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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1307.3040 | Mehul Bhatt | Mehul Bhatt | Between Sense and Sensibility: Declarative narrativisation of mental
models as a basis and benchmark for visuo-spatial cognition and computation
focussed collaborative cognitive systems | 5 pages, research statement summarising recent publications | null | null | null | cs.AI cs.CL cs.CV cs.HC cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | What lies between `\emph{sensing}' and `\emph{sensibility}'? In other words,
what kind of cognitive processes mediate sensing capability, and the formation
of sensible impressions ---e.g., abstractions, analogies, hypotheses and theory
formation, beliefs and their revision, argument formation--- in domain-specific
problem solving, or in regular activities of everyday living, working and
simply going around in the environment? How can knowledge and reasoning about
such capabilities, as exhibited by humans in particular problem contexts, be
used as a model and benchmark for the development of collaborative cognitive
(interaction) systems concerned with human assistance, assurance, and
empowerment?
We pose these questions in the context of a range of assistive technologies
concerned with \emph{visuo-spatial perception and cognition} tasks encompassing
aspects such as commonsense, creativity, and the application of specialist
domain knowledge and problem-solving thought processes. Assistive technologies
being considered include: (a) human activity interpretation; (b) high-level
cognitive rovotics; (c) people-centred creative design in domains such as
architecture & digital media creation, and (d) qualitative analyses geographic
information systems. Computational narratives not only provide a rich cognitive
basis, but they also serve as a benchmark of functional performance in our
development of computational cognitive assistance systems. We posit that
computational narrativisation pertaining to space, actions, and change provides
a useful model of \emph{visual} and \emph{spatio-temporal thinking} within a
wide-range of problem-solving tasks and application areas where collaborative
cognitive systems could serve an assistive and empowering function.
| [
{
"version": "v1",
"created": "Thu, 11 Jul 2013 10:01:29 GMT"
},
{
"version": "v2",
"created": "Mon, 31 Mar 2014 10:22:29 GMT"
}
] | 2014-04-01T00:00:00 | [
[
"Bhatt",
"Mehul",
""
]
] | TITLE: Between Sense and Sensibility: Declarative narrativisation of mental
models as a basis and benchmark for visuo-spatial cognition and computation
focussed collaborative cognitive systems
ABSTRACT: What lies between `\emph{sensing}' and `\emph{sensibility}'? In other words,
what kind of cognitive processes mediate sensing capability, and the formation
of sensible impressions ---e.g., abstractions, analogies, hypotheses and theory
formation, beliefs and their revision, argument formation--- in domain-specific
problem solving, or in regular activities of everyday living, working and
simply going around in the environment? How can knowledge and reasoning about
such capabilities, as exhibited by humans in particular problem contexts, be
used as a model and benchmark for the development of collaborative cognitive
(interaction) systems concerned with human assistance, assurance, and
empowerment?
We pose these questions in the context of a range of assistive technologies
concerned with \emph{visuo-spatial perception and cognition} tasks encompassing
aspects such as commonsense, creativity, and the application of specialist
domain knowledge and problem-solving thought processes. Assistive technologies
being considered include: (a) human activity interpretation; (b) high-level
cognitive rovotics; (c) people-centred creative design in domains such as
architecture & digital media creation, and (d) qualitative analyses geographic
information systems. Computational narratives not only provide a rich cognitive
basis, but they also serve as a benchmark of functional performance in our
development of computational cognitive assistance systems. We posit that
computational narrativisation pertaining to space, actions, and change provides
a useful model of \emph{visual} and \emph{spatio-temporal thinking} within a
wide-range of problem-solving tasks and application areas where collaborative
cognitive systems could serve an assistive and empowering function.
|
1307.7281 | Roopsha Samanta | Roopsha Samanta, Oswaldo Olivo and E. Allen Emerson | Cost-Aware Automatic Program Repair | null | null | null | null | cs.PL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a formal framework for repairing infinite-state, imperative,
sequential programs, with (possibly recursive) procedures and multiple
assertions; the framework can generate repaired programs by modifying the
original erroneous program in multiple program locations, and can ensure the
readability of the repaired program using user-defined expression templates;
the framework also generates a set of inductive assertions that serve as a
proof of correctness of the repaired program. As a step toward integrating
programmer intent and intuition in automated program repair, we present a "
cost-aware" formulation - given a cost function associated with permissible
statement modifications, the goal is to ensure that the total program
modification cost does not exceed a given repair budget. As part of our
predicate abstraction-based solution framework, we present a sound and complete
algorithm for repair of Boolean programs. We have developed a prototype tool
based on SMT solving and used it successfully to repair diverse errors in
benchmark C programs.
| [
{
"version": "v1",
"created": "Sat, 27 Jul 2013 16:57:03 GMT"
},
{
"version": "v2",
"created": "Mon, 31 Mar 2014 12:19:40 GMT"
}
] | 2014-04-01T00:00:00 | [
[
"Samanta",
"Roopsha",
""
],
[
"Olivo",
"Oswaldo",
""
],
[
"Emerson",
"E. Allen",
""
]
] | TITLE: Cost-Aware Automatic Program Repair
ABSTRACT: We present a formal framework for repairing infinite-state, imperative,
sequential programs, with (possibly recursive) procedures and multiple
assertions; the framework can generate repaired programs by modifying the
original erroneous program in multiple program locations, and can ensure the
readability of the repaired program using user-defined expression templates;
the framework also generates a set of inductive assertions that serve as a
proof of correctness of the repaired program. As a step toward integrating
programmer intent and intuition in automated program repair, we present a "
cost-aware" formulation - given a cost function associated with permissible
statement modifications, the goal is to ensure that the total program
modification cost does not exceed a given repair budget. As part of our
predicate abstraction-based solution framework, we present a sound and complete
algorithm for repair of Boolean programs. We have developed a prototype tool
based on SMT solving and used it successfully to repair diverse errors in
benchmark C programs.
|
1308.3892 | D\'aniel Kondor Mr | D\'aniel Kondor, M\'arton P\'osfai, Istv\'an Csabai and G\'abor Vattay | Do the rich get richer? An empirical analysis of the BitCoin transaction
network | Project website: http://www.vo.elte.hu/bitcoin/; updated after
publication | PLoS ONE 9(2): e86197, 2014 | 10.1371/journal.pone.0086197 | null | physics.soc-ph cs.SI q-fin.GN | http://creativecommons.org/licenses/by/3.0/ | The possibility to analyze everyday monetary transactions is limited by the
scarcity of available data, as this kind of information is usually considered
highly sensitive. Present econophysics models are usually employed on presumed
random networks of interacting agents, and only macroscopic properties (e.g.
the resulting wealth distribution) are compared to real-world data. In this
paper, we analyze BitCoin, which is a novel digital currency system, where the
complete list of transactions is publicly available. Using this dataset, we
reconstruct the network of transactions, and extract the time and amount of
each payment. We analyze the structure of the transaction network by measuring
network characteristics over time, such as the degree distribution, degree
correlations and clustering. We find that linear preferential attachment drives
the growth of the network. We also study the dynamics taking place on the
transaction network, i.e. the flow of money. We measure temporal patterns and
the wealth accumulation. Investigating the microscopic statistics of money
movement, we find that sublinear preferential attachment governs the evolution
of the wealth distribution. We report a scaling relation between the degree and
wealth associated to individual nodes.
| [
{
"version": "v1",
"created": "Sun, 18 Aug 2013 20:02:34 GMT"
},
{
"version": "v2",
"created": "Tue, 11 Feb 2014 10:17:56 GMT"
},
{
"version": "v3",
"created": "Mon, 31 Mar 2014 11:26:54 GMT"
}
] | 2014-04-01T00:00:00 | [
[
"Kondor",
"Dániel",
""
],
[
"Pósfai",
"Márton",
""
],
[
"Csabai",
"István",
""
],
[
"Vattay",
"Gábor",
""
]
] | TITLE: Do the rich get richer? An empirical analysis of the BitCoin transaction
network
ABSTRACT: The possibility to analyze everyday monetary transactions is limited by the
scarcity of available data, as this kind of information is usually considered
highly sensitive. Present econophysics models are usually employed on presumed
random networks of interacting agents, and only macroscopic properties (e.g.
the resulting wealth distribution) are compared to real-world data. In this
paper, we analyze BitCoin, which is a novel digital currency system, where the
complete list of transactions is publicly available. Using this dataset, we
reconstruct the network of transactions, and extract the time and amount of
each payment. We analyze the structure of the transaction network by measuring
network characteristics over time, such as the degree distribution, degree
correlations and clustering. We find that linear preferential attachment drives
the growth of the network. We also study the dynamics taking place on the
transaction network, i.e. the flow of money. We measure temporal patterns and
the wealth accumulation. Investigating the microscopic statistics of money
movement, we find that sublinear preferential attachment governs the evolution
of the wealth distribution. We report a scaling relation between the degree and
wealth associated to individual nodes.
|
1403.7654 | Anastasios Noulas Anastasios Noulas | Petko Georgiev, Anastasios Noulas and Cecilia Mascolo | Where Businesses Thrive: Predicting the Impact of the Olympic Games on
Local Retailers through Location-based Services Data | null | null | null | null | cs.SI physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The Olympic Games are an important sporting event with notable consequences
for the general economic landscape of the host city. Traditional economic
assessments focus on the aggregated impact of the event on the national income,
but fail to provide micro-scale insights on why local businesses will benefit
from the increased activity during the Games. In this paper we provide a novel
approach to modeling the impact of the Olympic Games on local retailers by
analyzing a dataset mined from a large location-based social service,
Foursquare. We hypothesize that the spatial positioning of businesses as well
as the mobility trends of visitors are primary indicators of whether retailers
will rise their popularity during the event. To confirm this we formulate a
retail winners prediction task in the context of which we evaluate a set of
geographic and mobility metrics. We find that the proximity to stadiums, the
diversity of activity in the neighborhood, the nearby area sociability, as well
as the probability of customer flows from and to event places such as stadiums
and parks are all vital factors. Through supervised learning techniques we
demonstrate that the success of businesses hinges on a combination of both
geographic and mobility factors. Our results suggest that location-based social
networks, where crowdsourced information about the dynamic interaction of users
with urban spaces becomes publicly available, present an alternative medium to
assess the economic impact of large scale events in a city.
| [
{
"version": "v1",
"created": "Sat, 29 Mar 2014 18:02:42 GMT"
}
] | 2014-04-01T00:00:00 | [
[
"Georgiev",
"Petko",
""
],
[
"Noulas",
"Anastasios",
""
],
[
"Mascolo",
"Cecilia",
""
]
] | TITLE: Where Businesses Thrive: Predicting the Impact of the Olympic Games on
Local Retailers through Location-based Services Data
ABSTRACT: The Olympic Games are an important sporting event with notable consequences
for the general economic landscape of the host city. Traditional economic
assessments focus on the aggregated impact of the event on the national income,
but fail to provide micro-scale insights on why local businesses will benefit
from the increased activity during the Games. In this paper we provide a novel
approach to modeling the impact of the Olympic Games on local retailers by
analyzing a dataset mined from a large location-based social service,
Foursquare. We hypothesize that the spatial positioning of businesses as well
as the mobility trends of visitors are primary indicators of whether retailers
will rise their popularity during the event. To confirm this we formulate a
retail winners prediction task in the context of which we evaluate a set of
geographic and mobility metrics. We find that the proximity to stadiums, the
diversity of activity in the neighborhood, the nearby area sociability, as well
as the probability of customer flows from and to event places such as stadiums
and parks are all vital factors. Through supervised learning techniques we
demonstrate that the success of businesses hinges on a combination of both
geographic and mobility factors. Our results suggest that location-based social
networks, where crowdsourced information about the dynamic interaction of users
with urban spaces becomes publicly available, present an alternative medium to
assess the economic impact of large scale events in a city.
|
1403.7726 | Ayman I. Madbouly | Ayman I. Madbouly, Amr M. Gody, Tamer M. Barakat | Relevant Feature Selection Model Using Data Mining for Intrusion
Detection System | 12 Pages, 3 figures, 5 tables, Published with "International Journal
of Engineering Trends and Technology (IJETT)". arXiv admin note: text overlap
with arXiv:1208.5997 by other authors without attribution | International Journal of Engineering Trends and Technology
(IJETT), V9(10),501-512 March 2014 | 10.14445/22315381/IJETT-V9P296 | null | cs.CR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Network intrusions have become a significant threat in recent years as a
result of the increased demand of computer networks for critical systems.
Intrusion detection system (IDS) has been widely deployed as a defense measure
for computer networks. Features extracted from network traffic can be used as
sign to detect anomalies. However with the huge amount of network traffic,
collected data contains irrelevant and redundant features that affect the
detection rate of the IDS, consumes high amount of system resources, and
slowdown the training and testing process of the IDS. In this paper, a new
feature selection model is proposed; this model can effectively select the most
relevant features for intrusion detection. Our goal is to build a lightweight
intrusion detection system by using a reduced features set. Deleting irrelevant
and redundant features helps to build a faster training and testing process, to
have less resource consumption as well as to maintain high detection rates. The
effectiveness and the feasibility of our feature selection model were verified
by several experiments on KDD intrusion detection dataset. The experimental
results strongly showed that our model is not only able to yield high detection
rates but also to speed up the detection process.
| [
{
"version": "v1",
"created": "Sun, 30 Mar 2014 09:41:17 GMT"
}
] | 2014-04-01T00:00:00 | [
[
"Madbouly",
"Ayman I.",
""
],
[
"Gody",
"Amr M.",
""
],
[
"Barakat",
"Tamer M.",
""
]
] | TITLE: Relevant Feature Selection Model Using Data Mining for Intrusion
Detection System
ABSTRACT: Network intrusions have become a significant threat in recent years as a
result of the increased demand of computer networks for critical systems.
Intrusion detection system (IDS) has been widely deployed as a defense measure
for computer networks. Features extracted from network traffic can be used as
sign to detect anomalies. However with the huge amount of network traffic,
collected data contains irrelevant and redundant features that affect the
detection rate of the IDS, consumes high amount of system resources, and
slowdown the training and testing process of the IDS. In this paper, a new
feature selection model is proposed; this model can effectively select the most
relevant features for intrusion detection. Our goal is to build a lightweight
intrusion detection system by using a reduced features set. Deleting irrelevant
and redundant features helps to build a faster training and testing process, to
have less resource consumption as well as to maintain high detection rates. The
effectiveness and the feasibility of our feature selection model were verified
by several experiments on KDD intrusion detection dataset. The experimental
results strongly showed that our model is not only able to yield high detection
rates but also to speed up the detection process.
|
1403.7793 | Xin-She Yang | Xin-She Yang, Christian Huyck, Mehmet Karamanoglu, Nawaz Khan | True Global Optimality of the Pressure Vessel Design Problem: A
Benchmark for Bio-Inspired Optimisation Algorithms | null | X.-S. Yang et al., Int. J. Bio-Inspired Computation, vol. 5, no.
6, pp. 329-335 (2013) | 10.4018/jdsst.2013040103 | null | math.OC cs.NE nlin.AO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The pressure vessel design problem is a well-known design benchmark for
validating bio-inspired optimization algorithms. However, its global optimality
is not clear and there has been no mathematical proof put forward. In this
paper, a detailed mathematical analysis of this problem is provided that proves
that 6059.714335048436 is the global minimum. The Lagrange multiplier method is
also used as an alternative proof and this method is extended to find the
global optimum of a cantilever beam design problem.
| [
{
"version": "v1",
"created": "Sun, 30 Mar 2014 18:11:44 GMT"
}
] | 2014-04-01T00:00:00 | [
[
"Yang",
"Xin-She",
""
],
[
"Huyck",
"Christian",
""
],
[
"Karamanoglu",
"Mehmet",
""
],
[
"Khan",
"Nawaz",
""
]
] | TITLE: True Global Optimality of the Pressure Vessel Design Problem: A
Benchmark for Bio-Inspired Optimisation Algorithms
ABSTRACT: The pressure vessel design problem is a well-known design benchmark for
validating bio-inspired optimization algorithms. However, its global optimality
is not clear and there has been no mathematical proof put forward. In this
paper, a detailed mathematical analysis of this problem is provided that proves
that 6059.714335048436 is the global minimum. The Lagrange multiplier method is
also used as an alternative proof and this method is extended to find the
global optimum of a cantilever beam design problem.
|
1403.7872 | Manzil Zaheer | Chenjie Gu, Manzil Zaheer and Xin Li | Multiple-Population Moment Estimation: Exploiting Inter-Population
Correlation for Efficient Moment Estimation in Analog/Mixed-Signal Validation | null | null | null | null | cs.OH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Moment estimation is an important problem during circuit validation, in both
pre-Silicon and post-Silicon stages. From the estimated moments, the
probability of failure and parametric yield can be estimated at each circuit
configuration and corner, and these metrics are used for design optimization
and making product qualification decisions. The problem is especially difficult
if only a very small sample size is allowed for measurement or simulation, as
is the case for complex analog/mixed-signal circuits. In this paper, we propose
an efficient moment estimation method, called Multiple-Population Moment
Estimation (MPME), that significantly improves estimation accuracy under small
sample size. The key idea is to leverage the data collected under different
corners/configurations to improve the accuracy of moment estimation at each
individual corner/configuration. Mathematically, we employ the hierarchical
Bayesian framework to exploit the underlying correlation in the data. We apply
the proposed method to several datasets including post-silicon measurements of
a commercial high-speed I/O link, and demonstrate an average error reduction of
up to 2$\times$, which can be equivalently translated to significant reduction
of validation time and cost.
| [
{
"version": "v1",
"created": "Mon, 31 Mar 2014 05:23:09 GMT"
}
] | 2014-04-01T00:00:00 | [
[
"Gu",
"Chenjie",
""
],
[
"Zaheer",
"Manzil",
""
],
[
"Li",
"Xin",
""
]
] | TITLE: Multiple-Population Moment Estimation: Exploiting Inter-Population
Correlation for Efficient Moment Estimation in Analog/Mixed-Signal Validation
ABSTRACT: Moment estimation is an important problem during circuit validation, in both
pre-Silicon and post-Silicon stages. From the estimated moments, the
probability of failure and parametric yield can be estimated at each circuit
configuration and corner, and these metrics are used for design optimization
and making product qualification decisions. The problem is especially difficult
if only a very small sample size is allowed for measurement or simulation, as
is the case for complex analog/mixed-signal circuits. In this paper, we propose
an efficient moment estimation method, called Multiple-Population Moment
Estimation (MPME), that significantly improves estimation accuracy under small
sample size. The key idea is to leverage the data collected under different
corners/configurations to improve the accuracy of moment estimation at each
individual corner/configuration. Mathematically, we employ the hierarchical
Bayesian framework to exploit the underlying correlation in the data. We apply
the proposed method to several datasets including post-silicon measurements of
a commercial high-speed I/O link, and demonstrate an average error reduction of
up to 2$\times$, which can be equivalently translated to significant reduction
of validation time and cost.
|
1403.8020 | Ashkan Tousimojarad Mr | Ashkan Tousimojarad and Wim Vanderbauwhede | An Efficient Thread Mapping Strategy for Multiprogramming on Manycore
Processors | ParCo Conference, Munich, Germany, 2013 | Parallel Computing: Accelerating Computational Science and
Engineering (CSE) 25 (2014) 63-71 | 10.3233/978-1-61499-381-0-63 | null | cs.DC cs.PF | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The emergence of multicore and manycore processors is set to change the
parallel computing world. Applications are shifting towards increased
parallelism in order to utilise these architectures efficiently. This leads to
a situation where every application creates its desirable number of threads,
based on its parallel nature and the system resources allowance. Task
scheduling in such a multithreaded multiprogramming environment is a
significant challenge. In task scheduling, not only the order of the execution,
but also the mapping of threads to the execution resources is of a great
importance. In this paper we state and discuss some fundamental rules based on
results obtained from selected applications of the BOTS benchmarks on the
64-core TILEPro64 processor. We demonstrate how previously efficient mapping
policies such as those of the SMP Linux scheduler become inefficient when the
number of threads and cores grows. We propose a novel, low-overhead technique,
a heuristic based on the amount of time spent by each CPU doing some useful
work, to fairly distribute the workloads amongst the cores in a
multiprogramming environment. Our novel approach could be implemented as a
pragma similar to those in the new task-based OpenMP versions, or can be
incorporated as a distributed thread mapping mechanism in future manycore
programming frameworks. We show that our thread mapping scheme can outperform
the native GNU/Linux thread scheduler in both single-programming and
multiprogramming environments.
| [
{
"version": "v1",
"created": "Mon, 31 Mar 2014 14:40:02 GMT"
}
] | 2014-04-01T00:00:00 | [
[
"Tousimojarad",
"Ashkan",
""
],
[
"Vanderbauwhede",
"Wim",
""
]
] | TITLE: An Efficient Thread Mapping Strategy for Multiprogramming on Manycore
Processors
ABSTRACT: The emergence of multicore and manycore processors is set to change the
parallel computing world. Applications are shifting towards increased
parallelism in order to utilise these architectures efficiently. This leads to
a situation where every application creates its desirable number of threads,
based on its parallel nature and the system resources allowance. Task
scheduling in such a multithreaded multiprogramming environment is a
significant challenge. In task scheduling, not only the order of the execution,
but also the mapping of threads to the execution resources is of a great
importance. In this paper we state and discuss some fundamental rules based on
results obtained from selected applications of the BOTS benchmarks on the
64-core TILEPro64 processor. We demonstrate how previously efficient mapping
policies such as those of the SMP Linux scheduler become inefficient when the
number of threads and cores grows. We propose a novel, low-overhead technique,
a heuristic based on the amount of time spent by each CPU doing some useful
work, to fairly distribute the workloads amongst the cores in a
multiprogramming environment. Our novel approach could be implemented as a
pragma similar to those in the new task-based OpenMP versions, or can be
incorporated as a distributed thread mapping mechanism in future manycore
programming frameworks. We show that our thread mapping scheme can outperform
the native GNU/Linux thread scheduler in both single-programming and
multiprogramming environments.
|
1403.8084 | Smriti Bhagat | Stratis Ioannidis, Andrea Montanari, Udi Weinsberg, Smriti Bhagat,
Nadia Fawaz, Nina Taft | Privacy Tradeoffs in Predictive Analytics | Extended version of the paper appearing in SIGMETRICS 2014 | null | null | null | cs.CR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Online services routinely mine user data to predict user preferences, make
recommendations, and place targeted ads. Recent research has demonstrated that
several private user attributes (such as political affiliation, sexual
orientation, and gender) can be inferred from such data. Can a
privacy-conscious user benefit from personalization while simultaneously
protecting her private attributes? We study this question in the context of a
rating prediction service based on matrix factorization. We construct a
protocol of interactions between the service and users that has remarkable
optimality properties: it is privacy-preserving, in that no inference algorithm
can succeed in inferring a user's private attribute with a probability better
than random guessing; it has maximal accuracy, in that no other
privacy-preserving protocol improves rating prediction; and, finally, it
involves a minimal disclosure, as the prediction accuracy strictly decreases
when the service reveals less information. We extensively evaluate our protocol
using several rating datasets, demonstrating that it successfully blocks the
inference of gender, age and political affiliation, while incurring less than
5% decrease in the accuracy of rating prediction.
| [
{
"version": "v1",
"created": "Mon, 31 Mar 2014 16:53:04 GMT"
}
] | 2014-04-01T00:00:00 | [
[
"Ioannidis",
"Stratis",
""
],
[
"Montanari",
"Andrea",
""
],
[
"Weinsberg",
"Udi",
""
],
[
"Bhagat",
"Smriti",
""
],
[
"Fawaz",
"Nadia",
""
],
[
"Taft",
"Nina",
""
]
] | TITLE: Privacy Tradeoffs in Predictive Analytics
ABSTRACT: Online services routinely mine user data to predict user preferences, make
recommendations, and place targeted ads. Recent research has demonstrated that
several private user attributes (such as political affiliation, sexual
orientation, and gender) can be inferred from such data. Can a
privacy-conscious user benefit from personalization while simultaneously
protecting her private attributes? We study this question in the context of a
rating prediction service based on matrix factorization. We construct a
protocol of interactions between the service and users that has remarkable
optimality properties: it is privacy-preserving, in that no inference algorithm
can succeed in inferring a user's private attribute with a probability better
than random guessing; it has maximal accuracy, in that no other
privacy-preserving protocol improves rating prediction; and, finally, it
involves a minimal disclosure, as the prediction accuracy strictly decreases
when the service reveals less information. We extensively evaluate our protocol
using several rating datasets, demonstrating that it successfully blocks the
inference of gender, age and political affiliation, while incurring less than
5% decrease in the accuracy of rating prediction.
|
1403.8105 | Jamie Portsmouth | David Koerner, Jamie Portsmouth, Filip Sadlo, Thomas Ertl, and Bernd
Eberhardt | Flux-Limited Diffusion for Multiple Scattering in Participating Media | Accepted in Computer Graphics Forum | null | 10.1111/cgf.12342 | null | cs.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | For the rendering of multiple scattering effects in participating media,
methods based on the diffusion approximation are an extremely efficient
alternative to Monte Carlo path tracing. However, in sufficiently transparent
regions, classical diffusion approximation suffers from non-physical radiative
fluxes which leads to a poor match to correct light transport. In particular,
this prevents the application of classical diffusion approximation to
heterogeneous media, where opaque material is embedded within transparent
regions. To address this limitation, we introduce flux-limited diffusion, a
technique from the astrophysics domain. This method provides a better
approximation to light transport than classical diffusion approximation,
particularly when applied to heterogeneous media, and hence broadens the
applicability of diffusion-based techniques. We provide an algorithm for
flux-limited diffusion, which is validated using the transport theory for a
point light source in an infinite homogeneous medium. We further demonstrate
that our implementation of flux-limited diffusion produces more accurate
renderings of multiple scattering in various heterogeneous datasets than
classical diffusion approximation, by comparing both methods to ground truth
renderings obtained via volumetric path tracing.
| [
{
"version": "v1",
"created": "Mon, 31 Mar 2014 17:54:34 GMT"
}
] | 2014-04-01T00:00:00 | [
[
"Koerner",
"David",
""
],
[
"Portsmouth",
"Jamie",
""
],
[
"Sadlo",
"Filip",
""
],
[
"Ertl",
"Thomas",
""
],
[
"Eberhardt",
"Bernd",
""
]
] | TITLE: Flux-Limited Diffusion for Multiple Scattering in Participating Media
ABSTRACT: For the rendering of multiple scattering effects in participating media,
methods based on the diffusion approximation are an extremely efficient
alternative to Monte Carlo path tracing. However, in sufficiently transparent
regions, classical diffusion approximation suffers from non-physical radiative
fluxes which leads to a poor match to correct light transport. In particular,
this prevents the application of classical diffusion approximation to
heterogeneous media, where opaque material is embedded within transparent
regions. To address this limitation, we introduce flux-limited diffusion, a
technique from the astrophysics domain. This method provides a better
approximation to light transport than classical diffusion approximation,
particularly when applied to heterogeneous media, and hence broadens the
applicability of diffusion-based techniques. We provide an algorithm for
flux-limited diffusion, which is validated using the transport theory for a
point light source in an infinite homogeneous medium. We further demonstrate
that our implementation of flux-limited diffusion produces more accurate
renderings of multiple scattering in various heterogeneous datasets than
classical diffusion approximation, by comparing both methods to ground truth
renderings obtained via volumetric path tracing.
|
1307.5161 | Xavier Boix | Gemma Roig, Xavier Boix, Luc Van Gool | Random Binary Mappings for Kernel Learning and Efficient SVM | null | null | null | null | cs.CV cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Support Vector Machines (SVMs) are powerful learners that have led to
state-of-the-art results in various computer vision problems. SVMs suffer from
various drawbacks in terms of selecting the right kernel, which depends on the
image descriptors, as well as computational and memory efficiency. This paper
introduces a novel kernel, which serves such issues well. The kernel is learned
by exploiting a large amount of low-complex, randomized binary mappings of the
input feature. This leads to an efficient SVM, while also alleviating the task
of kernel selection. We demonstrate the capabilities of our kernel on 6
standard vision benchmarks, in which we combine several common image
descriptors, namely histograms (Flowers17 and Daimler), attribute-like
descriptors (UCI, OSR, and a-VOC08), and Sparse Quantization (ImageNet).
