id
large_stringlengths 9
16
| submitter
large_stringlengths 3
64
⌀ | authors
large_stringlengths 5
20.2k
| title
large_stringlengths 5
245
| comments
large_stringlengths 1
482
⌀ | journal-ref
large_stringlengths 4
382
⌀ | doi
large_stringlengths 9
151
⌀ | report-no
large_stringlengths 2
509
⌀ | categories
large_stringlengths 5
78
| license
large_stringclasses 9
values | abstract
large_stringlengths 83
3.55k
| versions
large listlengths 1
21
| update_date
timestamp[ms]date 2007-05-23 00:00:00
2025-05-30 00:00:00
| authors_parsed
large listlengths 1
1.35k
| prompt
large_stringlengths 166
3.64k
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1407.3685 | Anthony Bagnall Dr | Anthony Bagnall, Jon Hills and Jason Lines | Finding Motif Sets in Time Series | null | null | null | CMPC14-03 | cs.LG cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Time-series motifs are representative subsequences that occur frequently in a
time series; a motif set is the set of subsequences deemed to be instances of a
given motif. We focus on finding motif sets. Our motivation is to detect motif
sets in household electricity-usage profiles, representing repeated patterns of
household usage.
We propose three algorithms for finding motif sets. Two are greedy algorithms
based on pairwise comparison, and the third uses a heuristic measure of set
quality to find the motif set directly. We compare these algorithms on
simulated datasets and on electricity-usage data. We show that Scan MK, the
simplest way of using the best-matching pair to find motif sets, is less
accurate on our synthetic data than Set Finder and Cluster MK, although the
latter is very sensitive to parameter settings. We qualitatively analyse the
outputs for the electricity-usage data and demonstrate that both Scan MK and
Set Finder can discover useful motif sets in such data.
| [
{
"version": "v1",
"created": "Mon, 14 Jul 2014 15:01:57 GMT"
}
] | 2014-07-15T00:00:00 | [
[
"Bagnall",
"Anthony",
""
],
[
"Hills",
"Jon",
""
],
[
"Lines",
"Jason",
""
]
] | TITLE: Finding Motif Sets in Time Series
ABSTRACT: Time-series motifs are representative subsequences that occur frequently in a
time series; a motif set is the set of subsequences deemed to be instances of a
given motif. We focus on finding motif sets. Our motivation is to detect motif
sets in household electricity-usage profiles, representing repeated patterns of
household usage.
We propose three algorithms for finding motif sets. Two are greedy algorithms
based on pairwise comparison, and the third uses a heuristic measure of set
quality to find the motif set directly. We compare these algorithms on
simulated datasets and on electricity-usage data. We show that Scan MK, the
simplest way of using the best-matching pair to find motif sets, is less
accurate on our synthetic data than Set Finder and Cluster MK, although the
latter is very sensitive to parameter settings. We qualitatively analyse the
outputs for the electricity-usage data and demonstrate that both Scan MK and
Set Finder can discover useful motif sets in such data.
|
1407.3686 | Alejandro Gonz\'alez Alzate | Alejandro Gonz\'alez and Sebastian Ramos and David V\'azquez and
Antonio M. L\'opez and Jaume Amores | Spatiotemporal Stacked Sequential Learning for Pedestrian Detection | 8 pages, 5 figure, 1 table | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Pedestrian classifiers decide which image windows contain a pedestrian. In
practice, such classifiers provide a relatively high response at neighbor
windows overlapping a pedestrian, while the responses around potential false
positives are expected to be lower. An analogous reasoning applies for image
sequences. If there is a pedestrian located within a frame, the same pedestrian
is expected to appear close to the same location in neighbor frames. Therefore,
such a location has chances of receiving high classification scores during
several frames, while false positives are expected to be more spurious. In this
paper we propose to exploit such correlations for improving the accuracy of
base pedestrian classifiers. In particular, we propose to use two-stage
classifiers which not only rely on the image descriptors required by the base
classifiers but also on the response of such base classifiers in a given
spatiotemporal neighborhood. More specifically, we train pedestrian classifiers
using a stacked sequential learning (SSL) paradigm. We use a new pedestrian
dataset we have acquired from a car to evaluate our proposal at different frame
rates. We also test on a well known dataset: Caltech. The obtained results show
that our SSL proposal boosts detection accuracy significantly with a minimal
impact on the computational cost. Interestingly, SSL improves more the accuracy
at the most dangerous situations, i.e. when a pedestrian is close to the
camera.
| [
{
"version": "v1",
"created": "Mon, 14 Jul 2014 15:03:01 GMT"
}
] | 2014-07-15T00:00:00 | [
[
"González",
"Alejandro",
""
],
[
"Ramos",
"Sebastian",
""
],
[
"Vázquez",
"David",
""
],
[
"López",
"Antonio M.",
""
],
[
"Amores",
"Jaume",
""
]
] | TITLE: Spatiotemporal Stacked Sequential Learning for Pedestrian Detection
ABSTRACT: Pedestrian classifiers decide which image windows contain a pedestrian. In
practice, such classifiers provide a relatively high response at neighbor
windows overlapping a pedestrian, while the responses around potential false
positives are expected to be lower. An analogous reasoning applies for image
sequences. If there is a pedestrian located within a frame, the same pedestrian
is expected to appear close to the same location in neighbor frames. Therefore,
such a location has chances of receiving high classification scores during
several frames, while false positives are expected to be more spurious. In this
paper we propose to exploit such correlations for improving the accuracy of
base pedestrian classifiers. In particular, we propose to use two-stage
classifiers which not only rely on the image descriptors required by the base
classifiers but also on the response of such base classifiers in a given
spatiotemporal neighborhood. More specifically, we train pedestrian classifiers
using a stacked sequential learning (SSL) paradigm. We use a new pedestrian
dataset we have acquired from a car to evaluate our proposal at different frame
rates. We also test on a well known dataset: Caltech. The obtained results show
that our SSL proposal boosts detection accuracy significantly with a minimal
impact on the computational cost. Interestingly, SSL improves more the accuracy
at the most dangerous situations, i.e. when a pedestrian is close to the
camera.
|
1407.3745 | Sutanay Choudhury | Sutanay Choudhury, Lawrence Holder, George Chin, Patrick Mackey,
Khushbu Agarwal, John Feo | Query Optimization for Dynamic Graphs | null | null | null | PNNL-SA-103238, Pacific Northwest National Laboratory, Richland, WA | cs.DB | http://creativecommons.org/licenses/publicdomain/ | Given a query graph that represents a pattern of interest, the emerging
pattern detection problem can be viewed as a continuous query problem on a
dynamic graph. We present an incremental algorithm for continuous query
processing on dynamic graphs. The algorithm is based on the concept of query
decomposition; we decompose a query graph into smaller subgraphs and assemble
the result of sub-queries to find complete matches with the specified query.
The novelty of our work lies in using the subgraph distributional statistics
collected from the dynamic graph to generate the decomposition. We introduce a
"Lazy Search" algorithm where the search strategy is decided on a
vertex-to-vertex basis depending on the likelihood of a match in the vertex
neighborhood. We also propose a metric named "Relative Selectivity" that is
used to select between different query decomposition strategies. Our
experiments performed on real online news, network traffic stream and a
synthetic social network benchmark demonstrate 10-100x speedups over competing
approaches.
| [
{
"version": "v1",
"created": "Mon, 14 Jul 2014 17:53:30 GMT"
}
] | 2014-07-15T00:00:00 | [
[
"Choudhury",
"Sutanay",
""
],
[
"Holder",
"Lawrence",
""
],
[
"Chin",
"George",
""
],
[
"Mackey",
"Patrick",
""
],
[
"Agarwal",
"Khushbu",
""
],
[
"Feo",
"John",
""
]
] | TITLE: Query Optimization for Dynamic Graphs
ABSTRACT: Given a query graph that represents a pattern of interest, the emerging
pattern detection problem can be viewed as a continuous query problem on a
dynamic graph. We present an incremental algorithm for continuous query
processing on dynamic graphs. The algorithm is based on the concept of query
decomposition; we decompose a query graph into smaller subgraphs and assemble
the result of sub-queries to find complete matches with the specified query.
The novelty of our work lies in using the subgraph distributional statistics
collected from the dynamic graph to generate the decomposition. We introduce a
"Lazy Search" algorithm where the search strategy is decided on a
vertex-to-vertex basis depending on the likelihood of a match in the vertex
neighborhood. We also propose a metric named "Relative Selectivity" that is
used to select between different query decomposition strategies. Our
experiments performed on real online news, network traffic stream and a
synthetic social network benchmark demonstrate 10-100x speedups over competing
approaches.
|
1407.3751 | Sutanay Choudhury | Sutanay Choudhury, Chase Dowling | Benchmarking Named Entity Disambiguation approaches for Streaming Graphs | null | null | null | PNNL-23455 2014-05, Pacific Northwest National Laboratory, Richland,
WA | cs.CL cs.IR | http://creativecommons.org/licenses/publicdomain/ | Named Entity Disambiaguation (NED) is a central task for applications dealing
with natural language text. Assume that we have a graph based knowledge base
(subsequently referred as Knowledge Graph) where nodes represent various real
world entities such as people, location, organization and concepts. Given data
sources such as social media streams and web pages Entity Linking is the task
of mapping named entities that are extracted from the data to those present in
the Knowledge Graph. This is an inherently difficult task due to several
reasons. Almost all these data sources are generated without any formal
ontology; the unstructured nature of the input, limited context and the
ambiguity involved when multiple entities are mapped to the same name make this
a hard task. This report looks at two state of the art systems employing two
distinctive approaches: graph based Accurate Online Disambiguation of Entities
(AIDA) and Mined Evidence Named Entity Disambiguation (MENED), which employs a
statistical inference approach. We compare both approaches using the data set
and queries provided by the Knowledge Base Population (KBP) track at 2011 NIST
Text Analytics Conference (TAC). This report begins with an overview of the
respective approaches, followed by detailed description of the experimental
setup. It concludes with our findings from the benchmarking exercise.
| [
{
"version": "v1",
"created": "Mon, 14 Jul 2014 18:01:08 GMT"
}
] | 2014-07-15T00:00:00 | [
[
"Choudhury",
"Sutanay",
""
],
[
"Dowling",
"Chase",
""
]
] | TITLE: Benchmarking Named Entity Disambiguation approaches for Streaming Graphs
ABSTRACT: Named Entity Disambiaguation (NED) is a central task for applications dealing
with natural language text. Assume that we have a graph based knowledge base
(subsequently referred as Knowledge Graph) where nodes represent various real
world entities such as people, location, organization and concepts. Given data
sources such as social media streams and web pages Entity Linking is the task
of mapping named entities that are extracted from the data to those present in
the Knowledge Graph. This is an inherently difficult task due to several
reasons. Almost all these data sources are generated without any formal
ontology; the unstructured nature of the input, limited context and the
ambiguity involved when multiple entities are mapped to the same name make this
a hard task. This report looks at two state of the art systems employing two
distinctive approaches: graph based Accurate Online Disambiguation of Entities
(AIDA) and Mined Evidence Named Entity Disambiguation (MENED), which employs a
statistical inference approach. We compare both approaches using the data set
and queries provided by the Knowledge Base Population (KBP) track at 2011 NIST
Text Analytics Conference (TAC). This report begins with an overview of the
respective approaches, followed by detailed description of the experimental
setup. It concludes with our findings from the benchmarking exercise.
|
1306.5794 | Zohar Nussinov | Richard K. Darst, David R. Reichman, Peter Ronhovde, Zohar Nussinov | Algorithm independent bounds on community detection problems and
associated transitions in stochastic block model graphs | 21 pages, 9 figures | null | null | null | cond-mat.stat-mech cs.SI physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We derive rigorous bounds for well-defined community structure in complex
networks for a stochastic block model (SBM) benchmark. In particular, we
analyze the effect of inter-community "noise" (inter-community edges) on any
"community detection" algorithm's ability to correctly group nodes assigned to
a planted partition, a problem which has been proven to be NP complete in a
standard rendition. Our result does not rely on the use of any one particular
algorithm nor on the analysis of the limitations of inference. Rather, we turn
the problem on its head and work backwards to examine when, in the first place,
well defined structure may exist in SBMs.The method that we introduce here
could potentially be applied to other computational problems. The objective of
community detection algorithms is to partition a given network into optimally
disjoint subgraphs (or communities). Similar to k-SAT and other combinatorial
optimization problems, "community detection" exhibits different phases.
Networks that lie in the "unsolvable phase" lack well-defined structure and
thus have no partition that is meaningful. Solvable systems splinter into two
disparate phases: those in the "hard" phase and those in the "easy" phase. As
befits its name, within the easy phase, a partition is easy to achieve by known
algorithms. When a network lies in the hard phase, it still has an underlying
structure yet finding a meaningful partition which can be checked in polynomial
time requires an exhaustive computational effort that rapidly increases with
the size of the graph. When taken together, (i) the rigorous results that we
report here on when graphs have an underlying structure and (ii) recent results
concerning the limits of rather general algorithms, suggest bounds on the hard
phase.
| [
{
"version": "v1",
"created": "Mon, 24 Jun 2013 22:10:50 GMT"
},
{
"version": "v2",
"created": "Fri, 28 Jun 2013 00:39:44 GMT"
},
{
"version": "v3",
"created": "Thu, 10 Jul 2014 21:20:32 GMT"
}
] | 2014-07-14T00:00:00 | [
[
"Darst",
"Richard K.",
""
],
[
"Reichman",
"David R.",
""
],
[
"Ronhovde",
"Peter",
""
],
[
"Nussinov",
"Zohar",
""
]
] | TITLE: Algorithm independent bounds on community detection problems and
associated transitions in stochastic block model graphs
ABSTRACT: We derive rigorous bounds for well-defined community structure in complex
networks for a stochastic block model (SBM) benchmark. In particular, we
analyze the effect of inter-community "noise" (inter-community edges) on any
"community detection" algorithm's ability to correctly group nodes assigned to
a planted partition, a problem which has been proven to be NP complete in a
standard rendition. Our result does not rely on the use of any one particular
algorithm nor on the analysis of the limitations of inference. Rather, we turn
the problem on its head and work backwards to examine when, in the first place,
well defined structure may exist in SBMs.The method that we introduce here
could potentially be applied to other computational problems. The objective of
community detection algorithms is to partition a given network into optimally
disjoint subgraphs (or communities). Similar to k-SAT and other combinatorial
optimization problems, "community detection" exhibits different phases.
Networks that lie in the "unsolvable phase" lack well-defined structure and
thus have no partition that is meaningful. Solvable systems splinter into two
disparate phases: those in the "hard" phase and those in the "easy" phase. As
befits its name, within the easy phase, a partition is easy to achieve by known
algorithms. When a network lies in the hard phase, it still has an underlying
structure yet finding a meaningful partition which can be checked in polynomial
time requires an exhaustive computational effort that rapidly increases with
the size of the graph. When taken together, (i) the rigorous results that we
report here on when graphs have an underlying structure and (ii) recent results
concerning the limits of rather general algorithms, suggest bounds on the hard
phase.
|
1407.2987 | Eren Golge | Eren Golge and Pinar Duygulu | FAME: Face Association through Model Evolution | Draft version of the study | null | null | null | cs.CV cs.AI cs.IR cs.LG | http://creativecommons.org/licenses/by-nc-sa/3.0/ | We attack the problem of learning face models for public faces from
weakly-labelled images collected from web through querying a name. The data is
very noisy even after face detection, with several irrelevant faces
corresponding to other people. We propose a novel method, Face Association
through Model Evolution (FAME), that is able to prune the data in an iterative
way, for the face models associated to a name to evolve. The idea is based on
capturing discriminativeness and representativeness of each instance and
eliminating the outliers. The final models are used to classify faces on novel
datasets with possibly different characteristics. On benchmark datasets, our
results are comparable to or better than state-of-the-art studies for the task
of face identification.
| [
{
"version": "v1",
"created": "Thu, 10 Jul 2014 23:52:44 GMT"
}
] | 2014-07-14T00:00:00 | [
[
"Golge",
"Eren",
""
],
[
"Duygulu",
"Pinar",
""
]
] | TITLE: FAME: Face Association through Model Evolution
ABSTRACT: We attack the problem of learning face models for public faces from
weakly-labelled images collected from web through querying a name. The data is
very noisy even after face detection, with several irrelevant faces
corresponding to other people. We propose a novel method, Face Association
through Model Evolution (FAME), that is able to prune the data in an iterative
way, for the face models associated to a name to evolve. The idea is based on
capturing discriminativeness and representativeness of each instance and
eliminating the outliers. The final models are used to classify faces on novel
datasets with possibly different characteristics. On benchmark datasets, our
results are comparable to or better than state-of-the-art studies for the task
of face identification.
|
1407.3262 | Brice Boyer | Brice Boyer, Jean-Guillaume Dumas (LJK), Pascal Giorgi (LIRMM),
Cl\'ement Pernet (INRIA Grenoble Rh\^one-Alpes / LIG Laboratoire
d'Informatique de Grenoble), B. David Saunders (CIS) | Elements of Design for Containers and Solutions in the LinBox Library | 8 pages, 4th International Congress on Mathematical Software, Seoul :
Korea, Republic Of (2014) | null | null | null | cs.MS cs.SC cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We describe in this paper new design techniques used in the \cpp exact linear
algebra library \linbox, intended to make the library safer and easier to use,
while keeping it generic and efficient. First, we review the new simplified
structure for containers, based on our \emph{founding scope allocation} model.
We explain design choices and their impact on coding: unification of our matrix
classes, clearer model for matrices and submatrices, \etc Then we present a
variation of the \emph{strategy} design pattern that is comprised of a
controller--plugin system: the controller (solution) chooses among plug-ins
(algorithms) that always call back the controllers for subtasks. We give
examples using the solution \mul. Finally we present a benchmark architecture
that serves two purposes: Providing the user with easier ways to produce
graphs; Creating a framework for automatically tuning the library and
supporting regression testing.
| [
{
"version": "v1",
"created": "Wed, 25 Jun 2014 19:38:09 GMT"
}
] | 2014-07-14T00:00:00 | [
[
"Boyer",
"Brice",
"",
"LJK"
],
[
"Dumas",
"Jean-Guillaume",
"",
"LJK"
],
[
"Giorgi",
"Pascal",
"",
"LIRMM"
],
[
"Pernet",
"Clément",
"",
"INRIA Grenoble Rhône-Alpes / LIG Laboratoire\n d'Informatique de Grenoble"
],
[
"Saunders",
"B. David",
"",
"CIS"
]
] | TITLE: Elements of Design for Containers and Solutions in the LinBox Library
ABSTRACT: We describe in this paper new design techniques used in the \cpp exact linear
algebra library \linbox, intended to make the library safer and easier to use,
while keeping it generic and efficient. First, we review the new simplified
structure for containers, based on our \emph{founding scope allocation} model.
We explain design choices and their impact on coding: unification of our matrix
classes, clearer model for matrices and submatrices, \etc Then we present a
variation of the \emph{strategy} design pattern that is comprised of a
controller--plugin system: the controller (solution) chooses among plug-ins
(algorithms) that always call back the controllers for subtasks. We give
examples using the solution \mul. Finally we present a benchmark architecture
that serves two purposes: Providing the user with easier ways to produce
graphs; Creating a framework for automatically tuning the library and
supporting regression testing.
|
1407.2649 | Alican Bozkurt | Alican Bozkurt, Pinar Duygulu, A. Enis Cetin | Classifying Fonts and Calligraphy Styles Using Complex Wavelet Transform | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/3.0/ | Recognizing fonts has become an important task in document analysis, due to
the increasing number of available digital documents in different fonts and
emphases. A generic font-recognition system independent of language, script and
content is desirable for processing various types of documents. At the same
time, categorizing calligraphy styles in handwritten manuscripts is important
for palaeographic analysis, but has not been studied sufficiently in the
literature. We address the font-recognition problem as analysis and
categorization of textures. We extract features using complex wavelet transform
and use support vector machines for classification. Extensive experimental
evaluations on different datasets in four languages and comparisons with
state-of-the-art studies show that our proposed method achieves higher
recognition accuracy while being computationally simpler. Furthermore, on a new
dataset generated from Ottoman manuscripts, we show that the proposed method
can also be used for categorizing Ottoman calligraphy with high accuracy.
| [
{
"version": "v1",
"created": "Wed, 9 Jul 2014 22:25:32 GMT"
}
] | 2014-07-11T00:00:00 | [
[
"Bozkurt",
"Alican",
""
],
[
"Duygulu",
"Pinar",
""
],
[
"Cetin",
"A. Enis",
""
]
] | TITLE: Classifying Fonts and Calligraphy Styles Using Complex Wavelet Transform
ABSTRACT: Recognizing fonts has become an important task in document analysis, due to
the increasing number of available digital documents in different fonts and
emphases. A generic font-recognition system independent of language, script and
content is desirable for processing various types of documents. At the same
time, categorizing calligraphy styles in handwritten manuscripts is important
for palaeographic analysis, but has not been studied sufficiently in the
literature. We address the font-recognition problem as analysis and
categorization of textures. We extract features using complex wavelet transform
and use support vector machines for classification. Extensive experimental
evaluations on different datasets in four languages and comparisons with
state-of-the-art studies show that our proposed method achieves higher
recognition accuracy while being computationally simpler. Furthermore, on a new
dataset generated from Ottoman manuscripts, we show that the proposed method
can also be used for categorizing Ottoman calligraphy with high accuracy.
|
1407.2683 | Jiaxing Shang | Jiaxing Shang, Lianchen Liu, Feng Xie, Zhen Chen, Jiajia Miao, Xuelin
Fang, Cheng Wu | A Real-Time Detecting Algorithm for Tracking Community Structure of
Dynamic Networks | 9 pages, 6 figures, 3 tables, 6th SNA-KDD Workshop (2012) | null | null | null | cs.SI physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper a simple but efficient real-time detecting algorithm is
proposed for tracking community structure of dynamic networks. Community
structure is intuitively characterized as divisions of network nodes into
subgroups, within which nodes are densely connected while between which they
are sparsely connected. To evaluate the quality of community structure of a
network, a metric called modularity is proposed and many algorithms are
developed on optimizing it. However, most of the modularity based algorithms
deal with static networks and cannot be performed frequently, due to their high
computing complexity. In order to track the community structure of dynamic
networks in a fine-grained way, we propose a modularity based algorithm that is
incremental and has very low computing complexity. In our algorithm we adopt a
two-step approach. Firstly we apply the algorithm of Blondel et al for
detecting static communities to obtain an initial community structure. Then,
apply our incremental updating strategies to track the dynamic communities. The
performance of our algorithm is measured in terms of the modularity. We test
the algorithm on tracking community structure of Enron Email and three other
real world datasets. The experimental results show that our algorithm can keep
track of community structure in time and outperform the well known CNM
algorithm in terms of modularity.
| [
{
"version": "v1",
"created": "Thu, 10 Jul 2014 04:08:29 GMT"
}
] | 2014-07-11T00:00:00 | [
[
"Shang",
"Jiaxing",
""
],
[
"Liu",
"Lianchen",
""
],
[
"Xie",
"Feng",
""
],
[
"Chen",
"Zhen",
""
],
[
"Miao",
"Jiajia",
""
],
[
"Fang",
"Xuelin",
""
],
[
"Wu",
"Cheng",
""
]
] | TITLE: A Real-Time Detecting Algorithm for Tracking Community Structure of
Dynamic Networks
ABSTRACT: In this paper a simple but efficient real-time detecting algorithm is
proposed for tracking community structure of dynamic networks. Community
structure is intuitively characterized as divisions of network nodes into
subgroups, within which nodes are densely connected while between which they
are sparsely connected. To evaluate the quality of community structure of a
network, a metric called modularity is proposed and many algorithms are
developed on optimizing it. However, most of the modularity based algorithms
deal with static networks and cannot be performed frequently, due to their high
computing complexity. In order to track the community structure of dynamic
networks in a fine-grained way, we propose a modularity based algorithm that is
incremental and has very low computing complexity. In our algorithm we adopt a
two-step approach. Firstly we apply the algorithm of Blondel et al for
detecting static communities to obtain an initial community structure. Then,
apply our incremental updating strategies to track the dynamic communities. The
performance of our algorithm is measured in terms of the modularity. We test
the algorithm on tracking community structure of Enron Email and three other
real world datasets. The experimental results show that our algorithm can keep
track of community structure in time and outperform the well known CNM
algorithm in terms of modularity.
|
1407.2697 | Aaron Defazio Mr | Aaron J. Defazio and Tiberio S. Caetano | A Convex Formulation for Learning Scale-Free Networks via Submodular
Relaxation | null | Advances in Neural Information Processing Systems 25 (NIPS 2012) | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A key problem in statistics and machine learning is the determination of
network structure from data. We consider the case where the structure of the
graph to be reconstructed is known to be scale-free. We show that in such cases
it is natural to formulate structured sparsity inducing priors using submodular
functions, and we use their Lov\'asz extension to obtain a convex relaxation.
For tractable classes such as Gaussian graphical models, this leads to a convex
optimization problem that can be efficiently solved. We show that our method
results in an improvement in the accuracy of reconstructed networks for
synthetic data. We also show how our prior encourages scale-free
reconstructions on a bioinfomatics dataset.
| [
{
"version": "v1",
"created": "Thu, 10 Jul 2014 05:45:17 GMT"
}
] | 2014-07-11T00:00:00 | [
[
"Defazio",
"Aaron J.",
""
],
[
"Caetano",
"Tiberio S.",
""
]
] | TITLE: A Convex Formulation for Learning Scale-Free Networks via Submodular
Relaxation
ABSTRACT: A key problem in statistics and machine learning is the determination of
network structure from data. We consider the case where the structure of the
graph to be reconstructed is known to be scale-free. We show that in such cases
it is natural to formulate structured sparsity inducing priors using submodular
functions, and we use their Lov\'asz extension to obtain a convex relaxation.
For tractable classes such as Gaussian graphical models, this leads to a convex
optimization problem that can be efficiently solved. We show that our method
results in an improvement in the accuracy of reconstructed networks for
synthetic data. We also show how our prior encourages scale-free
reconstructions on a bioinfomatics dataset.
|
1407.2736 | Hima Patel | Ramasubramanian Sundararajan, Hima Patel, Manisha Srivastava | A multi-instance learning algorithm based on a stacked ensemble of lazy
learners | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/3.0/ | This document describes a novel learning algorithm that classifies "bags" of
instances rather than individual instances. A bag is labeled positive if it
contains at least one positive instance (which may or may not be specifically
identified), and negative otherwise. This class of problems is known as
multi-instance learning problems, and is useful in situations where the class
label at an instance level may be unavailable or imprecise or difficult to
obtain, or in situations where the problem is naturally posed as one of
classifying instance groups. The algorithm described here is an ensemble-based
method, wherein the members of the ensemble are lazy learning classifiers
learnt using the Citation Nearest Neighbour method. Diversity among the
ensemble members is achieved by optimizing their parameters using a
multi-objective optimization method, with the objectives being to maximize
Class 1 accuracy and minimize false positive rate. The method has been found to
be effective on the Musk1 benchmark dataset.
| [
{
"version": "v1",
"created": "Thu, 10 Jul 2014 09:39:24 GMT"
}
] | 2014-07-11T00:00:00 | [
[
"Sundararajan",
"Ramasubramanian",
""
],
[
"Patel",
"Hima",
""
],
[
"Srivastava",
"Manisha",
""
]
] | TITLE: A multi-instance learning algorithm based on a stacked ensemble of lazy
learners
ABSTRACT: This document describes a novel learning algorithm that classifies "bags" of
instances rather than individual instances. A bag is labeled positive if it
contains at least one positive instance (which may or may not be specifically
identified), and negative otherwise. This class of problems is known as
multi-instance learning problems, and is useful in situations where the class
label at an instance level may be unavailable or imprecise or difficult to
obtain, or in situations where the problem is naturally posed as one of
classifying instance groups. The algorithm described here is an ensemble-based
method, wherein the members of the ensemble are lazy learning classifiers
learnt using the Citation Nearest Neighbour method. Diversity among the
ensemble members is achieved by optimizing their parameters using a
multi-objective optimization method, with the objectives being to maximize
Class 1 accuracy and minimize false positive rate. The method has been found to
be effective on the Musk1 benchmark dataset.
|
1407.2806 | Preux Philippe | J\'er\'emie Mary (INRIA Lille - Nord Europe, LIFL), Romaric Gaudel
(INRIA Lille - Nord Europe, LIFL), Preux Philippe (INRIA Lille - Nord Europe,
LIFL) | Bandits Warm-up Cold Recommender Systems | null | null | null | RR-8563 | cs.LG cs.IR stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We address the cold start problem in recommendation systems assuming no
contextual information is available neither about users, nor items. We consider
the case in which we only have access to a set of ratings of items by users.
Most of the existing works consider a batch setting, and use cross-validation
to tune parameters. The classical method consists in minimizing the root mean
square error over a training subset of the ratings which provides a
factorization of the matrix of ratings, interpreted as a latent representation
of items and users. Our contribution in this paper is 5-fold. First, we
explicit the issues raised by this kind of batch setting for users or items
with very few ratings. Then, we propose an online setting closer to the actual
use of recommender systems; this setting is inspired by the bandit framework.
The proposed methodology can be used to turn any recommender system dataset
(such as Netflix, MovieLens,...) into a sequential dataset. Then, we explicit a
strong and insightful link between contextual bandit algorithms and matrix
factorization; this leads us to a new algorithm that tackles the
exploration/exploitation dilemma associated to the cold start problem in a
strikingly new perspective. Finally, experimental evidence confirm that our
algorithm is effective in dealing with the cold start problem on publicly
available datasets. Overall, the goal of this paper is to bridge the gap
between recommender systems based on matrix factorizations and those based on
contextual bandits.
| [
{
"version": "v1",
"created": "Thu, 10 Jul 2014 14:32:37 GMT"
}
] | 2014-07-11T00:00:00 | [
[
"Mary",
"Jérémie",
"",
"INRIA Lille - Nord Europe, LIFL"
],
[
"Gaudel",
"Romaric",
"",
"INRIA Lille - Nord Europe, LIFL"
],
[
"Philippe",
"Preux",
"",
"INRIA Lille - Nord Europe,\n LIFL"
]
] | TITLE: Bandits Warm-up Cold Recommender Systems
ABSTRACT: We address the cold start problem in recommendation systems assuming no
contextual information is available neither about users, nor items. We consider
the case in which we only have access to a set of ratings of items by users.
Most of the existing works consider a batch setting, and use cross-validation
to tune parameters. The classical method consists in minimizing the root mean
square error over a training subset of the ratings which provides a
factorization of the matrix of ratings, interpreted as a latent representation
of items and users. Our contribution in this paper is 5-fold. First, we
explicit the issues raised by this kind of batch setting for users or items
with very few ratings. Then, we propose an online setting closer to the actual
use of recommender systems; this setting is inspired by the bandit framework.
