Our experiments demonstrate that the presented method is superior in its
ability of producing good quality word proposals when compared with
class-independent algorithms. We show impressive recall rates with a few
thousand proposals in different standard benchmarks, including focused or
incidental text datasets, and multi-language scenarios. Moreover, the
combination of our object proposals with existing whole-word recognizers shows
competitive performance in end-to-end word spotting, and, in some benchmarks,
outperforms previously published results. Concretely, in the challenging
ICDAR2015 Incidental Text dataset, we overcome in more than 10 percent f-score
the best-performing method in the last ICDAR Robust Reading Competition. Source
code of the complete end-to-end system is available at
https://github.com/lluisgomez/TextProposals
| new_dataset | 0.877161 |
1701.09049 | Amit Awekar | Panthadeep Bhattacharjee and Amit Awekar | Batch Incremental Shared Nearest Neighbor Density Based Clustering
Algorithm for Dynamic Datasets | 6 pages, Accepted at ECIR 2017 | null | null | null | cs.DB cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Incremental data mining algorithms process frequent updates to dynamic
datasets efficiently by avoiding redundant computation. Existing incremental
extension to shared nearest neighbor density based clustering (SNND) algorithm
cannot handle deletions to dataset and handles insertions only one point at a
time. We present an incremental algorithm to overcome both these bottlenecks by
efficiently identifying affected parts of clusters while processing updates to
dataset in batch mode. We show effectiveness of our algorithm by performing
experiments on large synthetic as well as real world datasets. Our algorithm is
up to four orders of magnitude faster than SNND and requires up to 60% extra
memory than SNND while providing output identical to SNND.
| [
{
"version": "v1",
"created": "Tue, 31 Jan 2017 14:19:18 GMT"
}
] | 2017-02-02T00:00:00 | [
[
"Bhattacharjee",
"Panthadeep",
""
],
[
"Awekar",
"Amit",
""
]
] | TITLE: Batch Incremental Shared Nearest Neighbor Density Based Clustering
Algorithm for Dynamic Datasets
ABSTRACT: Incremental data mining algorithms process frequent updates to dynamic
datasets efficiently by avoiding redundant computation. Existing incremental
extension to shared nearest neighbor density based clustering (SNND) algorithm
cannot handle deletions to dataset and handles insertions only one point at a
time. We present an incremental algorithm to overcome both these bottlenecks by
efficiently identifying affected parts of clusters while processing updates to
dataset in batch mode. We show effectiveness of our algorithm by performing
experiments on large synthetic as well as real world datasets. Our algorithm is
up to four orders of magnitude faster than SNND and requires up to 60% extra
memory than SNND while providing output identical to SNND.
| no_new_dataset | 0.949949 |
1702.00025 | Rainer Kelz | Rainer Kelz and Gerhard Widmer | An Experimental Analysis of the Entanglement Problem in
Neural-Network-based Music Transcription Systems | Submitted to AES Conference on Semantic Audio, Erlangen, Germany,
2017 June 22, 24 | null | null | null | cs.SD | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Several recent polyphonic music transcription systems have utilized deep
neural networks to achieve state of the art results on various benchmark
datasets, pushing the envelope on framewise and note-level performance
measures. Unfortunately we can observe a sort of glass ceiling effect. To
investigate this effect, we provide a detailed analysis of the particular kinds
of errors that state of the art deep neural transcription systems make, when
trained and tested on a piano transcription task. We are ultimately forced to
draw a rather disheartening conclusion: the networks seem to learn combinations
of notes, and have a hard time generalizing to unseen combinations of notes.
Furthermore, we speculate on various means to alleviate this situation.
| [
{
"version": "v1",
"created": "Tue, 31 Jan 2017 19:21:41 GMT"
}
] | 2017-02-02T00:00:00 | [
[
"Kelz",
"Rainer",
""
],
[
"Widmer",
"Gerhard",
""
]
] | TITLE: An Experimental Analysis of the Entanglement Problem in
Neural-Network-based Music Transcription Systems
ABSTRACT: Several recent polyphonic music transcription systems have utilized deep
neural networks to achieve state of the art results on various benchmark
datasets, pushing the envelope on framewise and note-level performance
measures. Unfortunately we can observe a sort of glass ceiling effect. To
investigate this effect, we provide a detailed analysis of the particular kinds
of errors that state of the art deep neural transcription systems make, when
trained and tested on a piano transcription task. We are ultimately forced to
draw a rather disheartening conclusion: the networks seem to learn combinations
of notes, and have a hard time generalizing to unseen combinations of notes.
Furthermore, we speculate on various means to alleviate this situation.
| no_new_dataset | 0.949482 |
1702.00045 | Le Lu | Holger R. Roth, Le Lu, Nathan Lay, Adam P. Harrison, Amal Farag,
Andrew Sohn, Ronald M. Summers | Spatial Aggregation of Holistically-Nested Convolutional Neural Networks
for Automated Pancreas Localization and Segmentation | This version was submitted to IEEE Trans. on Medical Imaging on Dec.
18th, 2016. The content of this article is covered by US Patent Applications
of 62/345,606# and 62/450,681# | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Accurate and automatic organ segmentation from 3D radiological scans is an
important yet challenging problem for medical image analysis. Specifically, the
pancreas demonstrates very high inter-patient anatomical variability in both
its shape and volume. In this paper, we present an automated system using 3D
computed tomography (CT) volumes via a two-stage cascaded approach: pancreas
localization and segmentation. For the first step, we localize the pancreas
from the entire 3D CT scan, providing a reliable bounding box for the more
refined segmentation step. We introduce a fully deep-learning approach, based
on an efficient application of holistically-nested convolutional networks
(HNNs) on the three orthogonal axial, sagittal, and coronal views. The
resulting HNN per-pixel probability maps are then fused using pooling to
reliably produce a 3D bounding box of the pancreas that maximizes the recall.
We show that our introduced localizer compares favorably to both a conventional
non-deep-learning method and a recent hybrid approach based on spatial
aggregation of superpixels using random forest classification. The second,
segmentation, phase operates within the computed bounding box and integrates
semantic mid-level cues of deeply-learned organ interior and boundary maps,
obtained by two additional and separate realizations of HNNs. By integrating
these two mid-level cues, our method is capable of generating
boundary-preserving pixel-wise class label maps that result in the final
pancreas segmentation. Quantitative evaluation is performed on a publicly
available dataset of 82 patient CT scans using 4-fold cross-validation (CV). We
achieve a Dice similarity coefficient (DSC) of 81.27+/-6.27% in validation,
which significantly outperforms previous state-of-the art methods that report
DSCs of 71.80+/-10.70% and 78.01+/-8.20%, respectively, using the same dataset.
| [
{
"version": "v1",
"created": "Tue, 31 Jan 2017 20:22:15 GMT"
}
] | 2017-02-02T00:00:00 | [
[
"Roth",
"Holger R.",
""
],
[
"Lu",
"Le",
""
],
[
"Lay",
"Nathan",
""
],
[
"Harrison",
"Adam P.",
""
],
[
"Farag",
"Amal",
""
],
[
"Sohn",
"Andrew",
""
],
[
"Summers",
"Ronald M.",
""
]
] | TITLE: Spatial Aggregation of Holistically-Nested Convolutional Neural Networks
for Automated Pancreas Localization and Segmentation
ABSTRACT: Accurate and automatic organ segmentation from 3D radiological scans is an
important yet challenging problem for medical image analysis. Specifically, the
pancreas demonstrates very high inter-patient anatomical variability in both
its shape and volume. In this paper, we present an automated system using 3D
computed tomography (CT) volumes via a two-stage cascaded approach: pancreas
localization and segmentation. For the first step, we localize the pancreas
from the entire 3D CT scan, providing a reliable bounding box for the more
refined segmentation step. We introduce a fully deep-learning approach, based
on an efficient application of holistically-nested convolutional networks
(HNNs) on the three orthogonal axial, sagittal, and coronal views. The
resulting HNN per-pixel probability maps are then fused using pooling to
reliably produce a 3D bounding box of the pancreas that maximizes the recall.