Results show that our kernel learning adapts well to the different descriptors
types, achieving the performance of the kernels specifically tuned for each
image descriptor, and with similar evaluation cost as efficient SVM methods.
| [
{
"version": "v1",
"created": "Fri, 19 Jul 2013 08:47:32 GMT"
},
{
"version": "v2",
"created": "Fri, 28 Mar 2014 08:49:17 GMT"
}
] | 2014-03-31T00:00:00 | [
[
"Roig",
"Gemma",
""
],
[
"Boix",
"Xavier",
""
],
[
"Van Gool",
"Luc",
""
]
] | TITLE: Random Binary Mappings for Kernel Learning and Efficient SVM
ABSTRACT: Support Vector Machines (SVMs) are powerful learners that have led to
state-of-the-art results in various computer vision problems. SVMs suffer from
various drawbacks in terms of selecting the right kernel, which depends on the
image descriptors, as well as computational and memory efficiency. This paper
introduces a novel kernel, which serves such issues well. The kernel is learned
by exploiting a large amount of low-complex, randomized binary mappings of the
input feature. This leads to an efficient SVM, while also alleviating the task
of kernel selection. We demonstrate the capabilities of our kernel on 6
standard vision benchmarks, in which we combine several common image
descriptors, namely histograms (Flowers17 and Daimler), attribute-like
descriptors (UCI, OSR, and a-VOC08), and Sparse Quantization (ImageNet).
Results show that our kernel learning adapts well to the different descriptors
types, achieving the performance of the kernels specifically tuned for each
image descriptor, and with similar evaluation cost as efficient SVM methods.
|
1312.5785 | Du Tran | Du Tran and Lorenzo Torresani | EXMOVES: Classifier-based Features for Scalable Action Recognition | In Proceedings of the 2nd International Conference on Learning
Representations, Banff, Canada, 2014 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper introduces EXMOVES, learned exemplar-based features for efficient
recognition of actions in videos. The entries in our descriptor are produced by
evaluating a set of movement classifiers over spatial-temporal volumes of the
input sequence. Each movement classifier is a simple exemplar-SVM trained on
low-level features, i.e., an SVM learned using a single annotated positive
space-time volume and a large number of unannotated videos.
Our representation offers two main advantages. First, since our mid-level
features are learned from individual video exemplars, they require minimal
amount of supervision. Second, we show that simple linear classification models
trained on our global video descriptor yield action recognition accuracy
approaching the state-of-the-art but at orders of magnitude lower cost, since
at test-time no sliding window is necessary and linear models are efficient to
train and test. This enables scalable action recognition, i.e., efficient
classification of a large number of different actions even in large video
databases. We show the generality of our approach by building our mid-level
descriptors from two different low-level feature representations. The accuracy
and efficiency of the approach are demonstrated on several large-scale action
recognition benchmarks.
| [
{
"version": "v1",
"created": "Fri, 20 Dec 2013 00:37:16 GMT"
},
{
"version": "v2",
"created": "Fri, 27 Dec 2013 19:58:48 GMT"
},
{
"version": "v3",
"created": "Fri, 28 Mar 2014 01:55:14 GMT"
}
] | 2014-03-31T00:00:00 | [
[
"Tran",
"Du",
""
],
[
"Torresani",
"Lorenzo",
""
]
] | TITLE: EXMOVES: Classifier-based Features for Scalable Action Recognition
ABSTRACT: This paper introduces EXMOVES, learned exemplar-based features for efficient
recognition of actions in videos. The entries in our descriptor are produced by
evaluating a set of movement classifiers over spatial-temporal volumes of the
input sequence. Each movement classifier is a simple exemplar-SVM trained on
low-level features, i.e., an SVM learned using a single annotated positive
space-time volume and a large number of unannotated videos.
Our representation offers two main advantages. First, since our mid-level
features are learned from individual video exemplars, they require minimal
amount of supervision. Second, we show that simple linear classification models
trained on our global video descriptor yield action recognition accuracy
approaching the state-of-the-art but at orders of magnitude lower cost, since
at test-time no sliding window is necessary and linear models are efficient to
train and test. This enables scalable action recognition, i.e., efficient
classification of a large number of different actions even in large video
databases. We show the generality of our approach by building our mid-level
descriptors from two different low-level feature representations. The accuracy
and efficiency of the approach are demonstrated on several large-scale action
recognition benchmarks.
|
1403.7291 | Roshan Ragel | R. G. Ragel, Swarnalatha Radhakrishnan and Angelo Ambrose | Instruction-set Selection for Multi-application based ASIP Design: An
Instruction-level Study | null | Information and Automation for Sustainability (ICIAfS), 2012 IEEE
6th International Conference on, 27-29 Sept 2012, pp 141-146, Beijing | 10.1109/ICIAFS.2012.6419895 | null | cs.AR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Efficiency in embedded systems is paramount to achieve high performance while
consuming less area and power. Processors in embedded systems have to be
designed carefully to achieve such design constraints. Application Specific
Instruction set Processors (ASIPs) exploit the nature of applications to design
an optimal instruction set. Despite being not general to execute any
application, ASIPs are highly preferred in the embedded systems industry where
the devices are produced to satisfy a certain type of application domain/s
(either intra-domain or inter-domain). Typically, ASIPs are designed from a
base-processor and functionalities are added for applications. This paper
studies the multi-application ASIPs and their instruction sets, extensively
analysing the instructions for inter-domain and intra-domain designs. Metrics
analysed are the reusable instructions and the extra cost to add a certain
application. A wide range of applications from various application benchmarks
(MiBench, MediaBench and SPEC2006) and domains are analysed for two different
architectures (ARM-Thumb and PISA). Our study shows that the intra-domain
applications contain larger number of common instructions, whereas the
inter-domain applications have very less common instructions, regardless of the
architecture (and therefore the ISA).
| [
{
"version": "v1",
"created": "Fri, 28 Mar 2014 07:35:09 GMT"
}
] | 2014-03-31T00:00:00 | [
[
"Ragel",
"R. G.",
""
],
[
"Radhakrishnan",
"Swarnalatha",
""
],
[
"Ambrose",
"Angelo",
""
]
] | TITLE: Instruction-set Selection for Multi-application based ASIP Design: An
Instruction-level Study
ABSTRACT: Efficiency in embedded systems is paramount to achieve high performance while
consuming less area and power. Processors in embedded systems have to be
designed carefully to achieve such design constraints. Application Specific
Instruction set Processors (ASIPs) exploit the nature of applications to design
an optimal instruction set. Despite being not general to execute any
application, ASIPs are highly preferred in the embedded systems industry where
the devices are produced to satisfy a certain type of application domain/s
(either intra-domain or inter-domain). Typically, ASIPs are designed from a
base-processor and functionalities are added for applications. This paper
studies the multi-application ASIPs and their instruction sets, extensively
analysing the instructions for inter-domain and intra-domain designs. Metrics
analysed are the reusable instructions and the extra cost to add a certain
application. A wide range of applications from various application benchmarks
(MiBench, MediaBench and SPEC2006) and domains are analysed for two different
architectures (ARM-Thumb and PISA). Our study shows that the intra-domain
applications contain larger number of common instructions, whereas the
inter-domain applications have very less common instructions, regardless of the
architecture (and therefore the ISA).
|
1403.7315 | Chuan Shi | Yitong Li, Chuan Shi, Philip S. Yu, and Qing Chen | HRank: A Path based Ranking Framework in Heterogeneous Information
Network | 12 pages, 11 figures | null | null | null | cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recently, there is a surge of interests on heterogeneous information network
analysis. As a newly emerging network model, heterogeneous information networks
have many unique features (e.g., complex structure and rich semantics) and a
number of interesting data mining tasks have been exploited in this kind of
networks, such as similarity measure, clustering, and classification. Although
evaluating the importance of objects has been well studied in homogeneous
networks, it is not yet exploited in heterogeneous networks. In this paper, we
study the ranking problem in heterogeneous networks and propose the HRank
framework to evaluate the importance of multiple types of objects and meta
paths. Since the importance of objects depends upon the meta paths in
heterogeneous networks, HRank develops a path based random walk process.
Moreover, a constrained meta path is proposed to subtly capture the rich
semantics in heterogeneous networks. Furthermore, HRank can simultaneously
determine the importance of objects and meta paths through applying the tensor
analysis. Extensive experiments on three real datasets show that HRank can
effectively evaluate the importance of objects and paths together. Moreover,
the constrained meta path shows its potential on mining subtle semantics by
obtaining more accurate ranking results.
| [
{
"version": "v1",
"created": "Fri, 28 Mar 2014 09:31:43 GMT"
}
] | 2014-03-31T00:00:00 | [
[
"Li",
"Yitong",
""
],
[
"Shi",
"Chuan",
""
],
[
"Yu",
"Philip S.",
""
],
[
"Chen",
"Qing",
""
]
] | TITLE: HRank: A Path based Ranking Framework in Heterogeneous Information
Network
ABSTRACT: Recently, there is a surge of interests on heterogeneous information network
analysis. As a newly emerging network model, heterogeneous information networks
have many unique features (e.g., complex structure and rich semantics) and a
number of interesting data mining tasks have been exploited in this kind of
networks, such as similarity measure, clustering, and classification. Although
evaluating the importance of objects has been well studied in homogeneous
networks, it is not yet exploited in heterogeneous networks. In this paper, we
study the ranking problem in heterogeneous networks and propose the HRank
framework to evaluate the importance of multiple types of objects and meta
paths. Since the importance of objects depends upon the meta paths in
heterogeneous networks, HRank develops a path based random walk process.
Moreover, a constrained meta path is proposed to subtly capture the rich
semantics in heterogeneous networks. Furthermore, HRank can simultaneously
determine the importance of objects and meta paths through applying the tensor
analysis. Extensive experiments on three real datasets show that HRank can
effectively evaluate the importance of objects and paths together. Moreover,
the constrained meta path shows its potential on mining subtle semantics by
obtaining more accurate ranking results.
|
1403.7373 | Radek Pel\'anek | Radek Pel\'anek | Difficulty Rating of Sudoku Puzzles: An Overview and Evaluation | 24 pages | null | null | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | How can we predict the difficulty of a Sudoku puzzle? We give an overview of
difficulty rating metrics and evaluate them on extensive dataset on human
problem solving (more then 1700 Sudoku puzzles, hundreds of solvers). The best
results are obtained using a computational model of human solving activity.
Using the model we show that there are two sources of the problem difficulty:
complexity of individual steps (logic operations) and structure of dependency
among steps. We also describe metrics based on analysis of solutions under
relaxed constraints -- a novel approach inspired by phase transition phenomenon
in the graph coloring problem. In our discussion we focus not just on the
performance of individual metrics on the Sudoku puzzle, but also on their
generalizability and applicability to other problems.
| [
{
"version": "v1",
"created": "Fri, 28 Mar 2014 13:43:50 GMT"
}
] | 2014-03-31T00:00:00 | [
[
"Pelánek",
"Radek",
""
]
] | TITLE: Difficulty Rating of Sudoku Puzzles: An Overview and Evaluation
ABSTRACT: How can we predict the difficulty of a Sudoku puzzle? We give an overview of
difficulty rating metrics and evaluate them on extensive dataset on human
problem solving (more then 1700 Sudoku puzzles, hundreds of solvers). The best
results are obtained using a computational model of human solving activity.
Using the model we show that there are two sources of the problem difficulty:
complexity of individual steps (logic operations) and structure of dependency
among steps. We also describe metrics based on analysis of solutions under
relaxed constraints -- a novel approach inspired by phase transition phenomenon
in the graph coloring problem. In our discussion we focus not just on the
performance of individual metrics on the Sudoku puzzle, but also on their
generalizability and applicability to other problems.
|
1403.6950 | Manuel Marin-Jimenez | F.M. Castro and M.J. Marin-Jimenez and R. Medina-Carnicer | Pyramidal Fisher Motion for Multiview Gait Recognition | Submitted to International Conference on Pattern Recognition, ICPR,
2014 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The goal of this paper is to identify individuals by analyzing their gait.
Instead of using binary silhouettes as input data (as done in many previous
works) we propose and evaluate the use of motion descriptors based on densely
sampled short-term trajectories. We take advantage of state-of-the-art people
detectors to define custom spatial configurations of the descriptors around the
target person. Thus, obtaining a pyramidal representation of the gait motion.
The local motion features (described by the Divergence-Curl-Shear descriptor)
extracted on the different spatial areas of the person are combined into a
single high-level gait descriptor by using the Fisher Vector encoding. The
proposed approach, coined Pyramidal Fisher Motion, is experimentally validated
on the recent `AVA Multiview Gait' dataset. The results show that this new
approach achieves promising results in the problem of gait recognition.
| [
{
"version": "v1",
"created": "Thu, 27 Mar 2014 08:39:31 GMT"
}
] | 2014-03-28T00:00:00 | [
[
"Castro",
"F. M.",
""
],
[
"Marin-Jimenez",
"M. J.",
""
],
[
"Medina-Carnicer",
"R.",
""
]
] | TITLE: Pyramidal Fisher Motion for Multiview Gait Recognition
ABSTRACT: The goal of this paper is to identify individuals by analyzing their gait.
Instead of using binary silhouettes as input data (as done in many previous
works) we propose and evaluate the use of motion descriptors based on densely
sampled short-term trajectories. We take advantage of state-of-the-art people
detectors to define custom spatial configurations of the descriptors around the
target person. Thus, obtaining a pyramidal representation of the gait motion.
The local motion features (described by the Divergence-Curl-Shear descriptor)
extracted on the different spatial areas of the person are combined into a
single high-level gait descriptor by using the Fisher Vector encoding. The
proposed approach, coined Pyramidal Fisher Motion, is experimentally validated
on the recent `AVA Multiview Gait' dataset. The results show that this new
approach achieves promising results in the problem of gait recognition.
|
1403.7057 | Alexander Kolesnikov | Alexander Kolesnikov, Matthieu Guillaumin, Vittorio Ferrari and
Christoph H. Lampert | Closed-Form Training of Conditional Random Fields for Large Scale Image
Segmentation | null | null | null | null | cs.LG cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present LS-CRF, a new method for very efficient large-scale training of
Conditional Random Fields (CRFs). It is inspired by existing closed-form
expressions for the maximum likelihood parameters of a generative graphical
model with tree topology. LS-CRF training requires only solving a set of
independent regression problems, for which closed-form expression as well as
efficient iterative solvers are available. This makes it orders of magnitude
faster than conventional maximum likelihood learning for CRFs that require
repeated runs of probabilistic inference. At the same time, the models learned
by our method still allow for joint inference at test time. We apply LS-CRF to
the task of semantic image segmentation, showing that it is highly efficient,
even for loopy models where probabilistic inference is problematic. It allows
the training of image segmentation models from significantly larger training
sets than had been used previously. We demonstrate this on two new datasets
that form a second contribution of this paper. They consist of over 180,000
images with figure-ground segmentation annotations. Our large-scale experiments
show that the possibilities of CRF-based image segmentation are far from
exhausted, indicating, for example, that semi-supervised learning and the use
of non-linear predictors are promising directions for achieving higher
segmentation accuracy in the future.
| [
{
"version": "v1",
"created": "Thu, 27 Mar 2014 14:38:23 GMT"
}
] | 2014-03-28T00:00:00 | [
[
"Kolesnikov",
"Alexander",
""
],
[
"Guillaumin",
"Matthieu",
""
],
[
"Ferrari",
"Vittorio",
""
],
[
"Lampert",
"Christoph H.",
""
]
] | TITLE: Closed-Form Training of Conditional Random Fields for Large Scale Image
Segmentation
ABSTRACT: We present LS-CRF, a new method for very efficient large-scale training of
Conditional Random Fields (CRFs). It is inspired by existing closed-form
expressions for the maximum likelihood parameters of a generative graphical
model with tree topology. LS-CRF training requires only solving a set of
independent regression problems, for which closed-form expression as well as
efficient iterative solvers are available. This makes it orders of magnitude
faster than conventional maximum likelihood learning for CRFs that require
repeated runs of probabilistic inference. At the same time, the models learned
by our method still allow for joint inference at test time. We apply LS-CRF to
the task of semantic image segmentation, showing that it is highly efficient,
even for loopy models where probabilistic inference is problematic. It allows
the training of image segmentation models from significantly larger training
sets than had been used previously. We demonstrate this on two new datasets
that form a second contribution of this paper. They consist of over 180,000
images with figure-ground segmentation annotations. Our large-scale experiments
show that the possibilities of CRF-based image segmentation are far from
exhausted, indicating, for example, that semi-supervised learning and the use
of non-linear predictors are promising directions for achieving higher
segmentation accuracy in the future.
|
1311.4082 | Joel Leibo | Qianli Liao, Joel Z Leibo, Youssef Mroueh, Tomaso Poggio | Can a biologically-plausible hierarchy effectively replace face
detection, alignment, and recognition pipelines? | 11 Pages, 4 Figures. Mar 26, (2014): Improved exposition. Added CBMM
memo cover page. No substantive changes | null | null | CBMM-003 | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The standard approach to unconstrained face recognition in natural
photographs is via a detection, alignment, recognition pipeline. While that
approach has achieved impressive results, there are several reasons to be
dissatisfied with it, among them is its lack of biological plausibility. A
recent theory of invariant recognition by feedforward hierarchical networks,
like HMAX, other convolutional networks, or possibly the ventral stream,
implies an alternative approach to unconstrained face recognition. This
approach accomplishes detection and alignment implicitly by storing
transformations of training images (called templates) rather than explicitly
detecting and aligning faces at test time. Here we propose a particular
locality-sensitive hashing based voting scheme which we call "consensus of
collisions" and show that it can be used to approximate the full 3-layer
hierarchy implied by the theory. The resulting end-to-end system for
unconstrained face recognition operates on photographs of faces taken under
natural conditions, e.g., Labeled Faces in the Wild (LFW), without aligning or
cropping them, as is normally done. It achieves a drastic improvement in the
state of the art on this end-to-end task, reaching the same level of
performance as the best systems operating on aligned, closely cropped images
(no outside training data). It also performs well on two newer datasets,
similar to LFW, but more difficult: LFW-jittered (new here) and SUFR-W.
| [
{
"version": "v1",
"created": "Sat, 16 Nov 2013 17:49:31 GMT"
},
{
"version": "v2",
"created": "Thu, 21 Nov 2013 10:25:29 GMT"
},
{
"version": "v3",
"created": "Wed, 26 Mar 2014 10:11:42 GMT"
}
] | 2014-03-27T00:00:00 | [
[
"Liao",
"Qianli",
""
],
[
"Leibo",
"Joel Z",
""
],
[
"Mroueh",
"Youssef",
""
],
[
"Poggio",
"Tomaso",
""
]
] | TITLE: Can a biologically-plausible hierarchy effectively replace face
detection, alignment, and recognition pipelines?
ABSTRACT: The standard approach to unconstrained face recognition in natural
photographs is via a detection, alignment, recognition pipeline. While that
approach has achieved impressive results, there are several reasons to be
dissatisfied with it, among them is its lack of biological plausibility. A
recent theory of invariant recognition by feedforward hierarchical networks,
like HMAX, other convolutional networks, or possibly the ventral stream,
implies an alternative approach to unconstrained face recognition. This
approach accomplishes detection and alignment implicitly by storing
transformations of training images (called templates) rather than explicitly
detecting and aligning faces at test time. Here we propose a particular
locality-sensitive hashing based voting scheme which we call "consensus of
collisions" and show that it can be used to approximate the full 3-layer
hierarchy implied by the theory. The resulting end-to-end system for
unconstrained face recognition operates on photographs of faces taken under
natural conditions, e.g., Labeled Faces in the Wild (LFW), without aligning or
cropping them, as is normally done. It achieves a drastic improvement in the
state of the art on this end-to-end task, reaching the same level of
performance as the best systems operating on aligned, closely cropped images
(no outside training data). It also performs well on two newer datasets,
similar to LFW, but more difficult: LFW-jittered (new here) and SUFR-W.
|
1311.4529 | Afroza Sultana | Afroza Sultana, Naeemul Hassan, Chengkai Li, Jun Yang, Cong Yu | Incremental Discovery of Prominent Situational Facts | null | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study the novel problem of finding new, prominent situational facts, which
are emerging statements about objects that stand out within certain contexts.
Many such facts are newsworthy---e.g., an athlete's outstanding performance in
a game, or a viral video's impressive popularity. Effective and efficient
identification of these facts assists journalists in reporting, one of the main
goals of computational journalism. Technically, we consider an ever-growing
table of objects with dimension and measure attributes. A situational fact is a
"contextual" skyline tuple that stands out against historical tuples in a
context, specified by a conjunctive constraint involving dimension attributes,
when a set of measure attributes are compared. New tuples are constantly added
to the table, reflecting events happening in the real world. Our goal is to
discover constraint-measure pairs that qualify a new tuple as a contextual
skyline tuple, and discover them quickly before the event becomes yesterday's
news. A brute-force approach requires exhaustive comparison with every tuple,
under every constraint, and in every measure subspace. We design algorithms in
response to these challenges using three corresponding ideas---tuple reduction,
constraint pruning, and sharing computation across measure subspaces. We also
adopt a simple prominence measure to rank the discovered facts when they are
numerous. Experiments over two real datasets validate the effectiveness and
efficiency of our techniques.
| [
{
"version": "v1",
"created": "Mon, 18 Nov 2013 20:44:13 GMT"
},
{
"version": "v2",
"created": "Wed, 26 Mar 2014 16:43:25 GMT"
}
] | 2014-03-27T00:00:00 | [
[
"Sultana",
"Afroza",
""
],
[
"Hassan",
"Naeemul",
""
],
[
"Li",
"Chengkai",
""
],
[
"Yang",
"Jun",
""
],
[
"Yu",
"Cong",
""
]
] | TITLE: Incremental Discovery of Prominent Situational Facts
ABSTRACT: We study the novel problem of finding new, prominent situational facts, which
are emerging statements about objects that stand out within certain contexts.
Many such facts are newsworthy---e.g., an athlete's outstanding performance in
a game, or a viral video's impressive popularity. Effective and efficient
identification of these facts assists journalists in reporting, one of the main
goals of computational journalism. Technically, we consider an ever-growing
table of objects with dimension and measure attributes. A situational fact is a
"contextual" skyline tuple that stands out against historical tuples in a
context, specified by a conjunctive constraint involving dimension attributes,
when a set of measure attributes are compared. New tuples are constantly added
to the table, reflecting events happening in the real world. Our goal is to
discover constraint-measure pairs that qualify a new tuple as a contextual
skyline tuple, and discover them quickly before the event becomes yesterday's
news. A brute-force approach requires exhaustive comparison with every tuple,
under every constraint, and in every measure subspace. We design algorithms in
response to these challenges using three corresponding ideas---tuple reduction,
constraint pruning, and sharing computation across measure subspaces. We also
adopt a simple prominence measure to rank the discovered facts when they are
numerous. Experiments over two real datasets validate the effectiveness and
efficiency of our techniques.
|
1311.2008 | Matus Medo | Matus Medo | Statistical validation of high-dimensional models of growing networks | 8 pages, 5 figures, 2 tables | Phys. Rev. E 89, 032801, 2014 | 10.1103/PhysRevE.89.032801 | null | physics.soc-ph cs.SI physics.data-an | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The abundance of models of complex networks and the current insufficient
validation standards make it difficult to judge which models are strongly
supported by data and which are not. We focus here on likelihood maximization
methods for models of growing networks with many parameters and compare their
performance on artificial and real datasets. While high dimensionality of the
parameter space harms the performance of direct likelihood maximization on
artificial data, this can be improved by introducing a suitable penalization
term. Likelihood maximization on real data shows that the presented approach is
able to discriminate among available network models. To make large-scale
datasets accessible to this kind of analysis, we propose a subset sampling
technique and show that it yields substantial model evidence in a fraction of
time necessary for the analysis of the complete data.
| [
{
"version": "v1",
"created": "Fri, 8 Nov 2013 16:09:08 GMT"
},
{
"version": "v2",
"created": "Thu, 30 Jan 2014 13:16:06 GMT"
}
] | 2014-03-26T00:00:00 | [
[
"Medo",
"Matus",
""
]
] | TITLE: Statistical validation of high-dimensional models of growing networks
ABSTRACT: The abundance of models of complex networks and the current insufficient
validation standards make it difficult to judge which models are strongly
supported by data and which are not. We focus here on likelihood maximization
methods for models of growing networks with many parameters and compare their
performance on artificial and real datasets. While high dimensionality of the
parameter space harms the performance of direct likelihood maximization on
artificial data, this can be improved by introducing a suitable penalization
term. Likelihood maximization on real data shows that the presented approach is
able to discriminate among available network models. To make large-scale
datasets accessible to this kind of analysis, we propose a subset sampling
technique and show that it yields substantial model evidence in a fraction of
time necessary for the analysis of the complete data.
|
1403.6270 | Alessio Guerrieri | Alessio Guerrieri, Alberto Montresor | Distributed Edge Partitioning for Graph Processing | null | null | null | null | cs.DC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The availability of larger and larger graph datasets, growing exponentially
over the years, has created several new algorithmic challenges to be addressed.