The proposed methodology can be used to turn any recommender system dataset
(such as Netflix, MovieLens,...) into a sequential dataset. Then, we explicit a
strong and insightful link between contextual bandit algorithms and matrix
factorization; this leads us to a new algorithm that tackles the
exploration/exploitation dilemma associated to the cold start problem in a
strikingly new perspective. Finally, experimental evidence confirm that our
algorithm is effective in dealing with the cold start problem on publicly
available datasets. Overall, the goal of this paper is to bridge the gap
between recommender systems based on matrix factorizations and those based on
contextual bandits.
|
1407.2889 | John-Alexander Assael | Charalampos S. Kouzinopoulos, John-Alexander M. Assael, Themistoklis
K. Pyrgiotis, and Konstantinos G. Margaritis | A Hybrid Parallel Implementation of the Aho-Corasick and Wu-Manber
Algorithms Using NVIDIA CUDA and MPI Evaluated on a Biological Sequence
Database | null | null | null | null | cs.DC cs.CE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multiple matching algorithms are used to locate the occurrences of patterns
from a finite pattern set in a large input string. Aho-Corasick and Wu-Manber,
two of the most well known algorithms for multiple matching require an
increased computing power, particularly in cases where large-size datasets must
be processed, as is common in computational biology applications. Over the past
years, Graphics Processing Units (GPUs) have evolved to powerful parallel
processors outperforming Central Processing Units (CPUs) in scientific
calculations. Moreover, multiple GPUs can be used in parallel, forming hybrid
computer cluster configurations to achieve an even higher processing
throughput. This paper evaluates the speedup of the parallel implementation of
the Aho-Corasick and Wu-Manber algorithms on a hybrid GPU cluster, when used to
process a snapshot of the Expressed Sequence Tags of the human genome and for
different problem parameters.
| [
{
"version": "v1",
"created": "Thu, 10 Jul 2014 18:15:18 GMT"
}
] | 2014-07-11T00:00:00 | [
[
"Kouzinopoulos",
"Charalampos S.",
""
],
[
"Assael",
"John-Alexander M.",
""
],
[
"Pyrgiotis",
"Themistoklis K.",
""
],
[
"Margaritis",
"Konstantinos G.",
""
]
] | TITLE: A Hybrid Parallel Implementation of the Aho-Corasick and Wu-Manber
Algorithms Using NVIDIA CUDA and MPI Evaluated on a Biological Sequence
Database
ABSTRACT: Multiple matching algorithms are used to locate the occurrences of patterns
from a finite pattern set in a large input string. Aho-Corasick and Wu-Manber,
two of the most well known algorithms for multiple matching require an
increased computing power, particularly in cases where large-size datasets must
be processed, as is common in computational biology applications. Over the past
years, Graphics Processing Units (GPUs) have evolved to powerful parallel
processors outperforming Central Processing Units (CPUs) in scientific
calculations. Moreover, multiple GPUs can be used in parallel, forming hybrid
computer cluster configurations to achieve an even higher processing
throughput. This paper evaluates the speedup of the parallel implementation of
the Aho-Corasick and Wu-Manber algorithms on a hybrid GPU cluster, when used to
process a snapshot of the Expressed Sequence Tags of the human genome and for
different problem parameters.
|
1407.2899 | Gabriela Montoya | Gabriela Montoya (LINA), Hala Skaf-Molli (LINA), Pascal Molli (LINA),
Maria-Esther Vidal | Fedra: Query Processing for SPARQL Federations with Divergence | null | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Data replication and deployment of local SPARQL endpoints improve scalability
and availability of public SPARQL endpoints, making the consumption of Linked
Data a reality. This solution requires synchronization and specific query
processing strategies to take advantage of replication. However, existing
replication aware techniques in federations of SPARQL endpoints do not consider
data dynamicity. We propose Fedra, an approach for querying federations of
endpoints that benefits from replication. Participants in Fedra federations can
copy fragments of data from several datasets, and describe them using
provenance and views. These descriptions enable Fedra to reduce the number of
selected endpoints while satisfying user divergence requirements. Experiments
on real-world datasets suggest savings of up to three orders of magnitude.
| [
{
"version": "v1",
"created": "Thu, 10 Jul 2014 18:39:47 GMT"
}
] | 2014-07-11T00:00:00 | [
[
"Montoya",
"Gabriela",
"",
"LINA"
],
[
"Skaf-Molli",
"Hala",
"",
"LINA"
],
[
"Molli",
"Pascal",
"",
"LINA"
],
[
"Vidal",
"Maria-Esther",
""
]
] | TITLE: Fedra: Query Processing for SPARQL Federations with Divergence
ABSTRACT: Data replication and deployment of local SPARQL endpoints improve scalability
and availability of public SPARQL endpoints, making the consumption of Linked
Data a reality. This solution requires synchronization and specific query
processing strategies to take advantage of replication. However, existing
replication aware techniques in federations of SPARQL endpoints do not consider
data dynamicity. We propose Fedra, an approach for querying federations of
endpoints that benefits from replication. Participants in Fedra federations can
copy fragments of data from several datasets, and describe them using
provenance and views. These descriptions enable Fedra to reduce the number of
selected endpoints while satisfying user divergence requirements. Experiments
on real-world datasets suggest savings of up to three orders of magnitude.
|
1407.2220 | Brian Thompson | Graham Cormode, Qiang Ma, S. Muthukrishnan, Brian Thompson | Modeling Collaboration in Academia: A Game Theoretic Approach | Presented at the 1st WWW Workshop on Big Scholarly Data (2014). 6
pages, 5 figures | Proceedings of the Companion Publication of the 23rd International
Conference on World Wide Web (WWW 2014), pgs 1177-1182 | null | null | cs.SI cs.DL cs.GT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work, we aim to understand the mechanisms driving academic
collaboration. We begin by building a model for how researchers split their
effort between multiple papers, and how collaboration affects the number of
citations a paper receives, supported by observations from a large real-world
publication and citation dataset, which we call the h-Reinvestment model. Using
tools from the field of Game Theory, we study researchers' collaborative
behavior over time under this model, with the premise that each researcher
wants to maximize his or her academic success. We find analytically that there
is a strong incentive to collaborate rather than work in isolation, and that
studying collaborative behavior through a game-theoretic lens is a promising
approach to help us better understand the nature and dynamics of academic
collaboration.
| [
{
"version": "v1",
"created": "Tue, 8 Jul 2014 19:09:31 GMT"
},
{
"version": "v2",
"created": "Wed, 9 Jul 2014 04:34:58 GMT"
}
] | 2014-07-10T00:00:00 | [
[
"Cormode",
"Graham",
""
],
[
"Ma",
"Qiang",
""
],
[
"Muthukrishnan",
"S.",
""
],
[
"Thompson",
"Brian",
""
]
] | TITLE: Modeling Collaboration in Academia: A Game Theoretic Approach
ABSTRACT: In this work, we aim to understand the mechanisms driving academic
collaboration. We begin by building a model for how researchers split their
effort between multiple papers, and how collaboration affects the number of
citations a paper receives, supported by observations from a large real-world
publication and citation dataset, which we call the h-Reinvestment model. Using
tools from the field of Game Theory, we study researchers' collaborative
behavior over time under this model, with the premise that each researcher
wants to maximize his or her academic success. We find analytically that there
is a strong incentive to collaborate rather than work in isolation, and that
studying collaborative behavior through a game-theoretic lens is a promising
approach to help us better understand the nature and dynamics of academic
collaboration.
|
1402.6882 | Lu Lu | Minglong Zhang and Lu Lu and Soung-Chang Liew | An Optimal Decoding Strategy for Physical-layer Network Coding over
Multipath Fading Channels | Submitted to IEEE Transactions on Vehicular Technology | null | null | null | cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present an optimal decoder for physical-layer network coding (PNC) in a
multipath fading channels. Previous studies on PNC have largely focused on the
single path case. For PNC, multipath not only introduces inter-symbol
interference (ISI), but also cross-symbol interference (Cross-SI) between
signals simultaneously transmitted by multiple users. In this paper, we assume
the transmitters do not have channel state information (CSI). The relay in the
PNC system, however, has CSI. The relay makes use of a belief propagation (BP)
algorithm to decode the multipath-distorted signals received from multiple
users into a network-coded packet. We refer to our multipath decoding algorithm
as MP-PNC. Our simulation results show that, benchmarked against synchronous
PNC over a one-path channel, the bit error rate (BER) performance penalty of
MP-PNC under a two-tap ITU channel model can be kept within 0.5dB. Moreover, it
outperforms a MUD-XOR algorithm by 3dB -- MUD-XOR decodes the individual
information from both users explicitly before performing the XOR network-coding
mapping. Although the framework of fading-channel PNC presented in this paper
is demonstrated based on two-path and three-path channel models, our algorithm
can be easily extended to cases with more than three paths.
| [
{
"version": "v1",
"created": "Thu, 27 Feb 2014 12:22:38 GMT"
},
{
"version": "v2",
"created": "Tue, 8 Jul 2014 06:06:36 GMT"
}
] | 2014-07-09T00:00:00 | [
[
"Zhang",
"Minglong",
""
],
[
"Lu",
"Lu",
""
],
[
"Liew",
"Soung-Chang",
""
]
] | TITLE: An Optimal Decoding Strategy for Physical-layer Network Coding over
Multipath Fading Channels
ABSTRACT: We present an optimal decoder for physical-layer network coding (PNC) in a
multipath fading channels. Previous studies on PNC have largely focused on the
single path case. For PNC, multipath not only introduces inter-symbol
interference (ISI), but also cross-symbol interference (Cross-SI) between
signals simultaneously transmitted by multiple users. In this paper, we assume
the transmitters do not have channel state information (CSI). The relay in the
PNC system, however, has CSI. The relay makes use of a belief propagation (BP)
algorithm to decode the multipath-distorted signals received from multiple
users into a network-coded packet. We refer to our multipath decoding algorithm
as MP-PNC. Our simulation results show that, benchmarked against synchronous
PNC over a one-path channel, the bit error rate (BER) performance penalty of
MP-PNC under a two-tap ITU channel model can be kept within 0.5dB. Moreover, it
outperforms a MUD-XOR algorithm by 3dB -- MUD-XOR decodes the individual
information from both users explicitly before performing the XOR network-coding
mapping. Although the framework of fading-channel PNC presented in this paper
is demonstrated based on two-path and three-path channel models, our algorithm
can be easily extended to cases with more than three paths.
|
1407.1976 | Shanta Phani | Shanta Phani, Shibamouli Lahiri and Arindam Biswas | Inter-Rater Agreement Study on Readability Assessment in Bengali | 6 pages, 4 tables, Accepted in ICCONAC, 2014 | International Journal on Natural Language Computing (IJNLC), 3(3),
2014 | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | An inter-rater agreement study is performed for readability assessment in
Bengali. A 1-7 rating scale was used to indicate different levels of
readability. We obtained moderate to fair agreement among seven independent
annotators on 30 text passages written by four eminent Bengali authors. As a by
product of our study, we obtained a readability-annotated ground truth dataset
in Bengali. .
| [
{
"version": "v1",
"created": "Tue, 8 Jul 2014 07:35:16 GMT"
}
] | 2014-07-09T00:00:00 | [
[
"Phani",
"Shanta",
""
],
[
"Lahiri",
"Shibamouli",
""
],
[
"Biswas",
"Arindam",
""
]
] | TITLE: Inter-Rater Agreement Study on Readability Assessment in Bengali
ABSTRACT: An inter-rater agreement study is performed for readability assessment in
Bengali. A 1-7 rating scale was used to indicate different levels of
readability. We obtained moderate to fair agreement among seven independent
annotators on 30 text passages written by four eminent Bengali authors. As a by
product of our study, we obtained a readability-annotated ground truth dataset
in Bengali. .
|
1407.2107 | Raghu Machiraju | Hao Ding, Chao Wang, Kun Huang and Raghu Machiraju | iGPSe: A Visual Analytic System for Integrative Genomic Based Cancer
Patient Stratification | BioVis 2014 conference | null | null | null | cs.GR cs.HC q-bio.QM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Background: Cancers are highly heterogeneous with different subtypes. These
subtypes often possess different genetic variants, present different
pathological phenotypes, and most importantly, show various clinical outcomes
such as varied prognosis and response to treatment and likelihood for
recurrence and metastasis. Recently, integrative genomics (or panomics)
approaches are often adopted with the goal of combining multiple types of omics
data to identify integrative biomarkers for stratification of patients into
groups with different clinical outcomes. Results: In this paper we present a
visual analytic system called Interactive Genomics Patient Stratification
explorer (iGPSe) which significantly reduces the computing burden for
biomedical researchers in the process of exploring complicated integrative
genomics data. Our system integrates unsupervised clustering with graph and
parallel sets visualization and allows direct comparison of clinical outcomes
via survival analysis. Using a breast cancer dataset obtained from the The
Cancer Genome Atlas (TCGA) project, we are able to quickly explore different
combinations of gene expression (mRNA) and microRNA features and identify
potential combined markers for survival prediction. Conclusions: Visualization
plays an important role in the process of stratifying given population
patients. Visual tools allowed for the selection of possibly features across
various datasets for the given patient population. We essentially made a case
for visualization for a very important problem in translational informatics.
| [
{
"version": "v1",
"created": "Tue, 8 Jul 2014 14:30:15 GMT"
}
] | 2014-07-09T00:00:00 | [
[
"Ding",
"Hao",
""
],
[
"Wang",
"Chao",
""
],
[
"Huang",
"Kun",
""
],
[
"Machiraju",
"Raghu",
""
]
] | TITLE: iGPSe: A Visual Analytic System for Integrative Genomic Based Cancer
Patient Stratification
ABSTRACT: Background: Cancers are highly heterogeneous with different subtypes. These
subtypes often possess different genetic variants, present different
pathological phenotypes, and most importantly, show various clinical outcomes
such as varied prognosis and response to treatment and likelihood for
recurrence and metastasis. Recently, integrative genomics (or panomics)
approaches are often adopted with the goal of combining multiple types of omics
data to identify integrative biomarkers for stratification of patients into
groups with different clinical outcomes. Results: In this paper we present a
visual analytic system called Interactive Genomics Patient Stratification
explorer (iGPSe) which significantly reduces the computing burden for
biomedical researchers in the process of exploring complicated integrative
genomics data. Our system integrates unsupervised clustering with graph and
parallel sets visualization and allows direct comparison of clinical outcomes
via survival analysis. Using a breast cancer dataset obtained from the The
Cancer Genome Atlas (TCGA) project, we are able to quickly explore different
combinations of gene expression (mRNA) and microRNA features and identify
potential combined markers for survival prediction. Conclusions: Visualization
plays an important role in the process of stratifying given population
patients. Visual tools allowed for the selection of possibly features across
various datasets for the given patient population. We essentially made a case
for visualization for a very important problem in translational informatics.
|
1404.1777 | Victor Lempitsky | Artem Babenko, Anton Slesarev, Alexandr Chigorin and Victor Lempitsky | Neural Codes for Image Retrieval | to appear at ECCV 2014 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | It has been shown that the activations invoked by an image within the top
layers of a large convolutional neural network provide a high-level descriptor
of the visual content of the image. In this paper, we investigate the use of
such descriptors (neural codes) within the image retrieval application. In the
experiments with several standard retrieval benchmarks, we establish that
neural codes perform competitively even when the convolutional neural network
has been trained for an unrelated classification task (e.g.\ Image-Net). We
also evaluate the improvement in the retrieval performance of neural codes,
when the network is retrained on a dataset of images that are similar to images
encountered at test time.
We further evaluate the performance of the compressed neural codes and show
that a simple PCA compression provides very good short codes that give
state-of-the-art accuracy on a number of datasets. In general, neural codes
turn out to be much more resilient to such compression in comparison other
state-of-the-art descriptors. Finally, we show that discriminative
dimensionality reduction trained on a dataset of pairs of matched photographs
improves the performance of PCA-compressed neural codes even further. Overall,
our quantitative experiments demonstrate the promise of neural codes as visual
descriptors for image retrieval.
| [
{
"version": "v1",
"created": "Mon, 7 Apr 2014 13:08:08 GMT"
},
{
"version": "v2",
"created": "Mon, 7 Jul 2014 07:51:04 GMT"
}
] | 2014-07-08T00:00:00 | [
[
"Babenko",
"Artem",
""
],
[
"Slesarev",
"Anton",
""
],
[
"Chigorin",
"Alexandr",
""
],
[
"Lempitsky",
"Victor",
""
]
] | TITLE: Neural Codes for Image Retrieval
ABSTRACT: It has been shown that the activations invoked by an image within the top
layers of a large convolutional neural network provide a high-level descriptor
of the visual content of the image. In this paper, we investigate the use of
such descriptors (neural codes) within the image retrieval application. In the
experiments with several standard retrieval benchmarks, we establish that
neural codes perform competitively even when the convolutional neural network
has been trained for an unrelated classification task (e.g.\ Image-Net). We
also evaluate the improvement in the retrieval performance of neural codes,
when the network is retrained on a dataset of images that are similar to images
encountered at test time.
We further evaluate the performance of the compressed neural codes and show
that a simple PCA compression provides very good short codes that give
state-of-the-art accuracy on a number of datasets. In general, neural codes
turn out to be much more resilient to such compression in comparison other
state-of-the-art descriptors. Finally, we show that discriminative
dimensionality reduction trained on a dataset of pairs of matched photographs
improves the performance of PCA-compressed neural codes even further. Overall,
our quantitative experiments demonstrate the promise of neural codes as visual
descriptors for image retrieval.
|
1407.0386 | Daniel Schall | Daniel Schall and Theo H\"arder | Energy and Performance-Can a Wimpy-Node Cluster Challenge a Brawny
Server? | arXiv admin note: substantial text overlap with arXiv:1407.0120 | null | null | null | cs.DB cs.DC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Traditional DBMS servers are usually over-provisioned for most of their daily
workloads and, because they do not show good energy proportionality, waste a
lot of energy while underutilized. A cluster of small (wimpy) servers, where
the number of nodes can dynamically adjust to the current workload, might offer
better energy characteristics for these workloads. Yet, clusters suffer from
"friction losses" and may not be able to quickly adapt to the workload, whereas
a single, brawny server delivers performance instantaneously. In this paper, we
compare a small cluster of lightweight nodes to a single server in terms of
performance and energy efficiency. We run several benchmarks, consisting of
OLTP and OLAP queries at variable utilization to test the system's ability to
adjust to the workloads. To quantify possible energy saving and its conceivable
drawback on query runtime, we evaluate our implementation on a cluster as well
as on a single, brawny server and compare the results w.r.t. performance and
energy consumption. Our findings confirm that - based on the workload - energy
can be saved without sacrificing too much performance.
| [
{
"version": "v1",
"created": "Tue, 1 Jul 2014 07:29:37 GMT"
},
{
"version": "v2",
"created": "Mon, 7 Jul 2014 07:32:53 GMT"
}
] | 2014-07-08T00:00:00 | [
[
"Schall",
"Daniel",
""
],
[
"Härder",
"Theo",
""
]
] | TITLE: Energy and Performance-Can a Wimpy-Node Cluster Challenge a Brawny
Server?
ABSTRACT: Traditional DBMS servers are usually over-provisioned for most of their daily
workloads and, because they do not show good energy proportionality, waste a
lot of energy while underutilized. A cluster of small (wimpy) servers, where
the number of nodes can dynamically adjust to the current workload, might offer
better energy characteristics for these workloads. Yet, clusters suffer from
"friction losses" and may not be able to quickly adapt to the workload, whereas
a single, brawny server delivers performance instantaneously. In this paper, we
compare a small cluster of lightweight nodes to a single server in terms of
performance and energy efficiency. We run several benchmarks, consisting of
OLTP and OLAP queries at variable utilization to test the system's ability to
adjust to the workloads. To quantify possible energy saving and its conceivable
drawback on query runtime, we evaluate our implementation on a cluster as well
as on a single, brawny server and compare the results w.r.t. performance and
energy consumption. Our findings confirm that - based on the workload - energy
can be saved without sacrificing too much performance.
|
1407.1538 | Xiangnan Kong | Xiangnan Kong and Zhaoming Wu and Li-Jia Li and Ruofei Zhang and
Philip S. Yu and Hang Wu and Wei Fan | Large-Scale Multi-Label Learning with Incomplete Label Assignments | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multi-label learning deals with the classification problems where each
instance can be assigned with multiple labels simultaneously. Conventional
multi-label learning approaches mainly focus on exploiting label correlations.
It is usually assumed, explicitly or implicitly, that the label sets for
training instances are fully labeled without any missing labels. However, in
many real-world multi-label datasets, the label assignments for training
instances can be incomplete. Some ground-truth labels can be missed by the
labeler from the label set. This problem is especially typical when the number
instances is very large, and the labeling cost is very high, which makes it
almost impossible to get a fully labeled training set. In this paper, we study
the problem of large-scale multi-label learning with incomplete label
assignments. We propose an approach, called MPU, based upon positive and
unlabeled stochastic gradient descent and stacked models. Unlike prior works,
our method can effectively and efficiently consider missing labels and label
correlations simultaneously, and is very scalable, that has linear time
complexities over the size of the data. Extensive experiments on two real-world
multi-label datasets show that our MPU model consistently outperform other
commonly-used baselines.
| [
{
"version": "v1",
"created": "Sun, 6 Jul 2014 20:13:48 GMT"
}
] | 2014-07-08T00:00:00 | [
[
"Kong",
"Xiangnan",
""
],
[
"Wu",
"Zhaoming",
""
],
[
"Li",
"Li-Jia",
""
],
[
"Zhang",
"Ruofei",
""
],
[
"Yu",
"Philip S.",
""
],
[
"Wu",
"Hang",
""
],
[
"Fan",
"Wei",
""
]
] | TITLE: Large-Scale Multi-Label Learning with Incomplete Label Assignments
ABSTRACT: Multi-label learning deals with the classification problems where each
instance can be assigned with multiple labels simultaneously. Conventional
multi-label learning approaches mainly focus on exploiting label correlations.
It is usually assumed, explicitly or implicitly, that the label sets for
training instances are fully labeled without any missing labels. However, in
many real-world multi-label datasets, the label assignments for training
instances can be incomplete. Some ground-truth labels can be missed by the
labeler from the label set. This problem is especially typical when the number
instances is very large, and the labeling cost is very high, which makes it
almost impossible to get a fully labeled training set. In this paper, we study
the problem of large-scale multi-label learning with incomplete label
assignments. We propose an approach, called MPU, based upon positive and
unlabeled stochastic gradient descent and stacked models. Unlike prior works,
our method can effectively and efficiently consider missing labels and label
correlations simultaneously, and is very scalable, that has linear time
complexities over the size of the data. Extensive experiments on two real-world
multi-label datasets show that our MPU model consistently outperform other
commonly-used baselines.
|
1407.1572 | Sivaram Ambikasaran | Sivaram Ambikasaran and Eric Darve | The Inverse Fast Multipole Method | 25 pages, 28 figures | null | null | null | math.NA cs.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This article introduces a new fast direct solver for linear systems arising
out of wide range of applications, integral equations, multivariate statistics,
radial basis interpolation, etc., to name a few. \emph{The highlight of this
new fast direct solver is that the solver scales linearly in the number of
unknowns in all dimensions.} The solver, termed as Inverse Fast Multipole
Method (abbreviated as IFMM), works on the same data-structure as the Fast
Multipole Method (abbreviated as FMM). More generally, the solver can be
immediately extended to the class of hierarchical matrices, denoted as
$\mathcal{H}^2$ matrices with strong admissibility criteria (weak low-rank
structure), i.e., \emph{the interaction between neighboring cluster of
particles is full-rank whereas the interaction between particles corresponding
to well-separated clusters can be efficiently represented as a low-rank
matrix}. The algorithm departs from existing approaches in the fact that
throughout the algorithm the interaction corresponding to neighboring clusters
are always treated as full-rank interactions. Our approach relies on two major
ideas: (i) The $N \times N$ matrix arising out of FMM (from now on termed as
FMM matrix) can be represented as an extended sparser matrix of size $M \times
M$, where $M \approx 3N$. (ii) While solving the larger extended sparser
matrix, \emph{the fill-in's that arise in the matrix blocks corresponding to
well-separated clusters are hierarchically compressed}. The ordering of the
equations and the unknowns in the extended sparser matrix is strongly related
to the local and multipole coefficients in the FMM~\cite{greengard1987fast} and
\emph{the order of elimination is different from the usual nested dissection
approach}. Numerical benchmarks on $2$D manifold confirm the linear scaling of
the algorithm.
| [
{
"version": "v1",
"created": "Mon, 7 Jul 2014 03:16:40 GMT"
}
] | 2014-07-08T00:00:00 | [
[
"Ambikasaran",
"Sivaram",
""
],
[
"Darve",
"Eric",
""
]
] | TITLE: The Inverse Fast Multipole Method
ABSTRACT: This article introduces a new fast direct solver for linear systems arising
out of wide range of applications, integral equations, multivariate statistics,
radial basis interpolation, etc., to name a few. \emph{The highlight of this
new fast direct solver is that the solver scales linearly in the number of
unknowns in all dimensions.} The solver, termed as Inverse Fast Multipole
Method (abbreviated as IFMM), works on the same data-structure as the Fast
Multipole Method (abbreviated as FMM). More generally, the solver can be
immediately extended to the class of hierarchical matrices, denoted as
$\mathcal{H}^2$ matrices with strong admissibility criteria (weak low-rank
structure), i.e., \emph{the interaction between neighboring cluster of
particles is full-rank whereas the interaction between particles corresponding
to well-separated clusters can be efficiently represented as a low-rank
matrix}. The algorithm departs from existing approaches in the fact that
throughout the algorithm the interaction corresponding to neighboring clusters
are always treated as full-rank interactions. Our approach relies on two major
ideas: (i) The $N \times N$ matrix arising out of FMM (from now on termed as
FMM matrix) can be represented as an extended sparser matrix of size $M \times
M$, where $M \approx 3N$. (ii) While solving the larger extended sparser
matrix, \emph{the fill-in's that arise in the matrix blocks corresponding to
well-separated clusters are hierarchically compressed}. The ordering of the
equations and the unknowns in the extended sparser matrix is strongly related
to the local and multipole coefficients in the FMM~\cite{greengard1987fast} and
\emph{the order of elimination is different from the usual nested dissection
approach}. Numerical benchmarks on $2$D manifold confirm the linear scaling of
the algorithm.
|
1407.1640 | Bin Gao | Bin Gao, Jiang Bian, and Tie-Yan Liu | WordRep: A Benchmark for Research on Learning Word Representations | null | null | null | null | cs.CL cs.LG | http://creativecommons.org/licenses/by/3.0/ | WordRep is a benchmark collection for the research on learning distributed
word representations (or word embeddings), released by Microsoft Research. In
this paper, we describe the details of the WordRep collection and show how to
use it in different types of machine learning research related to word
embedding. Specifically, we describe how the evaluation tasks in WordRep are
selected, how the data are sampled, and how the evaluation tool is built. We
then compare several state-of-the-art word representations on WordRep, report
their evaluation performance, and make discussions on the results. After that,
we discuss new potential research topics that can be supported by WordRep, in
addition to algorithm comparison. We hope that this paper can help people gain
deeper understanding of WordRep, and enable more interesting research on
learning distributed word representations and related topics.
| [
{
"version": "v1",
"created": "Mon, 7 Jul 2014 09:31:21 GMT"
}
] | 2014-07-08T00:00:00 | [
[
"Gao",
"Bin",
""
],
[
"Bian",
"Jiang",
""
],
[
"Liu",
"Tie-Yan",
""
]
] | TITLE: WordRep: A Benchmark for Research on Learning Word Representations
ABSTRACT: WordRep is a benchmark collection for the research on learning distributed
word representations (or word embeddings), released by Microsoft Research. In
this paper, we describe the details of the WordRep collection and show how to
use it in different types of machine learning research related to word
embedding. Specifically, we describe how the evaluation tasks in WordRep are
selected, how the data are sampled, and how the evaluation tool is built. We
then compare several state-of-the-art word representations on WordRep, report
their evaluation performance, and make discussions on the results. After that,
we discuss new potential research topics that can be supported by WordRep, in
addition to algorithm comparison. We hope that this paper can help people gain
deeper understanding of WordRep, and enable more interesting research on
learning distributed word representations and related topics.
|
1407.1772 | Senzhang Wang | Senzhang Wang and Sihong Xie and Xiaoming Zhang and Zhoujun Li and
Philip S. Yu and Xinyu Shu | Future Influence Ranking of Scientific Literature | 9 pages, Proceedings of the 2014 SIAM International Conference on
Data Mining | null | 10.1137/1.9781611973440.86 | null | cs.SI cs.DL physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Researchers or students entering a emerging research area are particularly
interested in what newly published papers will be most cited and which young
researchers will become influential in the future, so that they can catch the
most recent advances and find valuable research directions. However, predicting
the future importance of scientific articles and authors is challenging due to
the dynamic nature of literature networks and evolving research topics.
Different from most previous studies aiming to rank the current importance of
literatures and authors, we focus on \emph{ranking the future popularity of new
publications and young researchers} by proposing a unified ranking model to
combine various available information. Specifically, we first propose to
extract two kinds of text features, words and words co-occurrence to
characterize innovative papers and authors. Then, instead of using static and
un-weighted graphs, we construct time-aware weighted graphs to distinguish the
various importance of links established at different time. Finally, by
leveraging both the constructed text features and graphs, we propose a mutual
reinforcement ranking framework called \emph{MRFRank} to rank the future
importance of papers and authors simultaneously. Experimental results on the
ArnetMiner dataset show that the proposed approach significantly outperforms
the baselines on the metric \emph{recommendation intensity}.
| [
{
"version": "v1",
"created": "Mon, 7 Jul 2014 17:00:34 GMT"
}
] | 2014-07-08T00:00:00 | [
[
"Wang",
"Senzhang",
""
],
[
"Xie",
"Sihong",
""
],
[
"Zhang",
"Xiaoming",
""
],
[
"Li",
"Zhoujun",
""
],
[
"Yu",
"Philip S.",
""
],
[
"Shu",
"Xinyu",
""
]
] | TITLE: Future Influence Ranking of Scientific Literature
ABSTRACT: Researchers or students entering a emerging research area are particularly
interested in what newly published papers will be most cited and which young
researchers will become influential in the future, so that they can catch the
most recent advances and find valuable research directions. However, predicting
the future importance of scientific articles and authors is challenging due to
the dynamic nature of literature networks and evolving research topics.