We show that our introduced localizer compares favorably to both a conventional
non-deep-learning method and a recent hybrid approach based on spatial
aggregation of superpixels using random forest classification. The second,
segmentation, phase operates within the computed bounding box and integrates
semantic mid-level cues of deeply-learned organ interior and boundary maps,
obtained by two additional and separate realizations of HNNs. By integrating
these two mid-level cues, our method is capable of generating
boundary-preserving pixel-wise class label maps that result in the final
pancreas segmentation. Quantitative evaluation is performed on a publicly
available dataset of 82 patient CT scans using 4-fold cross-validation (CV). We
achieve a Dice similarity coefficient (DSC) of 81.27+/-6.27% in validation,
which significantly outperforms previous state-of-the art methods that report
DSCs of 71.80+/-10.70% and 78.01+/-8.20%, respectively, using the same dataset.
| no_new_dataset | 0.95096 |
1702.00158 | Xiaqing Pan | Xiaqing Pan, Yueru Chen, C.-C. Jay Kuo | Design, Analysis and Application of A Volumetric Convolutional Neural
Network | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The design, analysis and application of a volumetric convolutional neural
network (VCNN) are studied in this work. Although many CNNs have been proposed
in the literature, their design is empirical. In the design of the VCNN, we
propose a feed-forward K-means clustering algorithm to determine the filter
number and size at each convolutional layer systematically. For the analysis of
the VCNN, the cause of confusing classes in the output of the VCNN is explained
by analyzing the relationship between the filter weights (also known as anchor
vectors) from the last fully-connected layer to the output. Furthermore, a
hierarchical clustering method followed by a random forest classification
method is proposed to boost the classification performance among confusing
classes. For the application of the VCNN, we examine the 3D shape
classification problem and conduct experiments on a popular ModelNet40 dataset.
The proposed VCNN offers the state-of-the-art performance among all
volume-based CNN methods.
| [
{
"version": "v1",
"created": "Wed, 1 Feb 2017 08:32:11 GMT"
}
] | 2017-02-02T00:00:00 | [
[
"Pan",
"Xiaqing",
""
],
[
"Chen",
"Yueru",
""
],
[
"Kuo",
"C. -C. Jay",
""
]
] | TITLE: Design, Analysis and Application of A Volumetric Convolutional Neural
Network
ABSTRACT: The design, analysis and application of a volumetric convolutional neural
network (VCNN) are studied in this work. Although many CNNs have been proposed
in the literature, their design is empirical. In the design of the VCNN, we
propose a feed-forward K-means clustering algorithm to determine the filter
number and size at each convolutional layer systematically. For the analysis of
the VCNN, the cause of confusing classes in the output of the VCNN is explained
by analyzing the relationship between the filter weights (also known as anchor
vectors) from the last fully-connected layer to the output. Furthermore, a
hierarchical clustering method followed by a random forest classification
method is proposed to boost the classification performance among confusing
classes. For the application of the VCNN, we examine the 3D shape
classification problem and conduct experiments on a popular ModelNet40 dataset.
The proposed VCNN offers the state-of-the-art performance among all
volume-based CNN methods.
| no_new_dataset | 0.949763 |
1702.00196 | He Sun | Jiecao Chen and He Sun and David P. Woodruff and Qin Zhang | Communication-Optimal Distributed Clustering | A preliminary version of this paper appeared at the 30th Annual
Conference on Neural Information Processing Systems (NIPS), 2016 | null | null | null | cs.DS cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Clustering large datasets is a fundamental problem with a number of
applications in machine learning. Data is often collected on different sites
and clustering needs to be performed in a distributed manner with low
communication. We would like the quality of the clustering in the distributed
setting to match that in the centralized setting for which all the data resides
on a single site. In this work, we study both graph and geometric clustering
problems in two distributed models: (1) a point-to-point model, and (2) a model
with a broadcast channel. We give protocols in both models which we show are
nearly optimal by proving almost matching communication lower bounds. Our work
highlights the surprising power of a broadcast channel for clustering problems;
roughly speaking, to spectrally cluster $n$ points or $n$ vertices in a graph
distributed across $s$ servers, for a worst-case partitioning the communication
complexity in a point-to-point model is $n \cdot s$, while in the broadcast
model it is $n + s$. A similar phenomenon holds for the geometric setting as
well. We implement our algorithms and demonstrate this phenomenon on real life
datasets, showing that our algorithms are also very efficient in practice.
| [
{
"version": "v1",
"created": "Wed, 1 Feb 2017 10:30:32 GMT"
}
] | 2017-02-02T00:00:00 | [
[
"Chen",
"Jiecao",
""
],
[
"Sun",
"He",
""
],
[
"Woodruff",
"David P.",
""
],
[
"Zhang",
"Qin",
""
]
] | TITLE: Communication-Optimal Distributed Clustering
ABSTRACT: Clustering large datasets is a fundamental problem with a number of
applications in machine learning. Data is often collected on different sites
and clustering needs to be performed in a distributed manner with low
communication. We would like the quality of the clustering in the distributed
setting to match that in the centralized setting for which all the data resides
on a single site. In this work, we study both graph and geometric clustering
problems in two distributed models: (1) a point-to-point model, and (2) a model
with a broadcast channel. We give protocols in both models which we show are
nearly optimal by proving almost matching communication lower bounds. Our work
highlights the surprising power of a broadcast channel for clustering problems;
roughly speaking, to spectrally cluster $n$ points or $n$ vertices in a graph
distributed across $s$ servers, for a worst-case partitioning the communication
complexity in a point-to-point model is $n \cdot s$, while in the broadcast
model it is $n + s$. A similar phenomenon holds for the geometric setting as
well. We implement our algorithms and demonstrate this phenomenon on real life
datasets, showing that our algorithms are also very efficient in practice.
| no_new_dataset | 0.955152 |
1702.00338 | Eng-Jon Ong | Eng-Jon Ong and Sameed Husain and Miroslaw Bober | Siamese Network of Deep Fisher-Vector Descriptors for Image Retrieval | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper addresses the problem of large scale image retrieval, with the aim
of accurately ranking the similarity of a large number of images to a given
query image. To achieve this, we propose a novel Siamese network. This network
consists of two computational strands, each comprising of a CNN component
followed by a Fisher vector component. The CNN component produces dense, deep
convolutional descriptors that are then aggregated by the Fisher Vector method.
Crucially, we propose to simultaneously learn both the CNN filter weights and
Fisher Vector model parameters. This allows us to account for the evolving
distribution of deep descriptors over the course of the learning process. We
show that the proposed approach gives significant improvements over the
state-of-the-art methods on the Oxford and Paris image retrieval datasets.
Additionally, we provide a baseline performance measure for both these datasets
with the inclusion of 1 million distractors.
| [
{
"version": "v1",
"created": "Wed, 1 Feb 2017 16:20:00 GMT"
}
] | 2017-02-02T00:00:00 | [
[
"Ong",
"Eng-Jon",
""
],
[
"Husain",
"Sameed",
""
],
[
"Bober",
"Miroslaw",
""
]
] | TITLE: Siamese Network of Deep Fisher-Vector Descriptors for Image Retrieval
ABSTRACT: This paper addresses the problem of large scale image retrieval, with the aim
of accurately ranking the similarity of a large number of images to a given
query image. To achieve this, we propose a novel Siamese network. This network
consists of two computational strands, each comprising of a CNN component
followed by a Fisher vector component. The CNN component produces dense, deep
convolutional descriptors that are then aggregated by the Fisher Vector method.
Crucially, we propose to simultaneously learn both the CNN filter weights and
Fisher Vector model parameters. This allows us to account for the evolving
distribution of deep descriptors over the course of the learning process. We
show that the proposed approach gives significant improvements over the
state-of-the-art methods on the Oxford and Paris image retrieval datasets.