Sequential approaches have become unfeasible, while interest on parallel and
distributed algorithms has greatly increased.
Appropriately partitioning the graph as a preprocessing step can improve the
degree of parallelism of its analysis. A number of heuristic algorithms have
been developed to solve this problem, but many of them subdivide the graph on
its vertex set, thus obtaining a vertex-partitioned graph.
Aim of this paper is to explore a completely different approach based on edge
partitioning, in which edges, rather than vertices, are partitioned into
disjoint subsets. Contribution of this paper is twofold: first, we introduce a
graph processing framework based on edge partitioning, that is flexible enough
to be applied to several different graph problems. Second, we show the
feasibility of these ideas by presenting a distributed edge partitioning
algorithm called d-fep.
Our framework is thoroughly evaluated, using both simulations and an Hadoop
implementation running on the Amazon EC2 cloud. The experiments show that d-fep
is efficient, scalable and obtains consistently good partitions. The resulting
edge-partitioned graph can be exploited to obtain more efficient
implementations of graph analysis algorithms.
| [
{
"version": "v1",
"created": "Tue, 25 Mar 2014 09:38:12 GMT"
}
] | 2014-03-26T00:00:00 | [
[
"Guerrieri",
"Alessio",
""
],
[
"Montresor",
"Alberto",
""
]
] | TITLE: Distributed Edge Partitioning for Graph Processing
ABSTRACT: The availability of larger and larger graph datasets, growing exponentially
over the years, has created several new algorithmic challenges to be addressed.
Sequential approaches have become unfeasible, while interest on parallel and
distributed algorithms has greatly increased.
Appropriately partitioning the graph as a preprocessing step can improve the
degree of parallelism of its analysis. A number of heuristic algorithms have
been developed to solve this problem, but many of them subdivide the graph on
its vertex set, thus obtaining a vertex-partitioned graph.
Aim of this paper is to explore a completely different approach based on edge
partitioning, in which edges, rather than vertices, are partitioned into
disjoint subsets. Contribution of this paper is twofold: first, we introduce a
graph processing framework based on edge partitioning, that is flexible enough
to be applied to several different graph problems. Second, we show the
feasibility of these ideas by presenting a distributed edge partitioning
algorithm called d-fep.
Our framework is thoroughly evaluated, using both simulations and an Hadoop
implementation running on the Amazon EC2 cloud. The experiments show that d-fep
is efficient, scalable and obtains consistently good partitions. The resulting
edge-partitioned graph can be exploited to obtain more efficient
implementations of graph analysis algorithms.
|
1403.6275 | Vibhav Vineet Mr | Vibhav Vineet, Jonathan Warrell and Philip H.S. Torr | A Tiered Move-making Algorithm for General Non-submodular Pairwise
Energies | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A large number of problems in computer vision can be modelled as energy
minimization problems in a Markov Random Field (MRF) or Conditional Random
Field (CRF) framework. Graph-cuts based $\alpha$-expansion is a standard
move-making method to minimize the energy functions with sub-modular pairwise
terms. However, certain problems require more complex pairwise terms where the
$\alpha$-expansion method is generally not applicable.
In this paper, we propose an iterative {\em tiered move making algorithm}
which is able to handle general pairwise terms. Each move to the next
configuration is based on the current labeling and an optimal tiered move,
where each tiered move requires one application of the dynamic programming
based tiered labeling method introduced in Felzenszwalb et. al.
\cite{tiered_cvpr_felzenszwalbV10}. The algorithm converges to a local minimum
for any general pairwise potential, and we give a theoretical analysis of the
properties of the algorithm, characterizing the situations in which we can
expect good performance. We first evaluate our method on an object-class
segmentation problem using the Pascal VOC-11 segmentation dataset where we
learn general pairwise terms. Further we evaluate the algorithm on many other
benchmark labeling problems such as stereo, image segmentation, image stitching
and image denoising. Our method consistently gets better accuracy and energy
values than alpha-expansion, loopy belief propagation (LBP), quadratic
pseudo-boolean optimization (QPBO), and is competitive with TRWS.
| [
{
"version": "v1",
"created": "Tue, 25 Mar 2014 10:18:47 GMT"
}
] | 2014-03-26T00:00:00 | [
[
"Vineet",
"Vibhav",
""
],
[
"Warrell",
"Jonathan",
""
],
[
"Torr",
"Philip H. S.",
""
]
] | TITLE: A Tiered Move-making Algorithm for General Non-submodular Pairwise
Energies
ABSTRACT: A large number of problems in computer vision can be modelled as energy
minimization problems in a Markov Random Field (MRF) or Conditional Random
Field (CRF) framework. Graph-cuts based $\alpha$-expansion is a standard
move-making method to minimize the energy functions with sub-modular pairwise
terms. However, certain problems require more complex pairwise terms where the
$\alpha$-expansion method is generally not applicable.
In this paper, we propose an iterative {\em tiered move making algorithm}
which is able to handle general pairwise terms. Each move to the next
configuration is based on the current labeling and an optimal tiered move,
where each tiered move requires one application of the dynamic programming
based tiered labeling method introduced in Felzenszwalb et. al.
\cite{tiered_cvpr_felzenszwalbV10}. The algorithm converges to a local minimum
for any general pairwise potential, and we give a theoretical analysis of the
properties of the algorithm, characterizing the situations in which we can
expect good performance. We first evaluate our method on an object-class
segmentation problem using the Pascal VOC-11 segmentation dataset where we
learn general pairwise terms. Further we evaluate the algorithm on many other
benchmark labeling problems such as stereo, image segmentation, image stitching
and image denoising. Our method consistently gets better accuracy and energy
values than alpha-expansion, loopy belief propagation (LBP), quadratic
pseudo-boolean optimization (QPBO), and is competitive with TRWS.
|
1403.6426 | Elmar Peise | Elmar Peise (1), Diego Fabregat-Traver (1), Paolo Bientinesi (1) ((1)
AICES, RWTH Aachen) | High Performance Solutions for Big-data GWAS | Submitted to Parallel Computing. arXiv admin note: substantial text
overlap with arXiv:1304.2272 | null | null | AICES-2013/12-01 | q-bio.GN cs.CE cs.MS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In order to associate complex traits with genetic polymorphisms, genome-wide
association studies process huge datasets involving tens of thousands of
individuals genotyped for millions of polymorphisms. When handling these
datasets, which exceed the main memory of contemporary computers, one faces two
distinct challenges: 1) Millions of polymorphisms and thousands of phenotypes
come at the cost of hundreds of gigabytes of data, which can only be kept in
secondary storage; 2) the relatedness of the test population is represented by
a relationship matrix, which, for large populations, can only fit in the
combined main memory of a distributed architecture. In this paper, by using
distributed resources such as Cloud or clusters, we address both challenges:
The genotype and phenotype data is streamed from secondary storage using a
double buffer- ing technique, while the relationship matrix is kept across the
main memory of a distributed memory system. With the help of these solutions,
we develop separate algorithms for studies involving only one or a multitude of
traits. We show that these algorithms sustain high-performance and allow the
analysis of enormous datasets.
| [
{
"version": "v1",
"created": "Tue, 25 Mar 2014 17:21:55 GMT"
}
] | 2014-03-26T00:00:00 | [
[
"Peise",
"Elmar",
""
],
[
"Fabregat-Traver",
"Diego",
""
],
[
"Bientinesi",
"Paolo",
""
]
] | TITLE: High Performance Solutions for Big-data GWAS
ABSTRACT: In order to associate complex traits with genetic polymorphisms, genome-wide
association studies process huge datasets involving tens of thousands of
individuals genotyped for millions of polymorphisms. When handling these
datasets, which exceed the main memory of contemporary computers, one faces two
distinct challenges: 1) Millions of polymorphisms and thousands of phenotypes
come at the cost of hundreds of gigabytes of data, which can only be kept in
secondary storage; 2) the relatedness of the test population is represented by
a relationship matrix, which, for large populations, can only fit in the
combined main memory of a distributed architecture. In this paper, by using
distributed resources such as Cloud or clusters, we address both challenges:
The genotype and phenotype data is streamed from secondary storage using a
double buffer- ing technique, while the relationship matrix is kept across the
main memory of a distributed memory system. With the help of these solutions,
we develop separate algorithms for studies involving only one or a multitude of
traits. We show that these algorithms sustain high-performance and allow the
analysis of enormous datasets.
|
1312.5663 | Alireza Makhzani | Alireza Makhzani, Brendan Frey | k-Sparse Autoencoders | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recently, it has been observed that when representations are learnt in a way
that encourages sparsity, improved performance is obtained on classification
tasks. These methods involve combinations of activation functions, sampling
steps and different kinds of penalties. To investigate the effectiveness of
sparsity by itself, we propose the k-sparse autoencoder, which is an
autoencoder with linear activation function, where in hidden layers only the k
highest activities are kept. When applied to the MNIST and NORB datasets, we
find that this method achieves better classification results than denoising
autoencoders, networks trained with dropout, and RBMs. k-sparse autoencoders
are simple to train and the encoding stage is very fast, making them
well-suited to large problem sizes, where conventional sparse coding algorithms
cannot be applied.
| [
{
"version": "v1",
"created": "Thu, 19 Dec 2013 17:46:46 GMT"
},
{
"version": "v2",
"created": "Sat, 22 Mar 2014 17:12:07 GMT"
}
] | 2014-03-25T00:00:00 | [
[
"Makhzani",
"Alireza",
""
],
[
"Frey",
"Brendan",
""
]
] | TITLE: k-Sparse Autoencoders
ABSTRACT: Recently, it has been observed that when representations are learnt in a way
that encourages sparsity, improved performance is obtained on classification
tasks. These methods involve combinations of activation functions, sampling
steps and different kinds of penalties. To investigate the effectiveness of
sparsity by itself, we propose the k-sparse autoencoder, which is an
autoencoder with linear activation function, where in hidden layers only the k
highest activities are kept. When applied to the MNIST and NORB datasets, we
find that this method achieves better classification results than denoising
autoencoders, networks trained with dropout, and RBMs. k-sparse autoencoders
are simple to train and the encoding stage is very fast, making them
well-suited to large problem sizes, where conventional sparse coding algorithms
cannot be applied.
|
1402.6862 | Mohammad Reza Keshtkaran | Mohammad Reza Keshtkaran and Zhi Yang | A Fast, robust algorithm for power line interference cancellation in
neural recording | 18 pages, 21 figures, to appear in Journal of Neural Engineering | M. R. Keshtkaran and Z. Yang, "A fast, robust algorithm for power
line interference cancellation in neural recording," J. Neural Eng., vol. 11,
no. 2, p. 026017, Apr. 2014 | 10.1088/1741-2560/11/2/026017 | null | cs.SY physics.med-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Power line interference may severely corrupt neural recordings at 50/60 Hz
and harmonic frequencies. In this paper, we present a robust and
computationally efficient algorithm for removing power line interference from
neural recordings. The algorithm includes four steps. First, an adaptive notch
filter is used to estimate the fundamental frequency of the interference.
Subsequently, based on the estimated frequency, harmonics are generated by
using discrete-time oscillators, and then the amplitude and phase of each
harmonic are estimated through using a modified recursive least squares
algorithm. Finally, the estimated interference is subtracted from the recorded
data. The algorithm does not require any reference signal, and can track the
frequency, phase, and amplitude of each harmonic. When benchmarked with other
popular approaches, our algorithm performs better in terms of noise immunity,
convergence speed, and output signal-to-noise ratio (SNR). While minimally
affecting the signal bands of interest, the algorithm consistently yields fast
convergence and substantial interference rejection in different conditions of
interference strengths (input SNR from -30 dB to 30 dB), power line frequencies
(45-65 Hz), and phase and amplitude drifts. In addition, the algorithm features
a straightforward parameter adjustment since the parameters are independent of
the input SNR, input signal power, and the sampling rate. The proposed
algorithm features a highly robust operation, fast adaptation to interference
variations, significant SNR improvement, low computational complexity and
memory requirement, and straightforward parameter adjustment. These features
render the algorithm suitable for wearable and implantable sensor applications,
where reliable and real-time cancellation of the interference is desired.
| [
{
"version": "v1",
"created": "Thu, 27 Feb 2014 11:27:17 GMT"
},
{
"version": "v2",
"created": "Mon, 24 Mar 2014 13:03:10 GMT"
}
] | 2014-03-25T00:00:00 | [
[
"Keshtkaran",
"Mohammad Reza",
""
],
[
"Yang",
"Zhi",
""
]
] | TITLE: A Fast, robust algorithm for power line interference cancellation in
neural recording
ABSTRACT: Power line interference may severely corrupt neural recordings at 50/60 Hz
and harmonic frequencies. In this paper, we present a robust and
computationally efficient algorithm for removing power line interference from
neural recordings. The algorithm includes four steps. First, an adaptive notch
filter is used to estimate the fundamental frequency of the interference.
Subsequently, based on the estimated frequency, harmonics are generated by
using discrete-time oscillators, and then the amplitude and phase of each
harmonic are estimated through using a modified recursive least squares
algorithm. Finally, the estimated interference is subtracted from the recorded
data. The algorithm does not require any reference signal, and can track the
frequency, phase, and amplitude of each harmonic. When benchmarked with other
popular approaches, our algorithm performs better in terms of noise immunity,
convergence speed, and output signal-to-noise ratio (SNR). While minimally
affecting the signal bands of interest, the algorithm consistently yields fast
convergence and substantial interference rejection in different conditions of
interference strengths (input SNR from -30 dB to 30 dB), power line frequencies
(45-65 Hz), and phase and amplitude drifts. In addition, the algorithm features
a straightforward parameter adjustment since the parameters are independent of
the input SNR, input signal power, and the sampling rate. The proposed
algorithm features a highly robust operation, fast adaptation to interference
variations, significant SNR improvement, low computational complexity and
memory requirement, and straightforward parameter adjustment. These features
render the algorithm suitable for wearable and implantable sensor applications,
where reliable and real-time cancellation of the interference is desired.
|
1403.5590 | Petter Strandmark | Petter Strandmark and Sameer Agarwal | Continuous Optimization for Fields of Experts Denoising Works | 5 pages | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Several recent papers use image denoising with a Fields of Experts prior to
benchmark discrete optimization methods. We show that a non-linear least
squares solver significantly outperforms all known discrete methods on this
problem.
| [
{
"version": "v1",
"created": "Fri, 21 Mar 2014 23:26:47 GMT"
}
] | 2014-03-25T00:00:00 | [
[
"Strandmark",
"Petter",
""
],
[
"Agarwal",
"Sameer",
""
]
] | TITLE: Continuous Optimization for Fields of Experts Denoising Works
ABSTRACT: Several recent papers use image denoising with a Fields of Experts prior to
benchmark discrete optimization methods. We show that a non-linear least
squares solver significantly outperforms all known discrete methods on this
problem.
|
1403.5693 | Dougal Maclaurin | Dougal Maclaurin and Ryan P. Adams | Firefly Monte Carlo: Exact MCMC with Subsets of Data | null | null | null | null | stat.ML cs.LG stat.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Markov chain Monte Carlo (MCMC) is a popular and successful general-purpose
tool for Bayesian inference. However, MCMC cannot be practically applied to
large data sets because of the prohibitive cost of evaluating every likelihood
term at every iteration. Here we present Firefly Monte Carlo (FlyMC) an
auxiliary variable MCMC algorithm that only queries the likelihoods of a
potentially small subset of the data at each iteration yet simulates from the
exact posterior distribution, in contrast to recent proposals that are
approximate even in the asymptotic limit. FlyMC is compatible with a wide
variety of modern MCMC algorithms, and only requires a lower bound on the
per-datum likelihood factors. In experiments, we find that FlyMC generates
samples from the posterior more than an order of magnitude faster than regular
MCMC, opening up MCMC methods to larger datasets than were previously
considered feasible.
| [
{
"version": "v1",
"created": "Sat, 22 Mar 2014 18:21:29 GMT"
}
] | 2014-03-25T00:00:00 | [
[
"Maclaurin",
"Dougal",
""
],
[
"Adams",
"Ryan P.",
""
]
] | TITLE: Firefly Monte Carlo: Exact MCMC with Subsets of Data
ABSTRACT: Markov chain Monte Carlo (MCMC) is a popular and successful general-purpose
tool for Bayesian inference. However, MCMC cannot be practically applied to
large data sets because of the prohibitive cost of evaluating every likelihood
term at every iteration. Here we present Firefly Monte Carlo (FlyMC) an
auxiliary variable MCMC algorithm that only queries the likelihoods of a
potentially small subset of the data at each iteration yet simulates from the
exact posterior distribution, in contrast to recent proposals that are
approximate even in the asymptotic limit. FlyMC is compatible with a wide
variety of modern MCMC algorithms, and only requires a lower bound on the
per-datum likelihood factors. In experiments, we find that FlyMC generates
samples from the posterior more than an order of magnitude faster than regular
MCMC, opening up MCMC methods to larger datasets than were previously
considered feasible.
|
1403.5877 | Anastasios Kyrillidis | Anastasios Kyrillidis and Anastasios Zouzias | Non-uniform Feature Sampling for Decision Tree Ensembles | 7 pages, 7 figures, 1 table | null | null | null | stat.ML cs.IT cs.LG math.IT stat.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study the effectiveness of non-uniform randomized feature selection in
decision tree classification. We experimentally evaluate two feature selection
methodologies, based on information extracted from the provided dataset: $(i)$
\emph{leverage scores-based} and $(ii)$ \emph{norm-based} feature selection.
Experimental evaluation of the proposed feature selection techniques indicate
that such approaches might be more effective compared to naive uniform feature
selection and moreover having comparable performance to the random forest
algorithm [3]
| [
{
"version": "v1",
"created": "Mon, 24 Mar 2014 08:26:19 GMT"
}
] | 2014-03-25T00:00:00 | [
[
"Kyrillidis",
"Anastasios",
""
],
[
"Zouzias",
"Anastasios",
""
]
] | TITLE: Non-uniform Feature Sampling for Decision Tree Ensembles
ABSTRACT: We study the effectiveness of non-uniform randomized feature selection in
decision tree classification. We experimentally evaluate two feature selection
methodologies, based on information extracted from the provided dataset: $(i)$
\emph{leverage scores-based} and $(ii)$ \emph{norm-based} feature selection.
Experimental evaluation of the proposed feature selection techniques indicate
that such approaches might be more effective compared to naive uniform feature
selection and moreover having comparable performance to the random forest
algorithm [3]
|
1403.5299 | Arkadiusz Stopczynski Mr. | Arkadiusz Stopczynski, Riccardo Pietri, Alex Pentland, David Lazer,
Sune Lehmann | Privacy in Sensor-Driven Human Data Collection: A Guide for
Practitioners | null | null | null | null | cs.CY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In recent years, the amount of information collected about human beings has
increased dramatically. This development has been partially driven by
individuals posting and storing data about themselves and friends using online
social networks or collecting their data for self-tracking purposes
(quantified-self movement). Across the sciences, researchers conduct studies
collecting data with an unprecedented resolution and scale. Using computational
power combined with mathematical models, such rich datasets can be mined to
infer underlying patterns, thereby providing insights into human nature. Much
of the data collected is sensitive. It is private in the sense that most
individuals would feel uncomfortable sharing their collected personal data
publicly. For this reason, the need for solutions to ensure the privacy of the
individuals generating data has grown alongside the data collection efforts.
Out of all the massive data collection efforts, this paper focuses on efforts
directly instrumenting human behavior, and notes that -- in many cases -- the
privacy of participants is not sufficiently addressed. For example, study
purposes are often not explicit, informed consent is ill-defined, and security
and sharing protocols are only partially disclosed. This paper provides a
survey of the work related to addressing privacy issues in research studies
that collect detailed sensor data on human behavior. Reflections on the key
problems and recommendations for future work are included. We hope the overview
of the privacy-related practices in massive data collection studies can be used
as a frame of reference for practitioners in the field. Although focused on
data collection in an academic context, we believe that many of the challenges
and solutions we identify are also relevant and useful for other domains where
massive data collection takes place, including businesses and governments.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2014 21:16:55 GMT"
}
] | 2014-03-24T00:00:00 | [
[
"Stopczynski",
"Arkadiusz",
""
],
[
"Pietri",
"Riccardo",
""
],
[
"Pentland",
"Alex",
""
],
[
"Lazer",
"David",
""
],
[
"Lehmann",
"Sune",
""
]
] | TITLE: Privacy in Sensor-Driven Human Data Collection: A Guide for
Practitioners
ABSTRACT: In recent years, the amount of information collected about human beings has
increased dramatically. This development has been partially driven by
individuals posting and storing data about themselves and friends using online
social networks or collecting their data for self-tracking purposes
(quantified-self movement). Across the sciences, researchers conduct studies
collecting data with an unprecedented resolution and scale. Using computational
power combined with mathematical models, such rich datasets can be mined to
infer underlying patterns, thereby providing insights into human nature. Much
of the data collected is sensitive. It is private in the sense that most
individuals would feel uncomfortable sharing their collected personal data
publicly. For this reason, the need for solutions to ensure the privacy of the
individuals generating data has grown alongside the data collection efforts.