Different from most previous studies aiming to rank the current importance of
literatures and authors, we focus on \emph{ranking the future popularity of new
publications and young researchers} by proposing a unified ranking model to
combine various available information. Specifically, we first propose to
extract two kinds of text features, words and words co-occurrence to
characterize innovative papers and authors. Then, instead of using static and
un-weighted graphs, we construct time-aware weighted graphs to distinguish the
various importance of links established at different time. Finally, by
leveraging both the constructed text features and graphs, we propose a mutual
reinforcement ranking framework called \emph{MRFRank} to rank the future
importance of papers and authors simultaneously. Experimental results on the
ArnetMiner dataset show that the proposed approach significantly outperforms
the baselines on the metric \emph{recommendation intensity}.
|
1407.1165 | Prashant Borde | Prashant Bordea, Amarsinh Varpeb, Ramesh Manzac, Pravin Yannawara | Recognition of Isolated Words using Zernike and MFCC features for Audio
Visual Speech Recognition | null | null | null | null | cs.CV cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Automatic Speech Recognition (ASR) by machine is an attractive research topic
in signal processing domain and has attracted many researchers to contribute in
this area. In recent year, there have been many advances in automatic speech
reading system with the inclusion of audio and visual speech features to
recognize words under noisy conditions. The objective of audio-visual speech
recognition system is to improve recognition accuracy. In this paper we
computed visual features using Zernike moments and audio feature using Mel
Frequency Cepstral Coefficients (MFCC) on vVISWa (Visual Vocabulary of
Independent Standard Words) dataset which contains collection of isolated set
of city names of 10 speakers. The visual features were normalized and dimension
of features set was reduced by Principal Component Analysis (PCA) in order to
recognize the isolated word utterance on PCA space.The performance of
recognition of isolated words based on visual only and audio only features
results in 63.88 and 100 respectively.
| [
{
"version": "v1",
"created": "Fri, 4 Jul 2014 09:32:10 GMT"
}
] | 2014-07-07T00:00:00 | [
[
"Bordea",
"Prashant",
""
],
[
"Varpeb",
"Amarsinh",
""
],
[
"Manzac",
"Ramesh",
""
],
[
"Yannawara",
"Pravin",
""
]
] | TITLE: Recognition of Isolated Words using Zernike and MFCC features for Audio
Visual Speech Recognition
ABSTRACT: Automatic Speech Recognition (ASR) by machine is an attractive research topic
in signal processing domain and has attracted many researchers to contribute in
this area. In recent year, there have been many advances in automatic speech
reading system with the inclusion of audio and visual speech features to
recognize words under noisy conditions. The objective of audio-visual speech
recognition system is to improve recognition accuracy. In this paper we
computed visual features using Zernike moments and audio feature using Mel
Frequency Cepstral Coefficients (MFCC) on vVISWa (Visual Vocabulary of
Independent Standard Words) dataset which contains collection of isolated set
of city names of 10 speakers. The visual features were normalized and dimension
of features set was reduced by Principal Component Analysis (PCA) in order to
recognize the isolated word utterance on PCA space.The performance of
recognition of isolated words based on visual only and audio only features
results in 63.88 and 100 respectively.
|
1407.1176 | Felipe Llinares | Felipe Llinares, Mahito Sugiyama, Karsten M. Borgwardt | Identifying Higher-order Combinations of Binary Features | null | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Finding statistically significant interactions between binary variables is
computationally and statistically challenging in high-dimensional settings, due
to the combinatorial explosion in the number of hypotheses. Terada et al.
recently showed how to elegantly address this multiple testing problem by
excluding non-testable hypotheses. Still, it remains unclear how their approach
scales to large datasets.
We here proposed strategies to speed up the approach by Terada et al. and
evaluate them thoroughly in 11 real-world benchmark datasets. We observe that
one approach, incremental search with early stopping, is orders of magnitude
faster than the current state-of-the-art approach.
| [
{
"version": "v1",
"created": "Fri, 4 Jul 2014 10:17:43 GMT"
}
] | 2014-07-07T00:00:00 | [
[
"Llinares",
"Felipe",
""
],
[
"Sugiyama",
"Mahito",
""
],
[
"Borgwardt",
"Karsten M.",
""
]
] | TITLE: Identifying Higher-order Combinations of Binary Features
ABSTRACT: Finding statistically significant interactions between binary variables is
computationally and statistically challenging in high-dimensional settings, due
to the combinatorial explosion in the number of hypotheses. Terada et al.
recently showed how to elegantly address this multiple testing problem by
excluding non-testable hypotheses. Still, it remains unclear how their approach
scales to large datasets.
We here proposed strategies to speed up the approach by Terada et al. and
evaluate them thoroughly in 11 real-world benchmark datasets. We observe that
one approach, incremental search with early stopping, is orders of magnitude
faster than the current state-of-the-art approach.
|
1407.1208 | Piotr Bojanowski | Piotr Bojanowski, R\'emi Lajugie, Francis Bach, Ivan Laptev, Jean
Ponce, Cordelia Schmid, Josef Sivic | Weakly Supervised Action Labeling in Videos Under Ordering Constraints | 17 pages, completed version of a ECCV2014 conference paper | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We are given a set of video clips, each one annotated with an {\em ordered}
list of actions, such as "walk" then "sit" then "answer phone" extracted from,
for example, the associated text script. We seek to temporally localize the
individual actions in each clip as well as to learn a discriminative classifier
for each action. We formulate the problem as a weakly supervised temporal
assignment with ordering constraints. Each video clip is divided into small
time intervals and each time interval of each video clip is assigned one action
label, while respecting the order in which the action labels appear in the
given annotations. We show that the action label assignment can be determined
together with learning a classifier for each action in a discriminative manner.
We evaluate the proposed model on a new and challenging dataset of 937 video
clips with a total of 787720 frames containing sequences of 16 different
actions from 69 Hollywood movies.
| [
{
"version": "v1",
"created": "Fri, 4 Jul 2014 12:53:15 GMT"
}
] | 2014-07-07T00:00:00 | [
[
"Bojanowski",
"Piotr",
""
],
[
"Lajugie",
"Rémi",
""
],
[
"Bach",
"Francis",
""
],
[
"Laptev",
"Ivan",
""
],
[
"Ponce",
"Jean",
""
],
[
"Schmid",
"Cordelia",
""
],
[
"Sivic",
"Josef",
""
]
] | TITLE: Weakly Supervised Action Labeling in Videos Under Ordering Constraints
ABSTRACT: We are given a set of video clips, each one annotated with an {\em ordered}
list of actions, such as "walk" then "sit" then "answer phone" extracted from,
for example, the associated text script. We seek to temporally localize the
individual actions in each clip as well as to learn a discriminative classifier
for each action. We formulate the problem as a weakly supervised temporal
assignment with ordering constraints. Each video clip is divided into small
time intervals and each time interval of each video clip is assigned one action
label, while respecting the order in which the action labels appear in the
given annotations. We show that the action label assignment can be determined
together with learning a classifier for each action in a discriminative manner.
We evaluate the proposed model on a new and challenging dataset of 937 video
clips with a total of 787720 frames containing sequences of 16 different
actions from 69 Hollywood movies.
|
1406.6558 | Yaroslav Ganin | Yaroslav Ganin, Victor Lempitsky | $ N^4 $-Fields: Neural Network Nearest Neighbor Fields for Image
Transforms | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a new architecture for difficult image processing operations, such
as natural edge detection or thin object segmentation. The architecture is
based on a simple combination of convolutional neural networks with the nearest
neighbor search.
We focus our attention on the situations when the desired image
transformation is too hard for a neural network to learn explicitly. We show
that in such situations, the use of the nearest neighbor search on top of the
network output allows to improve the results considerably and to account for
the underfitting effect during the neural network training. The approach is
validated on three challenging benchmarks, where the performance of the
proposed architecture matches or exceeds the state-of-the-art.
| [
{
"version": "v1",
"created": "Wed, 25 Jun 2014 13:10:56 GMT"
},
{
"version": "v2",
"created": "Thu, 3 Jul 2014 08:07:52 GMT"
}
] | 2014-07-04T00:00:00 | [
[
"Ganin",
"Yaroslav",
""
],
[
"Lempitsky",
"Victor",
""
]
] | TITLE: $ N^4 $-Fields: Neural Network Nearest Neighbor Fields for Image
Transforms
ABSTRACT: We propose a new architecture for difficult image processing operations, such
as natural edge detection or thin object segmentation. The architecture is
based on a simple combination of convolutional neural networks with the nearest
neighbor search.
We focus our attention on the situations when the desired image
transformation is too hard for a neural network to learn explicitly. We show
that in such situations, the use of the nearest neighbor search on top of the
network output allows to improve the results considerably and to account for
the underfitting effect during the neural network training. The approach is
validated on three challenging benchmarks, where the performance of the
proposed architecture matches or exceeds the state-of-the-art.
|
1407.0717 | Lubomir Bourdev | Lubomir Bourdev, Fei Yang, Rob Fergus | Deep Poselets for Human Detection | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We address the problem of detecting people in natural scenes using a part
approach based on poselets. We propose a bootstrapping method that allows us to
collect millions of weakly labeled examples for each poselet type. We use these
examples to train a Convolutional Neural Net to discriminate different poselet
types and separate them from the background class. We then use the trained CNN
as a way to represent poselet patches with a Pose Discriminative Feature (PDF)
vector -- a compact 256-dimensional feature vector that is effective at
discriminating pose from appearance. We train the poselet model on top of PDF
features and combine them with object-level CNNs for detection and bounding box
prediction. The resulting model leads to state-of-the-art performance for human
detection on the PASCAL datasets.
| [
{
"version": "v1",
"created": "Wed, 2 Jul 2014 20:28:22 GMT"
}
] | 2014-07-04T00:00:00 | [
[
"Bourdev",
"Lubomir",
""
],
[
"Yang",
"Fei",
""
],
[
"Fergus",
"Rob",
""
]
] | TITLE: Deep Poselets for Human Detection
ABSTRACT: We address the problem of detecting people in natural scenes using a part
approach based on poselets. We propose a bootstrapping method that allows us to
collect millions of weakly labeled examples for each poselet type. We use these
examples to train a Convolutional Neural Net to discriminate different poselet
types and separate them from the background class. We then use the trained CNN
as a way to represent poselet patches with a Pose Discriminative Feature (PDF)
vector -- a compact 256-dimensional feature vector that is effective at
discriminating pose from appearance. We train the poselet model on top of PDF
features and combine them with object-level CNNs for detection and bounding box
prediction. The resulting model leads to state-of-the-art performance for human
detection on the PASCAL datasets.
|
1407.0786 | Chunhua Shen | Sakrapee Paisitkriangkrai, Chunhua Shen, Anton van den Hengel | Strengthening the Effectiveness of Pedestrian Detection with Spatially
Pooled Features | 16 pages. Appearing in Proc. European Conf. Computer Vision (ECCV)
2014 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a simple yet effective approach to the problem of pedestrian
detection which outperforms the current state-of-the-art. Our new features are
built on the basis of low-level visual features and spatial pooling.
Incorporating spatial pooling improves the translational invariance and thus
the robustness of the detection process. We then directly optimise the partial
area under the ROC curve (\pAUC) measure, which concentrates detection
performance in the range of most practical importance. The combination of these
factors leads to a pedestrian detector which outperforms all competitors on all
of the standard benchmark datasets. We advance state-of-the-art results by
lowering the average miss rate from $13\%$ to $11\%$ on the INRIA benchmark,
$41\%$ to $37\%$ on the ETH benchmark, $51\%$ to $42\%$ on the TUD-Brussels
benchmark and $36\%$ to $29\%$ on the Caltech-USA benchmark.
| [
{
"version": "v1",
"created": "Thu, 3 Jul 2014 05:39:30 GMT"
}
] | 2014-07-04T00:00:00 | [
[
"Paisitkriangkrai",
"Sakrapee",
""
],
[
"Shen",
"Chunhua",
""
],
[
"Hengel",
"Anton van den",
""
]
] | TITLE: Strengthening the Effectiveness of Pedestrian Detection with Spatially
Pooled Features
ABSTRACT: We propose a simple yet effective approach to the problem of pedestrian
detection which outperforms the current state-of-the-art. Our new features are
built on the basis of low-level visual features and spatial pooling.
Incorporating spatial pooling improves the translational invariance and thus
the robustness of the detection process. We then directly optimise the partial
area under the ROC curve (\pAUC) measure, which concentrates detection
performance in the range of most practical importance. The combination of these
factors leads to a pedestrian detector which outperforms all competitors on all
of the standard benchmark datasets. We advance state-of-the-art results by
lowering the average miss rate from $13\%$ to $11\%$ on the INRIA benchmark,
$41\%$ to $37\%$ on the ETH benchmark, $51\%$ to $42\%$ on the TUD-Brussels
benchmark and $36\%$ to $29\%$ on the Caltech-USA benchmark.
|
1407.0927 | Dominique Mery | Dominique M\'ery (LORIA), Neeraj Kumar Singh (McMaster University) | Modelling an Aircraft Landing System in Event-B (Full Report) | null | null | null | null | cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The failure of hardware or software in a critical system can lead to loss of
lives. The design errors can be main source of the failures that can be
introduced during system development process. Formal techniques are an
alternative approach to verify the correctness of critical systems, overcoming
limitations of the traditional validation techniques such as simulation and
testing. The increasing complexity and failure rate brings new challenges in
the area of verification and validation of avionic systems. Since the
reliability of the software cannot be quantified, the \textit{correct by
construction} approach can implement a reliable system. Refinement plays a
major role to build a large system incrementally from an abstract specification
to a concrete system. This paper contributes as a stepwise formal development
of the landing system of an aircraft. The formal models include the complex
behaviour, temporal behaviour and sequence of operations of the landing gear
system. The models are formalized in Event-B modelling language, which supports
stepwise refinement. This case study is considered as a benchmark for
techniques and tools dedicated to the verification of behavioural properties of
systems. The report is the full version of a paper published for the ABZ 2014
Case Study. is
| [
{
"version": "v1",
"created": "Thu, 3 Jul 2014 14:08:57 GMT"
}
] | 2014-07-04T00:00:00 | [
[
"Méry",
"Dominique",
"",
"LORIA"
],
[
"Singh",
"Neeraj Kumar",
"",
"McMaster University"
]
] | TITLE: Modelling an Aircraft Landing System in Event-B (Full Report)
ABSTRACT: The failure of hardware or software in a critical system can lead to loss of
lives. The design errors can be main source of the failures that can be
introduced during system development process. Formal techniques are an
alternative approach to verify the correctness of critical systems, overcoming
limitations of the traditional validation techniques such as simulation and
testing. The increasing complexity and failure rate brings new challenges in
the area of verification and validation of avionic systems. Since the
reliability of the software cannot be quantified, the \textit{correct by
construction} approach can implement a reliable system. Refinement plays a
major role to build a large system incrementally from an abstract specification
to a concrete system. This paper contributes as a stepwise formal development
of the landing system of an aircraft. The formal models include the complex
behaviour, temporal behaviour and sequence of operations of the landing gear
system. The models are formalized in Event-B modelling language, which supports
stepwise refinement. This case study is considered as a benchmark for
techniques and tools dedicated to the verification of behavioural properties of
systems. The report is the full version of a paper published for the ABZ 2014
Case Study. is
|
1407.0935 | Gopalkrishna MT | M. T Gopalakrishna, M. Ravishankar and D. R Rameshbabu | Multiple Moving Object Recognitions in video based on Log Gabor-PCA
Approach | 8,26,conference | null | 10.1007/978-3-319-01778-5_10 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Object recognition in the video sequence or images is one of the sub-field of
computer vision. Moving object recognition from a video sequence is an
appealing topic with applications in various areas such as airport safety,
intrusion surveillance, video monitoring, intelligent highway, etc. Moving
object recognition is the most challenging task in intelligent video
surveillance system. In this regard, many techniques have been proposed based
on different methods. Despite of its importance, moving object recognition in
complex environments is still far from being completely solved for low
resolution videos, foggy videos, and also dim video sequences. All in all,
these make it necessary to develop exceedingly robust techniques. This paper
introduces multiple moving object recognition in the video sequence based on
LoG Gabor-PCA approach and Angle based distance Similarity measures techniques
used to recognize the object as a human, vehicle etc. Number of experiments are
conducted for indoor and outdoor video sequences of standard datasets and also
our own collection of video sequences comprising of partial night vision video
sequences. Experimental results show that our proposed approach achieves an
excellent recognition rate. Results obtained are satisfactory and competent.
| [
{
"version": "v1",
"created": "Thu, 3 Jul 2014 14:52:56 GMT"
}
] | 2014-07-04T00:00:00 | [
[
"Gopalakrishna",
"M. T",
""
],
[
"Ravishankar",
"M.",
""
],
[
"Rameshbabu",
"D. R",
""
]
] | TITLE: Multiple Moving Object Recognitions in video based on Log Gabor-PCA
Approach
ABSTRACT: Object recognition in the video sequence or images is one of the sub-field of
computer vision. Moving object recognition from a video sequence is an
appealing topic with applications in various areas such as airport safety,
intrusion surveillance, video monitoring, intelligent highway, etc. Moving
object recognition is the most challenging task in intelligent video
surveillance system. In this regard, many techniques have been proposed based
on different methods. Despite of its importance, moving object recognition in
complex environments is still far from being completely solved for low
resolution videos, foggy videos, and also dim video sequences. All in all,
these make it necessary to develop exceedingly robust techniques. This paper
introduces multiple moving object recognition in the video sequence based on
LoG Gabor-PCA approach and Angle based distance Similarity measures techniques
used to recognize the object as a human, vehicle etc. Number of experiments are
conducted for indoor and outdoor video sequences of standard datasets and also
our own collection of video sequences comprising of partial night vision video
sequences. Experimental results show that our proposed approach achieves an
excellent recognition rate. Results obtained are satisfactory and competent.
|
1407.0977 | Robert Nowotniak | Robert Nowotniak and Jacek Kucharski | Higher-Order Quantum-Inspired Genetic Algorithms | null | null | null | null | cs.NE quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents a theory and an empirical evaluation of Higher-Order
Quantum-Inspired Genetic Algorithms. Fundamental notions of the theory have
been introduced, and a novel Order-2 Quantum-Inspired Genetic Algorithm (QIGA2)
has been presented. Contrary to all QIGA algorithms which represent quantum
genes as independent qubits, in higher-order QIGAs quantum registers are used
to represent genes strings which allows modelling of genes relations using
quantum phenomena. Performance comparison has been conducted on a benchmark of
20 deceptive combinatorial optimization problems. It has been presented that
using higher quantum orders is beneficial for genetic algorithm efficiency, and
the new QIGA2 algorithm outperforms the old QIGA algorithm which was tuned in
highly compute intensive metaoptimization process.
| [
{
"version": "v1",
"created": "Wed, 2 Jul 2014 15:22:11 GMT"
}
] | 2014-07-04T00:00:00 | [
[
"Nowotniak",
"Robert",
""
],
[
"Kucharski",
"Jacek",
""
]
] | TITLE: Higher-Order Quantum-Inspired Genetic Algorithms
ABSTRACT: This paper presents a theory and an empirical evaluation of Higher-Order
Quantum-Inspired Genetic Algorithms. Fundamental notions of the theory have
been introduced, and a novel Order-2 Quantum-Inspired Genetic Algorithm (QIGA2)
has been presented. Contrary to all QIGA algorithms which represent quantum
genes as independent qubits, in higher-order QIGAs quantum registers are used
to represent genes strings which allows modelling of genes relations using
quantum phenomena. Performance comparison has been conducted on a benchmark of
20 deceptive combinatorial optimization problems. It has been presented that
using higher quantum orders is beneficial for genetic algorithm efficiency, and
the new QIGA2 algorithm outperforms the old QIGA algorithm which was tuned in
highly compute intensive metaoptimization process.
|
1407.0455 | Yingyi Bu | Yingyi Bu, Vinayak Borkar, Jianfeng Jia, Michael J. Carey, Tyson
Condie | Pregelix: Big(ger) Graph Analytics on A Dataflow Engine | null | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | There is a growing need for distributed graph processing systems that are
capable of gracefully scaling to very large graph datasets. Unfortunately, this
challenge has not been easily met due to the intense memory pressure imposed by
process-centric, message passing designs that many graph processing systems
follow. Pregelix is a new open source distributed graph processing system that
is based on an iterative dataflow design that is better tuned to handle both
in-memory and out-of-core workloads. As such, Pregelix offers improved
performance characteristics and scaling properties over current open source
systems (e.g., we have seen up to 15x speedup compared to Apache Giraph and up
to 35x speedup compared to distributed GraphLab), and makes more effective use
of available machine resources to support Big(ger) Graph Analytics.
| [
{
"version": "v1",
"created": "Wed, 2 Jul 2014 05:04:28 GMT"
}
] | 2014-07-03T00:00:00 | [
[
"Bu",
"Yingyi",
""
],
[
"Borkar",
"Vinayak",
""
],
[
"Jia",
"Jianfeng",
""
],
[
"Carey",
"Michael J.",
""
],
[
"Condie",
"Tyson",
""
]
] | TITLE: Pregelix: Big(ger) Graph Analytics on A Dataflow Engine
ABSTRACT: There is a growing need for distributed graph processing systems that are
capable of gracefully scaling to very large graph datasets. Unfortunately, this
challenge has not been easily met due to the intense memory pressure imposed by
process-centric, message passing designs that many graph processing systems
follow. Pregelix is a new open source distributed graph processing system that
is based on an iterative dataflow design that is better tuned to handle both
in-memory and out-of-core workloads. As such, Pregelix offers improved
performance characteristics and scaling properties over current open source
systems (e.g., we have seen up to 15x speedup compared to Apache Giraph and up
to 35x speedup compared to distributed GraphLab), and makes more effective use
of available machine resources to support Big(ger) Graph Analytics.
|
1407.0547 | Mark Phillips | Mark Phillips, Lauren Ko | Understanding Repository Growth at the University of North Texas: A Case
Study | 5 pages | null | null | null | cs.DL | http://creativecommons.org/licenses/by/3.0/ | Over the past decade the University of North Texas Libraries (UNTL) has
developed a sizable digital library infrastructure for use in carrying out its
core mission to the students, faculty, staff and associated communities of the
university. This repository of content offers countless research possibilities
for end users across the Internet when it is discovered and used in research,
scholarship, entertainment, and lifelong learning. The characteristics of the
repository itself provide insight into the workings of a modern digital library
infrastructure, how it was created, how often it is updated, or how often it is
modified. In that vein, the authors created a dataset comprised of information
extracted from the UNT Libraries' archival repository Coda and analyzed this
dataset in order to demonstrate the value and insights that can be gained from
sharing repository characteristics more broadly. This case study presents the
findings from an analysis of this dataset.
| [
{
"version": "v1",
"created": "Wed, 2 Jul 2014 12:55:49 GMT"
}
] | 2014-07-03T00:00:00 | [
[
"Phillips",
"Mark",
""
],
[
"Ko",
"Lauren",
""
]
] | TITLE: Understanding Repository Growth at the University of North Texas: A Case
Study
ABSTRACT: Over the past decade the University of North Texas Libraries (UNTL) has
developed a sizable digital library infrastructure for use in carrying out its
core mission to the students, faculty, staff and associated communities of the
university. This repository of content offers countless research possibilities
for end users across the Internet when it is discovered and used in research,
scholarship, entertainment, and lifelong learning. The characteristics of the
repository itself provide insight into the workings of a modern digital library
infrastructure, how it was created, how often it is updated, or how often it is
modified. In that vein, the authors created a dataset comprised of information
extracted from the UNT Libraries' archival repository Coda and analyzed this
dataset in order to demonstrate the value and insights that can be gained from
sharing repository characteristics more broadly. This case study presents the
findings from an analysis of this dataset.
|
1209.5429 | Yasser Gonzalez Fernandez | Yasser Gonzalez-Fernandez, Marta Soto | copulaedas: An R Package for Estimation of Distribution Algorithms Based
on Copulas | null | Yasser Gonzalez-Fernandez, Marta Soto (2014). copulaedas: An R
Package for Estimation of Distribution Algorithms Based on Copulas. Journal
of Statistical Software, 58(9), 1-34. URL http://www.jstatsoft.org/v58/i09/ | null | null | cs.NE cs.MS | http://creativecommons.org/licenses/by/3.0/ | The use of copula-based models in EDAs (estimation of distribution
algorithms) is currently an active area of research. In this context, the
copulaedas package for R provides a platform where EDAs based on copulas can be
implemented and studied. The package offers complete implementations of various
EDAs based on copulas and vines, a group of well-known optimization problems,
and utility functions to study the performance of the algorithms. Newly
developed EDAs can be easily integrated into the package by extending an S4
class with generic functions for their main components. This paper presents
copulaedas by providing an overview of EDAs based on copulas, a description of
the implementation of the package, and an illustration of its use through
examples. The examples include running the EDAs defined in the package,
implementing new algorithms, and performing an empirical study to compare the
behavior of different algorithms on benchmark functions and a real-world
problem.
| [
{
"version": "v1",
"created": "Mon, 24 Sep 2012 21:24:17 GMT"
},
{
"version": "v2",
"created": "Tue, 30 Apr 2013 18:33:27 GMT"
},
{
"version": "v3",
"created": "Sat, 17 May 2014 20:52:15 GMT"
},
{
"version": "v4",
"created": "Tue, 1 Jul 2014 19:08:59 GMT"
}
] | 2014-07-02T00:00:00 | [
[
"Gonzalez-Fernandez",
"Yasser",
""
],
[
"Soto",
"Marta",
""
]
] | TITLE: copulaedas: An R Package for Estimation of Distribution Algorithms Based
on Copulas
ABSTRACT: The use of copula-based models in EDAs (estimation of distribution
algorithms) is currently an active area of research. In this context, the
copulaedas package for R provides a platform where EDAs based on copulas can be
implemented and studied. The package offers complete implementations of various
EDAs based on copulas and vines, a group of well-known optimization problems,
and utility functions to study the performance of the algorithms. Newly
developed EDAs can be easily integrated into the package by extending an S4
class with generic functions for their main components. This paper presents
copulaedas by providing an overview of EDAs based on copulas, a description of
the implementation of the package, and an illustration of its use through
examples. The examples include running the EDAs defined in the package,
implementing new algorithms, and performing an empirical study to compare the
behavior of different algorithms on benchmark functions and a real-world
problem.
|
1407.0179 | Novi Quadrianto | Daniel Hern\'andez-Lobato, Viktoriia Sharmanska, Kristian Kersting,
Christoph H. Lampert, Novi Quadrianto | Mind the Nuisance: Gaussian Process Classification using Privileged
Noise | 14 pages with figures | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The learning with privileged information setting has recently attracted a lot
of attention within the machine learning community, as it allows the
integration of additional knowledge into the training process of a classifier,
even when this comes in the form of a data modality that is not available at
test time. Here, we show that privileged information can naturally be treated
as noise in the latent function of a Gaussian Process classifier (GPC). That
is, in contrast to the standard GPC setting, the latent function is not just a
nuisance but a feature: it becomes a natural measure of confidence about the
training data by modulating the slope of the GPC sigmoid likelihood function.
Extensive experiments on public datasets show that the proposed GPC method
using privileged noise, called GPC+, improves over a standard GPC without
privileged knowledge, and also over the current state-of-the-art SVM-based
method, SVM+. Moreover, we show that advanced neural networks and deep learning
methods can be compressed as privileged information.
| [
{
"version": "v1",
"created": "Tue, 1 Jul 2014 10:44:49 GMT"
}
] | 2014-07-02T00:00:00 | [
[
"Hernández-Lobato",
"Daniel",
""
],
[
"Sharmanska",
"Viktoriia",
""
],
[
"Kersting",
"Kristian",
""
],
[
"Lampert",
"Christoph H.",
""
],
[
"Quadrianto",
"Novi",
""
]
] | TITLE: Mind the Nuisance: Gaussian Process Classification using Privileged
Noise
ABSTRACT: The learning with privileged information setting has recently attracted a lot
of attention within the machine learning community, as it allows the
integration of additional knowledge into the training process of a classifier,
even when this comes in the form of a data modality that is not available at
test time. Here, we show that privileged information can naturally be treated
as noise in the latent function of a Gaussian Process classifier (GPC). That
is, in contrast to the standard GPC setting, the latent function is not just a
nuisance but a feature: it becomes a natural measure of confidence about the
training data by modulating the slope of the GPC sigmoid likelihood function.