Additionally, we provide a baseline performance measure for both these datasets
with the inclusion of 1 million distractors.
| no_new_dataset | 0.948489 |
1702.00358 | Florin Rusu | Yu Cheng, Weijie Zhao, Florin Rusu | OLA-RAW: Scalable Exploration over Raw Data | null | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In-situ processing has been proposed as a novel data exploration solution in
many domains generating massive amounts of raw data, e.g., astronomy, since it
provides immediate SQL querying over raw files. The performance of in-situ
processing across a query workload is, however, limited by the speed of full
scan, tokenizing, and parsing of the entire data. Online aggregation (OLA) has
been introduced as an efficient method for data exploration that identifies
uninteresting patterns faster by continuously estimating the result of a
computation during the actual processing---the computation can be stopped as
early as the estimate is accurate enough to be deemed uninteresting. However,
existing OLA solutions have a high upfront cost of randomly shuffling and/or
sampling the data. In this paper, we present OLA-RAW, a bi-level sampling
scheme for parallel online aggregation over raw data. Sampling in OLA-RAW is
query-driven and performed exclusively in-situ during the runtime query
execution, without data reorganization. This is realized by a novel
resource-aware bi-level sampling algorithm that processes data in random chunks
concurrently and determines adaptively the number of sampled tuples inside a
chunk. In order to avoid the cost of repetitive conversion from raw data,
OLA-RAW builds and maintains a memory-resident bi-level sample synopsis
incrementally. We implement OLA-RAW inside a modern in-situ data processing
system and evaluate its performance across several real and synthetic datasets
and file formats. Our results show that OLA-RAW chooses the sampling plan that
minimizes the execution time and guarantees the required accuracy for each
query in a given workload. The end result is a focused data exploration process
that avoids unnecessary work and discards uninteresting data.
| [
{
"version": "v1",
"created": "Wed, 1 Feb 2017 17:07:56 GMT"
}
] | 2017-02-02T00:00:00 | [
[
"Cheng",
"Yu",
""
],
[
"Zhao",
"Weijie",
""
],
[
"Rusu",
"Florin",
""
]
] | TITLE: OLA-RAW: Scalable Exploration over Raw Data
ABSTRACT: In-situ processing has been proposed as a novel data exploration solution in
many domains generating massive amounts of raw data, e.g., astronomy, since it
provides immediate SQL querying over raw files. The performance of in-situ
processing across a query workload is, however, limited by the speed of full
scan, tokenizing, and parsing of the entire data. Online aggregation (OLA) has
been introduced as an efficient method for data exploration that identifies
uninteresting patterns faster by continuously estimating the result of a
computation during the actual processing---the computation can be stopped as
early as the estimate is accurate enough to be deemed uninteresting. However,
existing OLA solutions have a high upfront cost of randomly shuffling and/or
sampling the data. In this paper, we present OLA-RAW, a bi-level sampling
scheme for parallel online aggregation over raw data. Sampling in OLA-RAW is
query-driven and performed exclusively in-situ during the runtime query
execution, without data reorganization. This is realized by a novel
resource-aware bi-level sampling algorithm that processes data in random chunks
concurrently and determines adaptively the number of sampled tuples inside a
chunk. In order to avoid the cost of repetitive conversion from raw data,
OLA-RAW builds and maintains a memory-resident bi-level sample synopsis
incrementally. We implement OLA-RAW inside a modern in-situ data processing
system and evaluate its performance across several real and synthetic datasets
and file formats. Our results show that OLA-RAW chooses the sampling plan that
minimizes the execution time and guarantees the required accuracy for each
query in a given workload. The end result is a focused data exploration process
that avoids unnecessary work and discards uninteresting data.
| no_new_dataset | 0.951369 |
1608.07327 | Bo Qu | Bo Qu and Huijuan Wang | SIS Epidemic Spreading with Correlated Heterogeneous Infection Rates | null | null | 10.1016/j.physa.2016.12.077 | null | physics.soc-ph q-bio.PE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The epidemic spreading has been widely studied when each node may get
infected by an infected neighbor with the same rate. However, the infection
rate between a pair of nodes is usually heterogeneous and even correlated with
their nodal degrees in the contact network. We aim to understand how such
correlated heterogeneous infection rates influence the spreading on different
network topologies. Motivated by real-world datasets, we propose a correlated
heterogeneous Susceptible-Infected-Susceptible model which assumes that the
infection rate $\beta_{ij}(=\beta_{ji})$ between node $i$ and $j$ is correlated
with the degree of the two end nodes: $\beta_{ij}=c(d_id_j)^\alpha$, where
$\alpha$ indicates the strength of the correlation and $c$ is selected so that
the average infection rate is $1$. In order to understand the effect of such
correlation on epidemic spreading, we consider as well the corresponding
uncorrected but still heterogeneous infection rate scenario as a reference,
where the original correlated infection rates in our CSIS model are shuffled
and reallocated to the links of the same network topology. We compare these two
scenarios in the average fraction of infected nodes in the metastable state on
Erd{\"o}s-R{\'e}nyi (ER) and scale-free (SF) networks with a similar average
degree. Through the continuous-time simulations, we find that, when the
recovery rate is small, the negative correlation is more likely to help the
epidemic spread and the positive correlation prohibit the spreading; as the
recovery rate increases to be larger than a critical value, the positive but
not negative correlation tends to help the spreading. Our findings are further
analytically proved in a wheel network (one central node connects with each of
the nodes in a ring) and validated on real-world networks with correlated
heterogeneous interaction frequencies.
| [
{
"version": "v1",
"created": "Thu, 25 Aug 2016 22:33:50 GMT"
}
] | 2017-02-01T00:00:00 | [
[
"Qu",
"Bo",
""
],
[
"Wang",
"Huijuan",
""
]
] | TITLE: SIS Epidemic Spreading with Correlated Heterogeneous Infection Rates
ABSTRACT: The epidemic spreading has been widely studied when each node may get
infected by an infected neighbor with the same rate. However, the infection
rate between a pair of nodes is usually heterogeneous and even correlated with
their nodal degrees in the contact network. We aim to understand how such
correlated heterogeneous infection rates influence the spreading on different
network topologies. Motivated by real-world datasets, we propose a correlated
heterogeneous Susceptible-Infected-Susceptible model which assumes that the
infection rate $\beta_{ij}(=\beta_{ji})$ between node $i$ and $j$ is correlated
with the degree of the two end nodes: $\beta_{ij}=c(d_id_j)^\alpha$, where
$\alpha$ indicates the strength of the correlation and $c$ is selected so that
the average infection rate is $1$. In order to understand the effect of such
correlation on epidemic spreading, we consider as well the corresponding
uncorrected but still heterogeneous infection rate scenario as a reference,
where the original correlated infection rates in our CSIS model are shuffled
and reallocated to the links of the same network topology. We compare these two
scenarios in the average fraction of infected nodes in the metastable state on
Erd{\"o}s-R{\'e}nyi (ER) and scale-free (SF) networks with a similar average
degree. Through the continuous-time simulations, we find that, when the
recovery rate is small, the negative correlation is more likely to help the
epidemic spread and the positive correlation prohibit the spreading; as the
recovery rate increases to be larger than a critical value, the positive but
not negative correlation tends to help the spreading. Our findings are further
analytically proved in a wheel network (one central node connects with each of
the nodes in a ring) and validated on real-world networks with correlated
heterogeneous interaction frequencies.
| no_new_dataset | 0.954308 |
1609.09405 | Siva Reddy | Yonatan Bisk, Siva Reddy, John Blitzer, Julia Hockenmaier, Mark
Steedman | Evaluating Induced CCG Parsers on Grounded Semantic Parsing | EMNLP 2016, Table 2 erratum, Code and Freebase Semantic Parsing data
URL | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | We compare the effectiveness of four different syntactic CCG parsers for a
semantic slot-filling task to explore how much syntactic supervision is
required for downstream semantic analysis. This extrinsic, task-based
evaluation provides a unique window to explore the strengths and weaknesses of
semantics captured by unsupervised grammar induction systems. We release a new
Freebase semantic parsing dataset called SPADES (Semantic PArsing of
DEclarative Sentences) containing 93K cloze-style questions paired with
answers. We evaluate all our models on this dataset. Our code and data are
available at https://github.com/sivareddyg/graph-parser.
| [
{
"version": "v1",
"created": "Thu, 29 Sep 2016 16:09:29 GMT"
},
{
"version": "v2",
"created": "Tue, 31 Jan 2017 16:25:39 GMT"
}
] | 2017-02-01T00:00:00 | [
[
"Bisk",
"Yonatan",
""
],
[
"Reddy",
"Siva",
""
],
[
"Blitzer",
"John",
""
],
[
"Hockenmaier",
"Julia",
""
],
[
"Steedman",
"Mark",
""
]
] | TITLE: Evaluating Induced CCG Parsers on Grounded Semantic Parsing
ABSTRACT: We compare the effectiveness of four different syntactic CCG parsers for a
semantic slot-filling task to explore how much syntactic supervision is
required for downstream semantic analysis. This extrinsic, task-based
evaluation provides a unique window to explore the strengths and weaknesses of
semantics captured by unsupervised grammar induction systems. We release a new
Freebase semantic parsing dataset called SPADES (Semantic PArsing of
DEclarative Sentences) containing 93K cloze-style questions paired with
answers. We evaluate all our models on this dataset. Our code and data are
available at https://github.com/sivareddyg/graph-parser.