Out of all the massive data collection efforts, this paper focuses on efforts
directly instrumenting human behavior, and notes that -- in many cases -- the
privacy of participants is not sufficiently addressed. For example, study
purposes are often not explicit, informed consent is ill-defined, and security
and sharing protocols are only partially disclosed. This paper provides a
survey of the work related to addressing privacy issues in research studies
that collect detailed sensor data on human behavior. Reflections on the key
problems and recommendations for future work are included. We hope the overview
of the privacy-related practices in massive data collection studies can be used
as a frame of reference for practitioners in the field. Although focused on
data collection in an academic context, we believe that many of the challenges
and solutions we identify are also relevant and useful for other domains where
massive data collection takes place, including businesses and governments.
|
1403.5381 | Silu Huang | Silu Huang, Ada Wai-Chee Fu | ({\alpha}, k)-Minimal Sorting and Skew Join in MPI and MapReduce | 18 pages | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As computer clusters are found to be highly effective for handling massive
datasets, the design of efficient parallel algorithms for such a computing
model is of great interest. We consider ({\alpha}, k)-minimal algorithms for
such a purpose, where {\alpha} is the number of rounds in the algorithm, and k
is a bound on the deviation from perfect workload balance. We focus on new
({\alpha}, k)-minimal algorithms for sorting and skew equijoin operations for
computer clusters. To the best of our knowledge the proposed sorting and skew
join algorithms achieve the best workload balancing guarantee when compared to
previous works. Our empirical study shows that they are close to optimal in
workload balancing. In particular, our proposed sorting algorithm is around 25%
more efficient than the state-of-the-art Terasort algorithm and achieves
significantly more even workload distribution by over 50%.
| [
{
"version": "v1",
"created": "Fri, 21 Mar 2014 07:10:16 GMT"
}
] | 2014-03-24T00:00:00 | [
[
"Huang",
"Silu",
""
],
[
"Fu",
"Ada Wai-Chee",
""
]
] | TITLE: ({\alpha}, k)-Minimal Sorting and Skew Join in MPI and MapReduce
ABSTRACT: As computer clusters are found to be highly effective for handling massive
datasets, the design of efficient parallel algorithms for such a computing
model is of great interest. We consider ({\alpha}, k)-minimal algorithms for
such a purpose, where {\alpha} is the number of rounds in the algorithm, and k
is a bound on the deviation from perfect workload balance. We focus on new
({\alpha}, k)-minimal algorithms for sorting and skew equijoin operations for
computer clusters. To the best of our knowledge the proposed sorting and skew
join algorithms achieve the best workload balancing guarantee when compared to
previous works. Our empirical study shows that they are close to optimal in
workload balancing. In particular, our proposed sorting algorithm is around 25%
more efficient than the state-of-the-art Terasort algorithm and achieves
significantly more even workload distribution by over 50%.
|
1403.5488 | Tshilidzi Marwala | Collins Leke, Bhekisipho Twala, and T. Marwala | Missing Data Prediction and Classification: The Use of Auto-Associative
Neural Networks and Optimization Algorithms | null | null | null | null | cs.NE cs.LG | http://creativecommons.org/licenses/by-nc-sa/3.0/ | This paper presents methods which are aimed at finding approximations to
missing data in a dataset by using optimization algorithms to optimize the
network parameters after which prediction and classification tasks can be
performed. The optimization methods that are considered are genetic algorithm
(GA), simulated annealing (SA), particle swarm optimization (PSO), random
forest (RF) and negative selection (NS) and these methods are individually used
in combination with auto-associative neural networks (AANN) for missing data
estimation and the results obtained are compared. The methods suggested use the
optimization algorithms to minimize an error function derived from training the
auto-associative neural network during which the interrelationships between the
inputs and the outputs are obtained and stored in the weights connecting the
different layers of the network. The error function is expressed as the square
of the difference between the actual observations and predicted values from an
auto-associative neural network. In the event of missing data, all the values
of the actual observations are not known hence, the error function is
decomposed to depend on the known and unknown variable values. Multi-layer
perceptron (MLP) neural network is employed to train the neural networks using
the scaled conjugate gradient (SCG) method. Prediction accuracy is determined
by mean squared error (MSE), root mean squared error (RMSE), mean absolute
error (MAE), and correlation coefficient (r) computations. Accuracy in
classification is obtained by plotting ROC curves and calculating the areas
under these. Analysis of results depicts that the approach using RF with AANN
produces the most accurate predictions and classifications while on the other
end of the scale is the approach which entails using NS with AANN.
| [
{
"version": "v1",
"created": "Fri, 21 Mar 2014 15:11:52 GMT"
}
] | 2014-03-24T00:00:00 | [
[
"Leke",
"Collins",
""
],
[
"Twala",
"Bhekisipho",
""
],
[
"Marwala",
"T.",
""
]
] | TITLE: Missing Data Prediction and Classification: The Use of Auto-Associative
Neural Networks and Optimization Algorithms
ABSTRACT: This paper presents methods which are aimed at finding approximations to
missing data in a dataset by using optimization algorithms to optimize the
network parameters after which prediction and classification tasks can be
performed. The optimization methods that are considered are genetic algorithm
(GA), simulated annealing (SA), particle swarm optimization (PSO), random
forest (RF) and negative selection (NS) and these methods are individually used
in combination with auto-associative neural networks (AANN) for missing data
estimation and the results obtained are compared. The methods suggested use the
optimization algorithms to minimize an error function derived from training the
auto-associative neural network during which the interrelationships between the
inputs and the outputs are obtained and stored in the weights connecting the
different layers of the network. The error function is expressed as the square
of the difference between the actual observations and predicted values from an
auto-associative neural network. In the event of missing data, all the values
of the actual observations are not known hence, the error function is
decomposed to depend on the known and unknown variable values. Multi-layer
perceptron (MLP) neural network is employed to train the neural networks using
the scaled conjugate gradient (SCG) method. Prediction accuracy is determined
by mean squared error (MSE), root mean squared error (RMSE), mean absolute
error (MAE), and correlation coefficient (r) computations. Accuracy in
classification is obtained by plotting ROC curves and calculating the areas
under these. Analysis of results depicts that the approach using RF with AANN
produces the most accurate predictions and classifications while on the other
end of the scale is the approach which entails using NS with AANN.
|
1403.5115 | Ugo Louche | Ugo Louche (LIF), Liva Ralaivola (LIF) | Unconfused Ultraconservative Multiclass Algorithms | ACML, Australia (2013) | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We tackle the problem of learning linear classifiers from noisy datasets in a
multiclass setting. The two-class version of this problem was studied a few
years ago by, e.g. Bylander (1994) and Blum et al. (1996): in these
contributions, the proposed approaches to fight the noise revolve around a
Perceptron learning scheme fed with peculiar examples computed through a
weighted average of points from the noisy training set. We propose to build
upon these approaches and we introduce a new algorithm called UMA (for
Unconfused Multiclass additive Algorithm) which may be seen as a generalization
to the multiclass setting of the previous approaches. In order to characterize
the noise we use the confusion matrix as a multiclass extension of the
classification noise studied in the aforementioned literature. Theoretically
well-founded, UMA furthermore displays very good empirical noise robustness, as
evidenced by numerical simulations conducted on both synthetic and real data.
Keywords: Multiclass classification, Perceptron, Noisy labels, Confusion Matrix
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2014 12:46:33 GMT"
}
] | 2014-03-21T00:00:00 | [
[
"Louche",
"Ugo",
"",
"LIF"
],
[
"Ralaivola",
"Liva",
"",
"LIF"
]
] | TITLE: Unconfused Ultraconservative Multiclass Algorithms
ABSTRACT: We tackle the problem of learning linear classifiers from noisy datasets in a
multiclass setting. The two-class version of this problem was studied a few
years ago by, e.g. Bylander (1994) and Blum et al. (1996): in these
contributions, the proposed approaches to fight the noise revolve around a
Perceptron learning scheme fed with peculiar examples computed through a
weighted average of points from the noisy training set. We propose to build
upon these approaches and we introduce a new algorithm called UMA (for
Unconfused Multiclass additive Algorithm) which may be seen as a generalization
to the multiclass setting of the previous approaches. In order to characterize
the noise we use the confusion matrix as a multiclass extension of the
classification noise studied in the aforementioned literature. Theoretically
well-founded, UMA furthermore displays very good empirical noise robustness, as
evidenced by numerical simulations conducted on both synthetic and real data.
Keywords: Multiclass classification, Perceptron, Noisy labels, Confusion Matrix
|
1403.5118 | Robin Lovelace Dr | Robin Lovelace, Nick Malleson, Kirk Harland and Mark Birkin | Geotagged tweets to inform a spatial interaction model: a case study of
museums | A concise version of this article was submitted to GISRUK2014
conference | null | null | null | stat.ME cs.CY cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper explores the potential of volunteered geographical information
from social media for informing geographical models of behavior, based on a
case study of museums in Yorkshire, UK. A spatial interaction model of visitors
to 15 museums from 179 administrative zones is constructed to test this
potential. The main input dataset comprises geo-tagged messages harvested using
the Twitter Streaming Application Programming Interface (API), filtered,
analyzed and aggregated to allow direct comparison with the model's output.
Comparison between model output and tweet information allowed the calibration
of model parameters to optimize the fit between flows to museums inferred from
tweets and flow matrices generated by the spatial interaction model. We
conclude that volunteered geographic information from social media sites have
great potential for informing geographical models of behavior, especially if
the volume of geo-tagged social media messages continues to increase. However,
we caution that volunteered geographical information from social media has some
major limitations so should be used only as a supplement to more consistent
data sources or when official datasets are unavailable.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2014 12:48:24 GMT"
}
] | 2014-03-21T00:00:00 | [
[
"Lovelace",
"Robin",
""
],
[
"Malleson",
"Nick",
""
],
[
"Harland",
"Kirk",
""
],
[
"Birkin",
"Mark",
""
]
] | TITLE: Geotagged tweets to inform a spatial interaction model: a case study of
museums
ABSTRACT: This paper explores the potential of volunteered geographical information
from social media for informing geographical models of behavior, based on a
case study of museums in Yorkshire, UK. A spatial interaction model of visitors
to 15 museums from 179 administrative zones is constructed to test this
potential. The main input dataset comprises geo-tagged messages harvested using
the Twitter Streaming Application Programming Interface (API), filtered,
analyzed and aggregated to allow direct comparison with the model's output.
Comparison between model output and tweet information allowed the calibration
of model parameters to optimize the fit between flows to museums inferred from
tweets and flow matrices generated by the spatial interaction model. We
conclude that volunteered geographic information from social media sites have
great potential for informing geographical models of behavior, especially if
the volume of geo-tagged social media messages continues to increase. However,
we caution that volunteered geographical information from social media has some
major limitations so should be used only as a supplement to more consistent
data sources or when official datasets are unavailable.
|
1208.2908 | Georg Hager | Georg Hager, Jan Treibig, Johannes Habich and Gerhard Wellein | Exploring performance and power properties of modern multicore chips via
simple machine models | 23 pages, 10 figures. Typos corrected, DOI added | null | 10.1002/cpe.3180 | null | cs.PF cs.DC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Modern multicore chips show complex behavior with respect to performance and
power. Starting with the Intel Sandy Bridge processor, it has become possible
to directly measure the power dissipation of a CPU chip and correlate this data
with the performance properties of the running code. Going beyond a simple
bottleneck analysis, we employ the recently published Execution-Cache-Memory
(ECM) model to describe the single- and multi-core performance of streaming
kernels. The model refines the well-known roofline model, since it can predict
the scaling and the saturation behavior of bandwidth-limited loop kernels on a
multicore chip. The saturation point is especially relevant for considerations
of energy consumption. From power dissipation measurements of benchmark
programs with vastly different requirements to the hardware, we derive a
simple, phenomenological power model for the Sandy Bridge processor. Together
with the ECM model, we are able to explain many peculiarities in the
performance and power behavior of multicore processors, and derive guidelines
for energy-efficient execution of parallel programs. Finally, we show that the
ECM and power models can be successfully used to describe the scaling and power
behavior of a lattice-Boltzmann flow solver code.
| [
{
"version": "v1",
"created": "Tue, 14 Aug 2012 15:51:43 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Aug 2012 19:12:39 GMT"
},
{
"version": "v3",
"created": "Tue, 10 Sep 2013 10:17:43 GMT"
},
{
"version": "v4",
"created": "Wed, 19 Mar 2014 13:55:47 GMT"
}
] | 2014-03-20T00:00:00 | [
[
"Hager",
"Georg",
""
],
[
"Treibig",
"Jan",
""
],
[
"Habich",
"Johannes",
""
],
[
"Wellein",
"Gerhard",
""
]
] | TITLE: Exploring performance and power properties of modern multicore chips via
simple machine models
ABSTRACT: Modern multicore chips show complex behavior with respect to performance and
power. Starting with the Intel Sandy Bridge processor, it has become possible
to directly measure the power dissipation of a CPU chip and correlate this data
with the performance properties of the running code. Going beyond a simple
bottleneck analysis, we employ the recently published Execution-Cache-Memory
(ECM) model to describe the single- and multi-core performance of streaming
kernels. The model refines the well-known roofline model, since it can predict
the scaling and the saturation behavior of bandwidth-limited loop kernels on a
multicore chip. The saturation point is especially relevant for considerations
of energy consumption. From power dissipation measurements of benchmark
programs with vastly different requirements to the hardware, we derive a
simple, phenomenological power model for the Sandy Bridge processor. Together
with the ECM model, we are able to explain many peculiarities in the
performance and power behavior of multicore processors, and derive guidelines
for energy-efficient execution of parallel programs. Finally, we show that the
ECM and power models can be successfully used to describe the scaling and power
behavior of a lattice-Boltzmann flow solver code.
|
1403.4415 | Julia Perl | Julia Preusse, J\'er\^ome Kunegis, Matthias Thimm, Sergej Sizov | DecLiNe -- Models for Decay of Links in Networks | null | null | null | null | cs.SI physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The prediction of graph evolution is an important and challenging problem in
the analysis of networks and of the Web in particular. But while the appearance
of new links is part of virtually every model of Web growth, the disappearance
of links has received much less attention in the literature. To fill this gap,
our approach DecLiNe (an acronym for DECay of LInks in NEtworks) aims to
predict link decay in networks, based on structural analysis of corresponding
graph models. In analogy to the link prediction problem, we show that analysis
of graph structures can help to identify indicators for superfluous links under
consideration of common network models. In doing so, we introduce novel metrics
that denote the likelihood of certain links in social graphs to remain in the
network, and combine them with state-of-the-art machine learning methods for
predicting link decay. Our methods are independent of the underlying network
type, and can be applied to such diverse networks as the Web, social networks
and any other structure representable as a network, and can be easily combined
with case-specific content analysis and adopted for a variety of social network
mining, filtering and recommendation applications. In systematic evaluations
with large-scale datasets of Wikipedia we show the practical feasibility of the
proposed structure-based link decay prediction algorithms.
| [
{
"version": "v1",
"created": "Tue, 18 Mar 2014 11:44:36 GMT"
},
{
"version": "v2",
"created": "Wed, 19 Mar 2014 13:03:24 GMT"
}
] | 2014-03-20T00:00:00 | [
[
"Preusse",
"Julia",
""
],
[
"Kunegis",
"Jérôme",
""
],
[
"Thimm",
"Matthias",
""
],
[
"Sizov",
"Sergej",
""
]
] | TITLE: DecLiNe -- Models for Decay of Links in Networks
ABSTRACT: The prediction of graph evolution is an important and challenging problem in
the analysis of networks and of the Web in particular. But while the appearance
of new links is part of virtually every model of Web growth, the disappearance
of links has received much less attention in the literature. To fill this gap,
our approach DecLiNe (an acronym for DECay of LInks in NEtworks) aims to
predict link decay in networks, based on structural analysis of corresponding
graph models. In analogy to the link prediction problem, we show that analysis
of graph structures can help to identify indicators for superfluous links under
consideration of common network models. In doing so, we introduce novel metrics
that denote the likelihood of certain links in social graphs to remain in the
network, and combine them with state-of-the-art machine learning methods for
predicting link decay. Our methods are independent of the underlying network
type, and can be applied to such diverse networks as the Web, social networks
and any other structure representable as a network, and can be easily combined
with case-specific content analysis and adopted for a variety of social network
mining, filtering and recommendation applications. In systematic evaluations
with large-scale datasets of Wikipedia we show the practical feasibility of the
proposed structure-based link decay prediction algorithms.
|
1403.4781 | Subhadip Mukherjee | Subhadip Mukherjee and Chandra Sekhar Seelamantula | A Split-and-Merge Dictionary Learning Algorithm for Sparse
Representation | null | null | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In big data image/video analytics, we encounter the problem of learning an
overcomplete dictionary for sparse representation from a large training
dataset, which can not be processed at once because of storage and
computational constraints. To tackle the problem of dictionary learning in such
scenarios, we propose an algorithm for parallel dictionary learning. The
fundamental idea behind the algorithm is to learn a sparse representation in
two phases. In the first phase, the whole training dataset is partitioned into
small non-overlapping subsets, and a dictionary is trained independently on
each small database. In the second phase, the dictionaries are merged to form a
global dictionary. We show that the proposed algorithm is efficient in its
usage of memory and computational complexity, and performs on par with the
standard learning strategy operating on the entire data at a time. As an
application, we consider the problem of image denoising. We present a
comparative analysis of our algorithm with the standard learning techniques,
that use the entire database at a time, in terms of training and denoising
performance. We observe that the split-and-merge algorithm results in a
remarkable reduction of training time, without significantly affecting the
denoising performance.
| [
{
"version": "v1",
"created": "Wed, 19 Mar 2014 12:16:17 GMT"
}
] | 2014-03-20T00:00:00 | [
[
"Mukherjee",
"Subhadip",
""
],
[
"Seelamantula",
"Chandra Sekhar",
""
]
] | TITLE: A Split-and-Merge Dictionary Learning Algorithm for Sparse
Representation
ABSTRACT: In big data image/video analytics, we encounter the problem of learning an
overcomplete dictionary for sparse representation from a large training
dataset, which can not be processed at once because of storage and
computational constraints. To tackle the problem of dictionary learning in such
scenarios, we propose an algorithm for parallel dictionary learning. The
fundamental idea behind the algorithm is to learn a sparse representation in
two phases. In the first phase, the whole training dataset is partitioned into
small non-overlapping subsets, and a dictionary is trained independently on
each small database. In the second phase, the dictionaries are merged to form a
global dictionary. We show that the proposed algorithm is efficient in its
usage of memory and computational complexity, and performs on par with the
standard learning strategy operating on the entire data at a time. As an
application, we consider the problem of image denoising. We present a
comparative analysis of our algorithm with the standard learning techniques,
that use the entire database at a time, in terms of training and denoising
performance. We observe that the split-and-merge algorithm results in a
remarkable reduction of training time, without significantly affecting the
denoising performance.
|
1403.4540 | Llu\'is Belanche-Mu\~noz | Llu\'is Belanche and Jer\'onimo Hern\'andez | Similarity networks for classification: a case study in the Horse Colic
problem | 16 pages, 1 figure Universitat Polit\`ecnica de Catalunya preprint | null | null | Technical Report LSI-14-4-R | cs.LG cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper develops a two-layer neural network in which the neuron model
computes a user-defined similarity function between inputs and weights. The
neuron transfer function is formed by composition of an adapted logistic
function with the mean of the partial input-weight similarities. The resulting
neuron model is capable of dealing directly with variables of potentially
different nature (continuous, fuzzy, ordinal, categorical). There is also
provision for missing values. The network is trained using a two-stage
procedure very similar to that used to train a radial basis function (RBF)
neural network. The network is compared to two types of RBF networks in a
non-trivial dataset: the Horse Colic problem, taken as a case study and
analyzed in detail.
| [
{
"version": "v1",
"created": "Tue, 18 Mar 2014 17:15:21 GMT"
}
] | 2014-03-19T00:00:00 | [
[
"Belanche",
"Lluís",
""
],
[
"Hernández",
"Jerónimo",
""
]
] | TITLE: Similarity networks for classification: a case study in the Horse Colic
problem
ABSTRACT: This paper develops a two-layer neural network in which the neuron model
computes a user-defined similarity function between inputs and weights. The
neuron transfer function is formed by composition of an adapted logistic
function with the mean of the partial input-weight similarities. The resulting
neuron model is capable of dealing directly with variables of potentially
different nature (continuous, fuzzy, ordinal, categorical). There is also
provision for missing values. The network is trained using a two-stage
procedure very similar to that used to train a radial basis function (RBF)
neural network. The network is compared to two types of RBF networks in a
non-trivial dataset: the Horse Colic problem, taken as a case study and
analyzed in detail.
|
1206.5580 | John Moeller | John Moeller, Parasaran Raman, Avishek Saha, Suresh Venkatasubramanian | A Geometric Algorithm for Scalable Multiple Kernel Learning | 20 pages | null | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a geometric formulation of the Multiple Kernel Learning (MKL)
problem. To do so, we reinterpret the problem of learning kernel weights as
searching for a kernel that maximizes the minimum (kernel) distance between two
convex polytopes. This interpretation combined with novel structural insights
from our geometric formulation allows us to reduce the MKL problem to a simple
optimization routine that yields provable convergence as well as quality
guarantees. As a result our method scales efficiently to much larger data sets
than most prior methods can handle. Empirical evaluation on eleven datasets
shows that we are significantly faster and even compare favorably with a
uniform unweighted combination of kernels.
| [
{
"version": "v1",
"created": "Mon, 25 Jun 2012 05:57:29 GMT"
},
{
"version": "v2",
"created": "Sat, 15 Mar 2014 04:33:18 GMT"
}
] | 2014-03-18T00:00:00 | [
[
"Moeller",
"John",
""
],
[
"Raman",
"Parasaran",
""
],
[
"Saha",
"Avishek",
""
],
[
"Venkatasubramanian",
"Suresh",
""
]
] | TITLE: A Geometric Algorithm for Scalable Multiple Kernel Learning
ABSTRACT: We present a geometric formulation of the Multiple Kernel Learning (MKL)
problem. To do so, we reinterpret the problem of learning kernel weights as
searching for a kernel that maximizes the minimum (kernel) distance between two
convex polytopes. This interpretation combined with novel structural insights
from our geometric formulation allows us to reduce the MKL problem to a simple
optimization routine that yields provable convergence as well as quality
guarantees. As a result our method scales efficiently to much larger data sets
than most prior methods can handle. Empirical evaluation on eleven datasets
shows that we are significantly faster and even compare favorably with a
uniform unweighted combination of kernels.
|
1307.1289 | Miguel Angel Veganzones | Miguel Angel Veganzones (GIPSA), Mihai Datcu (DLR), Manuel Gra\~na
(GIC) | Further results on dissimilarity spaces for hyperspectral images RF-CBIR | In Pattern Recognition Letters (2013) | Pattern Recognition Letters 34, 14 (2013) 1659-1668 | 10.1016/j.patrec.2013.05.025 | veganzones_PRL2013 | cs.IR cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Content-Based Image Retrieval (CBIR) systems are powerful search tools in
image databases that have been little applied to hyperspectral images.
Relevance feedback (RF) is an iterative process that uses machine learning
techniques and user's feedback to improve the CBIR systems performance. We
pursued to expand previous research in hyperspectral CBIR systems built on
dissimilarity functions defined either on spectral and spatial features
extracted by spectral unmixing techniques, or on dictionaries extracted by
dictionary-based compressors. These dissimilarity functions were not suitable
for direct application in common machine learning techniques. We propose to use
a RF general approach based on dissimilarity spaces which is more appropriate
for the application of machine learning algorithms to the hyperspectral
RF-CBIR. We validate the proposed RF method for hyperspectral CBIR systems over
a real hyperspectral dataset.
| [
{
"version": "v1",
"created": "Thu, 4 Jul 2013 11:58:04 GMT"
}
] | 2014-03-18T00:00:00 | [
[
"Veganzones",
"Miguel Angel",
"",
"GIPSA"
],
[
"Datcu",
"Mihai",
"",
"DLR"
],
[
"Graña",
"Manuel",
"",
"GIC"
]
] | TITLE: Further results on dissimilarity spaces for hyperspectral images RF-CBIR
ABSTRACT: Content-Based Image Retrieval (CBIR) systems are powerful search tools in
image databases that have been little applied to hyperspectral images.
Relevance feedback (RF) is an iterative process that uses machine learning
techniques and user's feedback to improve the CBIR systems performance. We
pursued to expand previous research in hyperspectral CBIR systems built on
dissimilarity functions defined either on spectral and spatial features
extracted by spectral unmixing techniques, or on dictionaries extracted by
dictionary-based compressors. These dissimilarity functions were not suitable
for direct application in common machine learning techniques. We propose to use
a RF general approach based on dissimilarity spaces which is more appropriate
for the application of machine learning algorithms to the hyperspectral
RF-CBIR. We validate the proposed RF method for hyperspectral CBIR systems over
a real hyperspectral dataset.
|
1403.3829 | Wei Di | Zixuan Wang, Wei Di, Anurag Bhardwaj, Vignesh Jagadeesh, Robinson
Piramuthu | Geometric VLAD for Large Scale Image Search | 8 pages | null | null | null | cs.CV | http://creativecommons.org/licenses/by/3.0/ | We present a novel compact image descriptor for large scale image search. Our
proposed descriptor - Geometric VLAD (gVLAD) is an extension of VLAD (Vector of
Locally Aggregated Descriptors) that incorporates weak geometry information
into the VLAD framework. The proposed geometry cues are derived as a membership
function over keypoint angles which contain evident and informative information
but yet often discarded. A principled technique for learning the membership
function by clustering angles is also presented. Further, to address the
overhead of iterative codebook training over real-time datasets, a novel
codebook adaptation strategy is outlined. Finally, we demonstrate the efficacy
of proposed gVLAD based retrieval framework where we achieve more than 15%
improvement in mAP over existing benchmarks.
| [
{
"version": "v1",
"created": "Sat, 15 Mar 2014 17:35:26 GMT"
}
] | 2014-03-18T00:00:00 | [
[
"Wang",
"Zixuan",
""
],
[
"Di",
"Wei",
""
],
[
"Bhardwaj",
"Anurag",
""
],
[
"Jagadeesh",
"Vignesh",
""
],
[
"Piramuthu",
"Robinson",
""
]
] | TITLE: Geometric VLAD for Large Scale Image Search
ABSTRACT: We present a novel compact image descriptor for large scale image search. Our
proposed descriptor - Geometric VLAD (gVLAD) is an extension of VLAD (Vector of
Locally Aggregated Descriptors) that incorporates weak geometry information
into the VLAD framework. The proposed geometry cues are derived as a membership
function over keypoint angles which contain evident and informative information
but yet often discarded. A principled technique for learning the membership
function by clustering angles is also presented. Further, to address the
overhead of iterative codebook training over real-time datasets, a novel
codebook adaptation strategy is outlined. Finally, we demonstrate the efficacy
of proposed gVLAD based retrieval framework where we achieve more than 15%
improvement in mAP over existing benchmarks.
|
1403.3996 | Vineeth Kashyap | Vineeth Kashyap, Kyle Dewey, Ethan A. Kuefner, John Wagner, Kevin
Gibbons, John Sarracino, Ben Wiedermann, Ben Hardekopf | JSAI: Designing a Sound, Configurable, and Efficient Static Analyzer for
JavaScript | null | null | null | null | cs.PL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We describe JSAI, an abstract interpreter for JavaScript. JSAI uses novel
abstract domains to compute a reduced product of type inference, pointer
analysis, string analysis, integer and boolean constant propagation, and
control-flow analysis. In addition, JSAI allows for analysis control-flow
sensitivity (i.e., context-, path-, and heap-sensitivity) to be modularly
configured without requiring any changes to the analysis implementation. JSAI
is designed to be provably sound with respect to a specific concrete semantics
for JavaScript, which has been extensively tested against existing
production-quality JavaScript implementations.
We provide a comprehensive evaluation of JSAI's performance and precision
using an extensive benchmark suite. This benchmark suite includes real-world
JavaScript applications, machine-generated JavaScript code via Emscripten, and
browser addons. We use JSAI's configurability to evaluate a large number of
analysis sensitivities (some well-known, some novel) and observe some
surprising results. We believe that JSAI's configurability and its formal
specifications position it as a useful research platform to experiment on novel
sensitivities, abstract domains, and client analyses for JavaScript.
| [
{
"version": "v1",
"created": "Mon, 17 Mar 2014 04:29:25 GMT"
}
] | 2014-03-18T00:00:00 | [
[
"Kashyap",
"Vineeth",
""
],
[
"Dewey",
"Kyle",
""
],
[
"Kuefner",
"Ethan A.",
""
],
[
"Wagner",
"John",
""
],
[
"Gibbons",
"Kevin",
""
],
[
"Sarracino",
"John",
""
],
[
"Wiedermann",
"Ben",
""
],
[
"Hardekopf",
"Ben",
""
]
] | TITLE: JSAI: Designing a Sound, Configurable, and Efficient Static Analyzer for
JavaScript
ABSTRACT: We describe JSAI, an abstract interpreter for JavaScript. JSAI uses novel
abstract domains to compute a reduced product of type inference, pointer
analysis, string analysis, integer and boolean constant propagation, and
control-flow analysis. In addition, JSAI allows for analysis control-flow
sensitivity (i.e., context-, path-, and heap-sensitivity) to be modularly
configured without requiring any changes to the analysis implementation. JSAI
is designed to be provably sound with respect to a specific concrete semantics
for JavaScript, which has been extensively tested against existing
production-quality JavaScript implementations.