Extensive experiments on public datasets show that the proposed GPC method
using privileged noise, called GPC+, improves over a standard GPC without
privileged knowledge, and also over the current state-of-the-art SVM-based
method, SVM+. Moreover, we show that advanced neural networks and deep learning
methods can be compressed as privileged information.
|
1210.3456 | Mingjun Zhong | Mingjun Zhong, Rong Liu, Bo Liu | Bayesian Analysis for miRNA and mRNA Interactions Using Expression Data | 21 pages, 11 figures, 8 tables | null | null | null | stat.AP cs.LG q-bio.GN q-bio.MN stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | MicroRNAs (miRNAs) are small RNA molecules composed of 19-22 nt, which play
important regulatory roles in post-transcriptional gene regulation by
inhibiting the translation of the mRNA into proteins or otherwise cleaving the
target mRNA. Inferring miRNA targets provides useful information for
understanding the roles of miRNA in biological processes that are potentially
involved in complex diseases. Statistical methodologies for point estimation,
such as the Least Absolute Shrinkage and Selection Operator (LASSO) algorithm,
have been proposed to identify the interactions of miRNA and mRNA based on
sequence and expression data. In this paper, we propose using the Bayesian
LASSO (BLASSO) and the non-negative Bayesian LASSO (nBLASSO) to analyse the
interactions between miRNA and mRNA using expression data. The proposed
Bayesian methods explore the posterior distributions for those parameters
required to model the miRNA-mRNA interactions. These approaches can be used to
observe the inferred effects of the miRNAs on the targets by plotting the
posterior distributions of those parameters. For comparison purposes, the Least
Squares Regression (LSR), Ridge Regression (RR), LASSO, non-negative LASSO
(nLASSO), and the proposed Bayesian approaches were applied to four public
datasets. We concluded that nLASSO and nBLASSO perform best in terms of
sensitivity and specificity. Compared to the point estimate algorithms, which
only provide single estimates for those parameters, the Bayesian methods are
more meaningful and provide credible intervals, which take into account the
uncertainty of the inferred interactions of the miRNA and mRNA. Furthermore,
Bayesian methods naturally provide statistical significance to select
convincing inferred interactions, while point estimate algorithms require a
manually chosen threshold, which is less meaningful, to choose the possible
interactions.
| [
{
"version": "v1",
"created": "Fri, 12 Oct 2012 09:03:14 GMT"
},
{
"version": "v2",
"created": "Mon, 30 Jun 2014 10:16:51 GMT"
}
] | 2014-07-01T00:00:00 | [
[
"Zhong",
"Mingjun",
""
],
[
"Liu",
"Rong",
""
],
[
"Liu",
"Bo",
""
]
] | TITLE: Bayesian Analysis for miRNA and mRNA Interactions Using Expression Data
ABSTRACT: MicroRNAs (miRNAs) are small RNA molecules composed of 19-22 nt, which play
important regulatory roles in post-transcriptional gene regulation by
inhibiting the translation of the mRNA into proteins or otherwise cleaving the
target mRNA. Inferring miRNA targets provides useful information for
understanding the roles of miRNA in biological processes that are potentially
involved in complex diseases. Statistical methodologies for point estimation,
such as the Least Absolute Shrinkage and Selection Operator (LASSO) algorithm,
have been proposed to identify the interactions of miRNA and mRNA based on
sequence and expression data. In this paper, we propose using the Bayesian
LASSO (BLASSO) and the non-negative Bayesian LASSO (nBLASSO) to analyse the
interactions between miRNA and mRNA using expression data. The proposed
Bayesian methods explore the posterior distributions for those parameters
required to model the miRNA-mRNA interactions. These approaches can be used to
observe the inferred effects of the miRNAs on the targets by plotting the
posterior distributions of those parameters. For comparison purposes, the Least
Squares Regression (LSR), Ridge Regression (RR), LASSO, non-negative LASSO
(nLASSO), and the proposed Bayesian approaches were applied to four public
datasets. We concluded that nLASSO and nBLASSO perform best in terms of
sensitivity and specificity. Compared to the point estimate algorithms, which
only provide single estimates for those parameters, the Bayesian methods are
more meaningful and provide credible intervals, which take into account the
uncertainty of the inferred interactions of the miRNA and mRNA. Furthermore,
Bayesian methods naturally provide statistical significance to select
convincing inferred interactions, while point estimate algorithms require a
manually chosen threshold, which is less meaningful, to choose the possible
interactions.
|
1307.1424 | Phillippe Samer | Phillippe Samer, Sebasti\'an Urrutia | A branch and cut algorithm for minimum spanning trees under conflict
constraints | null | null | 10.1007/s11590-014-0750-x | null | cs.DS math.CO math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study approaches for the exact solution of the \NP--hard minimum spanning
tree problem under conflict constraints. Given a graph $G(V,E)$ and a set $C
\subset E \times E$ of conflicting edge pairs, the problem consists of finding
a conflict-free minimum spanning tree, i.e. feasible solutions are allowed to
include at most one of the edges from each pair in $C$. The problem was
introduced recently in the literature, with several results on its complexity
and approximability. Some formulations and both exact and heuristic algorithms
were also discussed, but computational results indicate considerably large
duality gaps and a lack of optimality certificates for benchmark instances. In
this paper, we build on the representation of conflict constraints using an
auxiliary conflict graph $\hat{G}(E,C)$, where stable sets correspond to
conflict-free subsets of $E$. We introduce a general preprocessing method and a
branch and cut algorithm using an IP formulation with exponentially sized
classes of valid inequalities for both the spanning tree and the stable set
polytopes. Encouraging computational results indicate that the dual bounds of
our approach are significantly stronger than those previously available,
already in the initial LP relaxation, and we are able to provide new
feasibility and optimality certificates.
| [
{
"version": "v1",
"created": "Thu, 4 Jul 2013 17:47:09 GMT"
},
{
"version": "v2",
"created": "Mon, 16 Dec 2013 18:08:25 GMT"
},
{
"version": "v3",
"created": "Mon, 30 Jun 2014 13:36:45 GMT"
}
] | 2014-07-01T00:00:00 | [
[
"Samer",
"Phillippe",
""
],
[
"Urrutia",
"Sebastián",
""
]
] | TITLE: A branch and cut algorithm for minimum spanning trees under conflict
constraints
ABSTRACT: We study approaches for the exact solution of the \NP--hard minimum spanning
tree problem under conflict constraints. Given a graph $G(V,E)$ and a set $C
\subset E \times E$ of conflicting edge pairs, the problem consists of finding
a conflict-free minimum spanning tree, i.e. feasible solutions are allowed to
include at most one of the edges from each pair in $C$. The problem was
introduced recently in the literature, with several results on its complexity
and approximability. Some formulations and both exact and heuristic algorithms
were also discussed, but computational results indicate considerably large
duality gaps and a lack of optimality certificates for benchmark instances. In
this paper, we build on the representation of conflict constraints using an
auxiliary conflict graph $\hat{G}(E,C)$, where stable sets correspond to
conflict-free subsets of $E$. We introduce a general preprocessing method and a
branch and cut algorithm using an IP formulation with exponentially sized
classes of valid inequalities for both the spanning tree and the stable set
polytopes. Encouraging computational results indicate that the dual bounds of
our approach are significantly stronger than those previously available,
already in the initial LP relaxation, and we are able to provide new
feasibility and optimality certificates.
|
1405.3518 | Yoon Kim | Yoon Kim, Owen Zhang | Credibility Adjusted Term Frequency: A Supervised Term Weighting Scheme
for Sentiment Analysis and Text Classification | null | Proceedings of the 5th Workshop on Computational Approaches to
Subjectivity, Sentiment and Social Media Analysis. June, 2014. 79--83 | null | null | cs.CL cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We provide a simple but novel supervised weighting scheme for adjusting term
frequency in tf-idf for sentiment analysis and text classification. We compare
our method to baseline weighting schemes and find that it outperforms them on
multiple benchmarks. The method is robust and works well on both snippets and
longer documents.
| [
{
"version": "v1",
"created": "Wed, 14 May 2014 14:50:59 GMT"
},
{
"version": "v2",
"created": "Sat, 28 Jun 2014 17:18:54 GMT"
}
] | 2014-07-01T00:00:00 | [
[
"Kim",
"Yoon",
""
],
[
"Zhang",
"Owen",
""
]
] | TITLE: Credibility Adjusted Term Frequency: A Supervised Term Weighting Scheme
for Sentiment Analysis and Text Classification
ABSTRACT: We provide a simple but novel supervised weighting scheme for adjusting term
frequency in tf-idf for sentiment analysis and text classification. We compare
our method to baseline weighting schemes and find that it outperforms them on
multiple benchmarks. The method is robust and works well on both snippets and
longer documents.
|
1406.7362 | KyungHyun Cho | Kyunghyun Cho and Yoshua Bengio | Exponentially Increasing the Capacity-to-Computation Ratio for
Conditional Computation in Deep Learning | null | null | null | null | stat.ML cs.LG cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Many state-of-the-art results obtained with deep networks are achieved with
the largest models that could be trained, and if more computation power was
available, we might be able to exploit much larger datasets in order to improve
generalization ability. Whereas in learning algorithms such as decision trees
the ratio of capacity (e.g., the number of parameters) to computation is very
favorable (up to exponentially more parameters than computation), the ratio is
essentially 1 for deep neural networks. Conditional computation has been
proposed as a way to increase the capacity of a deep neural network without
increasing the amount of computation required, by activating some parameters
and computation "on-demand", on a per-example basis. In this note, we propose a
novel parametrization of weight matrices in neural networks which has the
potential to increase up to exponentially the ratio of the number of parameters
to computation. The proposed approach is based on turning on some parameters
(weight matrices) when specific bit patterns of hidden unit activations are
obtained. In order to better control for the overfitting that might result, we
propose a parametrization that is tree-structured, where each node of the tree
corresponds to a prefix of a sequence of sign bits, or gating units, associated
with hidden units.
| [
{
"version": "v1",
"created": "Sat, 28 Jun 2014 06:45:51 GMT"
}
] | 2014-07-01T00:00:00 | [
[
"Cho",
"Kyunghyun",
""
],
[
"Bengio",
"Yoshua",
""
]
] | TITLE: Exponentially Increasing the Capacity-to-Computation Ratio for
Conditional Computation in Deep Learning
ABSTRACT: Many state-of-the-art results obtained with deep networks are achieved with
the largest models that could be trained, and if more computation power was
available, we might be able to exploit much larger datasets in order to improve
generalization ability. Whereas in learning algorithms such as decision trees
the ratio of capacity (e.g., the number of parameters) to computation is very
favorable (up to exponentially more parameters than computation), the ratio is
essentially 1 for deep neural networks. Conditional computation has been
proposed as a way to increase the capacity of a deep neural network without
increasing the amount of computation required, by activating some parameters
and computation "on-demand", on a per-example basis. In this note, we propose a
novel parametrization of weight matrices in neural networks which has the
potential to increase up to exponentially the ratio of the number of parameters
to computation. The proposed approach is based on turning on some parameters
(weight matrices) when specific bit patterns of hidden unit activations are
obtained. In order to better control for the overfitting that might result, we
propose a parametrization that is tree-structured, where each node of the tree
corresponds to a prefix of a sequence of sign bits, or gating units, associated
with hidden units.
|
1406.7429 | Jonathan Katzman | Jonathan Katzman and Diane Duros | Comparison of SVM Optimization Techniques in the Primal | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper examines the efficacy of different optimization techniques in a
primal formulation of a support vector machine (SVM). Three main techniques are
compared. The dataset used to compare all three techniques was the Sentiment
Analysis on Movie Reviews dataset, from kaggle.com.
| [
{
"version": "v1",
"created": "Sat, 28 Jun 2014 18:59:44 GMT"
}
] | 2014-07-01T00:00:00 | [
[
"Katzman",
"Jonathan",
""
],
[
"Duros",
"Diane",
""
]
] | TITLE: Comparison of SVM Optimization Techniques in the Primal
ABSTRACT: This paper examines the efficacy of different optimization techniques in a
primal formulation of a support vector machine (SVM). Three main techniques are
compared. The dataset used to compare all three techniques was the Sentiment
Analysis on Movie Reviews dataset, from kaggle.com.
|
1406.7525 | Wenqi Huang | Wenqi Huang, Xiaojin Gong | Fusion Based Holistic Road Scene Understanding | 14 pages,11 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper addresses the problem of holistic road scene understanding based
on the integration of visual and range data. To achieve the grand goal, we
propose an approach that jointly tackles object-level image segmentation and
semantic region labeling within a conditional random field (CRF) framework.
Specifically, we first generate semantic object hypotheses by clustering 3D
points, learning their prior appearance models, and using a deep learning
method for reasoning their semantic categories. The learned priors, together
with spatial and geometric contexts, are incorporated in CRF. With this
formulation, visual and range data are fused thoroughly, and moreover, the
coupled segmentation and semantic labeling problem can be inferred via Graph
Cuts. Our approach is validated on the challenging KITTI dataset that contains
diverse complicated road scenarios. Both quantitative and qualitative
evaluations demonstrate its effectiveness.
| [
{
"version": "v1",
"created": "Sun, 29 Jun 2014 17:11:25 GMT"
}
] | 2014-07-01T00:00:00 | [
[
"Huang",
"Wenqi",
""
],
[
"Gong",
"Xiaojin",
""
]
] | TITLE: Fusion Based Holistic Road Scene Understanding
ABSTRACT: This paper addresses the problem of holistic road scene understanding based
on the integration of visual and range data. To achieve the grand goal, we
propose an approach that jointly tackles object-level image segmentation and
semantic region labeling within a conditional random field (CRF) framework.
Specifically, we first generate semantic object hypotheses by clustering 3D
points, learning their prior appearance models, and using a deep learning
method for reasoning their semantic categories. The learned priors, together
with spatial and geometric contexts, are incorporated in CRF. With this
formulation, visual and range data are fused thoroughly, and moreover, the
coupled segmentation and semantic labeling problem can be inferred via Graph
Cuts. Our approach is validated on the challenging KITTI dataset that contains
diverse complicated road scenarios. Both quantitative and qualitative
evaluations demonstrate its effectiveness.
|
1406.7738 | Walter Lasecki | Sanmay Das, and Allen Lavoie | Home Is Where the Up-Votes Are: Behavior Changes in Response to Feedback
in Social Media | null | null | null | ci-2014/93 | cs.SI physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent research shows that humans are heavily influenced by online social
interactions: We are more likely to perform actions which, in the past, have
led to positive social feedback. We introduce a quantitative model of behavior
changes in response to such feedback, drawing on inverse reinforcement learning
and studies of human game playing. The model allows us to make predictions,
particularly in the context of social media, about which community a user will
select, and to quantify how future selections change based on the feedback a
user receives. We show that our model predicts real-world changes in behavior
on a dataset gathered from reddit. We also explore how this relatively simple
model of individual behavior can lead to complex collective dynamics when there
is a population of users, each individual learning in response to feedback and
in turn providing feedback to others.
| [
{
"version": "v1",
"created": "Mon, 30 Jun 2014 13:51:23 GMT"
}
] | 2014-07-01T00:00:00 | [
[
"Das",
"Sanmay",
""
],
[
"Lavoie",
"Allen",
""
]
] | TITLE: Home Is Where the Up-Votes Are: Behavior Changes in Response to Feedback
in Social Media
ABSTRACT: Recent research shows that humans are heavily influenced by online social
interactions: We are more likely to perform actions which, in the past, have
led to positive social feedback. We introduce a quantitative model of behavior
changes in response to such feedback, drawing on inverse reinforcement learning
and studies of human game playing. The model allows us to make predictions,
particularly in the context of social media, about which community a user will
select, and to quantify how future selections change based on the feedback a
user receives. We show that our model predicts real-world changes in behavior
on a dataset gathered from reddit. We also explore how this relatively simple
model of individual behavior can lead to complex collective dynamics when there
is a population of users, each individual learning in response to feedback and
in turn providing feedback to others.
|
1406.7799 | Pedram Mohammadi Mr. | Pedram Mohammadi, Abbas Ebrahimi-Moghadam, and Shahram Shirani | Subjective and Objective Quality Assessment of Image: A Survey | 50 pages, 12 figures, and 3 Tables. This work has been submitted to
Elsevier Journal of Visual Communication and Image Representation | null | null | null | cs.MM cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the increasing demand for image-based applications, the efficient and
reliable evaluation of image quality has increased in importance. Measuring the
image quality is of fundamental importance for numerous image processing
applications, where the goal of image quality assessment (IQA) methods is to
automatically evaluate the quality of images in agreement with human quality
judgments. Numerous IQA methods have been proposed over the past years to
fulfill this goal. In this paper, a survey of the quality assessment methods
for conventional image signals, as well as the newly emerged ones, which
includes the high dynamic range (HDR) and 3-D images, is presented. A
comprehensive explanation of the subjective and objective IQA and their
classification is provided. Six widely used subjective quality datasets, and
performance measures are reviewed. Emphasis is given to the full-reference
image quality assessment (FR-IQA) methods, and 9 often-used quality measures
(including mean squared error (MSE), structural similarity index (SSIM),
multi-scale structural similarity index (MS-SSIM), visual information fidelity
(VIF), most apparent distortion (MAD), feature similarity measure (FSIM),
feature similarity measure for color images (FSIMC), dynamic range independent
measure (DRIM), and tone-mapped images quality index (TMQI)) are carefully
described, and their performance and computation time on four subjective
quality datasets are evaluated. Furthermore, a brief introduction to 3-D IQA is
provided and the issues related to this area of research are reviewed.
| [
{
"version": "v1",
"created": "Mon, 30 Jun 2014 16:25:00 GMT"
}
] | 2014-07-01T00:00:00 | [
[
"Mohammadi",
"Pedram",
""
],
[
"Ebrahimi-Moghadam",
"Abbas",
""
],
[
"Shirani",
"Shahram",
""
]
] | TITLE: Subjective and Objective Quality Assessment of Image: A Survey
ABSTRACT: With the increasing demand for image-based applications, the efficient and
reliable evaluation of image quality has increased in importance. Measuring the
image quality is of fundamental importance for numerous image processing
applications, where the goal of image quality assessment (IQA) methods is to
automatically evaluate the quality of images in agreement with human quality
judgments. Numerous IQA methods have been proposed over the past years to
fulfill this goal. In this paper, a survey of the quality assessment methods
for conventional image signals, as well as the newly emerged ones, which
includes the high dynamic range (HDR) and 3-D images, is presented. A
comprehensive explanation of the subjective and objective IQA and their
classification is provided. Six widely used subjective quality datasets, and
performance measures are reviewed. Emphasis is given to the full-reference
image quality assessment (FR-IQA) methods, and 9 often-used quality measures
(including mean squared error (MSE), structural similarity index (SSIM),
multi-scale structural similarity index (MS-SSIM), visual information fidelity
(VIF), most apparent distortion (MAD), feature similarity measure (FSIM),
feature similarity measure for color images (FSIMC), dynamic range independent
measure (DRIM), and tone-mapped images quality index (TMQI)) are carefully
described, and their performance and computation time on four subjective
quality datasets are evaluated. Furthermore, a brief introduction to 3-D IQA is
provided and the issues related to this area of research are reviewed.
|
1406.7811 | Erik Cuevas | Erik Cuevas and Mauricio Gonzalez | An optimization algorithm for multimodal functions inspired by
collective animal behavior | 18 Pages. arXiv admin note: text overlap with arXiv:1405.5164 | Soft Computing 17 (3) , (2013), pp. 489-502 | null | null | cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Interest in multimodal function optimization is expanding rapidly since real
world optimization problems often demand locating multiple optima within a
search space. This article presents a new multimodal optimization algorithm
named as the Collective Animal Behavior (CAB). Animal groups, such as schools
of fish, flocks of birds, swarms of locusts and herds of wildebeest, exhibit a
variety of behaviors including swarming about a food source, milling around a
central location or migrating over large distances in aligned groups. These
collective behaviors are often advantageous to groups, allowing them to
increase their harvesting efficiency to follow better migration routes, to
improve their aerodynamic and to avoid predation. In the proposed algorithm,
searcher agents are a group of animals which interact to each other based on
the biological laws of collective motion. Experimental results demonstrate that
the proposed algorithm is capable of finding global and local optima of
benchmark multimodal optimization problems with a higher efficiency in
comparison to other methods reported in the literature.
| [
{
"version": "v1",
"created": "Mon, 30 Jun 2014 16:48:05 GMT"
}
] | 2014-07-01T00:00:00 | [
[
"Cuevas",
"Erik",
""
],
[
"Gonzalez",
"Mauricio",
""
]
] | TITLE: An optimization algorithm for multimodal functions inspired by
collective animal behavior
ABSTRACT: Interest in multimodal function optimization is expanding rapidly since real
world optimization problems often demand locating multiple optima within a
search space. This article presents a new multimodal optimization algorithm
named as the Collective Animal Behavior (CAB). Animal groups, such as schools
of fish, flocks of birds, swarms of locusts and herds of wildebeest, exhibit a
variety of behaviors including swarming about a food source, milling around a
central location or migrating over large distances in aligned groups. These
collective behaviors are often advantageous to groups, allowing them to
increase their harvesting efficiency to follow better migration routes, to
improve their aerodynamic and to avoid predation. In the proposed algorithm,
searcher agents are a group of animals which interact to each other based on
the biological laws of collective motion. Experimental results demonstrate that
the proposed algorithm is capable of finding global and local optima of
benchmark multimodal optimization problems with a higher efficiency in
comparison to other methods reported in the literature.
|
1312.0363 | Yuanming Shi | Yuanming Shi, Jun Zhang, Khaled B. Letaief | Optimal Stochastic Coordinated Beamforming for Wireless Cooperative
Networks with CSI Uncertainty | null | null | null | null | cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Transmit optimization and resource allocation for wireless cooperative
networks with channel state information (CSI) uncertainty are important but
challenging problems in terms of both the uncertainty modeling and performance
op- timization. In this paper, we establish a generic stochastic coordinated
beamforming (SCB) framework that provides flex- ibility in the channel
uncertainty modeling, while guaranteeing optimality in the transmission
strategies. We adopt a general stochastic model for the CSI uncertainty, which
is applicable for various practical scenarios. The SCB problem turns out to be
a joint chance constrained program (JCCP) and is known to be highly
intractable. In contrast to all the previous algo- rithms for JCCP that can
only find feasible but sub-optimal solutions, we propose a novel stochastic DC
(difference-of-convex) programming algorithm with optimality guarantee, which
can serve as the benchmark for evaluating heuristic and sub-optimal algorithms.
The key observation is that the highly intractable probability constraint can
be equivalently reformulated as a DC constraint. This further enables efficient
algorithms to achieve optimality. Simulation results will illustrate the
convergence, conservativeness, stability and performance gains of the proposed
algorithm.
| [
{
"version": "v1",
"created": "Mon, 2 Dec 2013 07:42:06 GMT"
},
{
"version": "v2",
"created": "Fri, 27 Jun 2014 01:19:50 GMT"
}
] | 2014-06-30T00:00:00 | [
[
"Shi",
"Yuanming",
""
],
[
"Zhang",
"Jun",
""
],
[
"Letaief",
"Khaled B.",
""
]
] | TITLE: Optimal Stochastic Coordinated Beamforming for Wireless Cooperative
Networks with CSI Uncertainty
ABSTRACT: Transmit optimization and resource allocation for wireless cooperative
networks with channel state information (CSI) uncertainty are important but
challenging problems in terms of both the uncertainty modeling and performance
op- timization. In this paper, we establish a generic stochastic coordinated
beamforming (SCB) framework that provides flex- ibility in the channel
uncertainty modeling, while guaranteeing optimality in the transmission
strategies. We adopt a general stochastic model for the CSI uncertainty, which
is applicable for various practical scenarios. The SCB problem turns out to be
a joint chance constrained program (JCCP) and is known to be highly
intractable. In contrast to all the previous algo- rithms for JCCP that can
only find feasible but sub-optimal solutions, we propose a novel stochastic DC
(difference-of-convex) programming algorithm with optimality guarantee, which
can serve as the benchmark for evaluating heuristic and sub-optimal algorithms.
The key observation is that the highly intractable probability constraint can
be equivalently reformulated as a DC constraint. This further enables efficient
algorithms to achieve optimality. Simulation results will illustrate the
convergence, conservativeness, stability and performance gains of the proposed
algorithm.
|
1406.7075 | \"Omer Faruk Ertu\u{g}rul | Omer Faruk Ertugrul | Adaptive texture energy measure method | null | International Journal of Intelligent Information Systems. Vol. 3,
No. 2, 2014, pp. 13-18 | 10.11648/j.ijiis.20140302.11 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent developments in image quality, data storage, and computational
capacity have heightened the need for texture analysis in image process. To
date various methods have been developed and introduced for assessing textures
in images. One of the most popular texture analysis methods is the Texture
Energy Measure (TEM) and it has been used for detecting edges, levels, waves,
spots and ripples by employing predefined TEM masks to images. Despite several
success- ful studies, TEM has a number of serious weaknesses in use. The major
drawback is; the masks are predefined therefore they cannot be adapted to
image. A new method, Adaptive Texture Energy Measure Method (aTEM), was offered
to over- come this disadvantage of TEM by using adaptive masks by adjusting the
contrast, sharpening and orientation angle of the mask. To assess the
applicability of aTEM, it is compared with TEM. The accuracy of the
classification of butterfly, flower seed and Brodatz datasets are 0.08, 0.3292
and 0.3343, respectively by TEM and 0.0053, 0.2417 and 0.3153, respectively by
aTEM. The results of this study indicate that aTEM is a successful method for
texture analysis.
| [
{
"version": "v1",
"created": "Fri, 27 Jun 2014 06:00:17 GMT"
}
] | 2014-06-30T00:00:00 | [
[
"Ertugrul",
"Omer Faruk",
""
]
] | TITLE: Adaptive texture energy measure method
ABSTRACT: Recent developments in image quality, data storage, and computational
capacity have heightened the need for texture analysis in image process. To
date various methods have been developed and introduced for assessing textures
in images. One of the most popular texture analysis methods is the Texture
Energy Measure (TEM) and it has been used for detecting edges, levels, waves,
spots and ripples by employing predefined TEM masks to images. Despite several
success- ful studies, TEM has a number of serious weaknesses in use. The major
drawback is; the masks are predefined therefore they cannot be adapted to
image. A new method, Adaptive Texture Energy Measure Method (aTEM), was offered
to over- come this disadvantage of TEM by using adaptive masks by adjusting the
contrast, sharpening and orientation angle of the mask. To assess the
applicability of aTEM, it is compared with TEM. The accuracy of the
classification of butterfly, flower seed and Brodatz datasets are 0.08, 0.3292
and 0.3343, respectively by TEM and 0.0053, 0.2417 and 0.3153, respectively by
aTEM. The results of this study indicate that aTEM is a successful method for
texture analysis.
|
1309.3132 | Xiao-Bo Jin | Xiao-Bo Jin, Guang-Gang Geng, Dexian Zhang | Combination of Multiple Bipartite Ranking for Web Content Quality
Evaluation | 17 pages, 8 figures, 2 tables | null | null | null | cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Web content quality estimation is crucial to various web content processing
applications. Our previous work applied Bagging + C4.5 to achive the best
results on the ECML/PKDD Discovery Challenge 2010, which is the comibination of
many point-wise rankinig models. In this paper, we combine multiple pair-wise
bipartite ranking learner to solve the multi-partite ranking problems for the
web quality estimation. In encoding stage, we present the ternary encoding and
the binary coding extending each rank value to $L - 1$ (L is the number of the
different ranking value). For the decoding, we discuss the combination of
multiple ranking results from multiple bipartite ranking models with the
predefined weighting and the adaptive weighting. The experiments on ECML/PKDD
2010 Discovery Challenge datasets show that \textit{binary coding} +
\textit{predefined weighting} yields the highest performance in all four
combinations and furthermore it is better than the best results reported in
ECML/PKDD 2010 Discovery Challenge competition.
| [
{
"version": "v1",
"created": "Thu, 12 Sep 2013 12:15:51 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2014 03:01:13 GMT"
}
] | 2014-06-27T00:00:00 | [
[
"Jin",
"Xiao-Bo",
""
],
[
"Geng",
"Guang-Gang",
""
],
[
"Zhang",
"Dexian",
""
]
] | TITLE: Combination of Multiple Bipartite Ranking for Web Content Quality
Evaluation
ABSTRACT: Web content quality estimation is crucial to various web content processing
applications. Our previous work applied Bagging + C4.5 to achive the best
results on the ECML/PKDD Discovery Challenge 2010, which is the comibination of
many point-wise rankinig models. In this paper, we combine multiple pair-wise
bipartite ranking learner to solve the multi-partite ranking problems for the
web quality estimation. In encoding stage, we present the ternary encoding and
the binary coding extending each rank value to $L - 1$ (L is the number of the
different ranking value). For the decoding, we discuss the combination of
multiple ranking results from multiple bipartite ranking models with the
predefined weighting and the adaptive weighting. The experiments on ECML/PKDD
2010 Discovery Challenge datasets show that \textit{binary coding} +
\textit{predefined weighting} yields the highest performance in all four
combinations and furthermore it is better than the best results reported in
ECML/PKDD 2010 Discovery Challenge competition.
|
1405.1328 | Emiliano De Cristofaro | Igor Bilogrevic, Julien Freudiger, Emiliano De Cristofaro, and Ersin
Uzun | What's the Gist? Privacy-Preserving Aggregation of User Profiles | To appear in the Proceedings of ESORICS 2014 | null | null | null | cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Over the past few years, online service providers have started gathering
increasing amounts of personal information to build user profiles and monetize
them with advertisers and data brokers. Users have little control of what
information is processed and are often left with an all-or-nothing decision
between receiving free services or refusing to be profiled. This paper explores
an alternative approach where users only disclose an aggregate model -- the
"gist" -- of their data. We aim to preserve data utility and simultaneously
provide user privacy. We show that this approach can be efficiently supported
by letting users contribute encrypted and differentially-private data to an
aggregator. The aggregator combines encrypted contributions and can only
extract an aggregate model of the underlying data. We evaluate our framework on
a dataset of 100,000 U.S. users obtained from the U.S. Census Bureau and show
that (i) it provides accurate aggregates with as little as 100 users, (ii) it
generates revenue for both users and data brokers, and (iii) its overhead is
appreciably low.
| [
{
"version": "v1",
"created": "Tue, 6 May 2014 15:49:48 GMT"
},
{
"version": "v2",
"created": "Wed, 25 Jun 2014 21:01:41 GMT"
}
] | 2014-06-27T00:00:00 | [
[
"Bilogrevic",
"Igor",
""
],
[
"Freudiger",
"Julien",
""
],
[
"De Cristofaro",
"Emiliano",
""
],
[
"Uzun",
"Ersin",
""
]
] | TITLE: What's the Gist? Privacy-Preserving Aggregation of User Profiles
ABSTRACT: Over the past few years, online service providers have started gathering
increasing amounts of personal information to build user profiles and monetize
them with advertisers and data brokers. Users have little control of what
information is processed and are often left with an all-or-nothing decision
between receiving free services or refusing to be profiled. This paper explores
an alternative approach where users only disclose an aggregate model -- the
"gist" -- of their data. We aim to preserve data utility and simultaneously
provide user privacy. We show that this approach can be efficiently supported
by letting users contribute encrypted and differentially-private data to an
aggregator. The aggregator combines encrypted contributions and can only
extract an aggregate model of the underlying data. We evaluate our framework on
a dataset of 100,000 U.S. users obtained from the U.S. Census Bureau and show
that (i) it provides accurate aggregates with as little as 100 users, (ii) it
generates revenue for both users and data brokers, and (iii) its overhead is
appreciably low.
|
1406.6832 | Michel Plantie | Michel Crampes and Michel Planti\'e | Overlapping Community Detection Optimization and Nash Equilibrium | Submitted to KDD | null | null | null | cs.SI physics.soc-ph stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Community detection using both graphs and social networks is the focus of
many algorithms. Recent methods aimed at optimizing the so-called modularity
function proceed by maximizing relations within communities while minimizing
inter-community relations.
However, given the NP-completeness of the problem, these algorithms are
heuristics that do not guarantee an optimum. In this paper, we introduce a new
algorithm along with a function that takes an approximate solution and modifies
it in order to reach an optimum. This reassignment function is considered a
'potential function' and becomes a necessary condition to asserting that the
computed optimum is indeed a Nash Equilibrium. We also use this function to
simultaneously show partitioning and overlapping communities, two detection and
visualization modes of great value in revealing interesting features of a
social network. Our approach is successfully illustrated through several
experiments on either real unipartite, multipartite or directed graphs of
medium and large-sized datasets.
| [
{
"version": "v1",
"created": "Thu, 26 Jun 2014 10:28:36 GMT"
}
] | 2014-06-27T00:00:00 | [
[
"Crampes",
"Michel",
""
],
[
"Plantié",
"Michel",
""
]
] | TITLE: Overlapping Community Detection Optimization and Nash Equilibrium
ABSTRACT: Community detection using both graphs and social networks is the focus of
many algorithms. Recent methods aimed at optimizing the so-called modularity
function proceed by maximizing relations within communities while minimizing
inter-community relations.
However, given the NP-completeness of the problem, these algorithms are
heuristics that do not guarantee an optimum. In this paper, we introduce a new
algorithm along with a function that takes an approximate solution and modifies
it in order to reach an optimum. This reassignment function is considered a
'potential function' and becomes a necessary condition to asserting that the
computed optimum is indeed a Nash Equilibrium. We also use this function to
simultaneously show partitioning and overlapping communities, two detection and
visualization modes of great value in revealing interesting features of a
social network. Our approach is successfully illustrated through several
experiments on either real unipartite, multipartite or directed graphs of
medium and large-sized datasets.
|
1406.6947 | Ping Luo | Zhenyao Zhu and Ping Luo and Xiaogang Wang and Xiaoou Tang | Deep Learning Multi-View Representation for Face Recognition | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Various factors, such as identities, views (poses), and illuminations, are
coupled in face images. Disentangling the identity and view representations is
a major challenge in face recognition. Existing face recognition systems either
use handcrafted features or learn features discriminatively to improve
recognition accuracy. This is different from the behavior of human brain.
Intriguingly, even without accessing 3D data, human not only can recognize face
identity, but can also imagine face images of a person under different
viewpoints given a single 2D image, making face perception in the brain robust
to view changes. In this sense, human brain has learned and encoded 3D face
models from 2D images. To take into account this instinct, this paper proposes
a novel deep neural net, named multi-view perceptron (MVP), which can untangle
the identity and view features, and infer a full spectrum of multi-view images
in the meanwhile, given a single 2D face image. The identity features of MVP
achieve superior performance on the MultiPIE dataset. MVP is also capable to
interpolate and predict images under viewpoints that are unobserved in the
training data.
| [
{
"version": "v1",
"created": "Thu, 26 Jun 2014 17:09:25 GMT"
}
] | 2014-06-27T00:00:00 | [
[
"Zhu",
"Zhenyao",
""
],
[
"Luo",
"Ping",
""
],
[
"Wang",
"Xiaogang",
""
],
[
"Tang",
"Xiaoou",
""
]
] | TITLE: Deep Learning Multi-View Representation for Face Recognition
ABSTRACT: Various factors, such as identities, views (poses), and illuminations, are
coupled in face images. Disentangling the identity and view representations is
a major challenge in face recognition. Existing face recognition systems either
use handcrafted features or learn features discriminatively to improve
recognition accuracy. This is different from the behavior of human brain.