| new_dataset | 0.954052 |
1701.07847 | Dmitry Petrov | Dmitry Petrov, Boris Gutman, Alexander Ivanov, Joshua Faskowitz, Neda
Jahanshad, Mikhail Belyaev, Paul Thompson | Structural Connectome Validation Using Pairwise Classification | Accepted for IEEE International Symposium on Biomedical Imaging 2017 | null | null | null | q-bio.NC cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work, we study the extent to which structural connectomes and
topological derivative measures are unique to individual changes within human
brains. To do so, we classify structural connectome pairs from two large
longitudinal datasets as either belonging to the same individual or not. Our
data is comprised of 227 individuals from the Alzheimer's Disease Neuroimaging
Initiative (ADNI) and 226 from the Parkinson's Progression Markers Initiative
(PPMI). We achieve 0.99 area under the ROC curve score for features which
represent either weights or network structure of the connectomes (node degrees,
PageRank and local efficiency). Our approach may be useful for eliminating
noisy features as a preprocessing step in brain aging studies and early
diagnosis classification problems.
| [
{
"version": "v1",
"created": "Thu, 26 Jan 2017 19:13:36 GMT"
},
{
"version": "v2",
"created": "Mon, 30 Jan 2017 19:55:15 GMT"
}
] | 2017-02-01T00:00:00 | [
[
"Petrov",
"Dmitry",
""
],
[
"Gutman",
"Boris",
""
],
[
"Ivanov",
"Alexander",
""
],
[
"Faskowitz",
"Joshua",
""
],
[
"Jahanshad",
"Neda",
""
],
[
"Belyaev",
"Mikhail",
""
],
[
"Thompson",
"Paul",
""
]
] | TITLE: Structural Connectome Validation Using Pairwise Classification
ABSTRACT: In this work, we study the extent to which structural connectomes and
topological derivative measures are unique to individual changes within human
brains. To do so, we classify structural connectome pairs from two large
longitudinal datasets as either belonging to the same individual or not. Our
data is comprised of 227 individuals from the Alzheimer's Disease Neuroimaging
Initiative (ADNI) and 226 from the Parkinson's Progression Markers Initiative
(PPMI). We achieve 0.99 area under the ROC curve score for features which
represent either weights or network structure of the connectomes (node degrees,
PageRank and local efficiency). Our approach may be useful for eliminating
noisy features as a preprocessing step in brain aging studies and early
diagnosis classification problems.
| no_new_dataset | 0.953579 |
1701.08302 | A Mani | Mani A and Rebeka Mukherjee | A Study of FOSS'2013 Survey Data Using Clustering Techniques | IEEE Women in Engineering Conference Paper: WIECON-ECE'2016
(Scheduled to appear in IEEE Xplore ) | null | null | null | cs.AI cs.CY cs.SE stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | FOSS is an acronym for Free and Open Source Software. The FOSS 2013 survey
primarily targets FOSS contributors and relevant anonymized dataset is publicly
available under CC by SA license. In this study, the dataset is analyzed from a
critical perspective using statistical and clustering techniques (especially
multiple correspondence analysis) with a strong focus on women contributors
towards discovering hidden trends and facts. Important inferences are drawn
about development practices and other facets of the free software and OSS
worlds.
| [
{
"version": "v1",
"created": "Sat, 28 Jan 2017 16:52:13 GMT"
},
{
"version": "v2",
"created": "Tue, 31 Jan 2017 17:18:01 GMT"
}
] | 2017-02-01T00:00:00 | [
[
"A",
"Mani",
""
],
[
"Mukherjee",
"Rebeka",
""
]
] | TITLE: A Study of FOSS'2013 Survey Data Using Clustering Techniques
ABSTRACT: FOSS is an acronym for Free and Open Source Software. The FOSS 2013 survey
primarily targets FOSS contributors and relevant anonymized dataset is publicly
available under CC by SA license. In this study, the dataset is analyzed from a
critical perspective using statistical and clustering techniques (especially
multiple correspondence analysis) with a strong focus on women contributors
towards discovering hidden trends and facts. Important inferences are drawn
about development practices and other facets of the free software and OSS
worlds.
| no_new_dataset | 0.950732 |
1701.08757 | Dmitry Ignatov | Mikhail V. Goubko and Sergey O. Kuznetsov and Alexey A. Neznanov and
Dmitry I. Ignatov | Bayesian Learning of Consumer Preferences for Residential Demand
Response | null | IFAC-PapersOnLine, 49(32), 2016, p. 24-29, ISSN 2405-8963 | 10.1016/j.ifacol.2016.12.184 | null | cs.LG cs.SY stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In coming years residential consumers will face real-time electricity tariffs
with energy prices varying day to day, and effective energy saving will require
automation - a recommender system, which learns consumer's preferences from her
actions. A consumer chooses a scenario of home appliance use to balance her
comfort level and the energy bill. We propose a Bayesian learning algorithm to
estimate the comfort level function from the history of appliance use. In
numeric experiments with datasets generated from a simulation model of a
consumer interacting with small home appliances the algorithm outperforms
popular regression analysis tools. Our approach can be extended to control an
air heating and conditioning system, which is responsible for up to half of a
household's energy bill.
| [
{
"version": "v1",
"created": "Fri, 27 Jan 2017 20:45:31 GMT"
}
] | 2017-02-01T00:00:00 | [
[
"Goubko",
"Mikhail V.",
""
],
[
"Kuznetsov",
"Sergey O.",
""
],
[
"Neznanov",
"Alexey A.",
""
],
[
"Ignatov",
"Dmitry I.",
""
]
] | TITLE: Bayesian Learning of Consumer Preferences for Residential Demand
Response
ABSTRACT: In coming years residential consumers will face real-time electricity tariffs
with energy prices varying day to day, and effective energy saving will require
automation - a recommender system, which learns consumer's preferences from her
actions. A consumer chooses a scenario of home appliance use to balance her
comfort level and the energy bill. We propose a Bayesian learning algorithm to
estimate the comfort level function from the history of appliance use. In
numeric experiments with datasets generated from a simulation model of a
consumer interacting with small home appliances the algorithm outperforms
popular regression analysis tools. Our approach can be extended to control an
air heating and conditioning system, which is responsible for up to half of a
household's energy bill.
| no_new_dataset | 0.942507 |
1701.08799 | Alan Kuhnle | Alan Kuhnle, Tianyi Pan, Md Abdul Alim, My T. Thai | Scalable Bicriteria Algorithms for the Threshold Activation Problem in
Online Social Networks | null | null | null | null | cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider the Threshold Activation Problem (TAP): given social network $G$
and positive threshold $T$, find a minimum-size seed set $A$ that can trigger
expected activation of at least $T$. We introduce the first scalable,
parallelizable algorithm with performance guarantee for TAP suitable for
datasets with millions of nodes and edges; we exploit the bicriteria nature of
solutions to TAP to allow the user to control the running time versus accuracy
of our algorithm through a parameter $\alpha \in (0,1)$: given $\eta > 0$, with
probability $1 - \eta$ our algorithm returns a solution $A$ with expected
activation greater than $T - 2 \alpha T$, and the size of the solution $A$ is
within factor $1 + 4 \alpha T + \log ( T )$ of the optimal size. The algorithm
runs in time $O \left( \alpha^{-2}\log \left( n / \eta \right) (n + m) |A|
\right)$, where $n$, $m$, refer to the number of nodes, edges in the network.
The performance guarantee holds for the general triggering model of internal
influence and also incorporates external influence, provided a certain
condition is met on the cost-effectivity of seed selection.
| [
{
"version": "v1",
"created": "Mon, 30 Jan 2017 19:52:25 GMT"
}
] | 2017-02-01T00:00:00 | [
[
"Kuhnle",
"Alan",
""
],
[
"Pan",
"Tianyi",
""
],
[
"Alim",
"Md Abdul",
""
],
[
"Thai",
"My T.",
""
]
] | TITLE: Scalable Bicriteria Algorithms for the Threshold Activation Problem in
Online Social Networks
ABSTRACT: We consider the Threshold Activation Problem (TAP): given social network $G$
and positive threshold $T$, find a minimum-size seed set $A$ that can trigger
expected activation of at least $T$. We introduce the first scalable,
parallelizable algorithm with performance guarantee for TAP suitable for
datasets with millions of nodes and edges; we exploit the bicriteria nature of
solutions to TAP to allow the user to control the running time versus accuracy
of our algorithm through a parameter $\alpha \in (0,1)$: given $\eta > 0$, with
probability $1 - \eta$ our algorithm returns a solution $A$ with expected
activation greater than $T - 2 \alpha T$, and the size of the solution $A$ is
within factor $1 + 4 \alpha T + \log ( T )$ of the optimal size. The algorithm
runs in time $O \left( \alpha^{-2}\log \left( n / \eta \right) (n + m) |A|
\right)$, where $n$, $m$, refer to the number of nodes, edges in the network.