We provide a comprehensive evaluation of JSAI's performance and precision
using an extensive benchmark suite. This benchmark suite includes real-world
JavaScript applications, machine-generated JavaScript code via Emscripten, and
browser addons. We use JSAI's configurability to evaluate a large number of
analysis sensitivities (some well-known, some novel) and observe some
surprising results. We believe that JSAI's configurability and its formal
specifications position it as a useful research platform to experiment on novel
sensitivities, abstract domains, and client analyses for JavaScript.
|
1403.4017 | Longqi Yang | Longqi Yang, Yibing Wang, Zhisong Pan and Guyu Hu | Multi-task Feature Selection based Anomaly Detection | 6 pages, 5 figures | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Network anomaly detection is still a vibrant research area. As the fast
growth of network bandwidth and the tremendous traffic on the network, there
arises an extremely challengeable question: How to efficiently and accurately
detect the anomaly on multiple traffic? In multi-task learning, the traffic
consisting of flows at different time periods is considered as a task. Multiple
tasks at different time periods performed simultaneously to detect anomalies.
In this paper, we apply the multi-task feature selection in network anomaly
detection area which provides a powerful method to gather information from
multiple traffic and detect anomalies on it simultaneously. In particular, the
multi-task feature selection includes the well-known l1-norm based feature
selection as a special case given only one task. Moreover, we show that the
multi-task feature selection is more accurate by utilizing more information
simultaneously than the l1-norm based method. At the evaluation stage, we
preprocess the raw data trace from trans-Pacific backbone link between Japan
and the United States, label with anomaly communities, and generate a
248-feature dataset. We show empirically that the multi-task feature selection
outperforms independent l1-norm based feature selection on real traffic
dataset.
| [
{
"version": "v1",
"created": "Mon, 17 Mar 2014 08:04:41 GMT"
}
] | 2014-03-18T00:00:00 | [
[
"Yang",
"Longqi",
""
],
[
"Wang",
"Yibing",
""
],
[
"Pan",
"Zhisong",
""
],
[
"Hu",
"Guyu",
""
]
] | TITLE: Multi-task Feature Selection based Anomaly Detection
ABSTRACT: Network anomaly detection is still a vibrant research area. As the fast
growth of network bandwidth and the tremendous traffic on the network, there
arises an extremely challengeable question: How to efficiently and accurately
detect the anomaly on multiple traffic? In multi-task learning, the traffic
consisting of flows at different time periods is considered as a task. Multiple
tasks at different time periods performed simultaneously to detect anomalies.
In this paper, we apply the multi-task feature selection in network anomaly
detection area which provides a powerful method to gather information from
multiple traffic and detect anomalies on it simultaneously. In particular, the
multi-task feature selection includes the well-known l1-norm based feature
selection as a special case given only one task. Moreover, we show that the
multi-task feature selection is more accurate by utilizing more information
simultaneously than the l1-norm based method. At the evaluation stage, we
preprocess the raw data trace from trans-Pacific backbone link between Japan
and the United States, label with anomaly communities, and generate a
248-feature dataset. We show empirically that the multi-task feature selection
outperforms independent l1-norm based feature selection on real traffic
dataset.
|
1205.1758 | Jonathan Ullman | Justin Thaler, Jonathan Ullman, Salil Vadhan | Faster Algorithms for Privately Releasing Marginals | null | null | null | null | cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study the problem of releasing $k$-way marginals of a database $D \in
(\{0,1\}^d)^n$, while preserving differential privacy. The answer to a $k$-way
marginal query is the fraction of $D$'s records $x \in \{0,1\}^d$ with a given
value in each of a given set of up to $k$ columns. Marginal queries enable a
rich class of statistical analyses of a dataset, and designing efficient
algorithms for privately releasing marginal queries has been identified as an
important open problem in private data analysis (cf. Barak et. al., PODS '07).
We give an algorithm that runs in time $d^{O(\sqrt{k})}$ and releases a
private summary capable of answering any $k$-way marginal query with at most
$\pm .01$ error on every query as long as $n \geq d^{O(\sqrt{k})}$. To our
knowledge, ours is the first algorithm capable of privately releasing marginal
queries with non-trivial worst-case accuracy guarantees in time substantially
smaller than the number of $k$-way marginal queries, which is $d^{\Theta(k)}$
(for $k \ll d$).
| [
{
"version": "v1",
"created": "Tue, 8 May 2012 17:43:11 GMT"
},
{
"version": "v2",
"created": "Mon, 25 Jun 2012 14:59:54 GMT"
},
{
"version": "v3",
"created": "Fri, 14 Mar 2014 13:56:07 GMT"
}
] | 2014-03-17T00:00:00 | [
[
"Thaler",
"Justin",
""
],
[
"Ullman",
"Jonathan",
""
],
[
"Vadhan",
"Salil",
""
]
] | TITLE: Faster Algorithms for Privately Releasing Marginals
ABSTRACT: We study the problem of releasing $k$-way marginals of a database $D \in
(\{0,1\}^d)^n$, while preserving differential privacy. The answer to a $k$-way
marginal query is the fraction of $D$'s records $x \in \{0,1\}^d$ with a given
value in each of a given set of up to $k$ columns. Marginal queries enable a
rich class of statistical analyses of a dataset, and designing efficient
algorithms for privately releasing marginal queries has been identified as an
important open problem in private data analysis (cf. Barak et. al., PODS '07).
We give an algorithm that runs in time $d^{O(\sqrt{k})}$ and releases a
private summary capable of answering any $k$-way marginal query with at most
$\pm .01$ error on every query as long as $n \geq d^{O(\sqrt{k})}$. To our
knowledge, ours is the first algorithm capable of privately releasing marginal
queries with non-trivial worst-case accuracy guarantees in time substantially
smaller than the number of $k$-way marginal queries, which is $d^{\Theta(k)}$
(for $k \ll d$).
|
1304.0869 | Conrad Sanderson | Yongkang Wong, Shaokang Chen, Sandra Mau, Conrad Sanderson, Brian C.
Lovell | Patch-based Probabilistic Image Quality Assessment for Face Selection
and Improved Video-based Face Recognition | null | IEEE Conference on Computer Vision and Pattern Recognition
Workshops (CVPRW), pp. 74-81, 2011 | 10.1109/CVPRW.2011.5981881 | null | cs.CV stat.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In video based face recognition, face images are typically captured over
multiple frames in uncontrolled conditions, where head pose, illumination,
shadowing, motion blur and focus change over the sequence. Additionally,
inaccuracies in face localisation can also introduce scale and alignment
variations. Using all face images, including images of poor quality, can
actually degrade face recognition performance. While one solution it to use
only the "best" subset of images, current face selection techniques are
incapable of simultaneously handling all of the abovementioned issues. We
propose an efficient patch-based face image quality assessment algorithm which
quantifies the similarity of a face image to a probabilistic face model,
representing an "ideal" face. Image characteristics that affect recognition are
taken into account, including variations in geometric alignment (shift,
rotation and scale), sharpness, head pose and cast shadows. Experiments on
FERET and PIE datasets show that the proposed algorithm is able to identify
images which are simultaneously the most frontal, aligned, sharp and well
illuminated. Further experiments on a new video surveillance dataset (termed
ChokePoint) show that the proposed method provides better face subsets than
existing face selection techniques, leading to significant improvements in
recognition accuracy.
| [
{
"version": "v1",
"created": "Wed, 3 Apr 2013 08:41:23 GMT"
},
{
"version": "v2",
"created": "Fri, 14 Mar 2014 15:53:31 GMT"
}
] | 2014-03-17T00:00:00 | [
[
"Wong",
"Yongkang",
""
],
[
"Chen",
"Shaokang",
""
],
[
"Mau",
"Sandra",
""
],
[
"Sanderson",
"Conrad",
""
],
[
"Lovell",
"Brian C.",
""
]
] | TITLE: Patch-based Probabilistic Image Quality Assessment for Face Selection
and Improved Video-based Face Recognition
ABSTRACT: In video based face recognition, face images are typically captured over
multiple frames in uncontrolled conditions, where head pose, illumination,
shadowing, motion blur and focus change over the sequence. Additionally,
inaccuracies in face localisation can also introduce scale and alignment
variations. Using all face images, including images of poor quality, can
actually degrade face recognition performance. While one solution it to use
only the "best" subset of images, current face selection techniques are
incapable of simultaneously handling all of the abovementioned issues. We
propose an efficient patch-based face image quality assessment algorithm which
quantifies the similarity of a face image to a probabilistic face model,
representing an "ideal" face. Image characteristics that affect recognition are
taken into account, including variations in geometric alignment (shift,
rotation and scale), sharpness, head pose and cast shadows. Experiments on
FERET and PIE datasets show that the proposed algorithm is able to identify
images which are simultaneously the most frontal, aligned, sharp and well
illuminated. Further experiments on a new video surveillance dataset (termed
ChokePoint) show that the proposed method provides better face subsets than
existing face selection techniques, leading to significant improvements in
recognition accuracy.
|
1403.3460 | Chi Wang | Chi Wang, Xueqing Liu, Yanglei Song, Jiawei Han | Scalable and Robust Construction of Topical Hierarchies | null | null | null | null | cs.LG cs.CL cs.DB cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Automated generation of high-quality topical hierarchies for a text
collection is a dream problem in knowledge engineering with many valuable
applications. In this paper a scalable and robust algorithm is proposed for
constructing a hierarchy of topics from a text collection. We divide and
conquer the problem using a top-down recursive framework, based on a tensor
orthogonal decomposition technique. We solve a critical challenge to perform
scalable inference for our newly designed hierarchical topic model. Experiments
with various real-world datasets illustrate its ability to generate robust,
high-quality hierarchies efficiently. Our method reduces the time of
construction by several orders of magnitude, and its robust feature renders it
possible for users to interactively revise the hierarchy.
| [
{
"version": "v1",
"created": "Thu, 13 Mar 2014 23:22:21 GMT"
}
] | 2014-03-17T00:00:00 | [
[
"Wang",
"Chi",
""
],
[
"Liu",
"Xueqing",
""
],
[
"Song",
"Yanglei",
""
],
[
"Han",
"Jiawei",
""
]
] | TITLE: Scalable and Robust Construction of Topical Hierarchies
ABSTRACT: Automated generation of high-quality topical hierarchies for a text
collection is a dream problem in knowledge engineering with many valuable
applications. In this paper a scalable and robust algorithm is proposed for
constructing a hierarchy of topics from a text collection. We divide and
conquer the problem using a top-down recursive framework, based on a tensor
orthogonal decomposition technique. We solve a critical challenge to perform
scalable inference for our newly designed hierarchical topic model. Experiments
with various real-world datasets illustrate its ability to generate robust,
high-quality hierarchies efficiently. Our method reduces the time of
construction by several orders of magnitude, and its robust feature renders it
possible for users to interactively revise the hierarchy.
|
1403.3480 | Fan Liang | Fan Liang, Chen Feng, Xiaoyi Lu, Zhiwei Xu | Performance Benefits of DataMPI: A Case Study with BigDataBench | null | null | null | null | cs.PF cs.DC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Apache Hadoop and Spark are gaining prominence in Big Data processing and
analytics. Both of them are widely deployed on Internet companies. On the other
hand, high-performance data analysis requirements are causing academical and
industrial communities to adopt state-of-the-art technologies in HPC to solve
Big Data problems. Recently, we have proposed a key-value pair based
communication library, DataMPI, which is extending MPI to support
Hadoop/Spark-like Big Data Computing jobs. In this paper, we use BigDataBench,
a Big Data benchmark suite, to do comprehensive studies on performance and
resource utilization characterizations of Hadoop, Spark and DataMPI. From our
experiments, we observe that the job execution time of DataMPI has up to 55%
and 39% speedups compared with those of Hadoop and Spark, respectively. Most of
the benefits come from the high-efficiency communication mechanisms in DataMPI.
We also notice that the resource (CPU, memory, disk and network I/O)
utilizations of DataMPI are also more efficient than those of the other two
frameworks.
| [
{
"version": "v1",
"created": "Fri, 14 Mar 2014 03:06:34 GMT"
}
] | 2014-03-17T00:00:00 | [
[
"Liang",
"Fan",
""
],
[
"Feng",
"Chen",
""
],
[
"Lu",
"Xiaoyi",
""
],
[
"Xu",
"Zhiwei",
""
]
] | TITLE: Performance Benefits of DataMPI: A Case Study with BigDataBench
ABSTRACT: Apache Hadoop and Spark are gaining prominence in Big Data processing and
analytics. Both of them are widely deployed on Internet companies. On the other
hand, high-performance data analysis requirements are causing academical and
industrial communities to adopt state-of-the-art technologies in HPC to solve
Big Data problems. Recently, we have proposed a key-value pair based
communication library, DataMPI, which is extending MPI to support
Hadoop/Spark-like Big Data Computing jobs. In this paper, we use BigDataBench,
a Big Data benchmark suite, to do comprehensive studies on performance and
resource utilization characterizations of Hadoop, Spark and DataMPI. From our
experiments, we observe that the job execution time of DataMPI has up to 55%
and 39% speedups compared with those of Hadoop and Spark, respectively. Most of
the benefits come from the high-efficiency communication mechanisms in DataMPI.
We also notice that the resource (CPU, memory, disk and network I/O)
utilizations of DataMPI are also more efficient than those of the other two
frameworks.
|
1403.3628 | Remi Flamary | R\'emi Flamary (LAGRANGE), Nisrine Jrad (GIPSA-lab), Ronald Phlypo
(GIPSA-lab), Marco Congedo (GIPSA-lab), Alain Rakotomamonjy (LITIS) | Mixed-norm Regularization for Brain Decoding | Computational and Mathematical Methods in Medicine (2014)
http://www.hindawi.com/journals/cmmm/ | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This work investigates the use of mixed-norm regularization for sensor
selection in Event-Related Potential (ERP) based Brain-Computer Interfaces
(BCI). The classification problem is cast as a discriminative optimization
framework where sensor selection is induced through the use of mixed-norms.
This framework is extended to the multi-task learning situation where several
similar classification tasks related to different subjects are learned
simultaneously. In this case, multi-task learning helps in leveraging data
scarcity issue yielding to more robust classifiers. For this purpose, we have
introduced a regularizer that induces both sensor selection and classifier
similarities. The different regularization approaches are compared on three ERP
datasets showing the interest of mixed-norm regularization in terms of sensor
selection. The multi-task approaches are evaluated when a small number of
learning examples are available yielding to significant performance
improvements especially for subjects performing poorly.
| [
{
"version": "v1",
"created": "Fri, 14 Mar 2014 16:15:24 GMT"
}
] | 2014-03-17T00:00:00 | [
[
"Flamary",
"Rémi",
"",
"LAGRANGE"
],
[
"Jrad",
"Nisrine",
"",
"GIPSA-lab"
],
[
"Phlypo",
"Ronald",
"",
"GIPSA-lab"
],
[
"Congedo",
"Marco",
"",
"GIPSA-lab"
],
[
"Rakotomamonjy",
"Alain",
"",
"LITIS"
]
] | TITLE: Mixed-norm Regularization for Brain Decoding
ABSTRACT: This work investigates the use of mixed-norm regularization for sensor
selection in Event-Related Potential (ERP) based Brain-Computer Interfaces
(BCI). The classification problem is cast as a discriminative optimization
framework where sensor selection is induced through the use of mixed-norms.
This framework is extended to the multi-task learning situation where several
similar classification tasks related to different subjects are learned
simultaneously. In this case, multi-task learning helps in leveraging data
scarcity issue yielding to more robust classifiers. For this purpose, we have
introduced a regularizer that induces both sensor selection and classifier
similarities. The different regularization approaches are compared on three ERP
datasets showing the interest of mixed-norm regularization in terms of sensor
selection. The multi-task approaches are evaluated when a small number of
learning examples are available yielding to significant performance
improvements especially for subjects performing poorly.
|
1403.2802 | Zhimin Cao | Haoqiang Fan, Zhimin Cao, Yuning Jiang, Qi Yin, Chinchilla Doudou | Learning Deep Face Representation | null | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Face representation is a crucial step of face recognition systems. An optimal
face representation should be discriminative, robust, compact, and very
easy-to-implement. While numerous hand-crafted and learning-based
representations have been proposed, considerable room for improvement is still
present. In this paper, we present a very easy-to-implement deep learning
framework for face representation. Our method bases on a new structure of deep
network (called Pyramid CNN). The proposed Pyramid CNN adopts a
greedy-filter-and-down-sample operation, which enables the training procedure
to be very fast and computation-efficient. In addition, the structure of
Pyramid CNN can naturally incorporate feature sharing across multi-scale face
representations, increasing the discriminative ability of resulting
representation. Our basic network is capable of achieving high recognition
accuracy ($85.8\%$ on LFW benchmark) with only 8 dimension representation. When
extended to feature-sharing Pyramid CNN, our system achieves the
state-of-the-art performance ($97.3\%$) on LFW benchmark. We also introduce a
new benchmark of realistic face images on social network and validate our
proposed representation has a good ability of generalization.
| [
{
"version": "v1",
"created": "Wed, 12 Mar 2014 03:47:18 GMT"
}
] | 2014-03-13T00:00:00 | [
[
"Fan",
"Haoqiang",
""
],
[
"Cao",
"Zhimin",
""
],
[
"Jiang",
"Yuning",
""
],
[
"Yin",
"Qi",
""
],
[
"Doudou",
"Chinchilla",
""
]
] | TITLE: Learning Deep Face Representation
ABSTRACT: Face representation is a crucial step of face recognition systems. An optimal
face representation should be discriminative, robust, compact, and very
easy-to-implement. While numerous hand-crafted and learning-based
representations have been proposed, considerable room for improvement is still
present. In this paper, we present a very easy-to-implement deep learning
framework for face representation. Our method bases on a new structure of deep
network (called Pyramid CNN). The proposed Pyramid CNN adopts a
greedy-filter-and-down-sample operation, which enables the training procedure
to be very fast and computation-efficient. In addition, the structure of
Pyramid CNN can naturally incorporate feature sharing across multi-scale face
representations, increasing the discriminative ability of resulting
representation. Our basic network is capable of achieving high recognition
accuracy ($85.8\%$ on LFW benchmark) with only 8 dimension representation. When
extended to feature-sharing Pyramid CNN, our system achieves the
state-of-the-art performance ($97.3\%$) on LFW benchmark. We also introduce a
new benchmark of realistic face images on social network and validate our
proposed representation has a good ability of generalization.
|
1403.3060 | Chanabasayya Vastrad M | Doreswamy, Chanabasayya M. Vastrad | Non linear Prediction of Antitubercular Activity Of Oxazolines and
Oxazoles derivatives Making Use of Compact TS-Fuzzy models Through Clustering
with orthogonal least sqaure technique and Fuzzy identification system | Published 2013 | null | null | null | cs.CE | http://creativecommons.org/licenses/by/3.0/ | The prediction of uncertain and predictive nonlinear systems is an important
and challenging problem. Fuzzy logic models are often a good choice to describe
such systems however in many cases these become complex soon. commonlly, too
less effort is put into descriptor selection and in the creation of suitable
local rules. Moreover, in common no model reduction is applied, while this may
analyze the model by removing redundant data. This paper suggests a combined
method that deal with these issues in order to create compact Takagi Sugeno
(TS) models that can be effectively used to represent complex predictive
systems. A new fuzzy clustering method is come up with for the identification
of compact TS-fuzzy models. The best relevant consequent variables of the TS
model are choosen by an orthogonal least squares technique based on the
obtained clusters.For the selection of the relevant antecedent (scheduling)
variables a new method has been developed based on Fisher's interclass
separability basis. This complete approach is demonstrated by means of the
Oxazolines and Oxazoles derivatives as antituberculosis agent for nonlinear
regression benchmark. The results are compared with results obtained by
neuro-fuzzy i.e. ANFIS algorithm and advanced fuzzyy clustering techniques i.e
FMID toolbox .
| [
{
"version": "v1",
"created": "Sat, 22 Feb 2014 02:40:40 GMT"
}
] | 2014-03-13T00:00:00 | [
[
"Doreswamy",
"",
""
],
[
"Vastrad",
"Chanabasayya M.",
""
]
] | TITLE: Non linear Prediction of Antitubercular Activity Of Oxazolines and
Oxazoles derivatives Making Use of Compact TS-Fuzzy models Through Clustering
with orthogonal least sqaure technique and Fuzzy identification system
ABSTRACT: The prediction of uncertain and predictive nonlinear systems is an important
and challenging problem. Fuzzy logic models are often a good choice to describe
such systems however in many cases these become complex soon. commonlly, too
less effort is put into descriptor selection and in the creation of suitable
local rules. Moreover, in common no model reduction is applied, while this may
analyze the model by removing redundant data. This paper suggests a combined
method that deal with these issues in order to create compact Takagi Sugeno
(TS) models that can be effectively used to represent complex predictive
systems. A new fuzzy clustering method is come up with for the identification
of compact TS-fuzzy models. The best relevant consequent variables of the TS
model are choosen by an orthogonal least squares technique based on the
obtained clusters.For the selection of the relevant antecedent (scheduling)
variables a new method has been developed based on Fisher's interclass
separability basis. This complete approach is demonstrated by means of the
Oxazolines and Oxazoles derivatives as antituberculosis agent for nonlinear
regression benchmark. The results are compared with results obtained by
neuro-fuzzy i.e. ANFIS algorithm and advanced fuzzyy clustering techniques i.e
FMID toolbox .
|
1403.2372 | Mehdi Naseriparsa | Mehdi Naseriparsa, Amir-Masoud Bidgoli, Touraj Varaee | A Hybrid Feature Selection Method to Improve Performance of a Group of
Classification Algorithms | 8 pages. arXiv admin note: substantial text overlap with
arXiv:1403.1946; and text overlap with arXiv:1106.1813 by other authors | International Journal of Computer Applications,Vol 69,No 17,pp
28-35,2013 | 10.5120/12065-8172 | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper a hybrid feature selection method is proposed which takes
advantages of wrapper subset evaluation with a lower cost and improves the
performance of a group of classifiers. The method uses combination of sample
domain filtering and resampling to refine the sample domain and two feature
subset evaluation methods to select reliable features. This method utilizes
both feature space and sample domain in two phases. The first phase filters and
resamples the sample domain and the second phase adopts a hybrid procedure by
information gain, wrapper subset evaluation and genetic search to find the
optimal feature space. Experiments carried out on different types of datasets
from UCI Repository of Machine Learning databases and the results show a rise
in the average performance of five classifiers (Naive Bayes, Logistic,
Multilayer Perceptron, Best First Decision Tree and JRIP) simultaneously and
the classification error for these classifiers decreases considerably. The
experiments also show that this method outperforms other feature selection
methods with a lower cost.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2014 08:04:29 GMT"
}
] | 2014-03-12T00:00:00 | [
[
"Naseriparsa",
"Mehdi",
""
],
[
"Bidgoli",
"Amir-Masoud",
""
],
[
"Varaee",
"Touraj",
""
]
] | TITLE: A Hybrid Feature Selection Method to Improve Performance of a Group of
Classification Algorithms
ABSTRACT: In this paper a hybrid feature selection method is proposed which takes
advantages of wrapper subset evaluation with a lower cost and improves the
performance of a group of classifiers. The method uses combination of sample
domain filtering and resampling to refine the sample domain and two feature
subset evaluation methods to select reliable features. This method utilizes
both feature space and sample domain in two phases. The first phase filters and
resamples the sample domain and the second phase adopts a hybrid procedure by
information gain, wrapper subset evaluation and genetic search to find the
optimal feature space. Experiments carried out on different types of datasets
from UCI Repository of Machine Learning databases and the results show a rise
in the average performance of five classifiers (Naive Bayes, Logistic,
Multilayer Perceptron, Best First Decision Tree and JRIP) simultaneously and
the classification error for these classifiers decreases considerably. The
experiments also show that this method outperforms other feature selection
methods with a lower cost.
|
1403.2404 | Long Cheng | Long Cheng, Avinash Malik, Spyros Kotoulas, Tomas E Ward, Georgios
Theodoropoulos | Scalable RDF Data Compression using X10 | null | null | null | null | cs.DC cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The Semantic Web comprises enormous volumes of semi-structured data elements.
For interoperability, these elements are represented by long strings. Such
representations are not efficient for the purposes of Semantic Web applications
that perform computations over large volumes of information. A typical method
for alleviating the impact of this problem is through the use of compression
methods that produce more compact representations of the data. The use of
dictionary encoding for this purpose is particularly prevalent in Semantic Web
database systems. However, centralized implementations present performance
bottlenecks, giving rise to the need for scalable, efficient distributed
encoding schemes. In this paper, we describe an encoding implementation based
on the asynchronous partitioned global address space (APGAS) parallel
programming model. We evaluate performance on a cluster of up to 384 cores and
datasets of up to 11 billion triples (1.9 TB). Compared to the state-of-art
MapReduce algorithm, we demonstrate a speedup of 2.6-7.4x and excellent
scalability. These results illustrate the strong potential of the APGAS model
for efficient implementation of dictionary encoding and contributes to the
engineering of larger scale Semantic Web applications.
| [
{
"version": "v1",
"created": "Mon, 10 Mar 2014 20:48:08 GMT"
}
] | 2014-03-12T00:00:00 | [
[
"Cheng",
"Long",
""
],
[
"Malik",
"Avinash",
""
],
[
"Kotoulas",
"Spyros",
""
],
[
"Ward",
"Tomas E",
""
],
[
"Theodoropoulos",
"Georgios",
""
]
] | TITLE: Scalable RDF Data Compression using X10
ABSTRACT: The Semantic Web comprises enormous volumes of semi-structured data elements.