Intriguingly, even without accessing 3D data, human not only can recognize face
identity, but can also imagine face images of a person under different
viewpoints given a single 2D image, making face perception in the brain robust
to view changes. In this sense, human brain has learned and encoded 3D face
models from 2D images. To take into account this instinct, this paper proposes
a novel deep neural net, named multi-view perceptron (MVP), which can untangle
the identity and view features, and infer a full spectrum of multi-view images
in the meanwhile, given a single 2D face image. The identity features of MVP
achieve superior performance on the MultiPIE dataset. MVP is also capable to
interpolate and predict images under viewpoints that are unobserved in the
training data.
|
1305.0763 | Matthew Crossley | Matthew Crossley, Andy Nisbet, Martyn Amos | Quantifying the Impact of Parameter Tuning on Nature-Inspired Algorithms | 8 pages, 7 figures. Accepted at the European Conference on Artificial
Life (ECAL) 2013, Taormina, Italy | null | 10.7551/978-0-262-31709-2-ch138 | null | cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The problem of parameterization is often central to the effective deployment
of nature-inspired algorithms. However, finding the optimal set of parameter
values for a combination of problem instance and solution method is highly
challenging, and few concrete guidelines exist on how and when such tuning may
be performed. Previous work tends to either focus on a specific algorithm or
use benchmark problems, and both of these restrictions limit the applicability
of any findings. Here, we examine a number of different algorithms, and study
them in a "problem agnostic" fashion (i.e., one that is not tied to specific
instances) by considering their performance on fitness landscapes with varying
characteristics. Using this approach, we make a number of observations on which
algorithms may (or may not) benefit from tuning, and in which specific
circumstances.
| [
{
"version": "v1",
"created": "Fri, 3 May 2013 16:15:20 GMT"
},
{
"version": "v2",
"created": "Mon, 1 Jul 2013 13:04:10 GMT"
}
] | 2014-06-26T00:00:00 | [
[
"Crossley",
"Matthew",
""
],
[
"Nisbet",
"Andy",
""
],
[
"Amos",
"Martyn",
""
]
] | TITLE: Quantifying the Impact of Parameter Tuning on Nature-Inspired Algorithms
ABSTRACT: The problem of parameterization is often central to the effective deployment
of nature-inspired algorithms. However, finding the optimal set of parameter
values for a combination of problem instance and solution method is highly
challenging, and few concrete guidelines exist on how and when such tuning may
be performed. Previous work tends to either focus on a specific algorithm or
use benchmark problems, and both of these restrictions limit the applicability
of any findings. Here, we examine a number of different algorithms, and study
them in a "problem agnostic" fashion (i.e., one that is not tied to specific
instances) by considering their performance on fitness landscapes with varying
characteristics. Using this approach, we make a number of observations on which
algorithms may (or may not) benefit from tuning, and in which specific
circumstances.
|
1406.6507 | Hyun Oh Song | Hyun Oh Song, Yong Jae Lee, Stefanie Jegelka, Trevor Darrell | Weakly-supervised Discovery of Visual Pattern Configurations | null | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The increasing prominence of weakly labeled data nurtures a growing demand
for object detection methods that can cope with minimal supervision. We propose
an approach that automatically identifies discriminative configurations of
visual patterns that are characteristic of a given object class. We formulate
the problem as a constrained submodular optimization problem and demonstrate
the benefits of the discovered configurations in remedying mislocalizations and
finding informative positive and negative training examples. Together, these
lead to state-of-the-art weakly-supervised detection results on the challenging
PASCAL VOC dataset.
| [
{
"version": "v1",
"created": "Wed, 25 Jun 2014 09:35:40 GMT"
}
] | 2014-06-26T00:00:00 | [
[
"Song",
"Hyun Oh",
""
],
[
"Lee",
"Yong Jae",
""
],
[
"Jegelka",
"Stefanie",
""
],
[
"Darrell",
"Trevor",
""
]
] | TITLE: Weakly-supervised Discovery of Visual Pattern Configurations
ABSTRACT: The increasing prominence of weakly labeled data nurtures a growing demand
for object detection methods that can cope with minimal supervision. We propose
an approach that automatically identifies discriminative configurations of
visual patterns that are characteristic of a given object class. We formulate
the problem as a constrained submodular optimization problem and demonstrate
the benefits of the discovered configurations in remedying mislocalizations and
finding informative positive and negative training examples. Together, these
lead to state-of-the-art weakly-supervised detection results on the challenging
PASCAL VOC dataset.
|
1406.6568 | Victor Miller | V. A. Miller, S. Erlien, J. Piersol | Support vector machine classification of dimensionally reduced
structural MRI images for dementia | technical note | null | null | null | cs.CV cs.LG physics.med-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We classify very-mild to moderate dementia in patients (CDR ranging from 0 to
2) using a support vector machine classifier acting on dimensionally reduced
feature set derived from MRI brain scans of the 416 subjects available in the
OASIS-Brains dataset. We use image segmentation and principal component
analysis to reduce the dimensionality of the data. Our resulting feature set
contains 11 features for each subject. Performance of the classifiers is
evaluated using 10-fold cross-validation. Using linear and (gaussian) kernels,
we obtain a training classification accuracy of 86.4% (90.1%), test accuracy of
85.0% (85.7%), test precision of 68.7% (68.5%), test recall of 68.0% (74.0%),
and test Matthews correlation coefficient of 0.594 (0.616).
| [
{
"version": "v1",
"created": "Wed, 25 Jun 2014 13:50:18 GMT"
}
] | 2014-06-26T00:00:00 | [
[
"Miller",
"V. A.",
""
],
[
"Erlien",
"S.",
""
],
[
"Piersol",
"J.",
""
]
] | TITLE: Support vector machine classification of dimensionally reduced
structural MRI images for dementia
ABSTRACT: We classify very-mild to moderate dementia in patients (CDR ranging from 0 to
2) using a support vector machine classifier acting on dimensionally reduced
feature set derived from MRI brain scans of the 416 subjects available in the
OASIS-Brains dataset. We use image segmentation and principal component
analysis to reduce the dimensionality of the data. Our resulting feature set
contains 11 features for each subject. Performance of the classifiers is
evaluated using 10-fold cross-validation. Using linear and (gaussian) kernels,
we obtain a training classification accuracy of 86.4% (90.1%), test accuracy of
85.0% (85.7%), test precision of 68.7% (68.5%), test recall of 68.0% (74.0%),
and test Matthews correlation coefficient of 0.594 (0.616).
|
1406.6651 | Ishanu Chattopadhyay | Ishanu Chattopadhyay | Causality Networks | 22 pages, 8 figures | null | null | null | cs.LG cs.IT math.IT q-fin.ST stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | While correlation measures are used to discern statistical relationships
between observed variables in almost all branches of data-driven scientific
inquiry, what we are really interested in is the existence of causal
dependence. Designing an efficient causality test, that may be carried out in
the absence of restrictive pre-suppositions on the underlying dynamical
structure of the data at hand, is non-trivial. Nevertheless, ability to
computationally infer statistical prima facie evidence of causal dependence may
yield a far more discriminative tool for data analysis compared to the
calculation of simple correlations. In the present work, we present a new
non-parametric test of Granger causality for quantized or symbolic data streams
generated by ergodic stationary sources. In contrast to state-of-art binary
tests, our approach makes precise and computes the degree of causal dependence
between data streams, without making any restrictive assumptions, linearity or
otherwise. Additionally, without any a priori imposition of specific dynamical
structure, we infer explicit generative models of causal cross-dependence,
which may be then used for prediction. These explicit models are represented as
generalized probabilistic automata, referred to crossed automata, and are shown
to be sufficient to capture a fairly general class of causal dependence. The
proposed algorithms are computationally efficient in the PAC sense; $i.e.$, we
find good models of cross-dependence with high probability, with polynomial
run-times and sample complexities. The theoretical results are applied to
weekly search-frequency data from Google Trends API for a chosen set of
socially "charged" keywords. The causality network inferred from this dataset
reveals, quite expectedly, the causal importance of certain keywords. It is
also illustrated that correlation analysis fails to gather such insight.
| [
{
"version": "v1",
"created": "Wed, 25 Jun 2014 17:46:32 GMT"
}
] | 2014-06-26T00:00:00 | [
[
"Chattopadhyay",
"Ishanu",
""
]
] | TITLE: Causality Networks
ABSTRACT: While correlation measures are used to discern statistical relationships
between observed variables in almost all branches of data-driven scientific
inquiry, what we are really interested in is the existence of causal
dependence. Designing an efficient causality test, that may be carried out in
the absence of restrictive pre-suppositions on the underlying dynamical
structure of the data at hand, is non-trivial. Nevertheless, ability to
computationally infer statistical prima facie evidence of causal dependence may
yield a far more discriminative tool for data analysis compared to the
calculation of simple correlations. In the present work, we present a new
non-parametric test of Granger causality for quantized or symbolic data streams
generated by ergodic stationary sources. In contrast to state-of-art binary
tests, our approach makes precise and computes the degree of causal dependence
between data streams, without making any restrictive assumptions, linearity or
otherwise. Additionally, without any a priori imposition of specific dynamical
structure, we infer explicit generative models of causal cross-dependence,
which may be then used for prediction. These explicit models are represented as
generalized probabilistic automata, referred to crossed automata, and are shown
to be sufficient to capture a fairly general class of causal dependence. The
proposed algorithms are computationally efficient in the PAC sense; $i.e.$, we
find good models of cross-dependence with high probability, with polynomial
run-times and sample complexities. The theoretical results are applied to
weekly search-frequency data from Google Trends API for a chosen set of
socially "charged" keywords. The causality network inferred from this dataset
reveals, quite expectedly, the causal importance of certain keywords. It is
also illustrated that correlation analysis fails to gather such insight.
|
1406.6037 | Sreepathi Pai | Sreepathi Pai, R. Govindarajan, Matthew J. Thazhuthaveetil | Preemptive Thread Block Scheduling with Online Structural Runtime
Prediction for Concurrent GPGPU Kernels | 14 pages, full pre-review version of PACT 2014 poster | null | null | null | cs.AR cs.OS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent NVIDIA Graphics Processing Units (GPUs) can execute multiple kernels
concurrently. On these GPUs, the thread block scheduler (TBS) uses the FIFO
policy to schedule their thread blocks. We show that FIFO leaves performance to
chance, resulting in significant loss of performance and fairness. To improve
performance and fairness, we propose use of the preemptive Shortest Remaining
Time First (SRTF) policy instead. Although SRTF requires an estimate of runtime
of GPU kernels, we show that such an estimate of the runtime can be easily
obtained using online profiling and exploiting a simple observation on GPU
kernels' grid structure. Specifically, we propose a novel Structural Runtime
Predictor. Using a simple Staircase model of GPU kernel execution, we show that
the runtime of a kernel can be predicted by profiling only the first few thread
blocks. We evaluate an online predictor based on this model on benchmarks from
ERCBench, and find that it can estimate the actual runtime reasonably well
after the execution of only a single thread block. Next, we design a thread
block scheduler that is both concurrent kernel-aware and uses this predictor.
We implement the SRTF policy and evaluate it on two-program workloads from
ERCBench. SRTF improves STP by 1.18x and ANTT by 2.25x over FIFO. When compared
to MPMax, a state-of-the-art resource allocation policy for concurrent kernels,
SRTF improves STP by 1.16x and ANTT by 1.3x. To improve fairness, we also
propose SRTF/Adaptive which controls resource usage of concurrently executing
kernels to maximize fairness. SRTF/Adaptive improves STP by 1.12x, ANTT by
2.23x and Fairness by 2.95x compared to FIFO. Overall, our implementation of
SRTF achieves system throughput to within 12.64% of Shortest Job First (SJF, an
oracle optimal scheduling policy), bridging 49% of the gap between FIFO and
SJF.
| [
{
"version": "v1",
"created": "Mon, 23 Jun 2014 19:44:03 GMT"
}
] | 2014-06-25T00:00:00 | [
[
"Pai",
"Sreepathi",
""
],
[
"Govindarajan",
"R.",
""
],
[
"Thazhuthaveetil",
"Matthew J.",
""
]
] | TITLE: Preemptive Thread Block Scheduling with Online Structural Runtime
Prediction for Concurrent GPGPU Kernels
ABSTRACT: Recent NVIDIA Graphics Processing Units (GPUs) can execute multiple kernels
concurrently. On these GPUs, the thread block scheduler (TBS) uses the FIFO
policy to schedule their thread blocks. We show that FIFO leaves performance to
chance, resulting in significant loss of performance and fairness. To improve
performance and fairness, we propose use of the preemptive Shortest Remaining
Time First (SRTF) policy instead. Although SRTF requires an estimate of runtime
of GPU kernels, we show that such an estimate of the runtime can be easily
obtained using online profiling and exploiting a simple observation on GPU
kernels' grid structure. Specifically, we propose a novel Structural Runtime
Predictor. Using a simple Staircase model of GPU kernel execution, we show that
the runtime of a kernel can be predicted by profiling only the first few thread
blocks. We evaluate an online predictor based on this model on benchmarks from
ERCBench, and find that it can estimate the actual runtime reasonably well
after the execution of only a single thread block. Next, we design a thread
block scheduler that is both concurrent kernel-aware and uses this predictor.
We implement the SRTF policy and evaluate it on two-program workloads from
ERCBench. SRTF improves STP by 1.18x and ANTT by 2.25x over FIFO. When compared
to MPMax, a state-of-the-art resource allocation policy for concurrent kernels,
SRTF improves STP by 1.16x and ANTT by 1.3x. To improve fairness, we also
propose SRTF/Adaptive which controls resource usage of concurrently executing
kernels to maximize fairness. SRTF/Adaptive improves STP by 1.12x, ANTT by
2.23x and Fairness by 2.95x compared to FIFO. Overall, our implementation of
SRTF achieves system throughput to within 12.64% of Shortest Job First (SJF, an
oracle optimal scheduling policy), bridging 49% of the gap between FIFO and
SJF.
|
1106.2229 | Fionn Murtagh | Pedro Contreras and Fionn Murtagh | Fast, Linear Time Hierarchical Clustering using the Baire Metric | 27 pages, 6 tables, 10 figures | Journal of Classification, July 2012, Volume 29, Issue 2, pp
118-143 | 10.1007/s00357-012-9106-3 | null | stat.ML cs.IR stat.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The Baire metric induces an ultrametric on a dataset and is of linear
computational complexity, contrasted with the standard quadratic time
agglomerative hierarchical clustering algorithm. In this work we evaluate
empirically this new approach to hierarchical clustering. We compare
hierarchical clustering based on the Baire metric with (i) agglomerative
hierarchical clustering, in terms of algorithm properties; (ii) generalized
ultrametrics, in terms of definition; and (iii) fast clustering through k-means
partititioning, in terms of quality of results. For the latter, we carry out an
in depth astronomical study. We apply the Baire distance to spectrometric and
photometric redshifts from the Sloan Digital Sky Survey using, in this work,
about half a million astronomical objects. We want to know how well the (more
costly to determine) spectrometric redshifts can predict the (more easily
obtained) photometric redshifts, i.e. we seek to regress the spectrometric on
the photometric redshifts, and we use clusterwise regression for this.
| [
{
"version": "v1",
"created": "Sat, 11 Jun 2011 12:05:43 GMT"
}
] | 2014-06-24T00:00:00 | [
[
"Contreras",
"Pedro",
""
],
[
"Murtagh",
"Fionn",
""
]
] | TITLE: Fast, Linear Time Hierarchical Clustering using the Baire Metric
ABSTRACT: The Baire metric induces an ultrametric on a dataset and is of linear
computational complexity, contrasted with the standard quadratic time
agglomerative hierarchical clustering algorithm. In this work we evaluate
empirically this new approach to hierarchical clustering. We compare
hierarchical clustering based on the Baire metric with (i) agglomerative
hierarchical clustering, in terms of algorithm properties; (ii) generalized
ultrametrics, in terms of definition; and (iii) fast clustering through k-means
partititioning, in terms of quality of results. For the latter, we carry out an
in depth astronomical study. We apply the Baire distance to spectrometric and
photometric redshifts from the Sloan Digital Sky Survey using, in this work,
about half a million astronomical objects. We want to know how well the (more
costly to determine) spectrometric redshifts can predict the (more easily
obtained) photometric redshifts, i.e. we seek to regress the spectrometric on
the photometric redshifts, and we use clusterwise regression for this.
|
1312.3913 | Xi He | Xi He and Ashwin Machanavajjhala and Bolin Ding | Blowfish Privacy: Tuning Privacy-Utility Trade-offs using Policies | Full version of the paper at SIGMOD'14 Snowbird, Utah USA | null | 10.1145/2588555.2588581 | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Privacy definitions provide ways for trading-off the privacy of individuals
in a statistical database for the utility of downstream analysis of the data.
In this paper, we present Blowfish, a class of privacy definitions inspired by
the Pufferfish framework, that provides a rich interface for this trade-off. In
particular, we allow data publishers to extend differential privacy using a
policy, which specifies (a) secrets, or information that must be kept secret,
and (b) constraints that may be known about the data. While the secret
specification allows increased utility by lessening protection for certain
individual properties, the constraint specification provides added protection
against an adversary who knows correlations in the data (arising from
constraints). We formalize policies and present novel algorithms that can
handle general specifications of sensitive information and certain count
constraints. We show that there are reasonable policies under which our privacy
mechanisms for k-means clustering, histograms and range queries introduce
significantly lesser noise than their differentially private counterparts. We
quantify the privacy-utility trade-offs for various policies analytically and
empirically on real datasets.
| [
{
"version": "v1",
"created": "Fri, 13 Dec 2013 19:23:12 GMT"
},
{
"version": "v2",
"created": "Sat, 28 Dec 2013 06:49:22 GMT"
},
{
"version": "v3",
"created": "Tue, 11 Feb 2014 15:55:15 GMT"
},
{
"version": "v4",
"created": "Tue, 8 Apr 2014 16:13:26 GMT"
},
{
"version": "v5",
"created": "Mon, 23 Jun 2014 05:09:12 GMT"
}
] | 2014-06-24T00:00:00 | [
[
"He",
"Xi",
""
],
[
"Machanavajjhala",
"Ashwin",
""
],
[
"Ding",
"Bolin",
""
]
] | TITLE: Blowfish Privacy: Tuning Privacy-Utility Trade-offs using Policies
ABSTRACT: Privacy definitions provide ways for trading-off the privacy of individuals
in a statistical database for the utility of downstream analysis of the data.
In this paper, we present Blowfish, a class of privacy definitions inspired by
the Pufferfish framework, that provides a rich interface for this trade-off. In
particular, we allow data publishers to extend differential privacy using a
policy, which specifies (a) secrets, or information that must be kept secret,
and (b) constraints that may be known about the data. While the secret
specification allows increased utility by lessening protection for certain
individual properties, the constraint specification provides added protection
against an adversary who knows correlations in the data (arising from
constraints). We formalize policies and present novel algorithms that can
handle general specifications of sensitive information and certain count
constraints. We show that there are reasonable policies under which our privacy
mechanisms for k-means clustering, histograms and range queries introduce
significantly lesser noise than their differentially private counterparts. We
quantify the privacy-utility trade-offs for various policies analytically and
empirically on real datasets.
|
1402.4238 | Shixin Luo | Shixin Luo, Rui Zhang, Teng Joon Lim | Downlink and Uplink Energy Minimization Through User Association and
Beamforming in Cloud RAN | 29 pages, 4 figures, submitted for possible journal publication | null | null | null | cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The cloud radio access network (C-RAN) concept, in which densely deployed
access points (APs) are empowered by cloud computing to cooperatively support
mobile users (MUs), to improve mobile data rates, has been recently proposed.
However, the high density of active ("on") APs results in severe interference
and also inefficient energy consumption. Moreover, the growing popularity of
highly interactive applications with stringent uplink (UL) requirements, e.g.
network gaming and real-time broadcasting by wireless users, means that the UL
transmission is becoming more crucial and requires special attention. Therefore
in this paper, we propose a joint downlink (DL) and UL MU-AP association and
beamforming design to coordinate interference in the C-RAN for energy
minimization, a problem which is shown to be NP hard. Due to the new
consideration of UL transmission, it is shown that the two state-of-the-art
approaches for finding computationally efficient solutions of joint MU-AP
association and beamforming considering only the DL, i.e., group-sparse
optimization and relaxed-integer programming, cannot be modified in a
straightforward way to solve our problem. Leveraging on the celebrated UL-DL
duality result, we show that by establishing a virtual DL transmission for the
original UL transmission, the joint DL and UL optimization problem can be
converted to an equivalent DL problem in C-RAN with two inter-related
subproblems for the original and virtual DL transmissions, respectively. Based
on this transformation, two efficient algorithms for joint DL and UL MU-AP
association and beamforming design are proposed, whose performances are
evaluated and compared with other benchmarking schemes through extensive
simulations.
| [
{
"version": "v1",
"created": "Tue, 18 Feb 2014 07:17:39 GMT"
},
{
"version": "v2",
"created": "Mon, 23 Jun 2014 02:14:51 GMT"
}
] | 2014-06-24T00:00:00 | [
[
"Luo",
"Shixin",
""
],
[
"Zhang",
"Rui",
""
],
[
"Lim",
"Teng Joon",
""
]
] | TITLE: Downlink and Uplink Energy Minimization Through User Association and
Beamforming in Cloud RAN
ABSTRACT: The cloud radio access network (C-RAN) concept, in which densely deployed
access points (APs) are empowered by cloud computing to cooperatively support
mobile users (MUs), to improve mobile data rates, has been recently proposed.
However, the high density of active ("on") APs results in severe interference
and also inefficient energy consumption. Moreover, the growing popularity of
highly interactive applications with stringent uplink (UL) requirements, e.g.
network gaming and real-time broadcasting by wireless users, means that the UL
transmission is becoming more crucial and requires special attention. Therefore
in this paper, we propose a joint downlink (DL) and UL MU-AP association and
beamforming design to coordinate interference in the C-RAN for energy
minimization, a problem which is shown to be NP hard. Due to the new
consideration of UL transmission, it is shown that the two state-of-the-art
approaches for finding computationally efficient solutions of joint MU-AP
association and beamforming considering only the DL, i.e., group-sparse
optimization and relaxed-integer programming, cannot be modified in a
straightforward way to solve our problem. Leveraging on the celebrated UL-DL
duality result, we show that by establishing a virtual DL transmission for the
original UL transmission, the joint DL and UL optimization problem can be
converted to an equivalent DL problem in C-RAN with two inter-related
subproblems for the original and virtual DL transmissions, respectively. Based
on this transformation, two efficient algorithms for joint DL and UL MU-AP
association and beamforming design are proposed, whose performances are
evaluated and compared with other benchmarking schemes through extensive
simulations.
|
1403.1528 | Andre Luckow | Shantenu Jha, Judy Qiu, Andre Luckow, Pradeep Mantha, Geoffrey C.Fox | A Tale of Two Data-Intensive Paradigms: Applications, Abstractions, and
Architectures | 8 pages, 2 figures | null | null | null | cs.DC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Scientific problems that depend on processing large amounts of data require
overcoming challenges in multiple areas: managing large-scale data
distribution, co-placement and scheduling of data with compute resources, and
storing and transferring large volumes of data. We analyze the ecosystems of
the two prominent paradigms for data-intensive applications, hereafter referred
to as the high-performance computing and the Apache-Hadoop paradigm. We propose
a basis, common terminology and functional factors upon which to analyze the
two approaches of both paradigms. We discuss the concept of "Big Data Ogres"
and their facets as means of understanding and characterizing the most common
application workloads found across the two paradigms. We then discuss the
salient features of the two paradigms, and compare and contrast the two
approaches. Specifically, we examine common implementation/approaches of these
paradigms, shed light upon the reasons for their current "architecture" and
discuss some typical workloads that utilize them. In spite of the significant
software distinctions, we believe there is architectural similarity. We discuss
the potential integration of different implementations, across the different
levels and components. Our comparison progresses from a fully qualitative
examination of the two paradigms, to a semi-quantitative methodology. We use a
simple and broadly used Ogre (K-means clustering), characterize its performance
on a range of representative platforms, covering several implementations from
both paradigms. Our experiments provide an insight into the relative strengths
of the two paradigms. We propose that the set of Ogres will serve as a
benchmark to evaluate the two paradigms along different dimensions.
| [
{
"version": "v1",
"created": "Thu, 6 Mar 2014 18:48:55 GMT"
},
{
"version": "v2",
"created": "Mon, 23 Jun 2014 01:56:34 GMT"
}
] | 2014-06-24T00:00:00 | [
[
"Jha",
"Shantenu",
""
],
[
"Qiu",
"Judy",
""
],
[
"Luckow",
"Andre",
""
],
[
"Mantha",
"Pradeep",
""
],
[
"Fox",
"Geoffrey C.",
""
]
] | TITLE: A Tale of Two Data-Intensive Paradigms: Applications, Abstractions, and
Architectures
ABSTRACT: Scientific problems that depend on processing large amounts of data require
overcoming challenges in multiple areas: managing large-scale data
distribution, co-placement and scheduling of data with compute resources, and
storing and transferring large volumes of data. We analyze the ecosystems of
the two prominent paradigms for data-intensive applications, hereafter referred
to as the high-performance computing and the Apache-Hadoop paradigm. We propose
a basis, common terminology and functional factors upon which to analyze the
two approaches of both paradigms. We discuss the concept of "Big Data Ogres"
and their facets as means of understanding and characterizing the most common
application workloads found across the two paradigms. We then discuss the
salient features of the two paradigms, and compare and contrast the two
approaches. Specifically, we examine common implementation/approaches of these
paradigms, shed light upon the reasons for their current "architecture" and
discuss some typical workloads that utilize them. In spite of the significant
software distinctions, we believe there is architectural similarity. We discuss
the potential integration of different implementations, across the different
levels and components. Our comparison progresses from a fully qualitative
examination of the two paradigms, to a semi-quantitative methodology. We use a
simple and broadly used Ogre (K-means clustering), characterize its performance
on a range of representative platforms, covering several implementations from
both paradigms. Our experiments provide an insight into the relative strengths
of the two paradigms. We propose that the set of Ogres will serve as a
benchmark to evaluate the two paradigms along different dimensions.
|
1405.1459 | Flavio Figueiredo | Flavio Figueiredo, Jussara M. Almeida, Yasuko Matsubara, Bruno
Ribeiro, Christos Faloutsos | Revisit Behavior in Social Media: The Phoenix-R Model and Discoveries | To appear on European Conference on Machine Learning and Principles
and Practice of Knowledge Discovery in Databases 2014 | null | null | null | cs.SI physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | How many listens will an artist receive on a online radio? How about plays on
a YouTube video? How many of these visits are new or returning users? Modeling
and mining popularity dynamics of social activity has important implications
for researchers, content creators and providers. We here investigate the effect
of revisits (successive visits from a single user) on content popularity. Using
four datasets of social activity, with up to tens of millions media objects
(e.g., YouTube videos, Twitter hashtags or LastFM artists), we show the effect
of revisits in the popularity evolution of such objects. Secondly, we propose
the Phoenix-R model which captures the popularity dynamics of individual
objects. Phoenix-R has the desired properties of being: (1) parsimonious, being
based on the minimum description length principle, and achieving lower root
mean squared error than state-of-the-art baselines; (2) applicable, the model
is effective for predicting future popularity values of objects.
| [
{
"version": "v1",
"created": "Tue, 6 May 2014 21:37:06 GMT"
},
{
"version": "v2",
"created": "Sun, 22 Jun 2014 19:13:29 GMT"
}
] | 2014-06-24T00:00:00 | [
[
"Figueiredo",
"Flavio",
""
],
[
"Almeida",
"Jussara M.",
""
],
[
"Matsubara",
"Yasuko",
""
],
[
"Ribeiro",
"Bruno",
""
],
[
"Faloutsos",
"Christos",
""
]
] | TITLE: Revisit Behavior in Social Media: The Phoenix-R Model and Discoveries
ABSTRACT: How many listens will an artist receive on a online radio? How about plays on
a YouTube video? How many of these visits are new or returning users? Modeling
and mining popularity dynamics of social activity has important implications
for researchers, content creators and providers. We here investigate the effect
of revisits (successive visits from a single user) on content popularity. Using
four datasets of social activity, with up to tens of millions media objects
(e.g., YouTube videos, Twitter hashtags or LastFM artists), we show the effect
of revisits in the popularity evolution of such objects. Secondly, we propose
the Phoenix-R model which captures the popularity dynamics of individual
objects. Phoenix-R has the desired properties of being: (1) parsimonious, being
based on the minimum description length principle, and achieving lower root
mean squared error than state-of-the-art baselines; (2) applicable, the model
is effective for predicting future popularity values of objects.
|
1406.5565 | Sam Keene | Kenneth D. Morton Jr., Peter Torrione, Leslie Collins, Sam Keene | An Open Source Pattern Recognition Toolbox for MATLAB | null | null | null | null | stat.ML cs.CV cs.LG cs.MS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Pattern recognition and machine learning are becoming integral parts of
algorithms in a wide range of applications. Different algorithms and approaches
for machine learning include different tradeoffs between performance and
computation, so during algorithm development it is often necessary to explore a
variety of different approaches to a given task. A toolbox with a unified
framework across multiple pattern recognition techniques enables algorithm
developers the ability to rapidly evaluate different choices prior to
deployment. MATLAB is a widely used environment for algorithm development and
prototyping, and although several MATLAB toolboxes for pattern recognition are
currently available these are either incomplete, expensive, or restrictively
licensed. In this work we describe a MATLAB toolbox for pattern recognition and
machine learning known as the PRT (Pattern Recognition Toolbox), licensed under
the permissive MIT license. The PRT includes many popular techniques for data
preprocessing, supervised learning, clustering, regression and feature
selection, as well as a methodology for combining these components using a
simple, uniform syntax. The resulting algorithms can be evaluated using
cross-validation and a variety of scoring metrics to ensure robust performance
when the algorithm is deployed. This paper presents an overview of the PRT as
well as an example of usage on Fisher's Iris dataset.
| [
{
"version": "v1",
"created": "Sat, 21 Jun 2014 01:50:54 GMT"
}
] | 2014-06-24T00:00:00 | [
[
"Morton",
"Kenneth D.",
"Jr."
],
[
"Torrione",
"Peter",
""
],
[
"Collins",
"Leslie",
""
],
[
"Keene",
"Sam",
""
]
] | TITLE: An Open Source Pattern Recognition Toolbox for MATLAB
ABSTRACT: Pattern recognition and machine learning are becoming integral parts of
algorithms in a wide range of applications. Different algorithms and approaches
for machine learning include different tradeoffs between performance and
computation, so during algorithm development it is often necessary to explore a
variety of different approaches to a given task. A toolbox with a unified
framework across multiple pattern recognition techniques enables algorithm
developers the ability to rapidly evaluate different choices prior to
deployment. MATLAB is a widely used environment for algorithm development and
prototyping, and although several MATLAB toolboxes for pattern recognition are
currently available these are either incomplete, expensive, or restrictively
licensed. In this work we describe a MATLAB toolbox for pattern recognition and
machine learning known as the PRT (Pattern Recognition Toolbox), licensed under
the permissive MIT license. The PRT includes many popular techniques for data
preprocessing, supervised learning, clustering, regression and feature
selection, as well as a methodology for combining these components using a
simple, uniform syntax. The resulting algorithms can be evaluated using
cross-validation and a variety of scoring metrics to ensure robust performance
when the algorithm is deployed. This paper presents an overview of the PRT as
well as an example of usage on Fisher's Iris dataset.
|
1406.5617 | S. K. Sahay | R.K. Roul, O. R. Devanand and S.K. Sahay | Web Document Clustering and Ranking using Tf-Idf based Apriori Approach | 5 Pages | IJCA Proceedings on ICACEA, No. 2, p. 34 (2014) | null | null | cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The dynamic web has increased exponentially over the past few years with more
than thousands of documents related to a subject available to the user now.