The performance guarantee holds for the general triggering model of internal
influence and also incorporates external influence, provided a certain
condition is met on the cost-effectivity of seed selection.
| no_new_dataset | 0.938632 |
1701.08869 | Xiaqing Pan | Xiaqing Pan, Yueru Chen, C.-C. Jay Kuo | 3D Shape Retrieval via Irrelevance Filtering and Similarity Ranking
(IF/SR) | arXiv admin note: text overlap with arXiv:1603.01942 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A novel solution for the content-based 3D shape retrieval problem using an
unsupervised clustering approach, which does not need any label information of
3D shapes, is presented in this work. The proposed shape retrieval system
consists of two modules in cascade: the irrelevance filtering (IF) module and
the similarity ranking (SR) module. The IF module attempts to cluster gallery
shapes that are similar to each other by examining global and local features
simultaneously. However, shapes that are close in the local feature space can
be distant in the global feature space, and vice versa. To resolve this issue,
we propose a joint cost function that strikes a balance between two distances.
Irrelevant samples that are close in the local feature space but distant in the
global feature space can be removed in this stage. The remaining gallery
samples are ranked in the SR module using the local feature. The superior
performance of the proposed IF/SR method is demonstrated by extensive
experiments conducted on the popular SHREC12 dataset.
| [
{
"version": "v1",
"created": "Mon, 30 Jan 2017 23:04:57 GMT"
}
] | 2017-02-01T00:00:00 | [
[
"Pan",
"Xiaqing",
""
],
[
"Chen",
"Yueru",
""
],
[
"Kuo",
"C. -C. Jay",
""
]
] | TITLE: 3D Shape Retrieval via Irrelevance Filtering and Similarity Ranking
(IF/SR)
ABSTRACT: A novel solution for the content-based 3D shape retrieval problem using an
unsupervised clustering approach, which does not need any label information of
3D shapes, is presented in this work. The proposed shape retrieval system
consists of two modules in cascade: the irrelevance filtering (IF) module and
the similarity ranking (SR) module. The IF module attempts to cluster gallery
shapes that are similar to each other by examining global and local features
simultaneously. However, shapes that are close in the local feature space can
be distant in the global feature space, and vice versa. To resolve this issue,
we propose a joint cost function that strikes a balance between two distances.
Irrelevant samples that are close in the local feature space but distant in the
global feature space can be removed in this stage. The remaining gallery
samples are ranked in the SR module using the local feature. The superior
performance of the proposed IF/SR method is demonstrated by extensive
experiments conducted on the popular SHREC12 dataset.
| no_new_dataset | 0.950732 |
1701.08886 | Moustafa Alzantot | Moustafa Alzantot, Supriyo Chakraborty, Mani B. Srivastava | SenseGen: A Deep Learning Architecture for Synthetic Sensor Data
Generation | null | null | null | null | cs.LG cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Our ability to synthesize sensory data that preserves specific statistical
properties of the real data has had tremendous implications on data privacy and
big data analytics. The synthetic data can be used as a substitute for
selective real data segments,that are sensitive to the user, thus protecting
privacy and resulting in improved analytics.However, increasingly adversarial
roles taken by data recipients such as mobile apps, or other cloud-based
analytics services, mandate that the synthetic data, in addition to preserving
statistical properties, should also be difficult to distinguish from the real
data. Typically, visual inspection has been used as a test to distinguish
between datasets. But more recently, sophisticated classifier models
(discriminators), corresponding to a set of events, have also been employed to
distinguish between synthesized and real data. The model operates on both
datasets and the respective event outputs are compared for consistency. In this
paper, we take a step towards generating sensory data that can pass a deep
learning based discriminator model test, and make two specific contributions:
first, we present a deep learning based architecture for synthesizing sensory
data. This architecture comprises of a generator model, which is a stack of
multiple Long-Short-Term-Memory (LSTM) networks and a Mixture Density Network.
second, we use another LSTM network based discriminator model for
distinguishing between the true and the synthesized data. Using a dataset of
accelerometer traces, collected using smartphones of users doing their daily
activities, we show that the deep learning based discriminator model can only
distinguish between the real and synthesized traces with an accuracy in the
neighborhood of 50%.
| [
{
"version": "v1",
"created": "Tue, 31 Jan 2017 01:59:58 GMT"
}
] | 2017-02-01T00:00:00 | [
[
"Alzantot",
"Moustafa",
""
],
[
"Chakraborty",
"Supriyo",
""
],
[
"Srivastava",
"Mani B.",
""
]
] | TITLE: SenseGen: A Deep Learning Architecture for Synthetic Sensor Data
Generation
ABSTRACT: Our ability to synthesize sensory data that preserves specific statistical
properties of the real data has had tremendous implications on data privacy and
big data analytics. The synthetic data can be used as a substitute for
selective real data segments,that are sensitive to the user, thus protecting
privacy and resulting in improved analytics.However, increasingly adversarial
roles taken by data recipients such as mobile apps, or other cloud-based
analytics services, mandate that the synthetic data, in addition to preserving
statistical properties, should also be difficult to distinguish from the real
data. Typically, visual inspection has been used as a test to distinguish
between datasets. But more recently, sophisticated classifier models
(discriminators), corresponding to a set of events, have also been employed to
distinguish between synthesized and real data. The model operates on both
datasets and the respective event outputs are compared for consistency. In this
paper, we take a step towards generating sensory data that can pass a deep
learning based discriminator model test, and make two specific contributions:
first, we present a deep learning based architecture for synthesizing sensory
data. This architecture comprises of a generator model, which is a stack of
multiple Long-Short-Term-Memory (LSTM) networks and a Mixture Density Network.
second, we use another LSTM network based discriminator model for
distinguishing between the true and the synthesized data. Using a dataset of
accelerometer traces, collected using smartphones of users doing their daily
activities, we show that the deep learning based discriminator model can only
distinguish between the real and synthesized traces with an accuracy in the
neighborhood of 50%.
| no_new_dataset | 0.839603 |
1701.08921 | Yasir Latif | Yasir Latif, Guoquan Huang, John Leonard, Jose Neira | Sparse Optimization for Robust and Efficient Loop Closing | null | null | null | null | cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | It is essential for a robot to be able to detect revisits or loop closures
for long-term visual navigation.A key insight explored in this work is that the
loop-closing event inherently occurs sparsely, that is, the image currently
being taken matches with only a small subset (if any) of previous images. Based
on this observation, we formulate the problem of loop-closure detection as a
sparse, convex $\ell_1$-minimization problem. By leveraging fast convex
optimization techniques, we are able to efficiently find loop closures, thus
enabling real-time robot navigation. This novel formulation requires no offline
dictionary learning, as required by most existing approaches, and thus allows
online incremental operation. Our approach ensures a unique hypothesis by
choosing only a single globally optimal match when making a loop-closure
decision. Furthermore, the proposed formulation enjoys a flexible
representation with no restriction imposed on how images should be represented,
while requiring only that the representations are "close" to each other when
the corresponding images are visually similar. The proposed algorithm is
validated extensively using real-world datasets.
| [
{
"version": "v1",
"created": "Tue, 31 Jan 2017 05:32:09 GMT"
}
] | 2017-02-01T00:00:00 | [
[
"Latif",
"Yasir",
""
],
[
"Huang",
"Guoquan",
""
],
[
"Leonard",
"John",
""
],
[
"Neira",
"Jose",
""
]
] | TITLE: Sparse Optimization for Robust and Efficient Loop Closing
ABSTRACT: It is essential for a robot to be able to detect revisits or loop closures
for long-term visual navigation.A key insight explored in this work is that the
loop-closing event inherently occurs sparsely, that is, the image currently
being taken matches with only a small subset (if any) of previous images. Based
on this observation, we formulate the problem of loop-closure detection as a
sparse, convex $\ell_1$-minimization problem. By leveraging fast convex
optimization techniques, we are able to efficiently find loop closures, thus
enabling real-time robot navigation. This novel formulation requires no offline
dictionary learning, as required by most existing approaches, and thus allows
online incremental operation. Our approach ensures a unique hypothesis by
choosing only a single globally optimal match when making a loop-closure
decision. Furthermore, the proposed formulation enjoys a flexible
representation with no restriction imposed on how images should be represented,
while requiring only that the representations are "close" to each other when
the corresponding images are visually similar. The proposed algorithm is
validated extensively using real-world datasets.