For interoperability, these elements are represented by long strings. Such
representations are not efficient for the purposes of Semantic Web applications
that perform computations over large volumes of information. A typical method
for alleviating the impact of this problem is through the use of compression
methods that produce more compact representations of the data. The use of
dictionary encoding for this purpose is particularly prevalent in Semantic Web
database systems. However, centralized implementations present performance
bottlenecks, giving rise to the need for scalable, efficient distributed
encoding schemes. In this paper, we describe an encoding implementation based
on the asynchronous partitioned global address space (APGAS) parallel
programming model. We evaluate performance on a cluster of up to 384 cores and
datasets of up to 11 billion triples (1.9 TB). Compared to the state-of-art
MapReduce algorithm, we demonstrate a speedup of 2.6-7.4x and excellent
scalability. These results illustrate the strong potential of the APGAS model
for efficient implementation of dictionary encoding and contributes to the
engineering of larger scale Semantic Web applications.
|
1309.2074 | Qiang Qiu | Qiang Qiu, Guillermo Sapiro | Learning Transformations for Clustering and Classification | arXiv admin note: substantial text overlap with arXiv:1308.0273,
arXiv:1308.0275 | null | null | null | cs.CV cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A low-rank transformation learning framework for subspace clustering and
classification is here proposed. Many high-dimensional data, such as face
images and motion sequences, approximately lie in a union of low-dimensional
subspaces. The corresponding subspace clustering problem has been extensively
studied in the literature to partition such high-dimensional data into clusters
corresponding to their underlying low-dimensional subspaces. However,
low-dimensional intrinsic structures are often violated for real-world
observations, as they can be corrupted by errors or deviate from ideal models.
We propose to address this by learning a linear transformation on subspaces
using matrix rank, via its convex surrogate nuclear norm, as the optimization
criteria. The learned linear transformation restores a low-rank structure for
data from the same subspace, and, at the same time, forces a a maximally
separated structure for data from different subspaces. In this way, we reduce
variations within subspaces, and increase separation between subspaces for a
more robust subspace clustering. This proposed learned robust subspace
clustering framework significantly enhances the performance of existing
subspace clustering methods. Basic theoretical results here presented help to
further support the underlying framework. To exploit the low-rank structures of
the transformed subspaces, we further introduce a fast subspace clustering
technique, which efficiently combines robust PCA with sparse modeling. When
class labels are present at the training stage, we show this low-rank
transformation framework also significantly enhances classification
performance. Extensive experiments using public datasets are presented, showing
that the proposed approach significantly outperforms state-of-the-art methods
for subspace clustering and classification.
| [
{
"version": "v1",
"created": "Mon, 9 Sep 2013 09:16:02 GMT"
},
{
"version": "v2",
"created": "Sun, 9 Mar 2014 18:50:35 GMT"
}
] | 2014-03-11T00:00:00 | [
[
"Qiu",
"Qiang",
""
],
[
"Sapiro",
"Guillermo",
""
]
] | TITLE: Learning Transformations for Clustering and Classification
ABSTRACT: A low-rank transformation learning framework for subspace clustering and
classification is here proposed. Many high-dimensional data, such as face
images and motion sequences, approximately lie in a union of low-dimensional
subspaces. The corresponding subspace clustering problem has been extensively
studied in the literature to partition such high-dimensional data into clusters
corresponding to their underlying low-dimensional subspaces. However,
low-dimensional intrinsic structures are often violated for real-world
observations, as they can be corrupted by errors or deviate from ideal models.
We propose to address this by learning a linear transformation on subspaces
using matrix rank, via its convex surrogate nuclear norm, as the optimization
criteria. The learned linear transformation restores a low-rank structure for
data from the same subspace, and, at the same time, forces a a maximally
separated structure for data from different subspaces. In this way, we reduce
variations within subspaces, and increase separation between subspaces for a
more robust subspace clustering. This proposed learned robust subspace
clustering framework significantly enhances the performance of existing
subspace clustering methods. Basic theoretical results here presented help to
further support the underlying framework. To exploit the low-rank structures of
the transformed subspaces, we further introduce a fast subspace clustering
technique, which efficiently combines robust PCA with sparse modeling. When
class labels are present at the training stage, we show this low-rank
transformation framework also significantly enhances classification
performance. Extensive experiments using public datasets are presented, showing
that the proposed approach significantly outperforms state-of-the-art methods
for subspace clustering and classification.
|
1312.4314 | David Eigen | David Eigen, Marc'Aurelio Ranzato, Ilya Sutskever | Learning Factored Representations in a Deep Mixture of Experts | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Mixtures of Experts combine the outputs of several "expert" networks, each of
which specializes in a different part of the input space. This is achieved by
training a "gating" network that maps each input to a distribution over the
experts. Such models show promise for building larger networks that are still
cheap to compute at test time, and more parallelizable at training time. In
this this work, we extend the Mixture of Experts to a stacked model, the Deep
Mixture of Experts, with multiple sets of gating and experts. This
exponentially increases the number of effective experts by associating each
input with a combination of experts at each layer, yet maintains a modest model
size. On a randomly translated version of the MNIST dataset, we find that the
Deep Mixture of Experts automatically learns to develop location-dependent
("where") experts at the first layer, and class-specific ("what") experts at
the second layer. In addition, we see that the different combinations are in
use when the model is applied to a dataset of speech monophones. These
demonstrate effective use of all expert combinations.
| [
{
"version": "v1",
"created": "Mon, 16 Dec 2013 11:15:10 GMT"
},
{
"version": "v2",
"created": "Wed, 19 Feb 2014 17:57:53 GMT"
},
{
"version": "v3",
"created": "Sun, 9 Mar 2014 20:15:03 GMT"
}
] | 2014-03-11T00:00:00 | [
[
"Eigen",
"David",
""
],
[
"Ranzato",
"Marc'Aurelio",
""
],
[
"Sutskever",
"Ilya",
""
]
] | TITLE: Learning Factored Representations in a Deep Mixture of Experts
ABSTRACT: Mixtures of Experts combine the outputs of several "expert" networks, each of
which specializes in a different part of the input space. This is achieved by
training a "gating" network that maps each input to a distribution over the
experts. Such models show promise for building larger networks that are still
cheap to compute at test time, and more parallelizable at training time. In
this this work, we extend the Mixture of Experts to a stacked model, the Deep
Mixture of Experts, with multiple sets of gating and experts. This
exponentially increases the number of effective experts by associating each
input with a combination of experts at each layer, yet maintains a modest model
size. On a randomly translated version of the MNIST dataset, we find that the
Deep Mixture of Experts automatically learns to develop location-dependent
("where") experts at the first layer, and class-specific ("what") experts at
the second layer. In addition, we see that the different combinations are in
use when the model is applied to a dataset of speech monophones. These
demonstrate effective use of all expert combinations.
|
1401.0509 | Yann Dauphin | Yann N. Dauphin, Gokhan Tur, Dilek Hakkani-Tur, Larry Heck | Zero-Shot Learning for Semantic Utterance Classification | null | null | null | null | cs.CL cs.LG | http://creativecommons.org/licenses/by/3.0/ | We propose a novel zero-shot learning method for semantic utterance
classification (SUC). It learns a classifier $f: X \to Y$ for problems where
none of the semantic categories $Y$ are present in the training set. The
framework uncovers the link between categories and utterances using a semantic
space. We show that this semantic space can be learned by deep neural networks
trained on large amounts of search engine query log data. More precisely, we
propose a novel method that can learn discriminative semantic features without
supervision. It uses the zero-shot learning framework to guide the learning of
the semantic features. We demonstrate the effectiveness of the zero-shot
semantic learning algorithm on the SUC dataset collected by (Tur, 2012).
Furthermore, we achieve state-of-the-art results by combining the semantic
features with a supervised method.
| [
{
"version": "v1",
"created": "Fri, 20 Dec 2013 17:08:26 GMT"
},
{
"version": "v2",
"created": "Tue, 18 Feb 2014 20:34:08 GMT"
},
{
"version": "v3",
"created": "Fri, 7 Mar 2014 23:31:02 GMT"
}
] | 2014-03-11T00:00:00 | [
[
"Dauphin",
"Yann N.",
""
],
[
"Tur",
"Gokhan",
""
],
[
"Hakkani-Tur",
"Dilek",
""
],
[
"Heck",
"Larry",
""
]
] | TITLE: Zero-Shot Learning for Semantic Utterance Classification
ABSTRACT: We propose a novel zero-shot learning method for semantic utterance
classification (SUC). It learns a classifier $f: X \to Y$ for problems where
none of the semantic categories $Y$ are present in the training set. The
framework uncovers the link between categories and utterances using a semantic
space. We show that this semantic space can be learned by deep neural networks
trained on large amounts of search engine query log data. More precisely, we
propose a novel method that can learn discriminative semantic features without
supervision. It uses the zero-shot learning framework to guide the learning of
the semantic features. We demonstrate the effectiveness of the zero-shot
semantic learning algorithm on the SUC dataset collected by (Tur, 2012).
Furthermore, we achieve state-of-the-art results by combining the semantic
features with a supervised method.
|
1403.1946 | Mehdi Naseriparsa | Mehdi Naseriparsa, Amir-masoud Bidgoli, Touraj Varaee | Improving Performance of a Group of Classification Algorithms Using
Resampling and Feature Selection | 7 pages | World of Computer Science and Information Technology Journal,Vol
3, No 4,pp 70-76,2013 | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In recent years the importance of finding a meaningful pattern from huge
datasets has become more challenging. Data miners try to adopt innovative
methods to face this problem by applying feature selection methods. In this
paper we propose a new hybrid method in which we use a combination of
resampling, filtering the sample domain and wrapper subset evaluation method
with genetic search to reduce dimensions of Lung-Cancer dataset that we
received from UCI Repository of Machine Learning databases. Finally, we apply
some well- known classification algorithms (Na\"ive Bayes, Logistic, Multilayer
Perceptron, Best First Decision Tree and JRIP) to the resulting dataset and
compare the results and prediction rates before and after the application of
our feature selection method on that dataset. The results show a substantial
progress in the average performance of five classification algorithms
simultaneously and the classification error for these classifiers decreases
considerably. The experiments also show that this method outperforms other
feature selection methods with a lower cost.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2014 07:47:44 GMT"
}
] | 2014-03-11T00:00:00 | [
[
"Naseriparsa",
"Mehdi",
""
],
[
"Bidgoli",
"Amir-masoud",
""
],
[
"Varaee",
"Touraj",
""
]
] | TITLE: Improving Performance of a Group of Classification Algorithms Using
Resampling and Feature Selection
ABSTRACT: In recent years the importance of finding a meaningful pattern from huge
datasets has become more challenging. Data miners try to adopt innovative
methods to face this problem by applying feature selection methods. In this
paper we propose a new hybrid method in which we use a combination of
resampling, filtering the sample domain and wrapper subset evaluation method
with genetic search to reduce dimensions of Lung-Cancer dataset that we
received from UCI Repository of Machine Learning databases. Finally, we apply
some well- known classification algorithms (Na\"ive Bayes, Logistic, Multilayer
Perceptron, Best First Decision Tree and JRIP) to the resulting dataset and
compare the results and prediction rates before and after the application of
our feature selection method on that dataset. The results show a substantial
progress in the average performance of five classification algorithms
simultaneously and the classification error for these classifiers decreases
considerably. The experiments also show that this method outperforms other
feature selection methods with a lower cost.
|
1403.1949 | Mehdi Naseriparsa | Mehdi Naseriparsa, Mohammad Mansour Riahi Kashani | Combination of PCA with SMOTE Resampling to Boost the Prediction Rate in
Lung Cancer Dataset | 6 pages. arXiv admin note: text overlap with arXiv:1106.1813,
arXiv:1001.1446 by other authors | International Journal of Computer Applications,Vol 77,No 3,pp
33-38,2013 | 10.5120/13376-0987 | null | cs.LG cs.CE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Classification algorithms are unable to make reliable models on the datasets
with huge sizes. These datasets contain many irrelevant and redundant features
that mislead the classifiers. Furthermore, many huge datasets have imbalanced
class distribution which leads to bias over majority class in the
classification process. In this paper combination of unsupervised
dimensionality reduction methods with resampling is proposed and the results
are tested on Lung-Cancer dataset. In the first step PCA is applied on
Lung-Cancer dataset to compact the dataset and eliminate irrelevant features
and in the second step SMOTE resampling is carried out to balance the class
distribution and increase the variety of sample domain. Finally, Naive Bayes
classifier is applied on the resulting dataset and the results are compared and
evaluation metrics are calculated. The experiments show the effectiveness of
the proposed method across four evaluation metrics: Overall accuracy, False
Positive Rate, Precision, Recall.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2014 08:12:54 GMT"
}
] | 2014-03-11T00:00:00 | [
[
"Naseriparsa",
"Mehdi",
""
],
[
"Kashani",
"Mohammad Mansour Riahi",
""
]
] | TITLE: Combination of PCA with SMOTE Resampling to Boost the Prediction Rate in
Lung Cancer Dataset
ABSTRACT: Classification algorithms are unable to make reliable models on the datasets
with huge sizes. These datasets contain many irrelevant and redundant features
that mislead the classifiers. Furthermore, many huge datasets have imbalanced
class distribution which leads to bias over majority class in the
classification process. In this paper combination of unsupervised
dimensionality reduction methods with resampling is proposed and the results
are tested on Lung-Cancer dataset. In the first step PCA is applied on
Lung-Cancer dataset to compact the dataset and eliminate irrelevant features
and in the second step SMOTE resampling is carried out to balance the class
distribution and increase the variety of sample domain. Finally, Naive Bayes
classifier is applied on the resulting dataset and the results are compared and
evaluation metrics are calculated. The experiments show the effectiveness of
the proposed method across four evaluation metrics: Overall accuracy, False
Positive Rate, Precision, Recall.
|
1403.2006 | Morteza Yousefi Kharaji | Morteza Yousefi Kharaji, Fatemeh Salehi Rizi | An IAC Approach for Detecting Profile Cloning in Online Social Networks | null | null | null | null | cs.SI cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Nowadays, Online Social Networks are popular websites on the internet, which
millions of users register on and share their own personal information with
others. Privacy threats and disclosing personal information are the most
important concerns of OSNs users. Recently, a new attack which is named
Identity Cloned Attack is detected on OSNs. In this attack the attacker tries
to make a fake identity of a real user in order to access to private
information of the users friends which they do not publish on the public
profiles. In today OSNs, there are some verification services, but they are not
active services and they are useful for users who are familiar with online
identity issues. In this paper, Identity cloned attacks are explained in more
details and a new and precise method to detect profile cloning in online social
networks is proposed. In this method, first, the social network is shown in a
form of graph, then, according to similarities among users, this graph is
divided into smaller communities. Afterwards, all of the similar profiles to
the real profile are gathered (from the same community), then strength of
relationship (among all selected profiles and the real profile) is calculated,
and those which have the less strength of relationship will be verified by
mutual friend system. In this study, in order to evaluate the effectiveness of
proposed method, all steps are applied on a dataset of Facebook, and finally
this work is compared with two previous works by applying them on the dataset.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2014 20:38:57 GMT"
}
] | 2014-03-11T00:00:00 | [
[
"Kharaji",
"Morteza Yousefi",
""
],
[
"Rizi",
"Fatemeh Salehi",
""
]
] | TITLE: An IAC Approach for Detecting Profile Cloning in Online Social Networks
ABSTRACT: Nowadays, Online Social Networks are popular websites on the internet, which
millions of users register on and share their own personal information with
others. Privacy threats and disclosing personal information are the most
important concerns of OSNs users. Recently, a new attack which is named
Identity Cloned Attack is detected on OSNs. In this attack the attacker tries
to make a fake identity of a real user in order to access to private
information of the users friends which they do not publish on the public
profiles. In today OSNs, there are some verification services, but they are not
active services and they are useful for users who are familiar with online
identity issues. In this paper, Identity cloned attacks are explained in more
details and a new and precise method to detect profile cloning in online social
networks is proposed. In this method, first, the social network is shown in a
form of graph, then, according to similarities among users, this graph is
divided into smaller communities. Afterwards, all of the similar profiles to
the real profile are gathered (from the same community), then strength of
relationship (among all selected profiles and the real profile) is calculated,
and those which have the less strength of relationship will be verified by
mutual friend system. In this study, in order to evaluate the effectiveness of
proposed method, all steps are applied on a dataset of Facebook, and finally
this work is compared with two previous works by applying them on the dataset.
|
1403.2024 | Pin-Yu Chen | Pin-Yu Chen and Alfred O. Hero III | Node Removal Vulnerability of the Largest Component of a Network | Published in IEEE GlobalSIP 2013 | null | null | null | cs.SI cs.NI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The connectivity structure of a network can be very sensitive to removal of
certain nodes in the network. In this paper, we study the sensitivity of the
largest component size to node removals. We prove that minimizing the largest
component size is equivalent to solving a matrix one-norm minimization problem
whose column vectors are orthogonal and sparse and they form a basis of the
null space of the associated graph Laplacian matrix. A greedy node removal
algorithm is then proposed based on the matrix one-norm minimization. In
comparison with other node centralities such as node degree and betweenness,
experimental results on US power grid dataset validate the effectiveness of the
proposed approach in terms of reduction of the largest component size with
relatively few node removals.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2014 02:52:51 GMT"
}
] | 2014-03-11T00:00:00 | [
[
"Chen",
"Pin-Yu",
""
],
[
"Hero",
"Alfred O.",
"III"
]
] | TITLE: Node Removal Vulnerability of the Largest Component of a Network
ABSTRACT: The connectivity structure of a network can be very sensitive to removal of
certain nodes in the network. In this paper, we study the sensitivity of the
largest component size to node removals. We prove that minimizing the largest
component size is equivalent to solving a matrix one-norm minimization problem
whose column vectors are orthogonal and sparse and they form a basis of the
null space of the associated graph Laplacian matrix. A greedy node removal
algorithm is then proposed based on the matrix one-norm minimization. In
comparison with other node centralities such as node degree and betweenness,
experimental results on US power grid dataset validate the effectiveness of the
proposed approach in terms of reduction of the largest component size with
relatively few node removals.
|
1310.2125 | Ritabrata Dutta | Ritabrata Dutta and Sohan Seth and Samuel Kaski | Retrieval of Experiments with Sequential Dirichlet Process Mixtures in
Model Space | null | null | null | null | stat.ML cs.IR stat.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We address the problem of retrieving relevant experiments given a query
experiment, motivated by the public databases of datasets in molecular biology
and other experimental sciences, and the need of scientists to relate to
earlier work on the level of actual measurement data. Since experiments are
inherently noisy and databases ever accumulating, we argue that a retrieval
engine should possess two particular characteristics. First, it should compare
models learnt from the experiments rather than the raw measurements themselves:
this allows incorporating experiment-specific prior knowledge to suppress noise
effects and focus on what is important. Second, it should be updated
sequentially from newly published experiments, without explicitly storing
either the measurements or the models, which is critical for saving storage
space and protecting data privacy: this promotes life long learning. We
formulate the retrieval as a ``supermodelling'' problem, of sequentially
learning a model of the set of posterior distributions, represented as sets of
MCMC samples, and suggest the use of Particle-Learning-based sequential
Dirichlet process mixture (DPM) for this purpose. The relevance measure for
retrieval is derived from the supermodel through the mixture representation. We
demonstrate the performance of the proposed retrieval method on simulated data
and molecular biological experiments.
| [
{
"version": "v1",
"created": "Tue, 8 Oct 2013 13:10:26 GMT"
},
{
"version": "v2",
"created": "Thu, 6 Mar 2014 22:04:33 GMT"
}
] | 2014-03-10T00:00:00 | [
[
"Dutta",
"Ritabrata",
""
],
[
"Seth",
"Sohan",
""
],
[
"Kaski",
"Samuel",
""
]
] | TITLE: Retrieval of Experiments with Sequential Dirichlet Process Mixtures in
Model Space
ABSTRACT: We address the problem of retrieving relevant experiments given a query
experiment, motivated by the public databases of datasets in molecular biology
and other experimental sciences, and the need of scientists to relate to
earlier work on the level of actual measurement data. Since experiments are
inherently noisy and databases ever accumulating, we argue that a retrieval
engine should possess two particular characteristics. First, it should compare
models learnt from the experiments rather than the raw measurements themselves:
this allows incorporating experiment-specific prior knowledge to suppress noise
effects and focus on what is important. Second, it should be updated
sequentially from newly published experiments, without explicitly storing
either the measurements or the models, which is critical for saving storage
space and protecting data privacy: this promotes life long learning. We
formulate the retrieval as a ``supermodelling'' problem, of sequentially
learning a model of the set of posterior distributions, represented as sets of
MCMC samples, and suggest the use of Particle-Learning-based sequential
Dirichlet process mixture (DPM) for this purpose. The relevance measure for
retrieval is derived from the supermodel through the mixture representation. We
demonstrate the performance of the proposed retrieval method on simulated data
and molecular biological experiments.
|
1403.1600 | Kai Zhu | Kai Zhu, Rui Wu, Lei Ying, R. Srikant | Collaborative Filtering with Information-Rich and Information-Sparse
Entities | null | null | null | null | stat.ML cs.IT cs.LG math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we consider a popular model for collaborative filtering in
recommender systems where some users of a website rate some items, such as
movies, and the goal is to recover the ratings of some or all of the unrated
items of each user. In particular, we consider both the clustering model, where
only users (or items) are clustered, and the co-clustering model, where both
users and items are clustered, and further, we assume that some users rate many
items (information-rich users) and some users rate only a few items
(information-sparse users). When users (or items) are clustered, our algorithm
can recover the rating matrix with $\omega(MK \log M)$ noisy entries while $MK$
entries are necessary, where $K$ is the number of clusters and $M$ is the
number of items. In the case of co-clustering, we prove that $K^2$ entries are
necessary for recovering the rating matrix, and our algorithm achieves this
lower bound within a logarithmic factor when $K$ is sufficiently large. We
compare our algorithms with a well-known algorithms called alternating
minimization (AM), and a similarity score-based algorithm known as the
popularity-among-friends (PAF) algorithm by applying all three to the MovieLens
and Netflix data sets. Our co-clustering algorithm and AM have similar overall
error rates when recovering the rating matrix, both of which are lower than the
error rate under PAF. But more importantly, the error rate of our co-clustering
algorithm is significantly lower than AM and PAF in the scenarios of interest
in recommender systems: when recommending a few items to each user or when
recommending items to users who only rated a few items (these users are the
majority of the total user population). The performance difference increases
even more when noise is added to the datasets.
| [
{
"version": "v1",
"created": "Thu, 6 Mar 2014 21:51:48 GMT"
}
] | 2014-03-10T00:00:00 | [
[
"Zhu",
"Kai",
""
],
[
"Wu",
"Rui",
""
],
[
"Ying",
"Lei",
""
],
[
"Srikant",
"R.",
""
]
] | TITLE: Collaborative Filtering with Information-Rich and Information-Sparse
Entities
ABSTRACT: In this paper, we consider a popular model for collaborative filtering in
recommender systems where some users of a website rate some items, such as
movies, and the goal is to recover the ratings of some or all of the unrated
items of each user. In particular, we consider both the clustering model, where
only users (or items) are clustered, and the co-clustering model, where both
users and items are clustered, and further, we assume that some users rate many
items (information-rich users) and some users rate only a few items
(information-sparse users). When users (or items) are clustered, our algorithm
can recover the rating matrix with $\omega(MK \log M)$ noisy entries while $MK$
entries are necessary, where $K$ is the number of clusters and $M$ is the
number of items. In the case of co-clustering, we prove that $K^2$ entries are
necessary for recovering the rating matrix, and our algorithm achieves this
lower bound within a logarithmic factor when $K$ is sufficiently large. We
compare our algorithms with a well-known algorithms called alternating
minimization (AM), and a similarity score-based algorithm known as the
popularity-among-friends (PAF) algorithm by applying all three to the MovieLens
and Netflix data sets. Our co-clustering algorithm and AM have similar overall
error rates when recovering the rating matrix, both of which are lower than the
error rate under PAF. But more importantly, the error rate of our co-clustering
algorithm is significantly lower than AM and PAF in the scenarios of interest
in recommender systems: when recommending a few items to each user or when
recommending items to users who only rated a few items (these users are the
majority of the total user population). The performance difference increases
even more when noise is added to the datasets.
|
1403.1706 | Johannes K\"oster | Johannes K\"oster, Sven Rahmann | Massively parallel read mapping on GPUs with PEANUT | null | null | null | null | cs.DS cs.DC q-bio.QM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present PEANUT (ParallEl AligNment UTility), a highly parallel GPU-based
read mapper with several distinguishing features, including a novel q-gram
index (called the q-group index) with small memory footprint built on-the-fly
over the reads and the possibility to output both the best hits or all hits of
a read. Designing the algorithm particularly for the GPU architecture, we were
able to reach maximum core occupancy for several key steps. Our benchmarks show
that PEANUT outperforms other state-of- the-art mappers in terms of speed and
sensitivity. The software is available at http://peanut.readthedocs.org.
| [
{
"version": "v1",
"created": "Fri, 7 Mar 2014 10:10:07 GMT"
}
] | 2014-03-10T00:00:00 | [
[
"Köster",
"Johannes",
""
],
[
"Rahmann",
"Sven",
""
]
] | TITLE: Massively parallel read mapping on GPUs with PEANUT
ABSTRACT: We present PEANUT (ParallEl AligNment UTility), a highly parallel GPU-based
read mapper with several distinguishing features, including a novel q-gram
index (called the q-group index) with small memory footprint built on-the-fly
over the reads and the possibility to output both the best hits or all hits of
a read. Designing the algorithm particularly for the GPU architecture, we were
able to reach maximum core occupancy for several key steps. Our benchmarks show
that PEANUT outperforms other state-of- the-art mappers in terms of speed and
sensitivity. The software is available at http://peanut.readthedocs.org.
|
1403.1749 | Saurabh Joshi | Saurabh Joshi and Akash Lal | Automatically finding atomic regions for fixing bugs in Concurrent
programs | 16 pages, 6 figures, 1 table | null | null | null | cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents a technique for automatically constructing a fix for
buggy concurrent programs: given a concurrent program that does not satisfy
user-provided assertions, we infer atomic blocks that fix the program. An
atomic block protects a piece of code and ensures that it runs without
interruption from other threads. Our technique uses a verification tool as a
subroutine to find the smallest atomic regions that remove all bugs in a given
program. Keeping the atomic regions small allows for maximum concurrency. We
have implemented our approach in a tool called AtomicInf. A user of AtomicInf
can choose between strong and weak atomicity semantics for the inferred fix.