Most of the web documents are unstructured and not in an organized manner and
hence user facing more difficult to find relevant documents. A more useful and
efficient mechanism is combining clustering with ranking, where clustering can
group the similar documents in one place and ranking can be applied to each
cluster for viewing the top documents at the beginning.. Besides the particular
clustering algorithm, the different term weighting functions applied to the
selected features to represent web document is a main aspect in clustering
task. Keeping this approach in mind, here we proposed a new mechanism called
Tf-Idf based Apriori for clustering the web documents. We then rank the
documents in each cluster using Tf-Idf and similarity factor of documents based
on the user query. This approach will helps the user to get all his relevant
documents in one place and can restrict his search to some top documents of his
choice. For experimental purpose, we have taken the Classic3 and Classic4
datasets of Cornell University having more than 10,000 documents and use gensim
toolkit to carry out our work. We have compared our approach with traditional
apriori algorithm and found that our approach is giving better results for
higher minimum support. Our ranking mechanism is also giving a good F-measure
of 78%.
| [
{
"version": "v1",
"created": "Sat, 21 Jun 2014 14:38:21 GMT"
}
] | 2014-06-24T00:00:00 | [
[
"Roul",
"R. K.",
""
],
[
"Devanand",
"O. R.",
""
],
[
"Sahay",
"S. K.",
""
]
] | TITLE: Web Document Clustering and Ranking using Tf-Idf based Apriori Approach
ABSTRACT: The dynamic web has increased exponentially over the past few years with more
than thousands of documents related to a subject available to the user now.
Most of the web documents are unstructured and not in an organized manner and
hence user facing more difficult to find relevant documents. A more useful and
efficient mechanism is combining clustering with ranking, where clustering can
group the similar documents in one place and ranking can be applied to each
cluster for viewing the top documents at the beginning.. Besides the particular
clustering algorithm, the different term weighting functions applied to the
selected features to represent web document is a main aspect in clustering
task. Keeping this approach in mind, here we proposed a new mechanism called
Tf-Idf based Apriori for clustering the web documents. We then rank the
documents in each cluster using Tf-Idf and similarity factor of documents based
on the user query. This approach will helps the user to get all his relevant
documents in one place and can restrict his search to some top documents of his
choice. For experimental purpose, we have taken the Classic3 and Classic4
datasets of Cornell University having more than 10,000 documents and use gensim
toolkit to carry out our work. We have compared our approach with traditional
apriori algorithm and found that our approach is giving better results for
higher minimum support. Our ranking mechanism is also giving a good F-measure
of 78%.
|
1406.5653 | Rushil Anirudh | Rushil Anirudh and Pavan Turaga | Interactively Test Driving an Object Detector: Estimating Performance on
Unlabeled Data | Published at Winter Conference on Applications of Computer Vision,
2014 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we study the problem of `test-driving' a detector, i.e.
allowing a human user to get a quick sense of how well the detector generalizes
to their specific requirement. To this end, we present the first system that
estimates detector performance interactively without extensive ground truthing
using a human in the loop. We approach this as a problem of estimating
proportions and show that it is possible to make accurate inferences on the
proportion of classes or groups within a large data collection by observing
only $5-10\%$ of samples from the data. In estimating the false detections (for
precision), the samples are chosen carefully such that the overall
characteristics of the data collection are preserved. Next, inspired by its use
in estimating disease propagation we apply pooled testing approaches to
estimate missed detections (for recall) from the dataset. The estimates thus
obtained are close to the ones obtained using ground truth, thus reducing the
need for extensive labeling which is expensive and time consuming.
| [
{
"version": "v1",
"created": "Sat, 21 Jun 2014 21:37:30 GMT"
}
] | 2014-06-24T00:00:00 | [
[
"Anirudh",
"Rushil",
""
],
[
"Turaga",
"Pavan",
""
]
] | TITLE: Interactively Test Driving an Object Detector: Estimating Performance on
Unlabeled Data
ABSTRACT: In this paper, we study the problem of `test-driving' a detector, i.e.
allowing a human user to get a quick sense of how well the detector generalizes
to their specific requirement. To this end, we present the first system that
estimates detector performance interactively without extensive ground truthing
using a human in the loop. We approach this as a problem of estimating
proportions and show that it is possible to make accurate inferences on the
proportion of classes or groups within a large data collection by observing
only $5-10\%$ of samples from the data. In estimating the false detections (for
precision), the samples are chosen carefully such that the overall
characteristics of the data collection are preserved. Next, inspired by its use
in estimating disease propagation we apply pooled testing approaches to
estimate missed detections (for recall) from the dataset. The estimates thus
obtained are close to the ones obtained using ground truth, thus reducing the
need for extensive labeling which is expensive and time consuming.
|
1406.5708 | Ylies Falcone | Hadil Charafeddine, Khalil El-Harake, Yli\`es Falcone, and Mohamad
Jaber | Runtime Enforcement for Component-Based Systems | arXiv admin note: text overlap with arXiv:1109.5505 by other authors | null | null | null | cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Runtime enforcement is an increasingly popular and effective dynamic
validation technique aiming to ensure the correct runtime behavior (w.r.t. a
formal specification) of systems using a so-called enforcement monitor. In this
paper we introduce runtime enforcement of specifications on component-based
systems (CBS) modeled in the BIP (Behavior, Interaction and Priority)
framework. BIP is a powerful and expressive component-based framework for
formal construction of heterogeneous systems. However, because of BIP
expressiveness, it remains difficult to enforce at design-time complex
behavioral properties.
First we propose a theoretical runtime enforcement framework for CBS where we
delineate a hierarchy of sets of enforceable properties (i.e., properties that
can be enforced) according to the number of observational steps a system is
allowed to deviate from the property (i.e., the notion of k-step
enforceability). To ensure the observational equivalence between the correct
executions of the initial system and the monitored system, we show that i) only
stutter-invariant properties should be enforced on CBS with our monitors, ii)
safety properties are 1-step enforceable. Given an abstract enforcement monitor
(as a finite-state machine) for some 1-step enforceable specification, we
formally instrument (at relevant locations) a given BIP system to integrate the
monitor. At runtime, the monitor observes and automatically avoids any error in
the behavior of the system w.r.t. the specification. Our approach is fully
implemented in an available tool that we used to i) avoid deadlock occurrences
on a dining philosophers benchmark, and ii) ensure the correct placement of
robots on a map.
| [
{
"version": "v1",
"created": "Sun, 22 Jun 2014 11:49:11 GMT"
}
] | 2014-06-24T00:00:00 | [
[
"Charafeddine",
"Hadil",
""
],
[
"El-Harake",
"Khalil",
""
],
[
"Falcone",
"Yliès",
""
],
[
"Jaber",
"Mohamad",
""
]
] | TITLE: Runtime Enforcement for Component-Based Systems
ABSTRACT: Runtime enforcement is an increasingly popular and effective dynamic
validation technique aiming to ensure the correct runtime behavior (w.r.t. a
formal specification) of systems using a so-called enforcement monitor. In this
paper we introduce runtime enforcement of specifications on component-based
systems (CBS) modeled in the BIP (Behavior, Interaction and Priority)
framework. BIP is a powerful and expressive component-based framework for
formal construction of heterogeneous systems. However, because of BIP
expressiveness, it remains difficult to enforce at design-time complex
behavioral properties.
First we propose a theoretical runtime enforcement framework for CBS where we
delineate a hierarchy of sets of enforceable properties (i.e., properties that
can be enforced) according to the number of observational steps a system is
allowed to deviate from the property (i.e., the notion of k-step
enforceability). To ensure the observational equivalence between the correct
executions of the initial system and the monitored system, we show that i) only
stutter-invariant properties should be enforced on CBS with our monitors, ii)
safety properties are 1-step enforceable. Given an abstract enforcement monitor
(as a finite-state machine) for some 1-step enforceable specification, we
formally instrument (at relevant locations) a given BIP system to integrate the
monitor. At runtime, the monitor observes and automatically avoids any error in
the behavior of the system w.r.t. the specification. Our approach is fully
implemented in an available tool that we used to i) avoid deadlock occurrences
on a dining philosophers benchmark, and ii) ensure the correct placement of
robots on a map.
|
1406.5752 | Tianyi Zhou | Tianyi Zhou and Jeff Bilmes and Carlos Guestrin | Divide-and-Conquer Learning by Anchoring a Conical Hull | 26 pages, long version, in updating | null | null | null | stat.ML cs.LG | http://creativecommons.org/licenses/by-nc-sa/3.0/ | We reduce a broad class of machine learning problems, usually addressed by EM
or sampling, to the problem of finding the $k$ extremal rays spanning the
conical hull of a data point set. These $k$ "anchors" lead to a global solution
and a more interpretable model that can even outperform EM and sampling on
generalization error. To find the $k$ anchors, we propose a novel
divide-and-conquer learning scheme "DCA" that distributes the problem to
$\mathcal O(k\log k)$ same-type sub-problems on different low-D random
hyperplanes, each can be solved by any solver. For the 2D sub-problem, we
present a non-iterative solver that only needs to compute an array of cosine
values and its max/min entries. DCA also provides a faster subroutine for other
methods to check whether a point is covered in a conical hull, which improves
algorithm design in multiple dimensions and brings significant speedup to
learning. We apply our method to GMM, HMM, LDA, NMF and subspace clustering,
then show its competitive performance and scalability over other methods on
rich datasets.
| [
{
"version": "v1",
"created": "Sun, 22 Jun 2014 19:16:20 GMT"
}
] | 2014-06-24T00:00:00 | [
[
"Zhou",
"Tianyi",
""
],
[
"Bilmes",
"Jeff",
""
],
[
"Guestrin",
"Carlos",
""
]
] | TITLE: Divide-and-Conquer Learning by Anchoring a Conical Hull
ABSTRACT: We reduce a broad class of machine learning problems, usually addressed by EM
or sampling, to the problem of finding the $k$ extremal rays spanning the
conical hull of a data point set. These $k$ "anchors" lead to a global solution
and a more interpretable model that can even outperform EM and sampling on
generalization error. To find the $k$ anchors, we propose a novel
divide-and-conquer learning scheme "DCA" that distributes the problem to
$\mathcal O(k\log k)$ same-type sub-problems on different low-D random
hyperplanes, each can be solved by any solver. For the 2D sub-problem, we
present a non-iterative solver that only needs to compute an array of cosine
values and its max/min entries. DCA also provides a faster subroutine for other
methods to check whether a point is covered in a conical hull, which improves
algorithm design in multiple dimensions and brings significant speedup to
learning. We apply our method to GMM, HMM, LDA, NMF and subspace clustering,
then show its competitive performance and scalability over other methods on
rich datasets.
|
1406.5824 | Serena Yeung | Serena Yeung, Alireza Fathi, and Li Fei-Fei | VideoSET: Video Summary Evaluation through Text | null | null | null | null | cs.CV cs.CL cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper we present VideoSET, a method for Video Summary Evaluation
through Text that can evaluate how well a video summary is able to retain the
semantic information contained in its original video. We observe that semantics
is most easily expressed in words, and develop a text-based approach for the
evaluation. Given a video summary, a text representation of the video summary
is first generated, and an NLP-based metric is then used to measure its
semantic distance to ground-truth text summaries written by humans. We show
that our technique has higher agreement with human judgment than pixel-based
distance metrics. We also release text annotations and ground-truth text
summaries for a number of publicly available video datasets, for use by the
computer vision community.
| [
{
"version": "v1",
"created": "Mon, 23 Jun 2014 07:56:23 GMT"
}
] | 2014-06-24T00:00:00 | [
[
"Yeung",
"Serena",
""
],
[
"Fathi",
"Alireza",
""
],
[
"Fei-Fei",
"Li",
""
]
] | TITLE: VideoSET: Video Summary Evaluation through Text
ABSTRACT: In this paper we present VideoSET, a method for Video Summary Evaluation
through Text that can evaluate how well a video summary is able to retain the
semantic information contained in its original video. We observe that semantics
is most easily expressed in words, and develop a text-based approach for the
evaluation. Given a video summary, a text representation of the video summary
is first generated, and an NLP-based metric is then used to measure its
semantic distance to ground-truth text summaries written by humans. We show
that our technique has higher agreement with human judgment than pixel-based
distance metrics. We also release text annotations and ground-truth text
summaries for a number of publicly available video datasets, for use by the
computer vision community.
|
1406.5910 | Roman Shapovalov | Roman Shapovalov, Dmitry Vetrov, Anton Osokin, Pushmeet Kohli | Multi-utility Learning: Structured-output Learning with Multiple
Annotation-specific Loss Functions | null | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Structured-output learning is a challenging problem; particularly so because
of the difficulty in obtaining large datasets of fully labelled instances for
training. In this paper we try to overcome this difficulty by presenting a
multi-utility learning framework for structured prediction that can learn from
training instances with different forms of supervision. We propose a unified
technique for inferring the loss functions most suitable for quantifying the
consistency of solutions with the given weak annotation. We demonstrate the
effectiveness of our framework on the challenging semantic image segmentation
problem for which a wide variety of annotations can be used. For instance, the
popular training datasets for semantic segmentation are composed of images with
hard-to-generate full pixel labellings, as well as images with easy-to-obtain
weak annotations, such as bounding boxes around objects, or image-level labels
that specify which object categories are present in an image. Experimental
evaluation shows that the use of annotation-specific loss functions
dramatically improves segmentation accuracy compared to the baseline system
where only one type of weak annotation is used.
| [
{
"version": "v1",
"created": "Mon, 23 Jun 2014 14:06:24 GMT"
}
] | 2014-06-24T00:00:00 | [
[
"Shapovalov",
"Roman",
""
],
[
"Vetrov",
"Dmitry",
""
],
[
"Osokin",
"Anton",
""
],
[
"Kohli",
"Pushmeet",
""
]
] | TITLE: Multi-utility Learning: Structured-output Learning with Multiple
Annotation-specific Loss Functions
ABSTRACT: Structured-output learning is a challenging problem; particularly so because
of the difficulty in obtaining large datasets of fully labelled instances for
training. In this paper we try to overcome this difficulty by presenting a
multi-utility learning framework for structured prediction that can learn from
training instances with different forms of supervision. We propose a unified
technique for inferring the loss functions most suitable for quantifying the
consistency of solutions with the given weak annotation. We demonstrate the
effectiveness of our framework on the challenging semantic image segmentation
problem for which a wide variety of annotations can be used. For instance, the
popular training datasets for semantic segmentation are composed of images with
hard-to-generate full pixel labellings, as well as images with easy-to-obtain
weak annotations, such as bounding boxes around objects, or image-level labels
that specify which object categories are present in an image. Experimental
evaluation shows that the use of annotation-specific loss functions
dramatically improves segmentation accuracy compared to the baseline system
where only one type of weak annotation is used.
|
1406.5947 | Thomas Martinetz | Bogdan Miclut, Thomas Kaester, Thomas Martinetz, Erhardt Barth | Committees of deep feedforward networks trained with few data | null | null | null | null | cs.CV cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep convolutional neural networks are known to give good results on image
classification tasks. In this paper we present a method to improve the
classification result by combining multiple such networks in a committee. We
adopt the STL-10 dataset which has very few training examples and show that our
method can achieve results that are better than the state of the art. The
networks are trained layer-wise and no backpropagation is used. We also explore
the effects of dataset augmentation by mirroring, rotation, and scaling.
| [
{
"version": "v1",
"created": "Mon, 23 Jun 2014 15:34:54 GMT"
}
] | 2014-06-24T00:00:00 | [
[
"Miclut",
"Bogdan",
""
],
[
"Kaester",
"Thomas",
""
],
[
"Martinetz",
"Thomas",
""
],
[
"Barth",
"Erhardt",
""
]
] | TITLE: Committees of deep feedforward networks trained with few data
ABSTRACT: Deep convolutional neural networks are known to give good results on image
classification tasks. In this paper we present a method to improve the
classification result by combining multiple such networks in a committee. We
adopt the STL-10 dataset which has very few training examples and show that our
method can achieve results that are better than the state of the art. The
networks are trained layer-wise and no backpropagation is used. We also explore
the effects of dataset augmentation by mirroring, rotation, and scaling.
|
1301.2628 | Xu-Cheng Yin | Xu-Cheng Yin, Xuwang Yin, Kaizhu Huang, Hong-Wei Hao | Robust Text Detection in Natural Scene Images | A Draft Version (Submitted to IEEE TPAMI) | IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 36,
no. 5, pp. 970-983, 2014 | 10.1109/TPAMI.2013.182 | null | cs.CV cs.IR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Text detection in natural scene images is an important prerequisite for many
content-based image analysis tasks. In this paper, we propose an accurate and
robust method for detecting texts in natural scene images. A fast and effective
pruning algorithm is designed to extract Maximally Stable Extremal Regions
(MSERs) as character candidates using the strategy of minimizing regularized
variations. Character candidates are grouped into text candidates by the
ingle-link clustering algorithm, where distance weights and threshold of the
clustering algorithm are learned automatically by a novel self-training
distance metric learning algorithm. The posterior probabilities of text
candidates corresponding to non-text are estimated with an character
classifier; text candidates with high probabilities are then eliminated and
finally texts are identified with a text classifier. The proposed system is
evaluated on the ICDAR 2011 Robust Reading Competition dataset; the f measure
is over 76% and is significantly better than the state-of-the-art performance
of 71%. Experimental results on a publicly available multilingual dataset also
show that our proposed method can outperform the other competitive method with
the f measure increase of over 9 percent. Finally, we have setup an online demo
of our proposed scene text detection system at
http://kems.ustb.edu.cn/learning/yin/dtext.
| [
{
"version": "v1",
"created": "Fri, 11 Jan 2013 23:08:15 GMT"
},
{
"version": "v2",
"created": "Wed, 23 Jan 2013 19:57:46 GMT"
},
{
"version": "v3",
"created": "Sun, 2 Jun 2013 16:27:49 GMT"
}
] | 2014-06-23T00:00:00 | [
[
"Yin",
"Xu-Cheng",
""
],
[
"Yin",
"Xuwang",
""
],
[
"Huang",
"Kaizhu",
""
],
[
"Hao",
"Hong-Wei",
""
]
] | TITLE: Robust Text Detection in Natural Scene Images
ABSTRACT: Text detection in natural scene images is an important prerequisite for many
content-based image analysis tasks. In this paper, we propose an accurate and
robust method for detecting texts in natural scene images. A fast and effective
pruning algorithm is designed to extract Maximally Stable Extremal Regions
(MSERs) as character candidates using the strategy of minimizing regularized
variations. Character candidates are grouped into text candidates by the
ingle-link clustering algorithm, where distance weights and threshold of the
clustering algorithm are learned automatically by a novel self-training
distance metric learning algorithm. The posterior probabilities of text
candidates corresponding to non-text are estimated with an character
classifier; text candidates with high probabilities are then eliminated and
finally texts are identified with a text classifier. The proposed system is
evaluated on the ICDAR 2011 Robust Reading Competition dataset; the f measure
is over 76% and is significantly better than the state-of-the-art performance
of 71%. Experimental results on a publicly available multilingual dataset also
show that our proposed method can outperform the other competitive method with
the f measure increase of over 9 percent. Finally, we have setup an online demo
of our proposed scene text detection system at
http://kems.ustb.edu.cn/learning/yin/dtext.
|
1406.4966 | Jingdong Wang | Chao Du, Jingdong Wang | Inner Product Similarity Search using Compositional Codes | The approach presented in this paper (ECCV14 submission) is closely
related to multi-stage vector quantization and residual quantization. Thanks
the reviewers (CVPR14 and ECCV14) for pointing out the relationship to the
two algorithms. Related paper:
http://sites.skoltech.ru/app/data/uploads/sites/2/2013/09/CVPR14.pdf, which
also adopts the summation of vectors for vector approximation | null | null | null | cs.CV cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper addresses the nearest neighbor search problem under inner product
similarity and introduces a compact code-based approach. The idea is to
approximate a vector using the composition of several elements selected from a
source dictionary and to represent this vector by a short code composed of the
indices of the selected elements. The inner product between a query vector and
a database vector is efficiently estimated from the query vector and the short
code of the database vector. We show the superior performance of the proposed
group $M$-selection algorithm that selects $M$ elements from $M$ source
dictionaries for vector approximation in terms of search accuracy and
efficiency for compact codes of the same length via theoretical and empirical
analysis. Experimental results on large-scale datasets ($1M$ and $1B$ SIFT
features, $1M$ linear models and Netflix) demonstrate the superiority of the
proposed approach.
| [
{
"version": "v1",
"created": "Thu, 19 Jun 2014 07:42:05 GMT"
},
{
"version": "v2",
"created": "Fri, 20 Jun 2014 02:13:56 GMT"
}
] | 2014-06-23T00:00:00 | [
[
"Du",
"Chao",
""
],
[
"Wang",
"Jingdong",
""
]
] | TITLE: Inner Product Similarity Search using Compositional Codes
ABSTRACT: This paper addresses the nearest neighbor search problem under inner product
similarity and introduces a compact code-based approach. The idea is to
approximate a vector using the composition of several elements selected from a
source dictionary and to represent this vector by a short code composed of the
indices of the selected elements. The inner product between a query vector and
a database vector is efficiently estimated from the query vector and the short
code of the database vector. We show the superior performance of the proposed
group $M$-selection algorithm that selects $M$ elements from $M$ source
dictionaries for vector approximation in terms of search accuracy and
efficiency for compact codes of the same length via theoretical and empirical
analysis. Experimental results on large-scale datasets ($1M$ and $1B$ SIFT
features, $1M$ linear models and Netflix) demonstrate the superiority of the
proposed approach.
|
1406.5212 | Georgia Gkioxari | Georgia Gkioxari, Bharath Hariharan, Ross Girshick, Jitendra Malik | R-CNNs for Pose Estimation and Action Detection | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present convolutional neural networks for the tasks of keypoint (pose)
prediction and action classification of people in unconstrained images. Our
approach involves training an R-CNN detector with loss functions depending on
the task being tackled. We evaluate our method on the challenging PASCAL VOC
dataset and compare it to previous leading approaches. Our method gives
state-of-the-art results for keypoint and action prediction. Additionally, we
introduce a new dataset for action detection, the task of simultaneously
localizing people and classifying their actions, and present results using our
approach.
| [
{
"version": "v1",
"created": "Thu, 19 Jun 2014 20:56:08 GMT"
}
] | 2014-06-23T00:00:00 | [
[
"Gkioxari",
"Georgia",
""
],
[
"Hariharan",
"Bharath",
""
],
[
"Girshick",
"Ross",
""
],
[
"Malik",
"Jitendra",
""
]
] | TITLE: R-CNNs for Pose Estimation and Action Detection
ABSTRACT: We present convolutional neural networks for the tasks of keypoint (pose)
prediction and action classification of people in unconstrained images. Our
approach involves training an R-CNN detector with loss functions depending on
the task being tackled. We evaluate our method on the challenging PASCAL VOC
dataset and compare it to previous leading approaches. Our method gives
state-of-the-art results for keypoint and action prediction. Additionally, we
introduce a new dataset for action detection, the task of simultaneously
localizing people and classifying their actions, and present results using our
approach.
|
1404.7796 | Emilie Morvant | Emilie Morvant (IST Austria), Amaury Habrard (LHC), St\'ephane Ayache
(LIF) | Majority Vote of Diverse Classifiers for Late Fusion | IAPR Joint International Workshops on Statistical Techniques in
Pattern Recognition and Structural and Syntactic Pattern Recignition, Joensuu
: Finland (2014) | null | null | null | stat.ML cs.LG cs.MM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In the past few years, a lot of attention has been devoted to multimedia
indexing by fusing multimodal informations. Two kinds of fusion schemes are
generally considered: The early fusion and the late fusion. We focus on late
classifier fusion, where one combines the scores of each modality at the
decision level. To tackle this problem, we investigate a recent and elegant
well-founded quadratic program named MinCq coming from the machine learning
PAC-Bayesian theory. MinCq looks for the weighted combination, over a set of
real-valued functions seen as voters, leading to the lowest misclassification
rate, while maximizing the voters' diversity. We propose an extension of MinCq
tailored to multimedia indexing. Our method is based on an order-preserving
pairwise loss adapted to ranking that allows us to improve Mean Averaged
Precision measure while taking into account the diversity of the voters that we
want to fuse. We provide evidence that this method is naturally adapted to late
fusion procedures and confirm the good behavior of our approach on the
challenging PASCAL VOC'07 benchmark.
| [
{
"version": "v1",
"created": "Wed, 30 Apr 2014 16:55:00 GMT"
},
{
"version": "v2",
"created": "Thu, 19 Jun 2014 08:06:24 GMT"
}
] | 2014-06-20T00:00:00 | [
[
"Morvant",
"Emilie",
"",
"IST Austria"
],
[
"Habrard",
"Amaury",
"",
"LHC"
],
[
"Ayache",
"Stéphane",
"",
"LIF"
]
] | TITLE: Majority Vote of Diverse Classifiers for Late Fusion
ABSTRACT: In the past few years, a lot of attention has been devoted to multimedia
indexing by fusing multimodal informations. Two kinds of fusion schemes are
generally considered: The early fusion and the late fusion. We focus on late
classifier fusion, where one combines the scores of each modality at the
decision level. To tackle this problem, we investigate a recent and elegant
well-founded quadratic program named MinCq coming from the machine learning
PAC-Bayesian theory. MinCq looks for the weighted combination, over a set of
real-valued functions seen as voters, leading to the lowest misclassification
rate, while maximizing the voters' diversity. We propose an extension of MinCq
tailored to multimedia indexing. Our method is based on an order-preserving
pairwise loss adapted to ranking that allows us to improve Mean Averaged
Precision measure while taking into account the diversity of the voters that we
want to fuse. We provide evidence that this method is naturally adapted to late
fusion procedures and confirm the good behavior of our approach on the
challenging PASCAL VOC'07 benchmark.
|
1406.4973 | Gaetan Marceau | Ga\'etan Marceau (LRI, INRIA Saclay - Ile de France), Marc Schoenauer
(LRI, INRIA Saclay - Ile de France) | Racing Multi-Objective Selection Probabilities | null | 13th International Conference on Parallel Problem Solving from
Nature, Ljubljana : France (2014) | null | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In the context of Noisy Multi-Objective Optimization, dealing with
uncertainties requires the decision maker to define some preferences about how
to handle them, through some statistics (e.g., mean, median) to be used to
evaluate the qualities of the solutions, and define the corresponding Pareto
set. Approximating these statistics requires repeated samplings of the
population, drastically increasing the overall computational cost. To tackle
this issue, this paper proposes to directly estimate the probability of each
individual to be selected, using some Hoeffding races to dynamically assign the
estimation budget during the selection step. The proposed racing approach is
validated against static budget approaches with NSGA-II on noisy versions of
the ZDT benchmark functions.
| [
{
"version": "v1",
"created": "Thu, 19 Jun 2014 08:07:47 GMT"
}
] | 2014-06-20T00:00:00 | [
[
"Marceau",
"Gaétan",
"",
"LRI, INRIA Saclay - Ile de France"
],
[
"Schoenauer",
"Marc",
"",
"LRI, INRIA Saclay - Ile de France"
]
] | TITLE: Racing Multi-Objective Selection Probabilities
ABSTRACT: In the context of Noisy Multi-Objective Optimization, dealing with
uncertainties requires the decision maker to define some preferences about how
to handle them, through some statistics (e.g., mean, median) to be used to
evaluate the qualities of the solutions, and define the corresponding Pareto
set. Approximating these statistics requires repeated samplings of the
population, drastically increasing the overall computational cost. To tackle
this issue, this paper proposes to directly estimate the probability of each
individual to be selected, using some Hoeffding races to dynamically assign the
estimation budget during the selection step. The proposed racing approach is
validated against static budget approaches with NSGA-II on noisy versions of
the ZDT benchmark functions.
|
1406.5059 | Abhishek Bhola | Abhishek Bhola | Twitter and Polls: Analyzing and estimating political orientation of
Twitter users in India General #Elections2014 | null | null | null | null | cs.SI physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This year (2014) in the month of May, the tenure of the 15th Lok Sabha was to
end and the elections to the 543 parliamentary seats were to be held. A
whooping $5 billion were spent on these elections, which made us stand second
only to the US Presidential elections in terms of money spent. Swelling number
of Internet users and Online Social Media (OSM) users could effect 3-4% of
urban population votes as per a report of IAMAI (Internet & Mobile Association
of India). Our count of tweets related to elections from September 2013 to May
2014, was close to 18.21 million. We analyzed the complete dataset and found
that the activity on Twitter peaked during important events. It was evident
from our data that the political behavior of the politicians affected their
followers count. Yet another aim of our work was to find an efficient way to
classify the political orientation of the users on Twitter. We used four
different techniques: two were based on the content of the tweets, one on the
user based features and another based on community detection algorithm on the
retweet and user mention networks. We found that the community detection
algorithm worked best. We built a portal to show the analysis of the tweets of
the last 24 hours. To the best of our knowledge, this is the first academic
pursuit to analyze the elections data and classify the users in the India
General Elections 2014.
| [
{
"version": "v1",
"created": "Thu, 19 Jun 2014 14:27:09 GMT"
}
] | 2014-06-20T00:00:00 | [
[
"Bhola",
"Abhishek",
""
]
] | TITLE: Twitter and Polls: Analyzing and estimating political orientation of
Twitter users in India General #Elections2014
ABSTRACT: This year (2014) in the month of May, the tenure of the 15th Lok Sabha was to
end and the elections to the 543 parliamentary seats were to be held. A
whooping $5 billion were spent on these elections, which made us stand second
only to the US Presidential elections in terms of money spent. Swelling number
of Internet users and Online Social Media (OSM) users could effect 3-4% of
urban population votes as per a report of IAMAI (Internet & Mobile Association
of India). Our count of tweets related to elections from September 2013 to May
2014, was close to 18.21 million. We analyzed the complete dataset and found
that the activity on Twitter peaked during important events. It was evident
from our data that the political behavior of the politicians affected their
followers count. Yet another aim of our work was to find an efficient way to
classify the political orientation of the users on Twitter. We used four
different techniques: two were based on the content of the tweets, one on the
user based features and another based on community detection algorithm on the
retweet and user mention networks. We found that the community detection
algorithm worked best. We built a portal to show the analysis of the tweets of
the last 24 hours. To the best of our knowledge, this is the first academic
pursuit to analyze the elections data and classify the users in the India
General Elections 2014.
|
1406.5074 | Vijendra Singh | Singh Vijendra and Pathak Shivani | Robust Outlier Detection Technique in Data Mining: A Univariate Approach | arXiv admin note: text overlap with arXiv:1402.6859 by other authors
without attribution | null | null | MT CS 2011 | cs.CV | http://creativecommons.org/licenses/by/3.0/ | Outliers are the points which are different from or inconsistent with the
rest of the data. They can be novel, new, abnormal, unusual or noisy
information. Outliers are sometimes more interesting than the majority of the
data. The main challenges of outlier detection with the increasing complexity,
size and variety of datasets, are how to catch similar outliers as a group, and
how to evaluate the outliers. This paper describes an approach which uses
Univariate outlier detection as a pre-processing step to detect the outlier and
then applies K-means algorithm hence to analyse the effects of the outliers on
the cluster analysis of dataset.