| no_new_dataset | 0.946794 |
1701.08931 | Hadar Averbuch-Elor | Hadar Averbuch-Elor, Johannes Kopf, Tamir Hazan and Daniel Cohen-Or | Co-segmentation for Space-Time Co-located Collections | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a co-segmentation technique for space-time co-located image
collections. These prevalent collections capture various dynamic events,
usually by multiple photographers, and may contain multiple co-occurring
objects which are not necessarily part of the intended foreground object,
resulting in ambiguities for traditional co-segmentation techniques. Thus, to
disambiguate what the common foreground object is, we introduce a
weakly-supervised technique, where we assume only a small seed, given in the
form of a single segmented image. We take a distributed approach, where local
belief models are propagated and reinforced with similar images. Our technique
progressively expands the foreground and background belief models across the
entire collection. The technique exploits the power of the entire set of image
without building a global model, and thus successfully overcomes large
variability in appearance of the common foreground object. We demonstrate that
our method outperforms previous co-segmentation techniques on challenging
space-time co-located collections, including dense benchmark datasets which
were adapted for our novel problem setting.
| [
{
"version": "v1",
"created": "Tue, 31 Jan 2017 07:05:58 GMT"
}
] | 2017-02-01T00:00:00 | [
[
"Averbuch-Elor",
"Hadar",
""
],
[
"Kopf",
"Johannes",
""
],
[
"Hazan",
"Tamir",
""
],
[
"Cohen-Or",
"Daniel",
""
]
] | TITLE: Co-segmentation for Space-Time Co-located Collections
ABSTRACT: We present a co-segmentation technique for space-time co-located image
collections. These prevalent collections capture various dynamic events,
usually by multiple photographers, and may contain multiple co-occurring
objects which are not necessarily part of the intended foreground object,
resulting in ambiguities for traditional co-segmentation techniques. Thus, to
disambiguate what the common foreground object is, we introduce a
weakly-supervised technique, where we assume only a small seed, given in the
form of a single segmented image. We take a distributed approach, where local
belief models are propagated and reinforced with similar images. Our technique
progressively expands the foreground and background belief models across the
entire collection. The technique exploits the power of the entire set of image
without building a global model, and thus successfully overcomes large
variability in appearance of the common foreground object. We demonstrate that
our method outperforms previous co-segmentation techniques on challenging
space-time co-located collections, including dense benchmark datasets which
were adapted for our novel problem setting.
| no_new_dataset | 0.951006 |
1701.08968 | Nhan Truong | Nhan Truong, Levin Kuhlmann, Mohammad Reza Bonyadi, Jiawei Yang,
Andrew Faulks, Omid Kavehei | Supervised Learning in Automatic Channel Selection for Epileptic Seizure
Detection | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Detecting seizure using brain neuroactivations recorded by intracranial
electroencephalogram (iEEG) has been widely used for monitoring, diagnosing,
and closed-loop therapy of epileptic patients, however, computational
efficiency gains are needed if state-of-the-art methods are to be implemented
in implanted devices. We present a novel method for automatic seizure detection
based on iEEG data that outperforms current state-of-the-art seizure detection
methods in terms of computational efficiency while maintaining the accuracy.
The proposed algorithm incorporates an automatic channel selection (ACS) engine
as a pre-processing stage to the seizure detection procedure. The ACS engine
consists of supervised classifiers which aim to find iEEGchannelswhich
contribute the most to a seizure. Seizure detection stage involves feature
extraction and classification. Feature extraction is performed in both
frequency and time domains where spectral power and correlation between channel
pairs are calculated. Random Forest is used in classification of interictal,
ictal and early ictal periods of iEEG signals. Seizure detection in this paper
is retrospective and patient-specific. iEEG data is accessed via Kaggle,
provided by International Epilepsy Electro-physiology Portal. The dataset
includes a training set of 6.5 hours of interictal data and 41 minin ictal data
and a test set of 9.14 hours. Compared to the state-of-the-art on the same
dataset, we achieve 49.4% increase in computational efficiency and 400 mins
better in average for detection delay. The proposed model is able to detect a
seizure onset at 91.95% sensitivity and 94.05% specificity with a mean
detection delay of 2.77 s. The area under the curve (AUC) is 96.44%, that is
comparable to the current state-of-the-art with AUC of 96.29%.
| [
{
"version": "v1",
"created": "Tue, 31 Jan 2017 10:01:45 GMT"
}
] | 2017-02-01T00:00:00 | [
[
"Truong",
"Nhan",
""
],
[
"Kuhlmann",
"Levin",
""
],
[
"Bonyadi",
"Mohammad Reza",
""
],
[
"Yang",
"Jiawei",
""
],
[
"Faulks",
"Andrew",
""
],
[
"Kavehei",
"Omid",
""
]
] | TITLE: Supervised Learning in Automatic Channel Selection for Epileptic Seizure
Detection
ABSTRACT: Detecting seizure using brain neuroactivations recorded by intracranial
electroencephalogram (iEEG) has been widely used for monitoring, diagnosing,
and closed-loop therapy of epileptic patients, however, computational
efficiency gains are needed if state-of-the-art methods are to be implemented
in implanted devices. We present a novel method for automatic seizure detection
based on iEEG data that outperforms current state-of-the-art seizure detection
methods in terms of computational efficiency while maintaining the accuracy.
The proposed algorithm incorporates an automatic channel selection (ACS) engine
as a pre-processing stage to the seizure detection procedure. The ACS engine
consists of supervised classifiers which aim to find iEEGchannelswhich
contribute the most to a seizure. Seizure detection stage involves feature
extraction and classification. Feature extraction is performed in both
frequency and time domains where spectral power and correlation between channel
pairs are calculated. Random Forest is used in classification of interictal,
ictal and early ictal periods of iEEG signals. Seizure detection in this paper
is retrospective and patient-specific. iEEG data is accessed via Kaggle,
provided by International Epilepsy Electro-physiology Portal. The dataset
includes a training set of 6.5 hours of interictal data and 41 minin ictal data
and a test set of 9.14 hours. Compared to the state-of-the-art on the same
dataset, we achieve 49.4% increase in computational efficiency and 400 mins
better in average for detection delay. The proposed model is able to detect a
seizure onset at 91.95% sensitivity and 94.05% specificity with a mean
detection delay of 2.77 s. The area under the curve (AUC) is 96.44%, that is
comparable to the current state-of-the-art with AUC of 96.29%.
| no_new_dataset | 0.947284 |
1701.08985 | Alin Popa | Alin-Ionut Popa, Mihai Zanfir, Cristian Sminchisescu | Deep Multitask Architecture for Integrated 2D and 3D Human Sensing | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a deep multitask architecture for \emph{fully automatic 2d and 3d
human sensing} (DMHS), including \emph{recognition and reconstruction}, in
\emph{monocular images}. The system computes the figure-ground segmentation,
semantically identifies the human body parts at pixel level, and estimates the
2d and 3d pose of the person. The model supports the joint training of all
components by means of multi-task losses where early processing stages
recursively feed into advanced ones for increasingly complex calculations,
accuracy and robustness. The design allows us to tie a complete training
protocol, by taking advantage of multiple datasets that would otherwise
restrictively cover only some of the model components: complex 2d image data
with no body part labeling and without associated 3d ground truth, or complex
3d data with limited 2d background variability. In detailed experiments based
on several challenging 2d and 3d datasets (LSP, HumanEva, Human3.6M), we
evaluate the sub-structures of the model, the effect of various types of
training data in the multitask loss, and demonstrate that state-of-the-art
results can be achieved at all processing levels. We also show that in the wild
our monocular RGB architecture is perceptually competitive to a state-of-the
art (commercial) Kinect system based on RGB-D data.