While the former is simpler to find, the latter provides more information about
the bugs that got fixed.
We ran AtomicInf on several benchmarks and came up with the smallest and the
most precise atomic regions in all of them. We implemented an earlier technique
to our setting and observed that AtomicInf is 1.7 times faster on an average as
compared to an earlier approach.
| [
{
"version": "v1",
"created": "Fri, 7 Mar 2014 13:46:25 GMT"
}
] | 2014-03-10T00:00:00 | [
[
"Joshi",
"Saurabh",
""
],
[
"Lal",
"Akash",
""
]
] | TITLE: Automatically finding atomic regions for fixing bugs in Concurrent
programs
ABSTRACT: This paper presents a technique for automatically constructing a fix for
buggy concurrent programs: given a concurrent program that does not satisfy
user-provided assertions, we infer atomic blocks that fix the program. An
atomic block protects a piece of code and ensures that it runs without
interruption from other threads. Our technique uses a verification tool as a
subroutine to find the smallest atomic regions that remove all bugs in a given
program. Keeping the atomic regions small allows for maximum concurrency. We
have implemented our approach in a tool called AtomicInf. A user of AtomicInf
can choose between strong and weak atomicity semantics for the inferred fix.
While the former is simpler to find, the latter provides more information about
the bugs that got fixed.
We ran AtomicInf on several benchmarks and came up with the smallest and the
most precise atomic regions in all of them. We implemented an earlier technique
to our setting and observed that AtomicInf is 1.7 times faster on an average as
compared to an earlier approach.
|
1311.6091 | Jianshu Chen | Jianshu Chen and Li Deng | A Primal-Dual Method for Training Recurrent Neural Networks Constrained
by the Echo-State Property | null | null | null | null | cs.LG cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present an architecture of a recurrent neural network (RNN) with a
fully-connected deep neural network (DNN) as its feature extractor. The RNN is
equipped with both causal temporal prediction and non-causal look-ahead, via
auto-regression (AR) and moving-average (MA), respectively. The focus of this
paper is a primal-dual training method that formulates the learning of the RNN
as a formal optimization problem with an inequality constraint that provides a
sufficient condition for the stability of the network dynamics. Experimental
results demonstrate the effectiveness of this new method, which achieves 18.86%
phone recognition error on the TIMIT benchmark for the core test set. The
result approaches the best result of 17.7%, which was obtained by using RNN
with long short-term memory (LSTM). The results also show that the proposed
primal-dual training method produces lower recognition errors than the popular
RNN methods developed earlier based on the carefully tuned threshold parameter
that heuristically prevents the gradient from exploding.
| [
{
"version": "v1",
"created": "Sun, 24 Nov 2013 08:04:41 GMT"
},
{
"version": "v2",
"created": "Thu, 19 Dec 2013 09:12:55 GMT"
},
{
"version": "v3",
"created": "Thu, 6 Mar 2014 03:06:36 GMT"
}
] | 2014-03-07T00:00:00 | [
[
"Chen",
"Jianshu",
""
],
[
"Deng",
"Li",
""
]
] | TITLE: A Primal-Dual Method for Training Recurrent Neural Networks Constrained
by the Echo-State Property
ABSTRACT: We present an architecture of a recurrent neural network (RNN) with a
fully-connected deep neural network (DNN) as its feature extractor. The RNN is
equipped with both causal temporal prediction and non-causal look-ahead, via
auto-regression (AR) and moving-average (MA), respectively. The focus of this
paper is a primal-dual training method that formulates the learning of the RNN
as a formal optimization problem with an inequality constraint that provides a
sufficient condition for the stability of the network dynamics. Experimental
results demonstrate the effectiveness of this new method, which achieves 18.86%
phone recognition error on the TIMIT benchmark for the core test set. The
result approaches the best result of 17.7%, which was obtained by using RNN
with long short-term memory (LSTM). The results also show that the proposed
primal-dual training method produces lower recognition errors than the popular
RNN methods developed earlier based on the carefully tuned threshold parameter
that heuristically prevents the gradient from exploding.
|
1403.1347 | Jian Zhou Zhou | Jian Zhou and Olga G. Troyanskaya | Deep Supervised and Convolutional Generative Stochastic Network for
Protein Secondary Structure Prediction | Accepted by ICML 2014 | null | null | null | q-bio.QM cs.CE cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Predicting protein secondary structure is a fundamental problem in protein
structure prediction. Here we present a new supervised generative stochastic
network (GSN) based method to predict local secondary structure with deep
hierarchical representations. GSN is a recently proposed deep learning
technique (Bengio & Thibodeau-Laufer, 2013) to globally train deep generative
model. We present the supervised extension of GSN, which learns a Markov chain
to sample from a conditional distribution, and applied it to protein structure
prediction. To scale the model to full-sized, high-dimensional data, like
protein sequences with hundreds of amino acids, we introduce a convolutional
architecture, which allows efficient learning across multiple layers of
hierarchical representations. Our architecture uniquely focuses on predicting
structured low-level labels informed with both low and high-level
representations learned by the model. In our application this corresponds to
labeling the secondary structure state of each amino-acid residue. We trained
and tested the model on separate sets of non-homologous proteins sharing less
than 30% sequence identity. Our model achieves 66.4% Q8 accuracy on the CB513
dataset, better than the previously reported best performance 64.9% (Wang et
al., 2011) for this challenging secondary structure prediction problem.
| [
{
"version": "v1",
"created": "Thu, 6 Mar 2014 05:18:26 GMT"
}
] | 2014-03-07T00:00:00 | [
[
"Zhou",
"Jian",
""
],
[
"Troyanskaya",
"Olga G.",
""
]
] | TITLE: Deep Supervised and Convolutional Generative Stochastic Network for
Protein Secondary Structure Prediction
ABSTRACT: Predicting protein secondary structure is a fundamental problem in protein
structure prediction. Here we present a new supervised generative stochastic
network (GSN) based method to predict local secondary structure with deep
hierarchical representations. GSN is a recently proposed deep learning
technique (Bengio & Thibodeau-Laufer, 2013) to globally train deep generative
model. We present the supervised extension of GSN, which learns a Markov chain
to sample from a conditional distribution, and applied it to protein structure
prediction. To scale the model to full-sized, high-dimensional data, like
protein sequences with hundreds of amino acids, we introduce a convolutional
architecture, which allows efficient learning across multiple layers of
hierarchical representations. Our architecture uniquely focuses on predicting
structured low-level labels informed with both low and high-level
representations learned by the model. In our application this corresponds to
labeling the secondary structure state of each amino-acid residue. We trained
and tested the model on separate sets of non-homologous proteins sharing less
than 30% sequence identity. Our model achieves 66.4% Q8 accuracy on the CB513
dataset, better than the previously reported best performance 64.9% (Wang et
al., 2011) for this challenging secondary structure prediction problem.
|
1403.1353 | Yang Wu | Yang Wu, Vansteenberge Jarich, Masayuki Mukunoki, and Michihiko Minoh | Collaborative Representation for Classification, Sparse or Non-sparse? | 8 pages, 1 figure | null | null | null | cs.CV cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Sparse representation based classification (SRC) has been proved to be a
simple, effective and robust solution to face recognition. As it gets popular,
doubts on the necessity of enforcing sparsity starts coming up, and primary
experimental results showed that simply changing the $l_1$-norm based
regularization to the computationally much more efficient $l_2$-norm based
non-sparse version would lead to a similar or even better performance. However,
that's not always the case. Given a new classification task, it's still unclear
which regularization strategy (i.e., making the coefficients sparse or
non-sparse) is a better choice without trying both for comparison. In this
paper, we present as far as we know the first study on solving this issue,
based on plenty of diverse classification experiments. We propose a scoring
function for pre-selecting the regularization strategy using only the dataset
size, the feature dimensionality and a discrimination score derived from a
given feature representation. Moreover, we show that when dictionary learning
is taking into account, non-sparse representation has a more significant
superiority to sparse representation. This work is expected to enrich our
understanding of sparse/non-sparse collaborative representation for
classification and motivate further research activities.
| [
{
"version": "v1",
"created": "Thu, 6 Mar 2014 05:44:32 GMT"
}
] | 2014-03-07T00:00:00 | [
[
"Wu",
"Yang",
""
],
[
"Jarich",
"Vansteenberge",
""
],
[
"Mukunoki",
"Masayuki",
""
],
[
"Minoh",
"Michihiko",
""
]
] | TITLE: Collaborative Representation for Classification, Sparse or Non-sparse?
ABSTRACT: Sparse representation based classification (SRC) has been proved to be a
simple, effective and robust solution to face recognition. As it gets popular,
doubts on the necessity of enforcing sparsity starts coming up, and primary
experimental results showed that simply changing the $l_1$-norm based
regularization to the computationally much more efficient $l_2$-norm based
non-sparse version would lead to a similar or even better performance. However,
that's not always the case. Given a new classification task, it's still unclear
which regularization strategy (i.e., making the coefficients sparse or
non-sparse) is a better choice without trying both for comparison. In this
paper, we present as far as we know the first study on solving this issue,
based on plenty of diverse classification experiments. We propose a scoring
function for pre-selecting the regularization strategy using only the dataset
size, the feature dimensionality and a discrimination score derived from a
given feature representation. Moreover, we show that when dictionary learning
is taking into account, non-sparse representation has a more significant
superiority to sparse representation. This work is expected to enrich our
understanding of sparse/non-sparse collaborative representation for
classification and motivate further research activities.
|
1302.4886 | Aleksandr Aravkin | Aleksandr Y. Aravkin and Rajiv Kumar and Hassan Mansour and Ben Recht
and Felix J. Herrmann | Fast methods for denoising matrix completion formulations, with
applications to robust seismic data interpolation | 26 pages, 13 figures | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent SVD-free matrix factorization formulations have enabled rank
minimization for systems with millions of rows and columns, paving the way for
matrix completion in extremely large-scale applications, such as seismic data
interpolation.
In this paper, we consider matrix completion formulations designed to hit a
target data-fitting error level provided by the user, and propose an algorithm
called LR-BPDN that is able to exploit factorized formulations to solve the
corresponding optimization problem. Since practitioners typically have strong
prior knowledge about target error level, this innovation makes it easy to
apply the algorithm in practice, leaving only the factor rank to be determined.
Within the established framework, we propose two extensions that are highly
relevant to solving practical challenges of data interpolation. First, we
propose a weighted extension that allows known subspace information to improve
the results of matrix completion formulations. We show how this weighting can
be used in the context of frequency continuation, an essential aspect to
seismic data interpolation. Second, we propose matrix completion formulations
that are robust to large measurement errors in the available data.
We illustrate the advantages of LR-BPDN on the collaborative filtering
problem using the MovieLens 1M, 10M, and Netflix 100M datasets. Then, we use
the new method, along with its robust and subspace re-weighted extensions, to
obtain high-quality reconstructions for large scale seismic interpolation
problems with real data, even in the presence of data contamination.
| [
{
"version": "v1",
"created": "Wed, 20 Feb 2013 12:31:30 GMT"
},
{
"version": "v2",
"created": "Wed, 1 May 2013 10:03:30 GMT"
},
{
"version": "v3",
"created": "Wed, 5 Mar 2014 10:29:18 GMT"
}
] | 2014-03-06T00:00:00 | [
[
"Aravkin",
"Aleksandr Y.",
""
],
[
"Kumar",
"Rajiv",
""
],
[
"Mansour",
"Hassan",
""
],
[
"Recht",
"Ben",
""
],
[
"Herrmann",
"Felix J.",
""
]
] | TITLE: Fast methods for denoising matrix completion formulations, with
applications to robust seismic data interpolation
ABSTRACT: Recent SVD-free matrix factorization formulations have enabled rank
minimization for systems with millions of rows and columns, paving the way for
matrix completion in extremely large-scale applications, such as seismic data
interpolation.
In this paper, we consider matrix completion formulations designed to hit a
target data-fitting error level provided by the user, and propose an algorithm
called LR-BPDN that is able to exploit factorized formulations to solve the
corresponding optimization problem. Since practitioners typically have strong
prior knowledge about target error level, this innovation makes it easy to
apply the algorithm in practice, leaving only the factor rank to be determined.
Within the established framework, we propose two extensions that are highly
relevant to solving practical challenges of data interpolation. First, we
propose a weighted extension that allows known subspace information to improve
the results of matrix completion formulations. We show how this weighting can
be used in the context of frequency continuation, an essential aspect to
seismic data interpolation. Second, we propose matrix completion formulations
that are robust to large measurement errors in the available data.
We illustrate the advantages of LR-BPDN on the collaborative filtering
problem using the MovieLens 1M, 10M, and Netflix 100M datasets. Then, we use
the new method, along with its robust and subspace re-weighted extensions, to
obtain high-quality reconstructions for large scale seismic interpolation
problems with real data, even in the presence of data contamination.
|
1311.6079 | Amirreza Shaban | Amirreza Shaban, Hamid R. Rabiee and Mahyar Najibi | Local Similarities, Global Coding: An Algorithm for Feature Coding and
its Applications | null | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Data coding as a building block of several image processing algorithms has
been received great attention recently. Indeed, the importance of the locality
assumption in coding approaches is studied in numerous works and several
methods are proposed based on this concept. We probe this assumption and claim
that taking the similarity between a data point and a more global set of anchor
points does not necessarily weaken the coding method as long as the underlying
structure of the anchor points are taken into account. Based on this fact, we
propose to capture this underlying structure by assuming a random walker over
the anchor points. We show that our method is a fast approximate learning
algorithm based on the diffusion map kernel. The experiments on various
datasets show that making different state-of-the-art coding algorithms aware of
this structure boosts them in different learning tasks.
| [
{
"version": "v1",
"created": "Sun, 24 Nov 2013 04:39:28 GMT"
},
{
"version": "v2",
"created": "Wed, 5 Mar 2014 20:30:13 GMT"
}
] | 2014-03-06T00:00:00 | [
[
"Shaban",
"Amirreza",
""
],
[
"Rabiee",
"Hamid R.",
""
],
[
"Najibi",
"Mahyar",
""
]
] | TITLE: Local Similarities, Global Coding: An Algorithm for Feature Coding and
its Applications
ABSTRACT: Data coding as a building block of several image processing algorithms has
been received great attention recently. Indeed, the importance of the locality
assumption in coding approaches is studied in numerous works and several
methods are proposed based on this concept. We probe this assumption and claim
that taking the similarity between a data point and a more global set of anchor
points does not necessarily weaken the coding method as long as the underlying
structure of the anchor points are taken into account. Based on this fact, we
propose to capture this underlying structure by assuming a random walker over
the anchor points. We show that our method is a fast approximate learning
algorithm based on the diffusion map kernel. The experiments on various
datasets show that making different state-of-the-art coding algorithms aware of
this structure boosts them in different learning tasks.
|
1402.7025 | Max Welling | Max Welling | Exploiting the Statistics of Learning and Inference | Proceedings of the NIPS workshop on "Probabilistic Models for Big
Data" | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | When dealing with datasets containing a billion instances or with simulations
that require a supercomputer to execute, computational resources become part of
the equation. We can improve the efficiency of learning and inference by
exploiting their inherent statistical nature. We propose algorithms that
exploit the redundancy of data relative to a model by subsampling data-cases
for every update and reasoning about the uncertainty created in this process.
In the context of learning we propose to test for the probability that a
stochastically estimated gradient points more than 180 degrees in the wrong
direction. In the context of MCMC sampling we use stochastic gradients to
improve the efficiency of MCMC updates, and hypothesis tests based on adaptive
mini-batches to decide whether to accept or reject a proposed parameter update.
Finally, we argue that in the context of likelihood free MCMC one needs to
store all the information revealed by all simulations, for instance in a
Gaussian process. We conclude that Bayesian methods will remain to play a
crucial role in the era of big data and big simulations, but only if we
overcome a number of computational challenges.
| [
{
"version": "v1",
"created": "Wed, 26 Feb 2014 10:47:09 GMT"
},
{
"version": "v2",
"created": "Tue, 4 Mar 2014 21:12:43 GMT"
}
] | 2014-03-06T00:00:00 | [
[
"Welling",
"Max",
""
]
] | TITLE: Exploiting the Statistics of Learning and Inference
ABSTRACT: When dealing with datasets containing a billion instances or with simulations
that require a supercomputer to execute, computational resources become part of
the equation. We can improve the efficiency of learning and inference by
exploiting their inherent statistical nature. We propose algorithms that
exploit the redundancy of data relative to a model by subsampling data-cases
for every update and reasoning about the uncertainty created in this process.
In the context of learning we propose to test for the probability that a
stochastically estimated gradient points more than 180 degrees in the wrong
direction. In the context of MCMC sampling we use stochastic gradients to
improve the efficiency of MCMC updates, and hypothesis tests based on adaptive
mini-batches to decide whether to accept or reject a proposed parameter update.
Finally, we argue that in the context of likelihood free MCMC one needs to
store all the information revealed by all simulations, for instance in a
Gaussian process. We conclude that Bayesian methods will remain to play a
crucial role in the era of big data and big simulations, but only if we
overcome a number of computational challenges.
|
1403.1056 | Conrad Sanderson | Andres Sanin, Conrad Sanderson, Mehrtash T. Harandi, Brian C. Lovell | K-Tangent Spaces on Riemannian Manifolds for Improved Pedestrian
Detection | IEEE International Conference on Image Processing (ICIP), 2012 | null | 10.1109/ICIP.2012.6466899 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | For covariance-based image descriptors, taking into account the curvature of
the corresponding feature space has been shown to improve discrimination
performance. This is often done through representing the descriptors as points
on Riemannian manifolds, with the discrimination accomplished on a tangent
space. However, such treatment is restrictive as distances between arbitrary
points on the tangent space do not represent true geodesic distances, and hence
do not represent the manifold structure accurately. In this paper we propose a
general discriminative model based on the combination of several tangent
spaces, in order to preserve more details of the structure. The model can be
used as a weak learner in a boosting-based pedestrian detection framework.
Experiments on the challenging INRIA and DaimlerChrysler datasets show that the
proposed model leads to considerably higher performance than methods based on
histograms of oriented gradients as well as previous Riemannian-based
techniques.
| [
{
"version": "v1",
"created": "Wed, 5 Mar 2014 09:44:41 GMT"
}
] | 2014-03-06T00:00:00 | [
[
"Sanin",
"Andres",
""
],
[
"Sanderson",
"Conrad",
""
],
[
"Harandi",
"Mehrtash T.",
""
],
[
"Lovell",
"Brian C.",
""
]
] | TITLE: K-Tangent Spaces on Riemannian Manifolds for Improved Pedestrian
Detection
ABSTRACT: For covariance-based image descriptors, taking into account the curvature of
the corresponding feature space has been shown to improve discrimination
performance. This is often done through representing the descriptors as points
on Riemannian manifolds, with the discrimination accomplished on a tangent
space. However, such treatment is restrictive as distances between arbitrary
points on the tangent space do not represent true geodesic distances, and hence
do not represent the manifold structure accurately. In this paper we propose a
general discriminative model based on the combination of several tangent
spaces, in order to preserve more details of the structure. The model can be
used as a weak learner in a boosting-based pedestrian detection framework.
Experiments on the challenging INRIA and DaimlerChrysler datasets show that the
proposed model leads to considerably higher performance than methods based on
histograms of oriented gradients as well as previous Riemannian-based
techniques.
|
1211.6664 | Fabien Campagne | Fabien Campagne, Kevin C. Dorff, Nyasha Chambwe, James T. Robinson,
Jill P. Mesirov and Thomas D. Wu | Compression of structured high-throughput sequencing data | main article: 2 figures, 2 tables. Supplementary material: 2 figures,
4 tables. Comment on this manuscript on Twitter or Google Plus using handle
#Goby2Paper | null | 10.1371/journal.pone.0079871 | null | q-bio.QM cs.DB q-bio.GN | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large biological datasets are being produced at a rapid pace and create
substantial storage challenges, particularly in the domain of high-throughput
sequencing (HTS). Most approaches currently used to store HTS data are either
unable to quickly adapt to the requirements of new sequencing or analysis
methods (because they do not support schema evolution), or fail to provide
state of the art compression of the datasets. We have devised new approaches to
store HTS data that support seamless data schema evolution and compress
datasets substantially better than existing approaches. Building on these new
approaches, we discuss and demonstrate how a multi-tier data organization can
dramatically reduce the storage, computational and network burden of
collecting, analyzing, and archiving large sequencing datasets. For instance,
we show that spliced RNA-Seq alignments can be stored in less than 4% the size
of a BAM file with perfect data fidelity. Compared to the previous compression
state of the art, these methods reduce dataset size more than 20% when storing
gene expression and epigenetic datasets. The approaches have been integrated in
a comprehensive suite of software tools (http://goby.campagnelab.org) that
support common analyses for a range of high-throughput sequencing assays.
| [
{
"version": "v1",
"created": "Wed, 28 Nov 2012 17:11:54 GMT"
}
] | 2014-03-05T00:00:00 | [
[
"Campagne",
"Fabien",
""
],
[
"Dorff",
"Kevin C.",
""
],
[
"Chambwe",
"Nyasha",
""
],
[
"Robinson",
"James T.",
""
],
[
"Mesirov",
"Jill P.",
""
],
[
"Wu",
"Thomas D.",
""
]
] | TITLE: Compression of structured high-throughput sequencing data
ABSTRACT: Large biological datasets are being produced at a rapid pace and create
substantial storage challenges, particularly in the domain of high-throughput
sequencing (HTS). Most approaches currently used to store HTS data are either
unable to quickly adapt to the requirements of new sequencing or analysis
methods (because they do not support schema evolution), or fail to provide
state of the art compression of the datasets. We have devised new approaches to
store HTS data that support seamless data schema evolution and compress
datasets substantially better than existing approaches. Building on these new
approaches, we discuss and demonstrate how a multi-tier data organization can
dramatically reduce the storage, computational and network burden of
collecting, analyzing, and archiving large sequencing datasets. For instance,
we show that spliced RNA-Seq alignments can be stored in less than 4% the size
of a BAM file with perfect data fidelity. Compared to the previous compression
state of the art, these methods reduce dataset size more than 20% when storing
gene expression and epigenetic datasets. The approaches have been integrated in
a comprehensive suite of software tools (http://goby.campagnelab.org) that
support common analyses for a range of high-throughput sequencing assays.
|
1306.0196 | Peng Bao | Peng Bao, Hua-Wei Shen, Wei Chen, Xue-Qi Cheng | Cumulative Effect in Information Diffusion: A Comprehensive Empirical
Study on Microblogging Network | null | null | 10.1371/journal.pone.0076027 | null | cs.SI physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Cumulative effect in social contagions underlies many studies on the spread
of innovation, behaviors, and influence. However, few large-scale empirical
studies are conducted to validate the existence of cumulative effect in the
information diffusion on social networks. In this paper, using the
population-scale dataset from the largest Chinese microblogging website, we
conduct a comprehensive study on the cumulative effect in information
diffusion. We base our study on the diffusion network of each message, where
nodes are the involved users and links are the following relationships among
them. We find that multiple exposures to the same message indeed increase the
possibility of forwarding it. However, additional exposures cannot further
improve the chance of forwarding when the number of exposures crosses its peak
at two. This finding questions the cumulative effect hypothesis in information
diffusion. Furthermore, to clarify the forwarding preference among users, we
investigate both the structural motif of the diffusion network and the temporal
pattern of information diffusion process among users. The patterns provide
vital insight for understanding the variation of message popularity and explain
the characteristics of diffusion networks.
| [
{
"version": "v1",
"created": "Sun, 2 Jun 2013 11:31:51 GMT"
}
] | 2014-03-05T00:00:00 | [
[
"Bao",
"Peng",
""
],
[
"Shen",
"Hua-Wei",
""
],
[
"Chen",
"Wei",
""
],
[
"Cheng",
"Xue-Qi",
""
]
] | TITLE: Cumulative Effect in Information Diffusion: A Comprehensive Empirical
Study on Microblogging Network
ABSTRACT: Cumulative effect in social contagions underlies many studies on the spread
of innovation, behaviors, and influence. However, few large-scale empirical
studies are conducted to validate the existence of cumulative effect in the
information diffusion on social networks. In this paper, using the
population-scale dataset from the largest Chinese microblogging website, we
conduct a comprehensive study on the cumulative effect in information
diffusion. We base our study on the diffusion network of each message, where
nodes are the involved users and links are the following relationships among
them. We find that multiple exposures to the same message indeed increase the
possibility of forwarding it. However, additional exposures cannot further
improve the chance of forwarding when the number of exposures crosses its peak
at two. This finding questions the cumulative effect hypothesis in information
diffusion. Furthermore, to clarify the forwarding preference among users, we
investigate both the structural motif of the diffusion network and the temporal
pattern of information diffusion process among users. The patterns provide
vital insight for understanding the variation of message popularity and explain
the characteristics of diffusion networks.