| [
{
"version": "v1",
"created": "Thu, 19 Jun 2014 15:12:49 GMT"
}
] | 2014-06-20T00:00:00 | [
[
"Vijendra",
"Singh",
""
],
[
"Shivani",
"Pathak",
""
]
] | TITLE: Robust Outlier Detection Technique in Data Mining: A Univariate Approach
ABSTRACT: Outliers are the points which are different from or inconsistent with the
rest of the data. They can be novel, new, abnormal, unusual or noisy
information. Outliers are sometimes more interesting than the majority of the
data. The main challenges of outlier detection with the increasing complexity,
size and variety of datasets, are how to catch similar outliers as a group, and
how to evaluate the outliers. This paper describes an approach which uses
Univariate outlier detection as a pre-processing step to detect the outlier and
then applies K-means algorithm hence to analyse the effects of the outliers on
the cluster analysis of dataset.
|
1406.5095 | Conrad Sanderson | Vikas Reddy, Conrad Sanderson, Andres Sanin, Brian C. Lovell | MRF-based Background Initialisation for Improved Foreground Detection in
Cluttered Surveillance Videos | arXiv admin note: substantial text overlap with arXiv:1303.2465 | null | 10.1007/978-3-642-19318-7_43 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Robust foreground object segmentation via background modelling is a difficult
problem in cluttered environments, where obtaining a clear view of the
background to model is almost impossible. In this paper, we propose a method
capable of robustly estimating the background and detecting regions of interest
in such environments. In particular, we propose to extend the background
initialisation component of a recent patch-based foreground detection algorithm
with an elaborate technique based on Markov Random Fields, where the optimal
labelling solution is computed using iterated conditional modes. Rather than
relying purely on local temporal statistics, the proposed technique takes into
account the spatial continuity of the entire background. Experiments with
several tracking algorithms on the CAVIAR dataset indicate that the proposed
method leads to considerable improvements in object tracking accuracy, when
compared to methods based on Gaussian mixture models and feature histograms.
| [
{
"version": "v1",
"created": "Thu, 19 Jun 2014 16:06:53 GMT"
}
] | 2014-06-20T00:00:00 | [
[
"Reddy",
"Vikas",
""
],
[
"Sanderson",
"Conrad",
""
],
[
"Sanin",
"Andres",
""
],
[
"Lovell",
"Brian C.",
""
]
] | TITLE: MRF-based Background Initialisation for Improved Foreground Detection in
Cluttered Surveillance Videos
ABSTRACT: Robust foreground object segmentation via background modelling is a difficult
problem in cluttered environments, where obtaining a clear view of the
background to model is almost impossible. In this paper, we propose a method
capable of robustly estimating the background and detecting regions of interest
in such environments. In particular, we propose to extend the background
initialisation component of a recent patch-based foreground detection algorithm
with an elaborate technique based on Markov Random Fields, where the optimal
labelling solution is computed using iterated conditional modes. Rather than
relying purely on local temporal statistics, the proposed technique takes into
account the spatial continuity of the entire background. Experiments with
several tracking algorithms on the CAVIAR dataset indicate that the proposed
method leads to considerable improvements in object tracking accuracy, when
compared to methods based on Gaussian mixture models and feature histograms.
|
1406.5106 | David Van Horn | J. Ian Johnson, Ilya Sergey, Christopher Earl, Matthew Might, David
Van Horn | Pushdown flow analysis with abstract garbage collection | null | Journal of Functional Programming, Volume 24, Special Issue 2-3,
May 2014, pp 218-283 | 10.1017/S0956796814000100 10.1017/S0956796814000100
10.1017/S0956796814000100 10.1017/S0956796814000100 10.1017/S0956796814000100 | null | cs.PL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In the static analysis of functional programs, pushdown flow analysis and
abstract garbage collection push the boundaries of what we can learn about
programs statically. This work illuminates and poses solutions to theoretical
and practical challenges that stand in the way of combining the power of these
techniques. Pushdown flow analysis grants unbounded yet computable polyvariance
to the analysis of return-flow in higher-order programs. Abstract garbage
collection grants unbounded polyvariance to abstract addresses which become
unreachable between invocations of the abstract contexts in which they were
created. Pushdown analysis solves the problem of precisely analyzing recursion
in higher-order languages; abstract garbage collection is essential in solving
the "stickiness" problem. Alone, our benchmarks demonstrate that each method
can reduce analysis times and boost precision by orders of magnitude. We
combine these methods. The challenge in marrying these techniques is not
subtle: computing the reachable control states of a pushdown system relies on
limiting access during transition to the top of the stack; abstract garbage
collection, on the other hand, needs full access to the entire stack to compute
a root set, just as concrete collection does. Conditional pushdown systems were
developed for just such a conundrum, but existing methods are ill-suited for
the dynamic nature of garbage collection. We show fully precise and approximate
solutions to the feasible paths problem for pushdown garbage-collecting
control-flow analysis. Experiments reveal synergistic interplay between garbage
collection and pushdown techniques, and the fusion demonstrates
"better-than-both-worlds" precision.
| [
{
"version": "v1",
"created": "Thu, 19 Jun 2014 16:51:12 GMT"
}
] | 2014-06-20T00:00:00 | [
[
"Johnson",
"J. Ian",
""
],
[
"Sergey",
"Ilya",
""
],
[
"Earl",
"Christopher",
""
],
[
"Might",
"Matthew",
""
],
[
"Van Horn",
"David",
""
]
] | TITLE: Pushdown flow analysis with abstract garbage collection
ABSTRACT: In the static analysis of functional programs, pushdown flow analysis and
abstract garbage collection push the boundaries of what we can learn about
programs statically. This work illuminates and poses solutions to theoretical
and practical challenges that stand in the way of combining the power of these
techniques. Pushdown flow analysis grants unbounded yet computable polyvariance
to the analysis of return-flow in higher-order programs. Abstract garbage
collection grants unbounded polyvariance to abstract addresses which become
unreachable between invocations of the abstract contexts in which they were
created. Pushdown analysis solves the problem of precisely analyzing recursion
in higher-order languages; abstract garbage collection is essential in solving
the "stickiness" problem. Alone, our benchmarks demonstrate that each method
can reduce analysis times and boost precision by orders of magnitude. We
combine these methods. The challenge in marrying these techniques is not
subtle: computing the reachable control states of a pushdown system relies on
limiting access during transition to the top of the stack; abstract garbage
collection, on the other hand, needs full access to the entire stack to compute
a root set, just as concrete collection does. Conditional pushdown systems were
developed for just such a conundrum, but existing methods are ill-suited for
the dynamic nature of garbage collection. We show fully precise and approximate
solutions to the feasible paths problem for pushdown garbage-collecting
control-flow analysis. Experiments reveal synergistic interplay between garbage
collection and pushdown techniques, and the fusion demonstrates
"better-than-both-worlds" precision.
|
1406.5161 | Jeyanthi Salem Narasimhan | Jeyanthi Narasimhan, Abhinav Vishnu, Lawrence Holder, Adolfy Hoisie | Fast Support Vector Machines Using Parallel Adaptive Shrinking on
Distributed Systems | 10 pages, 9 figures, 3 tables | null | null | null | cs.DC cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Support Vector Machines (SVM), a popular machine learning technique, has been
applied to a wide range of domains such as science, finance, and social
networks for supervised learning. Whether it is identifying high-risk patients
by health-care professionals, or potential high-school students to enroll in
college by school districts, SVMs can play a major role for social good. This
paper undertakes the challenge of designing a scalable parallel SVM training
algorithm for large scale systems, which includes commodity multi-core
machines, tightly connected supercomputers and cloud computing systems.
Intuitive techniques for improving the time-space complexity including adaptive
elimination of samples for faster convergence and sparse format representation
are proposed. Under sample elimination, several heuristics for {\em earliest
possible} to {\em lazy} elimination of non-contributing samples are proposed.
In several cases, where an early sample elimination might result in a false
positive, low overhead mechanisms for reconstruction of key data structures are
proposed. The algorithm and heuristics are implemented and evaluated on various
publicly available datasets. Empirical evaluation shows up to 26x speed
improvement on some datasets against the sequential baseline, when evaluated on
multiple compute nodes, and an improvement in execution time up to 30-60\% is
readily observed on a number of other datasets against our parallel baseline.
| [
{
"version": "v1",
"created": "Thu, 19 Jun 2014 19:22:28 GMT"
}
] | 2014-06-20T00:00:00 | [
[
"Narasimhan",
"Jeyanthi",
""
],
[
"Vishnu",
"Abhinav",
""
],
[
"Holder",
"Lawrence",
""
],
[
"Hoisie",
"Adolfy",
""
]
] | TITLE: Fast Support Vector Machines Using Parallel Adaptive Shrinking on
Distributed Systems
ABSTRACT: Support Vector Machines (SVM), a popular machine learning technique, has been
applied to a wide range of domains such as science, finance, and social
networks for supervised learning. Whether it is identifying high-risk patients
by health-care professionals, or potential high-school students to enroll in
college by school districts, SVMs can play a major role for social good. This
paper undertakes the challenge of designing a scalable parallel SVM training
algorithm for large scale systems, which includes commodity multi-core
machines, tightly connected supercomputers and cloud computing systems.
Intuitive techniques for improving the time-space complexity including adaptive
elimination of samples for faster convergence and sparse format representation
are proposed. Under sample elimination, several heuristics for {\em earliest
possible} to {\em lazy} elimination of non-contributing samples are proposed.
In several cases, where an early sample elimination might result in a false
positive, low overhead mechanisms for reconstruction of key data structures are
proposed. The algorithm and heuristics are implemented and evaluated on various
publicly available datasets. Empirical evaluation shows up to 26x speed
improvement on some datasets against the sequential baseline, when evaluated on
multiple compute nodes, and an improvement in execution time up to 30-60\% is
readily observed on a number of other datasets against our parallel baseline.
|
1311.4336 | Junming Huang | Junming Huang, Chao Li, Wen-Qiang Wang, Hua-Wei Shen, Guojie Li,
Xue-Qi Cheng | Temporal scaling in information propagation | 13 pages, 2 figures. published on Scientific Reports | Scientific Reports 4, 5334, (2014) | 10.1038/srep05334 | null | cs.SI physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | For the study of information propagation, one fundamental problem is
uncovering universal laws governing the dynamics of information propagation.
This problem, from the microscopic perspective, is formulated as estimating the
propagation probability that a piece of information propagates from one
individual to another. Such a propagation probability generally depends on two
major classes of factors: the intrinsic attractiveness of information and the
interactions between individuals. Despite the fact that the temporal effect of
attractiveness is widely studied, temporal laws underlying individual
interactions remain unclear, causing inaccurate prediction of information
propagation on evolving social networks. In this report, we empirically study
the dynamics of information propagation, using the dataset from a
population-scale social media website. We discover a temporal scaling in
information propagation: the probability a message propagates between two
individuals decays with the length of time latency since their latest
interaction, obeying a power-law rule. Leveraging the scaling law, we further
propose a temporal model to estimate future propagation probabilities between
individuals, reducing the error rate of information propagation prediction from
6.7% to 2.6% and improving viral marketing with 9.7% incremental customers.
| [
{
"version": "v1",
"created": "Mon, 18 Nov 2013 11:15:26 GMT"
},
{
"version": "v2",
"created": "Fri, 22 Nov 2013 02:15:14 GMT"
},
{
"version": "v3",
"created": "Wed, 18 Jun 2014 09:55:29 GMT"
}
] | 2014-06-19T00:00:00 | [
[
"Huang",
"Junming",
""
],
[
"Li",
"Chao",
""
],
[
"Wang",
"Wen-Qiang",
""
],
[
"Shen",
"Hua-Wei",
""
],
[
"Li",
"Guojie",
""
],
[
"Cheng",
"Xue-Qi",
""
]
] | TITLE: Temporal scaling in information propagation
ABSTRACT: For the study of information propagation, one fundamental problem is
uncovering universal laws governing the dynamics of information propagation.
This problem, from the microscopic perspective, is formulated as estimating the
propagation probability that a piece of information propagates from one
individual to another. Such a propagation probability generally depends on two
major classes of factors: the intrinsic attractiveness of information and the
interactions between individuals. Despite the fact that the temporal effect of
attractiveness is widely studied, temporal laws underlying individual
interactions remain unclear, causing inaccurate prediction of information
propagation on evolving social networks. In this report, we empirically study
the dynamics of information propagation, using the dataset from a
population-scale social media website. We discover a temporal scaling in
information propagation: the probability a message propagates between two
individuals decays with the length of time latency since their latest
interaction, obeying a power-law rule. Leveraging the scaling law, we further
propose a temporal model to estimate future propagation probabilities between
individuals, reducing the error rate of information propagation prediction from
6.7% to 2.6% and improving viral marketing with 9.7% incremental customers.
|
1402.5516 | Peng Zhang | Peng Zhang, Wei Chen, Xiaoming Sun, Yajun Wang, Jialin Zhang | Minimizing Seed Set Selection with Probabilistic Coverage Guarantee in a
Social Network | Conference version will appear in KDD 2014 | null | null | null | cs.SI cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A topic propagating in a social network reaches its tipping point if the
number of users discussing it in the network exceeds a critical threshold such
that a wide cascade on the topic is likely to occur. In this paper, we consider
the task of selecting initial seed users of a topic with minimum size so that
with a guaranteed probability the number of users discussing the topic would
reach a given threshold. We formulate the task as an optimization problem
called seed minimization with probabilistic coverage guarantee (SM-PCG). This
problem departs from the previous studies on social influence maximization or
seed minimization because it considers influence coverage with probabilistic
guarantees instead of guarantees on expected influence coverage. We show that
the problem is not submodular, and thus is harder than previously studied
problems based on submodular function optimization. We provide an approximation
algorithm and show that it approximates the optimal solution with both a
multiplicative ratio and an additive error. The multiplicative ratio is tight
while the additive error would be small if influence coverage distributions of
certain seed sets are well concentrated. For one-way bipartite graphs we
analytically prove the concentration condition and obtain an approximation
algorithm with an $O(\log n)$ multiplicative ratio and an $O(\sqrt{n})$
additive error, where $n$ is the total number of nodes in the social graph.
Moreover, we empirically verify the concentration condition in real-world
networks and experimentally demonstrate the effectiveness of our proposed
algorithm comparing to commonly adopted benchmark algorithms.
| [
{
"version": "v1",
"created": "Sat, 22 Feb 2014 15:24:12 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2014 17:41:41 GMT"
}
] | 2014-06-19T00:00:00 | [
[
"Zhang",
"Peng",
""
],
[
"Chen",
"Wei",
""
],
[
"Sun",
"Xiaoming",
""
],
[
"Wang",
"Yajun",
""
],
[
"Zhang",
"Jialin",
""
]
] | TITLE: Minimizing Seed Set Selection with Probabilistic Coverage Guarantee in a
Social Network
ABSTRACT: A topic propagating in a social network reaches its tipping point if the
number of users discussing it in the network exceeds a critical threshold such
that a wide cascade on the topic is likely to occur. In this paper, we consider
the task of selecting initial seed users of a topic with minimum size so that
with a guaranteed probability the number of users discussing the topic would
reach a given threshold. We formulate the task as an optimization problem
called seed minimization with probabilistic coverage guarantee (SM-PCG). This
problem departs from the previous studies on social influence maximization or
seed minimization because it considers influence coverage with probabilistic
guarantees instead of guarantees on expected influence coverage. We show that
the problem is not submodular, and thus is harder than previously studied
problems based on submodular function optimization. We provide an approximation
algorithm and show that it approximates the optimal solution with both a
multiplicative ratio and an additive error. The multiplicative ratio is tight
while the additive error would be small if influence coverage distributions of
certain seed sets are well concentrated. For one-way bipartite graphs we
analytically prove the concentration condition and obtain an approximation
algorithm with an $O(\log n)$ multiplicative ratio and an $O(\sqrt{n})$
additive error, where $n$ is the total number of nodes in the social graph.
Moreover, we empirically verify the concentration condition in real-world
networks and experimentally demonstrate the effectiveness of our proposed
algorithm comparing to commonly adopted benchmark algorithms.
|
1406.4296 | Adrien Gaidon | Adrien Gaidon (Xerox Research Center Europe, France), Gloria Zen
(University of Trento, Italy), Jose A. Rodriguez-Serrano (Xerox Research
Center Europe, France) | Self-Learning Camera: Autonomous Adaptation of Object Detectors to
Unlabeled Video Streams | null | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Learning object detectors requires massive amounts of labeled training
samples from the specific data source of interest. This is impractical when
dealing with many different sources (e.g., in camera networks), or constantly
changing ones such as mobile cameras (e.g., in robotics or driving assistant
systems). In this paper, we address the problem of self-learning detectors in
an autonomous manner, i.e. (i) detectors continuously updating themselves to
efficiently adapt to streaming data sources (contrary to transductive
algorithms), (ii) without any labeled data strongly related to the target data
stream (contrary to self-paced learning), and (iii) without manual intervention
to set and update hyper-parameters. To that end, we propose an unsupervised,
on-line, and self-tuning learning algorithm to optimize a multi-task learning
convex objective. Our method uses confident but laconic oracles (high-precision
but low-recall off-the-shelf generic detectors), and exploits the structure of
the problem to jointly learn on-line an ensemble of instance-level trackers,
from which we derive an adapted category-level object detector. Our approach is
validated on real-world publicly available video object datasets.
| [
{
"version": "v1",
"created": "Tue, 17 Jun 2014 09:51:18 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2014 12:33:22 GMT"
}
] | 2014-06-19T00:00:00 | [
[
"Gaidon",
"Adrien",
"",
"Xerox Research Center Europe, France"
],
[
"Zen",
"Gloria",
"",
"University of Trento, Italy"
],
[
"Rodriguez-Serrano",
"Jose A.",
"",
"Xerox Research\n Center Europe, France"
]
] | TITLE: Self-Learning Camera: Autonomous Adaptation of Object Detectors to
Unlabeled Video Streams
ABSTRACT: Learning object detectors requires massive amounts of labeled training
samples from the specific data source of interest. This is impractical when
dealing with many different sources (e.g., in camera networks), or constantly
changing ones such as mobile cameras (e.g., in robotics or driving assistant
systems). In this paper, we address the problem of self-learning detectors in
an autonomous manner, i.e. (i) detectors continuously updating themselves to
efficiently adapt to streaming data sources (contrary to transductive
algorithms), (ii) without any labeled data strongly related to the target data
stream (contrary to self-paced learning), and (iii) without manual intervention
to set and update hyper-parameters. To that end, we propose an unsupervised,
on-line, and self-tuning learning algorithm to optimize a multi-task learning
convex objective. Our method uses confident but laconic oracles (high-precision
but low-recall off-the-shelf generic detectors), and exploits the structure of
the problem to jointly learn on-line an ensemble of instance-level trackers,
from which we derive an adapted category-level object detector. Our approach is
validated on real-world publicly available video object datasets.
|
1406.4518 | Erfan Khaji Mr. | Erfan Khaji, Amin Satlikh Mohammadi | A Heuristic Method to Generate Better Initial Population for
Evolutionary Methods | null | null | null | null | cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Initial population plays an important role in heuristic algorithms such as GA
as it help to decrease the time those algorithms need to achieve an acceptable
result. Furthermore, it may influence the quality of the final answer given by
evolutionary algorithms. In this paper, we shall introduce a heuristic method
to generate a target based initial population which possess two mentioned
characteristics. The efficiency of the proposed method has been shown by
presenting the results of our tests on the benchmarks.
| [
{
"version": "v1",
"created": "Sun, 15 Jun 2014 15:20:17 GMT"
}
] | 2014-06-19T00:00:00 | [
[
"Khaji",
"Erfan",
""
],
[
"Mohammadi",
"Amin Satlikh",
""
]
] | TITLE: A Heuristic Method to Generate Better Initial Population for
Evolutionary Methods
ABSTRACT: Initial population plays an important role in heuristic algorithms such as GA
as it help to decrease the time those algorithms need to achieve an acceptable
result. Furthermore, it may influence the quality of the final answer given by
evolutionary algorithms. In this paper, we shall introduce a heuristic method
to generate a target based initial population which possess two mentioned
characteristics. The efficiency of the proposed method has been shown by
presenting the results of our tests on the benchmarks.
|
1406.4773 | Yi Sun | Yi Sun, Xiaogang Wang, Xiaoou Tang | Deep Learning Face Representation by Joint Identification-Verification | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The key challenge of face recognition is to develop effective feature
representations for reducing intra-personal variations while enlarging
inter-personal differences. In this paper, we show that it can be well solved
with deep learning and using both face identification and verification signals
as supervision. The Deep IDentification-verification features (DeepID2) are
learned with carefully designed deep convolutional networks. The face
identification task increases the inter-personal variations by drawing DeepID2
extracted from different identities apart, while the face verification task
reduces the intra-personal variations by pulling DeepID2 extracted from the
same identity together, both of which are essential to face recognition. The
learned DeepID2 features can be well generalized to new identities unseen in
the training data. On the challenging LFW dataset, 99.15% face verification
accuracy is achieved. Compared with the best deep learning result on LFW, the
error rate has been significantly reduced by 67%.
| [
{
"version": "v1",
"created": "Wed, 18 Jun 2014 15:42:16 GMT"
}
] | 2014-06-19T00:00:00 | [
[
"Sun",
"Yi",
""
],
[
"Wang",
"Xiaogang",
""
],
[
"Tang",
"Xiaoou",
""
]
] | TITLE: Deep Learning Face Representation by Joint Identification-Verification
ABSTRACT: The key challenge of face recognition is to develop effective feature
representations for reducing intra-personal variations while enlarging
inter-personal differences. In this paper, we show that it can be well solved
with deep learning and using both face identification and verification signals
as supervision. The Deep IDentification-verification features (DeepID2) are
learned with carefully designed deep convolutional networks. The face
identification task increases the inter-personal variations by drawing DeepID2
extracted from different identities apart, while the face verification task
reduces the intra-personal variations by pulling DeepID2 extracted from the
same identity together, both of which are essential to face recognition. The
learned DeepID2 features can be well generalized to new identities unseen in
the training data. On the challenging LFW dataset, 99.15% face verification
accuracy is achieved. Compared with the best deep learning result on LFW, the
error rate has been significantly reduced by 67%.
|
1406.4775 | Andrea Montanari | Andrea Montanari and Emile Richard | Non-negative Principal Component Analysis: Message Passing Algorithms
and Sharp Asymptotics | 51 pages, 7 pdf figures | null | null | null | cs.IT math.IT math.ST stat.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Principal component analysis (PCA) aims at estimating the direction of
maximal variability of a high-dimensional dataset. A natural question is: does
this task become easier, and estimation more accurate, when we exploit
additional knowledge on the principal vector? We study the case in which the
principal vector is known to lie in the positive orthant. Similar constraints
arise in a number of applications, ranging from analysis of gene expression
data to spike sorting in neural signal processing.
In the unconstrained case, the estimation performances of PCA has been
precisely characterized using random matrix theory, under a statistical model
known as the `spiked model.' It is known that the estimation error undergoes a
phase transition as the signal-to-noise ratio crosses a certain threshold.
Unfortunately, tools from random matrix theory have no bearing on the
constrained problem. Despite this challenge, we develop an analogous
characterization in the constrained case, within a one-spike model.
In particular: $(i)$~We prove that the estimation error undergoes a similar
phase transition, albeit at a different threshold in signal-to-noise ratio that
we determine exactly; $(ii)$~We prove that --unlike in the unconstrained case--
estimation error depends on the spike vector, and characterize the least
favorable vectors; $(iii)$~We show that a non-negative principal component can
be approximately computed --under the spiked model-- in nearly linear time.
This despite the fact that the problem is non-convex and, in general, NP-hard
to solve exactly.
| [
{
"version": "v1",
"created": "Wed, 18 Jun 2014 15:47:33 GMT"
}
] | 2014-06-19T00:00:00 | [
[
"Montanari",
"Andrea",
""
],
[
"Richard",
"Emile",
""
]
] | TITLE: Non-negative Principal Component Analysis: Message Passing Algorithms
and Sharp Asymptotics
ABSTRACT: Principal component analysis (PCA) aims at estimating the direction of
maximal variability of a high-dimensional dataset. A natural question is: does
this task become easier, and estimation more accurate, when we exploit
additional knowledge on the principal vector? We study the case in which the
principal vector is known to lie in the positive orthant. Similar constraints
arise in a number of applications, ranging from analysis of gene expression
data to spike sorting in neural signal processing.
In the unconstrained case, the estimation performances of PCA has been
precisely characterized using random matrix theory, under a statistical model
known as the `spiked model.' It is known that the estimation error undergoes a
phase transition as the signal-to-noise ratio crosses a certain threshold.
Unfortunately, tools from random matrix theory have no bearing on the
constrained problem. Despite this challenge, we develop an analogous
characterization in the constrained case, within a one-spike model.
In particular: $(i)$~We prove that the estimation error undergoes a similar
phase transition, albeit at a different threshold in signal-to-noise ratio that
we determine exactly; $(ii)$~We prove that --unlike in the unconstrained case--
estimation error depends on the spike vector, and characterize the least
favorable vectors; $(iii)$~We show that a non-negative principal component can
be approximately computed --under the spiked model-- in nearly linear time.
This despite the fact that the problem is non-convex and, in general, NP-hard
to solve exactly.
|
1406.4784 | Ping Li | Anshumali Shrivastava and Ping Li | Improved Densification of One Permutation Hashing | null | null | null | null | stat.ME cs.DS cs.IR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The existing work on densification of one permutation hashing reduces the
query processing cost of the $(K,L)$-parameterized Locality Sensitive Hashing
(LSH) algorithm with minwise hashing, from $O(dKL)$ to merely $O(d + KL)$,
where $d$ is the number of nonzeros of the data vector, $K$ is the number of
hashes in each hash table, and $L$ is the number of hash tables. While that is
a substantial improvement, our analysis reveals that the existing densification
scheme is sub-optimal. In particular, there is no enough randomness in that
procedure, which affects its accuracy on very sparse datasets.
In this paper, we provide a new densification procedure which is provably
better than the existing scheme. This improvement is more significant for very
sparse datasets which are common over the web. The improved technique has the
same cost of $O(d + KL)$ for query processing, thereby making it strictly
preferable over the existing procedure. Experimental evaluations on public
datasets, in the task of hashing based near neighbor search, support our
theoretical findings.
| [
{
"version": "v1",
"created": "Wed, 18 Jun 2014 16:16:22 GMT"
}
] | 2014-06-19T00:00:00 | [
[
"Shrivastava",
"Anshumali",
""
],
[
"Li",
"Ping",
""
]
] | TITLE: Improved Densification of One Permutation Hashing
ABSTRACT: The existing work on densification of one permutation hashing reduces the
query processing cost of the $(K,L)$-parameterized Locality Sensitive Hashing
(LSH) algorithm with minwise hashing, from $O(dKL)$ to merely $O(d + KL)$,
where $d$ is the number of nonzeros of the data vector, $K$ is the number of
hashes in each hash table, and $L$ is the number of hash tables. While that is
a substantial improvement, our analysis reveals that the existing densification
scheme is sub-optimal. In particular, there is no enough randomness in that
procedure, which affects its accuracy on very sparse datasets.