| [
{
"version": "v1",
"created": "Tue, 31 Jan 2017 10:52:48 GMT"
}
] | 2017-02-01T00:00:00 | [
[
"Popa",
"Alin-Ionut",
""
],
[
"Zanfir",
"Mihai",
""
],
[
"Sminchisescu",
"Cristian",
""
]
] | TITLE: Deep Multitask Architecture for Integrated 2D and 3D Human Sensing
ABSTRACT: We propose a deep multitask architecture for \emph{fully automatic 2d and 3d
human sensing} (DMHS), including \emph{recognition and reconstruction}, in
\emph{monocular images}. The system computes the figure-ground segmentation,
semantically identifies the human body parts at pixel level, and estimates the
2d and 3d pose of the person. The model supports the joint training of all
components by means of multi-task losses where early processing stages
recursively feed into advanced ones for increasingly complex calculations,
accuracy and robustness. The design allows us to tie a complete training
protocol, by taking advantage of multiple datasets that would otherwise
restrictively cover only some of the model components: complex 2d image data
with no body part labeling and without associated 3d ground truth, or complex
3d data with limited 2d background variability. In detailed experiments based
on several challenging 2d and 3d datasets (LSP, HumanEva, Human3.6M), we
evaluate the sub-structures of the model, the effect of various types of
training data in the multitask loss, and demonstrate that state-of-the-art
results can be achieved at all processing levels. We also show that in the wild
our monocular RGB architecture is perceptually competitive to a state-of-the
art (commercial) Kinect system based on RGB-D data.
| no_new_dataset | 0.950869 |
1701.09039 | Bryan Perozzi | Aria Rezaei, Bryan Perozzi, Leman Akoglu | Ties That Bind - Characterizing Classes by Attributes and Social Ties | WWW'17 Web Science, 9 pages | null | null | null | cs.SI cs.IR physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Given a set of attributed subgraphs known to be from different classes, how
can we discover their differences? There are many cases where collections of
subgraphs may be contrasted against each other. For example, they may be
assigned ground truth labels (spam/not-spam), or it may be desired to directly
compare the biological networks of different species or compound networks of
different chemicals.
In this work we introduce the problem of characterizing the differences
between attributed subgraphs that belong to different classes. We define this
characterization problem as one of partitioning the attributes into as many
groups as the number of classes, while maximizing the total attributed quality
score of all the given subgraphs.
We show that our attribute-to-class assignment problem is NP-hard and an
optimal $(1 - 1/e)$-approximation algorithm exists. We also propose two
different faster heuristics that are linear-time in the number of attributes
and subgraphs. Unlike previous work where only attributes were taken into
account for characterization, here we exploit both attributes and social ties
(i.e. graph structure).
Through extensive experiments, we compare our proposed algorithms, show
findings that agree with human intuition on datasets from Amazon co-purchases,
Congressional bill sponsorships, and DBLP co-authorships. We also show that our
approach of characterizing subgraphs is better suited for sense-making than
discriminating classification approaches.
| [
{
"version": "v1",
"created": "Tue, 31 Jan 2017 14:01:04 GMT"
}
] | 2017-02-01T00:00:00 | [
[
"Rezaei",
"Aria",
""
],
[
"Perozzi",
"Bryan",
""
],
[
"Akoglu",
"Leman",
""
]
] | TITLE: Ties That Bind - Characterizing Classes by Attributes and Social Ties
ABSTRACT: Given a set of attributed subgraphs known to be from different classes, how
can we discover their differences? There are many cases where collections of
subgraphs may be contrasted against each other. For example, they may be
assigned ground truth labels (spam/not-spam), or it may be desired to directly
compare the biological networks of different species or compound networks of
different chemicals.
In this work we introduce the problem of characterizing the differences
between attributed subgraphs that belong to different classes. We define this
characterization problem as one of partitioning the attributes into as many
groups as the number of classes, while maximizing the total attributed quality
score of all the given subgraphs.
We show that our attribute-to-class assignment problem is NP-hard and an
optimal $(1 - 1/e)$-approximation algorithm exists. We also propose two
different faster heuristics that are linear-time in the number of attributes
and subgraphs. Unlike previous work where only attributes were taken into
account for characterization, here we exploit both attributes and social ties
(i.e. graph structure).
Through extensive experiments, we compare our proposed algorithms, show
findings that agree with human intuition on datasets from Amazon co-purchases,
Congressional bill sponsorships, and DBLP co-authorships. We also show that our
approach of characterizing subgraphs is better suited for sense-making than
discriminating classification approaches.
| no_new_dataset | 0.943348 |
1701.09042 | Jeff Heaton | Jeff Heaton | Comparing Dataset Characteristics that Favor the Apriori, Eclat or
FP-Growth Frequent Itemset Mining Algorithms | null | null | null | null | cs.DB cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Frequent itemset mining is a popular data mining technique. Apriori, Eclat,
and FP-Growth are among the most common algorithms for frequent itemset mining.
Considerable research has been performed to compare the relative performance
between these three algorithms, by evaluating the scalability of each algorithm
as the dataset size increases. While scalability as data size increases is
important, previous papers have not examined the performance impact of
similarly sized datasets that contain different itemset characteristics. This
paper explores the effects that two dataset characteristics can have on the
performance of these three frequent itemset algorithms. To perform this
empirical analysis, a dataset generator is created to measure the effects of
frequent item density and the maximum transaction size on performance. The
generated datasets contain the same number of rows. This provides some insight
into dataset characteristics that are conducive to each algorithm. The results
of this paper's research demonstrate Eclat and FP-Growth both handle increases
in maximum transaction size and frequent itemset density considerably better
than the Apriori algorithm.
This paper explores the effects that two dataset characteristics can have on
the performance of these three frequent itemset algorithms. To perform this
empirical analysis, a dataset generator is created to measure the effects of
frequent item density and the maximum transaction size on performance. The
generated datasets contain the same number of rows. This provides some insight
into dataset characteristics that are conducive to each algorithm. The results
of this paper's research demonstrate Eclat and FP-Growth both handle increases
in maximum transaction size and frequent itemset density considerably better
than the Apriori algorithm.
| [
{
"version": "v1",
"created": "Mon, 30 Jan 2017 12:34:02 GMT"
}
] | 2017-02-01T00:00:00 | [
[
"Heaton",
"Jeff",
""
]
] | TITLE: Comparing Dataset Characteristics that Favor the Apriori, Eclat or
FP-Growth Frequent Itemset Mining Algorithms
ABSTRACT: Frequent itemset mining is a popular data mining technique. Apriori, Eclat,
and FP-Growth are among the most common algorithms for frequent itemset mining.
Considerable research has been performed to compare the relative performance
between these three algorithms, by evaluating the scalability of each algorithm
as the dataset size increases. While scalability as data size increases is
important, previous papers have not examined the performance impact of
similarly sized datasets that contain different itemset characteristics. This
paper explores the effects that two dataset characteristics can have on the
performance of these three frequent itemset algorithms. To perform this
empirical analysis, a dataset generator is created to measure the effects of
frequent item density and the maximum transaction size on performance. The
generated datasets contain the same number of rows. This provides some insight
into dataset characteristics that are conducive to each algorithm. The results
of this paper's research demonstrate Eclat and FP-Growth both handle increases
in maximum transaction size and frequent itemset density considerably better
than the Apriori algorithm.
This paper explores the effects that two dataset characteristics can have on
the performance of these three frequent itemset algorithms. To perform this
empirical analysis, a dataset generator is created to measure the effects of
frequent item density and the maximum transaction size on performance. The
generated datasets contain the same number of rows. This provides some insight
into dataset characteristics that are conducive to each algorithm. The results
of this paper's research demonstrate Eclat and FP-Growth both handle increases
in maximum transaction size and frequent itemset density considerably better
than the Apriori algorithm.
| new_dataset | 0.973569 |
1111.4171 | Arian Ojeda Gonz\'alez | G. A. Ojeda, O. Mendes, M. A. Calzadilla, M. O. Domingues | The Entropy Index (EI): an Auxiliary Tool to Identify the Occurrence of
Interplanetary Magnetic Clouds | This paper has been withdrawn by the author due to great amount of
modifications done, and low quality of figures | null | null | null | physics.space-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | By the study of the dynamical processes related to entropy, this work aims to
create a mathematical tool to identify magnetic clouds (MCs) in the
interplanetary space using only interplanetary magnetic field (IMF) data. Used
as basis for an analysis methodology, the spatio-temporal entropy (STE)
measures the image (recurrence plots) "structuredness" in both space and time
domains. Initially we worked with the Huttunen et al. 2005's dataset and
studied the 41 MCs presenting a shock wave identified before the cloud. The STE
values for each Bx, By, Bz IMF time series, with dimension and time delay equal
to one, were respectively calculated. We found higher STE values in the sheaths
and zero STE values in some of the three components in most of the MCs (30
among 41 events). In a physically consistent manner, data windows of 2500
magnetic records were selected as the calculation interval for the time series.
As not all MCs have zero STE simultaneously, we created a standardization index
(an entropy index, called as EI) to allow joining the result of the three
components. With the use of EI three not known MCs were indeed identified and
then the MVA method allowed calculating their boundaries. Thus the EI is
proposed as an auxiliary tool to identify MC candidates based only on IMF
analysis. In a promissor condition, this methodology implemented gives basis
for an automatic MC identification procedure and surely useful for space
weather purposes.
| [
{
"version": "v1",
"created": "Thu, 17 Nov 2011 18:14:10 GMT"
},
{
"version": "v2",
"created": "Mon, 30 Jan 2017 12:49:41 GMT"
}
] | 2017-01-31T00:00:00 | [
[
"Ojeda",
"G. A.",
""
],
[
"Mendes",
"O.",
""
],
[
"Calzadilla",
"M. A.",
""
],
[
"Domingues",
"M. O.",
""
]
] | TITLE: The Entropy Index (EI): an Auxiliary Tool to Identify the Occurrence of
Interplanetary Magnetic Clouds
ABSTRACT: By the study of the dynamical processes related to entropy, this work aims to
create a mathematical tool to identify magnetic clouds (MCs) in the
interplanetary space using only interplanetary magnetic field (IMF) data. Used
as basis for an analysis methodology, the spatio-temporal entropy (STE)
measures the image (recurrence plots) "structuredness" in both space and time
domains. Initially we worked with the Huttunen et al. 2005's dataset and
studied the 41 MCs presenting a shock wave identified before the cloud. The STE
values for each Bx, By, Bz IMF time series, with dimension and time delay equal
to one, were respectively calculated. We found higher STE values in the sheaths
and zero STE values in some of the three components in most of the MCs (30
among 41 events). In a physically consistent manner, data windows of 2500
magnetic records were selected as the calculation interval for the time series.
As not all MCs have zero STE simultaneously, we created a standardization index
(an entropy index, called as EI) to allow joining the result of the three
components. With the use of EI three not known MCs were indeed identified and
then the MVA method allowed calculating their boundaries. Thus the EI is
proposed as an auxiliary tool to identify MC candidates based only on IMF
analysis. In a promissor condition, this methodology implemented gives basis
for an automatic MC identification procedure and surely useful for space
weather purposes.
| no_new_dataset | 0.947575 |
1409.0272 | Andre Goncalves | Andre R. Goncalves, Puja Das, Soumyadeep Chatterjee, Vidyashankar
Sivakumar, Fernando J. Von Zuben, Arindam Banerjee | Multi-task Sparse Structure Learning | 23rd ACM International Conference on Information and Knowledge
Management - CIKM 2014 | null | 10.1145/2661829.2662091 | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multi-task learning (MTL) aims to improve generalization performance by
learning multiple related tasks simultaneously. While sometimes the underlying
task relationship structure is known, often the structure needs to be estimated
from data at hand. In this paper, we present a novel family of models for MTL,
applicable to regression and classification problems, capable of learning the
structure of task relationships. In particular, we consider a joint estimation
problem of the task relationship structure and the individual task parameters,
which is solved using alternating minimization. The task relationship structure
learning component builds on recent advances in structure learning of Gaussian
graphical models based on sparse estimators of the precision (inverse
covariance) matrix. We illustrate the effectiveness of the proposed model on a
variety of synthetic and benchmark datasets for regression and classification.
We also consider the problem of combining climate model outputs for better
projections of future climate, with focus on temperature in South America, and
show that the proposed model outperforms several existing methods for the
problem.
| [
{
"version": "v1",
"created": "Mon, 1 Sep 2014 00:33:38 GMT"
},
{
"version": "v2",
"created": "Tue, 2 Sep 2014 00:33:35 GMT"
}
] | 2017-01-31T00:00:00 | [
[
"Goncalves",
"Andre R.",
""
],
[
"Das",
"Puja",
""
],
[
"Chatterjee",
"Soumyadeep",
""
],
[
"Sivakumar",
"Vidyashankar",
""
],
[
"Von Zuben",
"Fernando J.",
""
],
[
"Banerjee",
"Arindam",
""
]
] | TITLE: Multi-task Sparse Structure Learning
ABSTRACT: Multi-task learning (MTL) aims to improve generalization performance by
learning multiple related tasks simultaneously. While sometimes the underlying
task relationship structure is known, often the structure needs to be estimated
from data at hand. In this paper, we present a novel family of models for MTL,
applicable to regression and classification problems, capable of learning the
structure of task relationships. In particular, we consider a joint estimation
problem of the task relationship structure and the individual task parameters,
which is solved using alternating minimization. The task relationship structure
learning component builds on recent advances in structure learning of Gaussian
graphical models based on sparse estimators of the precision (inverse
covariance) matrix. We illustrate the effectiveness of the proposed model on a
variety of synthetic and benchmark datasets for regression and classification.
We also consider the problem of combining climate model outputs for better
projections of future climate, with focus on temperature in South America, and
show that the proposed model outperforms several existing methods for the
problem.
| no_new_dataset | 0.938688 |
1511.07118 | Dong-Hyun Lee | Dong-Hyun Lee | Cascading Denoising Auto-Encoder as a Deep Directed Generative Model | not completed | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent work (Bengio et al., 2013) has shown howDenoising Auto-Encoders(DAE)
become gener-ative models as a density estimator. However,in practice, the
framework suffers from a mixingproblem in the MCMC sampling process and
nodirect method to estimate the test log-likelihood.We consider a directed
model with an stochas-tic identity mapping (simple corruption pro-cess) as an
inference model and a DAE as agenerative model. By cascading these mod-els, we
propose Cascading Denoising Auto-Encoders(CDAE) which can generate samples
ofdata distribution from tractable prior distributionunder the assumption that
probabilistic distribu-tion of corrupted data approaches tractable
priordistribution as the level of corruption increases.This work tries to
answer two questions. On theone hand, can deep directed models be success-fully
trained without intractable posterior infer-ence and difficult optimization of
very deep neu-ral networks in inference and generative mod-els? These are
unavoidable when recent suc-cessful directed model like VAE (Kingma &Welling,
2014) is trained on complex dataset likereal images. On the other hand, can
DAEs getclean samples of data distribution from heavilycorrupted samples which
can be considered oftractable prior distribution far from data mani-fold?
so-called global denoising scheme.Our results show positive responses of
thesequestions and this work can provide fairly simpleframework for generative
models of very com-plex dataset.
| [
{
"version": "v1",
"created": "Mon, 23 Nov 2015 06:32:57 GMT"
},
{
"version": "v2",
"created": "Fri, 27 Jan 2017 19:09:52 GMT"
}
] | 2017-01-31T00:00:00 | [
[
"Lee",
"Dong-Hyun",
""
]
] | TITLE: Cascading Denoising Auto-Encoder as a Deep Directed Generative Model
ABSTRACT: Recent work (Bengio et al., 2013) has shown howDenoising Auto-Encoders(DAE)
become gener-ative models as a density estimator. However,in practice, the
framework suffers from a mixingproblem in the MCMC sampling process and
nodirect method to estimate the test log-likelihood.We consider a directed
model with an stochas-tic identity mapping (simple corruption pro-cess) as an
inference model and a DAE as agenerative model. By cascading these mod-els, we
propose Cascading Denoising Auto-Encoders(CDAE) which can generate samples
ofdata distribution from tractable prior distributionunder the assumption that
probabilistic distribu-tion of corrupted data approaches tractable
priordistribution as the level of corruption increases.This work tries to
answer two questions. On theone hand, can deep directed models be success-fully
trained without intractable posterior infer-ence and difficult optimization of
very deep neu-ral networks in inference and generative mod-els? These are
unavoidable when recent suc-cessful directed model like VAE (Kingma &Welling,
2014) is trained on complex dataset likereal images. On the other hand, can
DAEs getclean samples of data distribution from heavilycorrupted samples which
can be considered oftractable prior distribution far from data mani-fold?
so-called global denoising scheme.Our results show positive responses of
thesequestions and this work can provide fairly simpleframework for generative
models of very com-plex dataset.
| no_new_dataset | 0.949763 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.