In this paper, we provide a new densification procedure which is provably
better than the existing scheme. This improvement is more significant for very
sparse datasets which are common over the web. The improved technique has the
same cost of $O(d + KL)$ for query processing, thereby making it strictly
preferable over the existing procedure. Experimental evaluations on public
datasets, in the task of hashing based near neighbor search, support our
theoretical findings.
|
1012.3951 | Alex Bronstein | Roee Litman, Alex M. Bronstein, Michael M. Bronstein | Diffusion-geometric maximally stable component detection in deformable
shapes | null | null | 10.1016/j.cag.2011.03.011 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Maximally stable component detection is a very popular method for feature
analysis in images, mainly due to its low computation cost and high
repeatability. With the recent advance of feature-based methods in geometric
shape analysis, there is significant interest in finding analogous approaches
in the 3D world. In this paper, we formulate a diffusion-geometric framework
for stable component detection in non-rigid 3D shapes, which can be used for
geometric feature detection and description. A quantitative evaluation of our
method on the SHREC'10 feature detection benchmark shows its potential as a
source of high-quality features.
| [
{
"version": "v1",
"created": "Fri, 17 Dec 2010 18:23:35 GMT"
}
] | 2014-06-18T00:00:00 | [
[
"Litman",
"Roee",
""
],
[
"Bronstein",
"Alex M.",
""
],
[
"Bronstein",
"Michael M.",
""
]
] | TITLE: Diffusion-geometric maximally stable component detection in deformable
shapes
ABSTRACT: Maximally stable component detection is a very popular method for feature
analysis in images, mainly due to its low computation cost and high
repeatability. With the recent advance of feature-based methods in geometric
shape analysis, there is significant interest in finding analogous approaches
in the 3D world. In this paper, we formulate a diffusion-geometric framework
for stable component detection in non-rigid 3D shapes, which can be used for
geometric feature detection and description. A quantitative evaluation of our
method on the SHREC'10 feature detection benchmark shows its potential as a
source of high-quality features.
|
1103.5188 | Catuscia Palamidessi | M\'ario S. Alvim, Miguel E. Andr\'es, Konstantinos Chatzikokolakis,
Pierpaolo Degano, Catuscia Palamidessi | Differential Privacy: on the trade-off between Utility and Information
Leakage | 30 pages; HAL repository | Proceedings of the 8th International Workshop on Formal Aspects of
Security & Trust (FAST'11), Springer, LNCS 7140, pp. 39-54, 2011 | 10.1007/978-3-642-29420-4_3 | inria-00580122 | cs.CR cs.DB cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Differential privacy is a notion of privacy that has become very popular in
the database community. Roughly, the idea is that a randomized query mechanism
provides sufficient privacy protection if the ratio between the probabilities
that two adjacent datasets give the same answer is bound by e^epsilon. In the
field of information flow there is a similar concern for controlling
information leakage, i.e. limiting the possibility of inferring the secret
information from the observables. In recent years, researchers have proposed to
quantify the leakage in terms of R\'enyi min mutual information, a notion
strictly related to the Bayes risk. In this paper, we show how to model the
query system in terms of an information-theoretic channel, and we compare the
notion of differential privacy with that of mutual information. We show that
differential privacy implies a bound on the mutual information (but not
vice-versa). Furthermore, we show that our bound is tight. Then, we consider
the utility of the randomization mechanism, which represents how close the
randomized answers are, in average, to the real ones. We show that the notion
of differential privacy implies a bound on utility, also tight, and we propose
a method that under certain conditions builds an optimal randomization
mechanism, i.e. a mechanism which provides the best utility while guaranteeing
differential privacy.
| [
{
"version": "v1",
"created": "Sun, 27 Mar 2011 06:41:12 GMT"
},
{
"version": "v2",
"created": "Mon, 9 May 2011 00:04:26 GMT"
},
{
"version": "v3",
"created": "Thu, 25 Aug 2011 04:12:17 GMT"
}
] | 2014-06-18T00:00:00 | [
[
"Alvim",
"Mário S.",
""
],
[
"Andrés",
"Miguel E.",
""
],
[
"Chatzikokolakis",
"Konstantinos",
""
],
[
"Degano",
"Pierpaolo",
""
],
[
"Palamidessi",
"Catuscia",
""
]
] | TITLE: Differential Privacy: on the trade-off between Utility and Information
Leakage
ABSTRACT: Differential privacy is a notion of privacy that has become very popular in
the database community. Roughly, the idea is that a randomized query mechanism
provides sufficient privacy protection if the ratio between the probabilities
that two adjacent datasets give the same answer is bound by e^epsilon. In the
field of information flow there is a similar concern for controlling
information leakage, i.e. limiting the possibility of inferring the secret
information from the observables. In recent years, researchers have proposed to
quantify the leakage in terms of R\'enyi min mutual information, a notion
strictly related to the Bayes risk. In this paper, we show how to model the
query system in terms of an information-theoretic channel, and we compare the
notion of differential privacy with that of mutual information. We show that
differential privacy implies a bound on the mutual information (but not
vice-versa). Furthermore, we show that our bound is tight. Then, we consider
the utility of the randomization mechanism, which represents how close the
randomized answers are, in average, to the real ones. We show that the notion
of differential privacy implies a bound on utility, also tight, and we propose
a method that under certain conditions builds an optimal randomization
mechanism, i.e. a mechanism which provides the best utility while guaranteeing
differential privacy.
|
1307.0345 | Tobias Sutter | Peyman Mohajerin Esfahani, Tobias Sutter, John Lygeros | Performance Bounds for the Scenario Approach and an Extension to a Class
of Non-convex Programs | 19 pages, revised version | null | 10.1109/TAC.2014.2330702 | null | math.OC cs.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider the Scenario Convex Program (SCP) for two classes of optimization
problems that are not tractable in general: Robust Convex Programs (RCPs) and
Chance-Constrained Programs (CCPs). We establish a probabilistic bridge from
the optimal value of SCP to the optimal values of RCP and CCP in which the
uncertainty takes values in a general, possibly infinite dimensional, metric
space. We then extend our results to a certain class of non-convex problems
that includes, for example, binary decision variables. In the process, we also
settle a measurability issue for a general class of scenario programs, which to
date has been addressed by an assumption. Finally, we demonstrate the
applicability of our results on a benchmark problem and a problem in fault
detection and isolation.
| [
{
"version": "v1",
"created": "Mon, 1 Jul 2013 12:30:58 GMT"
},
{
"version": "v2",
"created": "Fri, 6 Dec 2013 11:27:53 GMT"
}
] | 2014-06-18T00:00:00 | [
[
"Esfahani",
"Peyman Mohajerin",
""
],
[
"Sutter",
"Tobias",
""
],
[
"Lygeros",
"John",
""
]
] | TITLE: Performance Bounds for the Scenario Approach and an Extension to a Class
of Non-convex Programs
ABSTRACT: We consider the Scenario Convex Program (SCP) for two classes of optimization
problems that are not tractable in general: Robust Convex Programs (RCPs) and
Chance-Constrained Programs (CCPs). We establish a probabilistic bridge from
the optimal value of SCP to the optimal values of RCP and CCP in which the
uncertainty takes values in a general, possibly infinite dimensional, metric
space. We then extend our results to a certain class of non-convex problems
that includes, for example, binary decision variables. In the process, we also
settle a measurability issue for a general class of scenario programs, which to
date has been addressed by an assumption. Finally, we demonstrate the
applicability of our results on a benchmark problem and a problem in fault
detection and isolation.
|
1405.5047 | Michael Burke Mr | Michael Burke and Joan Lasenby | Single camera pose estimation using Bayesian filtering and Kinect motion
priors | 25 pages, Technical report, related to Burke and Lasenby, AMDO 2014
conference paper. Code sample: https://github.com/mgb45/SignerBodyPose Video:
https://www.youtube.com/watch?v=dJMTSo7-uFE | null | null | null | cs.CV cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Traditional approaches to upper body pose estimation using monocular vision
rely on complex body models and a large variety of geometric constraints. We
argue that this is not ideal and somewhat inelegant as it results in large
processing burdens, and instead attempt to incorporate these constraints
through priors obtained directly from training data. A prior distribution
covering the probability of a human pose occurring is used to incorporate
likely human poses. This distribution is obtained offline, by fitting a
Gaussian mixture model to a large dataset of recorded human body poses, tracked
using a Kinect sensor. We combine this prior information with a random walk
transition model to obtain an upper body model, suitable for use within a
recursive Bayesian filtering framework. Our model can be viewed as a mixture of
discrete Ornstein-Uhlenbeck processes, in that states behave as random walks,
but drift towards a set of typically observed poses. This model is combined
with measurements of the human head and hand positions, using recursive
Bayesian estimation to incorporate temporal information. Measurements are
obtained using face detection and a simple skin colour hand detector, trained
using the detected face. The suggested model is designed with analytical
tractability in mind and we show that the pose tracking can be
Rao-Blackwellised using the mixture Kalman filter, allowing for computational
efficiency while still incorporating bio-mechanical properties of the upper
body. In addition, the use of the proposed upper body model allows reliable
three-dimensional pose estimates to be obtained indirectly for a number of
joints that are often difficult to detect using traditional object recognition
strategies. Comparisons with Kinect sensor results and the state of the art in
2D pose estimation highlight the efficacy of the proposed approach.
| [
{
"version": "v1",
"created": "Tue, 20 May 2014 11:54:04 GMT"
},
{
"version": "v2",
"created": "Tue, 17 Jun 2014 12:15:42 GMT"
}
] | 2014-06-18T00:00:00 | [
[
"Burke",
"Michael",
""
],
[
"Lasenby",
"Joan",
""
]
] | TITLE: Single camera pose estimation using Bayesian filtering and Kinect motion
priors
ABSTRACT: Traditional approaches to upper body pose estimation using monocular vision
rely on complex body models and a large variety of geometric constraints. We
argue that this is not ideal and somewhat inelegant as it results in large
processing burdens, and instead attempt to incorporate these constraints
through priors obtained directly from training data. A prior distribution
covering the probability of a human pose occurring is used to incorporate
likely human poses. This distribution is obtained offline, by fitting a
Gaussian mixture model to a large dataset of recorded human body poses, tracked
using a Kinect sensor. We combine this prior information with a random walk
transition model to obtain an upper body model, suitable for use within a
recursive Bayesian filtering framework. Our model can be viewed as a mixture of
discrete Ornstein-Uhlenbeck processes, in that states behave as random walks,
but drift towards a set of typically observed poses. This model is combined
with measurements of the human head and hand positions, using recursive
Bayesian estimation to incorporate temporal information. Measurements are
obtained using face detection and a simple skin colour hand detector, trained
using the detected face. The suggested model is designed with analytical
tractability in mind and we show that the pose tracking can be
Rao-Blackwellised using the mixture Kalman filter, allowing for computational
efficiency while still incorporating bio-mechanical properties of the upper
body. In addition, the use of the proposed upper body model allows reliable
three-dimensional pose estimates to be obtained indirectly for a number of
joints that are often difficult to detect using traditional object recognition
strategies. Comparisons with Kinect sensor results and the state of the art in
2D pose estimation highlight the efficacy of the proposed approach.
|
1406.4462 | Erfan Khaji Mr. | Erfan Khaji | Soccer League Optimization: A heuristic Algorithm Inspired by the
Football System in European Countries | 6 Pages, 12 Figures, 4 Tables, Accepted in GEM 2014, but rejected due
to lack of money | null | null | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper a new heuristic optimization algorithm has been introduced
based on the performance of the major football leagues within each season in EU
countries. The algorithm starts with an initial population including three
different groups of teams: the wealthiest (strongest), the regular, the poorest
(weakest). Each individual of population constitute a football team while each
player is an indication of a player in a post. The optimization can hopefully
occurs when the competition among the teams in all the leagues is imitated as
the strongest teams usually purchase the best players of the regular teams and
in turn, regular teams purchase the best players of the weakest who should
always discover young players instead of buying professionals. It has been
shown that the algorithm can hopefully converge to an acceptable solution
solving various benchmarks. Key words: Heuristic Algorithms
| [
{
"version": "v1",
"created": "Sun, 15 Jun 2014 15:10:20 GMT"
}
] | 2014-06-18T00:00:00 | [
[
"Khaji",
"Erfan",
""
]
] | TITLE: Soccer League Optimization: A heuristic Algorithm Inspired by the
Football System in European Countries
ABSTRACT: In this paper a new heuristic optimization algorithm has been introduced
based on the performance of the major football leagues within each season in EU
countries. The algorithm starts with an initial population including three
different groups of teams: the wealthiest (strongest), the regular, the poorest
(weakest). Each individual of population constitute a football team while each
player is an indication of a player in a post. The optimization can hopefully
occurs when the competition among the teams in all the leagues is imitated as
the strongest teams usually purchase the best players of the regular teams and
in turn, regular teams purchase the best players of the weakest who should
always discover young players instead of buying professionals. It has been
shown that the algorithm can hopefully converge to an acceptable solution
solving various benchmarks. Key words: Heuristic Algorithms
|
1209.0738 | Bernardino Romera Paredes | Andreas Maurer, Massimiliano Pontil, Bernardino Romera-Paredes | Sparse coding for multitask and transfer learning | International Conference on Machine Learning 2013 | null | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We investigate the use of sparse coding and dictionary learning in the
context of multitask and transfer learning. The central assumption of our
learning method is that the tasks parameters are well approximated by sparse
linear combinations of the atoms of a dictionary on a high or infinite
dimensional space. This assumption, together with the large quantity of
available data in the multitask and transfer learning settings, allows a
principled choice of the dictionary. We provide bounds on the generalization
error of this approach, for both settings. Numerical experiments on one
synthetic and two real datasets show the advantage of our method over single
task learning, a previous method based on orthogonal and dense representation
of the tasks and a related method learning task grouping.
| [
{
"version": "v1",
"created": "Tue, 4 Sep 2012 19:06:51 GMT"
},
{
"version": "v2",
"created": "Sat, 23 Mar 2013 19:35:27 GMT"
},
{
"version": "v3",
"created": "Mon, 16 Jun 2014 15:06:48 GMT"
}
] | 2014-06-17T00:00:00 | [
[
"Maurer",
"Andreas",
""
],
[
"Pontil",
"Massimiliano",
""
],
[
"Romera-Paredes",
"Bernardino",
""
]
] | TITLE: Sparse coding for multitask and transfer learning
ABSTRACT: We investigate the use of sparse coding and dictionary learning in the
context of multitask and transfer learning. The central assumption of our
learning method is that the tasks parameters are well approximated by sparse
linear combinations of the atoms of a dictionary on a high or infinite
dimensional space. This assumption, together with the large quantity of
available data in the multitask and transfer learning settings, allows a
principled choice of the dictionary. We provide bounds on the generalization
error of this approach, for both settings. Numerical experiments on one
synthetic and two real datasets show the advantage of our method over single
task learning, a previous method based on orthogonal and dense representation
of the tasks and a related method learning task grouping.
|
1307.6239 | David Van Horn | Phuc C. Nguyen, Sam Tobin-Hochstadt, David Van Horn | Soft Contract Verification | ICFP '14, September 1-6, 2014, Gothenburg, Sweden | null | 10.1145/2628136.2628156 | null | cs.PL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Behavioral software contracts are a widely used mechanism for governing the
flow of values between components. However, run-time monitoring and enforcement
of contracts imposes significant overhead and delays discovery of faulty
components to run-time.
To overcome these issues, we present soft contract verification, which aims
to statically prove either complete or partial contract correctness of
components, written in an untyped, higher-order language with first-class
contracts. Our approach uses higher-order symbolic execution, leveraging
contracts as a source of symbolic values including unknown behavioral values,
and employs an updatable heap of contract invariants to reason about
flow-sensitive facts. We prove the symbolic execution soundly approximates the
dynamic semantics and that verified programs can't be blamed.
The approach is able to analyze first-class contracts, recursive data
structures, unknown functions, and control-flow-sensitive refinements of
values, which are all idiomatic in dynamic languages. It makes effective use of
an off-the-shelf solver to decide problems without heavy encodings. The
approach is competitive with a wide range of existing tools---including type
systems, flow analyzers, and model checkers---on their own benchmarks.
| [
{
"version": "v1",
"created": "Tue, 23 Jul 2013 20:30:46 GMT"
},
{
"version": "v2",
"created": "Sun, 17 Nov 2013 19:55:36 GMT"
},
{
"version": "v3",
"created": "Mon, 3 Mar 2014 21:40:18 GMT"
},
{
"version": "v4",
"created": "Mon, 16 Jun 2014 18:57:38 GMT"
}
] | 2014-06-17T00:00:00 | [
[
"Nguyen",
"Phuc C.",
""
],
[
"Tobin-Hochstadt",
"Sam",
""
],
[
"Van Horn",
"David",
""
]
] | TITLE: Soft Contract Verification
ABSTRACT: Behavioral software contracts are a widely used mechanism for governing the
flow of values between components. However, run-time monitoring and enforcement
of contracts imposes significant overhead and delays discovery of faulty
components to run-time.
To overcome these issues, we present soft contract verification, which aims
to statically prove either complete or partial contract correctness of
components, written in an untyped, higher-order language with first-class
contracts. Our approach uses higher-order symbolic execution, leveraging
contracts as a source of symbolic values including unknown behavioral values,
and employs an updatable heap of contract invariants to reason about
flow-sensitive facts. We prove the symbolic execution soundly approximates the
dynamic semantics and that verified programs can't be blamed.
The approach is able to analyze first-class contracts, recursive data
structures, unknown functions, and control-flow-sensitive refinements of
values, which are all idiomatic in dynamic languages. It makes effective use of
an off-the-shelf solver to decide problems without heavy encodings. The
approach is competitive with a wide range of existing tools---including type
systems, flow analyzers, and model checkers---on their own benchmarks.
|
1405.1131 | Ali Bou Nassif | Ali Bou Nassif, Luiz Fernando Capretz, Danny Ho | Analyzing the Non-Functional Requirements in the Desharnais Dataset for
Software Effort Estimation | 6 pages | null | null | null | cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Studying the quality requirements (aka Non-Functional Requirements (NFR)) of
a system is crucial in Requirements Engineering. Many software projects fail
because of neglecting or failing to incorporate the NFR during the software
life development cycle. This paper focuses on analyzing the importance of the
quality requirements attributes in software effort estimation models based on
the Desharnais dataset. The Desharnais dataset is a collection of eighty one
software projects of twelve attributes developed by a Canadian software house.
The analysis includes studying the influence of each of the quality
requirements attributes, as well as the influence of all quality requirements
attributes combined when calculating software effort using regression and
Artificial Neural Network (ANN) models. The evaluation criteria used in this
investigation include the Mean of the Magnitude of Relative Error (MMRE), the
Prediction Level (PRED), Root Mean Squared Error (RMSE), Mean Error and the
Coefficient of determination (R2). Results show that the quality attribute
Language is the most statistically significant when calculating software
effort. Moreover, if all quality requirements attributes are eliminated in the
training stage and software effort is predicted based on software size only,
the value of the error (MMRE) is doubled.
| [
{
"version": "v1",
"created": "Tue, 6 May 2014 02:32:41 GMT"
},
{
"version": "v2",
"created": "Sat, 14 Jun 2014 03:19:18 GMT"
}
] | 2014-06-17T00:00:00 | [
[
"Nassif",
"Ali Bou",
""
],
[
"Capretz",
"Luiz Fernando",
""
],
[
"Ho",
"Danny",
""
]
] | TITLE: Analyzing the Non-Functional Requirements in the Desharnais Dataset for
Software Effort Estimation
ABSTRACT: Studying the quality requirements (aka Non-Functional Requirements (NFR)) of
a system is crucial in Requirements Engineering. Many software projects fail
because of neglecting or failing to incorporate the NFR during the software
life development cycle. This paper focuses on analyzing the importance of the
quality requirements attributes in software effort estimation models based on
the Desharnais dataset. The Desharnais dataset is a collection of eighty one
software projects of twelve attributes developed by a Canadian software house.
The analysis includes studying the influence of each of the quality
requirements attributes, as well as the influence of all quality requirements
attributes combined when calculating software effort using regression and
Artificial Neural Network (ANN) models. The evaluation criteria used in this
investigation include the Mean of the Magnitude of Relative Error (MMRE), the
Prediction Level (PRED), Root Mean Squared Error (RMSE), Mean Error and the
Coefficient of determination (R2). Results show that the quality attribute
Language is the most statistically significant when calculating software
effort. Moreover, if all quality requirements attributes are eliminated in the
training stage and software effort is predicted based on software size only,
the value of the error (MMRE) is doubled.
|
1405.7452 | Tim Althoff | Tim Althoff, Damian Borth, J\"orn Hees, Andreas Dengel | Analysis and Forecasting of Trending Topics in Online Media Streams | ACM Multimedia 2013 | null | null | null | cs.SI cs.MM physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Among the vast information available on the web, social media streams capture
what people currently pay attention to and how they feel about certain topics.
Awareness of such trending topics plays a crucial role in multimedia systems
such as trend aware recommendation and automatic vocabulary selection for video
concept detection systems.
Correctly utilizing trending topics requires a better understanding of their
various characteristics in different social media streams. To this end, we
present the first comprehensive study across three major online and social
media streams, Twitter, Google, and Wikipedia, covering thousands of trending
topics during an observation period of an entire year. Our results indicate
that depending on one's requirements one does not necessarily have to turn to
Twitter for information about current events and that some media streams
strongly emphasize content of specific categories. As our second key
contribution, we further present a novel approach for the challenging task of
forecasting the life cycle of trending topics in the very moment they emerge.
Our fully automated approach is based on a nearest neighbor forecasting
technique exploiting our assumption that semantically similar topics exhibit
similar behavior.
We demonstrate on a large-scale dataset of Wikipedia page view statistics
that forecasts by the proposed approach are about 9-48k views closer to the
actual viewing statistics compared to baseline methods and achieve a mean
average percentage error of 45-19% for time periods of up to 14 days.
| [
{
"version": "v1",
"created": "Thu, 29 May 2014 03:43:41 GMT"
},
{
"version": "v2",
"created": "Sat, 14 Jun 2014 20:14:07 GMT"
}
] | 2014-06-17T00:00:00 | [
[
"Althoff",
"Tim",
""
],
[
"Borth",
"Damian",
""
],
[
"Hees",
"Jörn",
""
],
[
"Dengel",
"Andreas",
""
]
] | TITLE: Analysis and Forecasting of Trending Topics in Online Media Streams
ABSTRACT: Among the vast information available on the web, social media streams capture
what people currently pay attention to and how they feel about certain topics.
Awareness of such trending topics plays a crucial role in multimedia systems
such as trend aware recommendation and automatic vocabulary selection for video
concept detection systems.
Correctly utilizing trending topics requires a better understanding of their
various characteristics in different social media streams. To this end, we
present the first comprehensive study across three major online and social
media streams, Twitter, Google, and Wikipedia, covering thousands of trending
topics during an observation period of an entire year. Our results indicate
that depending on one's requirements one does not necessarily have to turn to
Twitter for information about current events and that some media streams
strongly emphasize content of specific categories. As our second key
contribution, we further present a novel approach for the challenging task of
forecasting the life cycle of trending topics in the very moment they emerge.
Our fully automated approach is based on a nearest neighbor forecasting
technique exploiting our assumption that semantically similar topics exhibit
similar behavior.
We demonstrate on a large-scale dataset of Wikipedia page view statistics
that forecasts by the proposed approach are about 9-48k views closer to the
actual viewing statistics compared to baseline methods and achieve a mean
average percentage error of 45-19% for time periods of up to 14 days.
|
1406.1774 | Toufiq Parag | Toufiq Parag, Stephen Plaza, Louis Scheffer (Janelia Farm Research
Campus- HHMI) | Small Sample Learning of Superpixel Classifiers for EM Segmentation-
Extended Version | Accepted for MICCAI 2014 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Pixel and superpixel classifiers have become essential tools for EM
segmentation algorithms. Training these classifiers remains a major bottleneck
primarily due to the requirement of completely annotating the dataset which is
tedious, error-prone and costly. In this paper, we propose an interactive
learning scheme for the superpixel classifier for EM segmentation. Our
algorithm is "active semi-supervised" because it requests the labels of a small
number of examples from user and applies label propagation technique to
generate these queries. Using only a small set ($<20\%$) of all datapoints, the
proposed algorithm consistently generates a classifier almost as accurate as
that estimated from a complete groundtruth. We provide segmentation results on
multiple datasets to show the strength of these classifiers.
| [
{
"version": "v1",
"created": "Fri, 6 Jun 2014 18:59:58 GMT"
},
{
"version": "v2",
"created": "Fri, 13 Jun 2014 22:05:57 GMT"
}
] | 2014-06-17T00:00:00 | [
[
"Parag",
"Toufiq",
"",
"Janelia Farm Research\n Campus- HHMI"
],
[
"Plaza",
"Stephen",
"",
"Janelia Farm Research\n Campus- HHMI"
],
[
"Scheffer",
"Louis",
"",
"Janelia Farm Research\n Campus- HHMI"
]
] | TITLE: Small Sample Learning of Superpixel Classifiers for EM Segmentation-
Extended Version
ABSTRACT: Pixel and superpixel classifiers have become essential tools for EM
segmentation algorithms. Training these classifiers remains a major bottleneck
primarily due to the requirement of completely annotating the dataset which is
tedious, error-prone and costly. In this paper, we propose an interactive
learning scheme for the superpixel classifier for EM segmentation. Our
algorithm is "active semi-supervised" because it requests the labels of a small
number of examples from user and applies label propagation technique to
generate these queries. Using only a small set ($<20\%$) of all datapoints, the
proposed algorithm consistently generates a classifier almost as accurate as
that estimated from a complete groundtruth. We provide segmentation results on
multiple datasets to show the strength of these classifiers.
|
1406.3682 | Srishti Gupta | Srishti Gupta, Ponnurangam Kumaraguru | Emerging Phishing Trends and Effectiveness of the Anti-Phishing Landing
Page | null | null | null | null | cs.CY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Each month, more attacks are launched with the aim of making web users
believe that they are communicating with a trusted entity which compels them to
share their personal, financial information. Phishing costs Internet users
billions of dollars every year. Researchers at Carnegie Mellon University (CMU)
created an anti-phishing landing page supported by Anti-Phishing Working Group
(APWG) with the aim to train users on how to prevent themselves from phishing
attacks. It is used by financial institutions, phish site take down vendors,
government organizations, and online merchants. When a potential victim clicks
on a phishing link that has been taken down, he / she is redirected to the
landing page. In this paper, we present the comparative analysis on two
datasets that we obtained from APWG's landing page log files; one, from
September 7, 2008 - November 11, 2009, and other from January 1, 2014 - April
30, 2014. We found that the landing page has been successful in training users
against phishing. Forty six percent users clicked lesser number of phishing
URLs from January 2014 to April 2014 which shows that training from the landing
page helped users not to fall for phishing attacks. Our analysis shows that
phishers have started to modify their techniques by creating more legitimate
looking URLs and buying large number of domains to increase their activity. We
observed that phishers are exploiting ICANN accredited registrars to launch
their attacks even after strict surveillance. We saw that phishers are trying
to exploit free subdomain registration services to carry out attacks. In this
paper, we also compared the phishing e-mails used by phishers to lure victims
in 2008 and 2014. We found that the phishing e-mails have changed considerably
over time. Phishers have adopted new techniques like sending promotional
e-mails and emotionally targeting users in clicking phishing URLs.
| [
{
"version": "v1",
"created": "Sat, 14 Jun 2014 04:19:16 GMT"
}
] | 2014-06-17T00:00:00 | [
[
"Gupta",
"Srishti",
""
],
[
"Kumaraguru",
"Ponnurangam",
""
]
] | TITLE: Emerging Phishing Trends and Effectiveness of the Anti-Phishing Landing
Page
ABSTRACT: Each month, more attacks are launched with the aim of making web users
believe that they are communicating with a trusted entity which compels them to
share their personal, financial information. Phishing costs Internet users
billions of dollars every year. Researchers at Carnegie Mellon University (CMU)
created an anti-phishing landing page supported by Anti-Phishing Working Group
(APWG) with the aim to train users on how to prevent themselves from phishing
attacks. It is used by financial institutions, phish site take down vendors,
government organizations, and online merchants. When a potential victim clicks
on a phishing link that has been taken down, he / she is redirected to the
landing page. In this paper, we present the comparative analysis on two
datasets that we obtained from APWG's landing page log files; one, from
September 7, 2008 - November 11, 2009, and other from January 1, 2014 - April
30, 2014. We found that the landing page has been successful in training users
against phishing. Forty six percent users clicked lesser number of phishing
URLs from January 2014 to April 2014 which shows that training from the landing
page helped users not to fall for phishing attacks. Our analysis shows that
phishers have started to modify their techniques by creating more legitimate
looking URLs and buying large number of domains to increase their activity. We
observed that phishers are exploiting ICANN accredited registrars to launch
their attacks even after strict surveillance. We saw that phishers are trying
to exploit free subdomain registration services to carry out attacks. In this
paper, we also compared the phishing e-mails used by phishers to lure victims
in 2008 and 2014. We found that the phishing e-mails have changed considerably
over time. Phishers have adopted new techniques like sending promotional
e-mails and emotionally targeting users in clicking phishing URLs.
|
1406.3687 | Neha Gupta | Neha Gupta, Anupama Aggarwal, Ponnurangam Kumaraguru | bit.ly/malicious: Deep Dive into Short URL based e-Crime Detection | arXiv admin note: substantial text overlap with arXiv:1405.1511 | null | null | null | cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Existence of spam URLs over emails and Online Social Media (OSM) has become a
massive e-crime. To counter the dissemination of long complex URLs in emails
and character limit imposed on various OSM (like Twitter), the concept of URL
shortening has gained a lot of traction. URL shorteners take as input a long
URL and output a short URL with the same landing page (as in the long URL) in
return. With their immense popularity over time, URL shorteners have become a
prime target for the attackers giving them an advantage to conceal malicious
content. Bitly, a leading service among all shortening services is being
exploited heavily to carry out phishing attacks, work-from-home scams,
pornographic content propagation, etc. This imposes additional performance
pressure on Bitly and other URL shorteners to be able to detect and take a
timely action against the illegitimate content. In this study, we analyzed a
dataset of 763,160 short URLs marked suspicious by Bitly in the month of
October 2013. Our results reveal that Bitly is not using its claimed spam
detection services very effectively. We also show how a suspicious Bitly
account goes unnoticed despite of a prolonged recurrent illegitimate activity.
Bitly displays a warning page on identification of suspicious links, but we
observed this approach to be weak in controlling the overall propagation of
spam. We also identified some short URL based features and coupled them with
two domain specific features to classify a Bitly URL as malicious or benign and
achieved an accuracy of 86.41%. The feature set identified can be generalized
to other URL shortening services as well. To the best of our knowledge, this is
the first large scale study to highlight the issues with the implementation of
Bitly's spam detection policies and proposing suitable countermeasures.
| [
{
"version": "v1",
"created": "Sat, 14 Jun 2014 06:22:16 GMT"
}
] | 2014-06-17T00:00:00 | [
[
"Gupta",
"Neha",
""
],
[
"Aggarwal",
"Anupama",
""
],
[
"Kumaraguru",
"Ponnurangam",
""
]
] | TITLE: bit.ly/malicious: Deep Dive into Short URL based e-Crime Detection
ABSTRACT: Existence of spam URLs over emails and Online Social Media (OSM) has become a
massive e-crime. To counter the dissemination of long complex URLs in emails
and character limit imposed on various OSM (like Twitter), the concept of URL
shortening has gained a lot of traction. URL shorteners take as input a long
URL and output a short URL with the same landing page (as in the long URL) in
return. With their immense popularity over time, URL shorteners have become a
prime target for the attackers giving them an advantage to conceal malicious
content. Bitly, a leading service among all shortening services is being
exploited heavily to carry out phishing attacks, work-from-home scams,
pornographic content propagation, etc. This imposes additional performance
pressure on Bitly and other URL shorteners to be able to detect and take a
timely action against the illegitimate content. In this study, we analyzed a
dataset of 763,160 short URLs marked suspicious by Bitly in the month of
October 2013. Our results reveal that Bitly is not using its claimed spam
detection services very effectively. We also show how a suspicious Bitly
account goes unnoticed despite of a prolonged recurrent illegitimate activity.
Bitly displays a warning page on identification of suspicious links, but we
observed this approach to be weak in controlling the overall propagation of
spam. We also identified some short URL based features and coupled them with
two domain specific features to classify a Bitly URL as malicious or benign and
achieved an accuracy of 86.41%. The feature set identified can be generalized
to other URL shortening services as well. To the best of our knowledge, this is
the first large scale study to highlight the issues with the implementation of
Bitly's spam detection policies and proposing suitable countermeasures.
|
1406.3692 | Prateek Dewan | Prateek Dewan and Anand Kashyap and Ponnurangam Kumaraguru | Analyzing Social and Stylometric Features to Identify Spear phishing
Emails | Detection of spear phishing using social media features | null | null | null | cs.CY cs.LG cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Spear phishing is a complex targeted attack in which, an attacker harvests
information about the victim prior to the attack. This information is then used
to create sophisticated, genuine-looking attack vectors, drawing the victim to
compromise confidential information. What makes spear phishing different, and
more powerful than normal phishing, is this contextual information about the
victim. Online social media services can be one such source for gathering vital
information about an individual. In this paper, we characterize and examine a
true positive dataset of spear phishing, spam, and normal phishing emails from
Symantec's enterprise email scanning service. We then present a model to detect
spear phishing emails sent to employees of 14 international organizations, by
using social features extracted from LinkedIn. Our dataset consists of 4,742
targeted attack emails sent to 2,434 victims, and 9,353 non targeted attack
emails sent to 5,912 non victims; and publicly available information from their
LinkedIn profiles. We applied various machine learning algorithms to this
labeled data, and achieved an overall maximum accuracy of 97.76% in identifying
spear phishing emails. We used a combination of social features from LinkedIn
profiles, and stylometric features extracted from email subjects, bodies, and
attachments. However, we achieved a slightly better accuracy of 98.28% without
the social features. Our analysis revealed that social features extracted from
LinkedIn do not help in identifying spear phishing emails. To the best of our
knowledge, this is one of the first attempts to make use of a combination of
stylometric features extracted from emails, and social features extracted from
an online social network to detect targeted spear phishing emails.
| [
{
"version": "v1",
"created": "Sat, 14 Jun 2014 07:01:03 GMT"
}
] | 2014-06-17T00:00:00 | [
[
"Dewan",
"Prateek",
""
],
[
"Kashyap",
"Anand",
""
],
[
"Kumaraguru",
"Ponnurangam",
""
]
] | TITLE: Analyzing Social and Stylometric Features to Identify Spear phishing
Emails
ABSTRACT: Spear phishing is a complex targeted attack in which, an attacker harvests
information about the victim prior to the attack. This information is then used
to create sophisticated, genuine-looking attack vectors, drawing the victim to
compromise confidential information. What makes spear phishing different, and
more powerful than normal phishing, is this contextual information about the
victim. Online social media services can be one such source for gathering vital
information about an individual. In this paper, we characterize and examine a
true positive dataset of spear phishing, spam, and normal phishing emails from
Symantec's enterprise email scanning service. We then present a model to detect
spear phishing emails sent to employees of 14 international organizations, by
using social features extracted from LinkedIn. Our dataset consists of 4,742
targeted attack emails sent to 2,434 victims, and 9,353 non targeted attack
emails sent to 5,912 non victims; and publicly available information from their
LinkedIn profiles. We applied various machine learning algorithms to this
labeled data, and achieved an overall maximum accuracy of 97.76% in identifying
spear phishing emails. We used a combination of social features from LinkedIn
profiles, and stylometric features extracted from email subjects, bodies, and
attachments. However, we achieved a slightly better accuracy of 98.28% without
the social features. Our analysis revealed that social features extracted from
LinkedIn do not help in identifying spear phishing emails. To the best of our
knowledge, this is one of the first attempts to make use of a combination of
stylometric features extracted from emails, and social features extracted from
an online social network to detect targeted spear phishing emails.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.