|
1308.3060 | Wei Zeng | Wei Zeng, An Zeng, Ming-Sheng Shang and Yi-Cheng Zhang | Information filtering in sparse online systems: recommendation via
semi-local diffusion | 8 figures | null | 10.1371/journal.pone.0079354 | null | cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the rapid growth of the Internet and overwhelming amount of information
and choices that people are confronted with, recommender systems have been
developed to effectively support users' decision-making process in the online
systems. However, many recommendation algorithms suffer from the data sparsity
problem, i.e. the user-object bipartite networks are so sparse that algorithms
cannot accurately recommend objects for users. This data sparsity problem makes
many well-known recommendation algorithms perform poorly. To solve the problem,
we propose a recommendation algorithm based on the semi-local diffusion process
on a user-object bipartite network. The numerical simulation on two sparse
datasets, Amazon and Bookcross, show that our method significantly outperforms
the state-of-the-art methods especially for those small-degree users. Two
personalized semi-local diffusion methods are proposed which further improve
the recommendation accuracy. Finally, our work indicates that sparse online
systems are essentially different from the dense online systems, all the
algorithms and conclusions based on dense data should be rechecked again in
sparse data.
| [
{
"version": "v1",
"created": "Wed, 14 Aug 2013 08:29:41 GMT"
}
] | 2014-03-05T00:00:00 | [
[
"Zeng",
"Wei",
""
],
[
"Zeng",
"An",
""
],
[
"Shang",
"Ming-Sheng",
""
],
[
"Zhang",
"Yi-Cheng",
""
]
] | TITLE: Information filtering in sparse online systems: recommendation via
semi-local diffusion
ABSTRACT: With the rapid growth of the Internet and overwhelming amount of information
and choices that people are confronted with, recommender systems have been
developed to effectively support users' decision-making process in the online
systems. However, many recommendation algorithms suffer from the data sparsity
problem, i.e. the user-object bipartite networks are so sparse that algorithms
cannot accurately recommend objects for users. This data sparsity problem makes
many well-known recommendation algorithms perform poorly. To solve the problem,
we propose a recommendation algorithm based on the semi-local diffusion process
on a user-object bipartite network. The numerical simulation on two sparse
datasets, Amazon and Bookcross, show that our method significantly outperforms
the state-of-the-art methods especially for those small-degree users. Two
personalized semi-local diffusion methods are proposed which further improve
the recommendation accuracy. Finally, our work indicates that sparse online
systems are essentially different from the dense online systems, all the
algorithms and conclusions based on dense data should be rechecked again in
sparse data.
|
1308.5703 | Gonzalo Diaz | Marcelo Arenas, Gonzalo I. Diaz, Achille Fokoue, Anastasios
Kementsietsidis, Kavitha Srinivas | A Principled Approach to Bridging the Gap between Graph Data and their
Schemas | 18 pages, 8 figures. To be published in PVLDB Vol. 8, No. 9 | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Although RDF graphs have schema information associated with them, in practice
it is very common to find cases in which data do not fully conform to their
schema. A prominent example of this is DBpedia, which is RDF data extracted
from Wikipedia, a publicly editable source of information. In such situations,
it becomes interesting to study the structural properties of the actual data,
because the schema gives an incomplete description of the organization of a
dataset. In this paper we have approached the study of the structuredness of an
RDF graph in a principled way: we propose a framework for specifying
structuredness functions, which gauge the degree to which an RDF graph conforms
to a schema. In particular, we first define a formal language for specifying
structuredness functions with expressions we call rules. This language allows a
user or a database administrator to state a rule to which an RDF graph may
fully or partially conform. Then we consider the issue of discovering a
refinement of a sort (type) by partitioning the dataset into subsets whose
structuredness is over a specified threshold. In particular, we prove that the
natural decision problem associated to this refinement problem is NP-complete,
and we provide a natural translation of this problem into Integer Linear
Programming (ILP). Finally, we test this ILP solution with two real world
datasets, DBpedia Persons and WordNet Nouns, and 4 different and intuitive
rules, which gauge the structuredness in different ways. The rules give
meaningful refinements of the datasets, showing that our language can be a
powerful tool for understanding the structure of RDF data.
| [
{
"version": "v1",
"created": "Mon, 26 Aug 2013 21:26:00 GMT"
},
{
"version": "v2",
"created": "Tue, 4 Mar 2014 14:01:46 GMT"
}
] | 2014-03-05T00:00:00 | [
[
"Arenas",
"Marcelo",
""
],
[
"Diaz",
"Gonzalo I.",
""
],
[
"Fokoue",
"Achille",
""
],
[
"Kementsietsidis",
"Anastasios",
""
],
[
"Srinivas",
"Kavitha",
""
]
] | TITLE: A Principled Approach to Bridging the Gap between Graph Data and their
Schemas
ABSTRACT: Although RDF graphs have schema information associated with them, in practice
it is very common to find cases in which data do not fully conform to their
schema. A prominent example of this is DBpedia, which is RDF data extracted
from Wikipedia, a publicly editable source of information. In such situations,
it becomes interesting to study the structural properties of the actual data,
because the schema gives an incomplete description of the organization of a
dataset. In this paper we have approached the study of the structuredness of an
RDF graph in a principled way: we propose a framework for specifying
structuredness functions, which gauge the degree to which an RDF graph conforms
to a schema. In particular, we first define a formal language for specifying
structuredness functions with expressions we call rules. This language allows a
user or a database administrator to state a rule to which an RDF graph may
fully or partially conform. Then we consider the issue of discovering a
refinement of a sort (type) by partitioning the dataset into subsets whose
structuredness is over a specified threshold. In particular, we prove that the
natural decision problem associated to this refinement problem is NP-complete,
and we provide a natural translation of this problem into Integer Linear
Programming (ILP). Finally, we test this ILP solution with two real world
datasets, DBpedia Persons and WordNet Nouns, and 4 different and intuitive
rules, which gauge the structuredness in different ways. The rules give
meaningful refinements of the datasets, showing that our language can be a
powerful tool for understanding the structure of RDF data.
|
1312.3005 | Ciprian Chelba | Ciprian Chelba, Tomas Mikolov, Mike Schuster, Qi Ge, Thorsten Brants,
Phillipp Koehn, Tony Robinson | One Billion Word Benchmark for Measuring Progress in Statistical
Language Modeling | Accompanied by a code.google.com project allowing anyone to generate
the benchmark data, and use it to compare their language model against the
ones described in the paper | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a new benchmark corpus to be used for measuring progress in
statistical language modeling. With almost one billion words of training data,
we hope this benchmark will be useful to quickly evaluate novel language
modeling techniques, and to compare their contribution when combined with other
advanced techniques. We show performance of several well-known types of
language models, with the best results achieved with a recurrent neural network
based language model. The baseline unpruned Kneser-Ney 5-gram model achieves
perplexity 67.6; a combination of techniques leads to 35% reduction in
perplexity, or 10% reduction in cross-entropy (bits), over that baseline.
The benchmark is available as a code.google.com project; besides the scripts
needed to rebuild the training/held-out data, it also makes available
log-probability values for each word in each of ten held-out data sets, for
each of the baseline n-gram models.
| [
{
"version": "v1",
"created": "Wed, 11 Dec 2013 00:25:57 GMT"
},
{
"version": "v2",
"created": "Fri, 28 Feb 2014 22:26:59 GMT"
},
{
"version": "v3",
"created": "Tue, 4 Mar 2014 18:30:26 GMT"
}
] | 2014-03-05T00:00:00 | [
[
"Chelba",
"Ciprian",
""
],
[
"Mikolov",
"Tomas",
""
],
[
"Schuster",
"Mike",
""
],
[
"Ge",
"Qi",
""
],
[
"Brants",
"Thorsten",
""
],
[
"Koehn",
"Phillipp",
""
],
[
"Robinson",
"Tony",
""
]
] | TITLE: One Billion Word Benchmark for Measuring Progress in Statistical
Language Modeling
ABSTRACT: We propose a new benchmark corpus to be used for measuring progress in
statistical language modeling. With almost one billion words of training data,
we hope this benchmark will be useful to quickly evaluate novel language
modeling techniques, and to compare their contribution when combined with other
advanced techniques. We show performance of several well-known types of
language models, with the best results achieved with a recurrent neural network
based language model. The baseline unpruned Kneser-Ney 5-gram model achieves
perplexity 67.6; a combination of techniques leads to 35% reduction in
perplexity, or 10% reduction in cross-entropy (bits), over that baseline.
The benchmark is available as a code.google.com project; besides the scripts
needed to rebuild the training/held-out data, it also makes available
log-probability values for each word in each of ten held-out data sets, for
each of the baseline n-gram models.
|
1312.4400 | Min Lin | Min Lin, Qiang Chen, Shuicheng Yan | Network In Network | 10 pages, 4 figures, for iclr2014 | null | null | null | cs.NE cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a novel deep network structure called "Network In Network" (NIN)
to enhance model discriminability for local patches within the receptive field.
The conventional convolutional layer uses linear filters followed by a
nonlinear activation function to scan the input. Instead, we build micro neural
networks with more complex structures to abstract the data within the receptive
field. We instantiate the micro neural network with a multilayer perceptron,
which is a potent function approximator. The feature maps are obtained by
sliding the micro networks over the input in a similar manner as CNN; they are
then fed into the next layer. Deep NIN can be implemented by stacking mutiple
of the above described structure. With enhanced local modeling via the micro
network, we are able to utilize global average pooling over feature maps in the
classification layer, which is easier to interpret and less prone to
overfitting than traditional fully connected layers. We demonstrated the
state-of-the-art classification performances with NIN on CIFAR-10 and
CIFAR-100, and reasonable performances on SVHN and MNIST datasets.
| [
{
"version": "v1",
"created": "Mon, 16 Dec 2013 15:34:13 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Dec 2013 09:30:27 GMT"
},
{
"version": "v3",
"created": "Tue, 4 Mar 2014 05:15:42 GMT"
}
] | 2014-03-05T00:00:00 | [
[
"Lin",
"Min",
""
],
[
"Chen",
"Qiang",
""
],
[
"Yan",
"Shuicheng",
""
]
] | TITLE: Network In Network
ABSTRACT: We propose a novel deep network structure called "Network In Network" (NIN)
to enhance model discriminability for local patches within the receptive field.
The conventional convolutional layer uses linear filters followed by a
nonlinear activation function to scan the input. Instead, we build micro neural
networks with more complex structures to abstract the data within the receptive
field. We instantiate the micro neural network with a multilayer perceptron,
which is a potent function approximator. The feature maps are obtained by
sliding the micro networks over the input in a similar manner as CNN; they are
then fed into the next layer. Deep NIN can be implemented by stacking mutiple
of the above described structure. With enhanced local modeling via the micro
network, we are able to utilize global average pooling over feature maps in the
classification layer, which is easier to interpret and less prone to
overfitting than traditional fully connected layers. We demonstrated the
state-of-the-art classification performances with NIN on CIFAR-10 and
CIFAR-100, and reasonable performances on SVHN and MNIST datasets.
|
1403.0598 | Pinar Yanardag | Pinar Yanardag, S.V.N. Vishwanathan | The Structurally Smoothed Graphlet Kernel | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A commonly used paradigm for representing graphs is to use a vector that
contains normalized frequencies of occurrence of certain motifs or sub-graphs.
This vector representation can be used in a variety of applications, such as,
for computing similarity between graphs. The graphlet kernel of Shervashidze et
al. [32] uses induced sub-graphs of k nodes (christened as graphlets by Przulj
[28]) as motifs in the vector representation, and computes the kernel via a dot
product between these vectors. One can easily show that this is a valid kernel
between graphs. However, such a vector representation suffers from a few
drawbacks. As k becomes larger we encounter the sparsity problem; most higher
order graphlets will not occur in a given graph. This leads to diagonal
dominance, that is, a given graph is similar to itself but not to any other
graph in the dataset. On the other hand, since lower order graphlets tend to be
more numerous, using lower values of k does not provide enough discrimination
ability. We propose a smoothing technique to tackle the above problems. Our
method is based on a novel extension of Kneser-Ney and Pitman-Yor smoothing
techniques from natural language processing to graphs. We use the relationships
between lower order and higher order graphlets in order to derive our method.
Consequently, our smoothing algorithm not only respects the dependency between
sub-graphs but also tackles the diagonal dominance problem by distributing the
probability mass across graphlets. In our experiments, the smoothed graphlet
kernel outperforms graph kernels based on raw frequency counts.
| [
{
"version": "v1",
"created": "Mon, 3 Mar 2014 21:20:14 GMT"
}
] | 2014-03-05T00:00:00 | [
[
"Yanardag",
"Pinar",
""
],
[
"Vishwanathan",
"S. V. N.",
""
]
] | TITLE: The Structurally Smoothed Graphlet Kernel
ABSTRACT: A commonly used paradigm for representing graphs is to use a vector that
contains normalized frequencies of occurrence of certain motifs or sub-graphs.
This vector representation can be used in a variety of applications, such as,
for computing similarity between graphs. The graphlet kernel of Shervashidze et
al. [32] uses induced sub-graphs of k nodes (christened as graphlets by Przulj
[28]) as motifs in the vector representation, and computes the kernel via a dot
product between these vectors. One can easily show that this is a valid kernel
between graphs. However, such a vector representation suffers from a few
drawbacks. As k becomes larger we encounter the sparsity problem; most higher
order graphlets will not occur in a given graph. This leads to diagonal
dominance, that is, a given graph is similar to itself but not to any other
graph in the dataset. On the other hand, since lower order graphlets tend to be
more numerous, using lower values of k does not provide enough discrimination
ability. We propose a smoothing technique to tackle the above problems. Our
method is based on a novel extension of Kneser-Ney and Pitman-Yor smoothing
techniques from natural language processing to graphs. We use the relationships
between lower order and higher order graphlets in order to derive our method.
Consequently, our smoothing algorithm not only respects the dependency between
sub-graphs but also tackles the diagonal dominance problem by distributing the
probability mass across graphlets. In our experiments, the smoothed graphlet
kernel outperforms graph kernels based on raw frequency counts.
|
1403.0699 | Conrad Sanderson | Azadeh Alavi, Yan Yang, Mehrtash Harandi, Conrad Sanderson | Multi-Shot Person Re-Identification via Relational Stein Divergence | IEEE International Conference on Image Processing (ICIP), 2013 | null | 10.1109/ICIP.2013.6738731 | null | cs.CV stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Person re-identification is particularly challenging due to significant
appearance changes across separate camera views. In order to re-identify
people, a representative human signature should effectively handle differences
in illumination, pose and camera parameters. While general appearance-based
methods are modelled in Euclidean spaces, it has been argued that some
applications in image and video analysis are better modelled via non-Euclidean
manifold geometry. To this end, recent approaches represent images as
covariance matrices, and interpret such matrices as points on Riemannian
manifolds. As direct classification on such manifolds can be difficult, in this
paper we propose to represent each manifold point as a vector of similarities
to class representers, via a recently introduced form of Bregman matrix
divergence known as the Stein divergence. This is followed by using a
discriminative mapping of similarity vectors for final classification. The use
of similarity vectors is in contrast to the traditional approach of embedding
manifolds into tangent spaces, which can suffer from representing the manifold
structure inaccurately. Comparative evaluations on benchmark ETHZ and iLIDS
datasets for the person re-identification task show that the proposed approach
obtains better performance than recent techniques such as Histogram Plus
Epitome, Partial Least Squares, and Symmetry-Driven Accumulation of Local
Features.
| [
{
"version": "v1",
"created": "Tue, 4 Mar 2014 06:44:17 GMT"
}
] | 2014-03-05T00:00:00 | [
[
"Alavi",
"Azadeh",
""
],
[
"Yang",
"Yan",
""
],
[
"Harandi",
"Mehrtash",
""
],
[
"Sanderson",
"Conrad",
""
]
] | TITLE: Multi-Shot Person Re-Identification via Relational Stein Divergence
ABSTRACT: Person re-identification is particularly challenging due to significant
appearance changes across separate camera views. In order to re-identify
people, a representative human signature should effectively handle differences
in illumination, pose and camera parameters. While general appearance-based
methods are modelled in Euclidean spaces, it has been argued that some
applications in image and video analysis are better modelled via non-Euclidean
manifold geometry. To this end, recent approaches represent images as
covariance matrices, and interpret such matrices as points on Riemannian
manifolds. As direct classification on such manifolds can be difficult, in this
paper we propose to represent each manifold point as a vector of similarities
to class representers, via a recently introduced form of Bregman matrix
divergence known as the Stein divergence. This is followed by using a
discriminative mapping of similarity vectors for final classification. The use
of similarity vectors is in contrast to the traditional approach of embedding
manifolds into tangent spaces, which can suffer from representing the manifold
structure inaccurately. Comparative evaluations on benchmark ETHZ and iLIDS
datasets for the person re-identification task show that the proposed approach
obtains better performance than recent techniques such as Histogram Plus
Epitome, Partial Least Squares, and Symmetry-Driven Accumulation of Local
Features.
|
1403.0829 | Weifeng Liu | W. Liu, H. Liu, D. Tao, Y. Wang, Ke Lu | Multiview Hessian regularized logistic regression for action recognition | 13 pages,2 figures, submitted to signal processing | null | null | null | cs.CV cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the rapid development of social media sharing, people often need to
manage the growing volume of multimedia data such as large scale video
classification and annotation, especially to organize those videos containing
human activities. Recently, manifold regularized semi-supervised learning
(SSL), which explores the intrinsic data probability distribution and then
improves the generalization ability with only a small number of labeled data,
has emerged as a promising paradigm for semiautomatic video classification. In
addition, human action videos often have multi-modal content and different
representations. To tackle the above problems, in this paper we propose
multiview Hessian regularized logistic regression (mHLR) for human action
recognition. Compared with existing work, the advantages of mHLR lie in three
folds: (1) mHLR combines multiple Hessian regularization, each of which
obtained from a particular representation of instance, to leverage the
exploring of local geometry; (2) mHLR naturally handle multi-view instances
with multiple representations; (3) mHLR employs a smooth loss function and then
can be effectively optimized. We carefully conduct extensive experiments on the
unstructured social activity attribute (USAA) dataset and the experimental
results demonstrate the effectiveness of the proposed multiview Hessian
regularized logistic regression for human action recognition.
| [
{
"version": "v1",
"created": "Mon, 3 Mar 2014 01:11:40 GMT"
}
] | 2014-03-05T00:00:00 | [
[
"Liu",
"W.",
""
],
[
"Liu",
"H.",
""
],
[
"Tao",
"D.",
""
],
[
"Wang",
"Y.",
""
],
[
"Lu",
"Ke",
""
]
] | TITLE: Multiview Hessian regularized logistic regression for action recognition
ABSTRACT: With the rapid development of social media sharing, people often need to
manage the growing volume of multimedia data such as large scale video
classification and annotation, especially to organize those videos containing
human activities. Recently, manifold regularized semi-supervised learning
(SSL), which explores the intrinsic data probability distribution and then
improves the generalization ability with only a small number of labeled data,
has emerged as a promising paradigm for semiautomatic video classification. In
addition, human action videos often have multi-modal content and different
representations. To tackle the above problems, in this paper we propose
multiview Hessian regularized logistic regression (mHLR) for human action
recognition. Compared with existing work, the advantages of mHLR lie in three
folds: (1) mHLR combines multiple Hessian regularization, each of which
obtained from a particular representation of instance, to leverage the
exploring of local geometry; (2) mHLR naturally handle multi-view instances
with multiple representations; (3) mHLR employs a smooth loss function and then
can be effectively optimized. We carefully conduct extensive experiments on the
unstructured social activity attribute (USAA) dataset and the experimental
results demonstrate the effectiveness of the proposed multiview Hessian
regularized logistic regression for human action recognition.
|
1310.6887 | Filipe Brand\~ao M.Sc | Filipe Brand\~ao, Jo\~ao Pedro Pedroso | Bin Packing and Related Problems: General Arc-flow Formulation with
Graph Compression | null | null | null | DCC-2013-08 | math.OC cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present an exact method, based on an arc-flow formulation with side
constraints, for solving bin packing and cutting stock problems --- including
multi-constraint variants --- by simply representing all the patterns in a very
compact graph. Our method includes a graph compression algorithm that usually
reduces the size of the underlying graph substantially without weakening the
model. As opposed to our method, which provides strong models, conventional
models are usually highly symmetric and provide very weak lower bounds.
Our formulation is equivalent to Gilmore and Gomory's, thus providing a very
strong linear relaxation. However, instead of using column-generation in an
iterative process, the method constructs a graph, where paths from the source
to the target node represent every valid packing pattern.
The same method, without any problem-specific parameterization, was used to
solve a large variety of instances from several different cutting and packing
problems. In this paper, we deal with vector packing, graph coloring, bin
packing, cutting stock, cardinality constrained bin packing, cutting stock with
cutting knife limitation, cutting stock with binary patterns, bin packing with
conflicts, and cutting stock with binary patterns and forbidden pairs. We
report computational results obtained with many benchmark test data sets, all
of them showing a large advantage of this formulation with respect to the
traditional ones.
| [
{
"version": "v1",
"created": "Fri, 25 Oct 2013 11:55:30 GMT"
}
] | 2014-03-04T00:00:00 | [
[
"Brandão",
"Filipe",
""
],
[
"Pedroso",
"João Pedro",
""
]
] | TITLE: Bin Packing and Related Problems: General Arc-flow Formulation with
Graph Compression
ABSTRACT: We present an exact method, based on an arc-flow formulation with side
constraints, for solving bin packing and cutting stock problems --- including
multi-constraint variants --- by simply representing all the patterns in a very
compact graph. Our method includes a graph compression algorithm that usually
reduces the size of the underlying graph substantially without weakening the
model. As opposed to our method, which provides strong models, conventional
models are usually highly symmetric and provide very weak lower bounds.
Our formulation is equivalent to Gilmore and Gomory's, thus providing a very
strong linear relaxation. However, instead of using column-generation in an
iterative process, the method constructs a graph, where paths from the source
to the target node represent every valid packing pattern.
The same method, without any problem-specific parameterization, was used to
solve a large variety of instances from several different cutting and packing
problems. In this paper, we deal with vector packing, graph coloring, bin
packing, cutting stock, cardinality constrained bin packing, cutting stock with
cutting knife limitation, cutting stock with binary patterns, bin packing with
conflicts, and cutting stock with binary patterns and forbidden pairs. We
report computational results obtained with many benchmark test data sets, all
of them showing a large advantage of this formulation with respect to the
traditional ones.
|
1403.0224 | Rakesh Mohanty | Mitali Sinha, Suchismita Pattanaik, Rakesh Mohanty and Prachi Tripathy | Experimental Study of A Novel Variant of Fiduccia Mattheyses(FM)
Partitioning Algorithm | null | null | null | null | cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Partitioning is a well studied research problem in the area of VLSI physical
design automation. In this problem, input is an integrated circuit and output
is a set of almost equal disjoint blocks. The main objective of partitioning is
to assign the components of circuit to blocks in order to minimize the numbers
of inter-block connections. A partitioning algorithm using hypergraph was
proposed by Fiduccia and Mattheyses with linear time complexity which has been
popularly known as FM algorithm. Most of the hypergraph based partitioning
algorithms proposed in the literature are variants of FM algorithm. In this
paper, we have proposed a novel variant of FM algorithm by using pair wise
swapping technique. We have performed a comparative experimental study of FM
algorithm and our proposed algorithm using two dataset such as ISPD98 and
ISPD99. Experimental results show that performance of our proposed algorithm is
better than the FM algorithm using the above dataset.
| [
{
"version": "v1",
"created": "Sun, 2 Mar 2014 15:34:48 GMT"
}
] | 2014-03-04T00:00:00 | [
[
"Sinha",
"Mitali",
""
],
[
"Pattanaik",
"Suchismita",
""
],
[
"Mohanty",
"Rakesh",
""
],
[
"Tripathy",
"Prachi",
""
]
] | TITLE: Experimental Study of A Novel Variant of Fiduccia Mattheyses(FM)
Partitioning Algorithm
ABSTRACT: Partitioning is a well studied research problem in the area of VLSI physical
design automation. In this problem, input is an integrated circuit and output
is a set of almost equal disjoint blocks. The main objective of partitioning is
to assign the components of circuit to blocks in order to minimize the numbers
of inter-block connections. A partitioning algorithm using hypergraph was
proposed by Fiduccia and Mattheyses with linear time complexity which has been
popularly known as FM algorithm. Most of the hypergraph based partitioning
algorithms proposed in the literature are variants of FM algorithm. In this
paper, we have proposed a novel variant of FM algorithm by using pair wise
swapping technique. We have performed a comparative experimental study of FM
algorithm and our proposed algorithm using two dataset such as ISPD98 and
ISPD99. Experimental results show that performance of our proposed algorithm is
better than the FM algorithm using the above dataset.
|
1403.0240 | Ivo Sbalzarini | Ivo F. Sbalzarini, Sophie Schneider, Janick Cardinale | Particle methods enable fast and simple approximation of Sobolev
gradients in image segmentation | 21 pages, 10 figures | null | null | null | cs.CV cs.CE cs.NA q-bio.QM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Bio-image analysis is challenging due to inhomogeneous intensity
distributions and high levels of noise in the images. Bayesian inference
provides a principled way for regularizing the problem using prior knowledge. A
fundamental choice is how one measures "distances" between shapes in an image.
It has been shown that the straightforward geometric L2 distance is degenerate
and leads to pathological situations. This is avoided when using Sobolev
gradients, rendering the segmentation problem less ill-posed. The high
computational cost and implementation overhead of Sobolev gradients, however,
have hampered practical applications. We show how particle methods as applied
to image segmentation allow for a simple and computationally efficient
implementation of Sobolev gradients. We show that the evaluation of Sobolev
gradients amounts to particle-particle interactions along the contour in an
image. We extend an existing particle-based segmentation algorithm to using
Sobolev gradients. Using synthetic and real-world images, we benchmark the
results for both 2D and 3D images using piecewise smooth and piecewise constant
region models. The present particle approximation of Sobolev gradients is 2.8
to 10 times faster than the previous reference implementation, but retains the
known favorable properties of Sobolev gradients. This speedup is achieved by
using local particle-particle interactions instead of solving a global Poisson
equation at each iteration. The computational time per iteration is higher for
Sobolev gradients than for L2 gradients. Since Sobolev gradients precondition
the optimization problem, however, a smaller number of overall iterations may
be necessary for the algorithm to converge, which can in some cases amortize
the higher per-iteration cost.
| [
{
"version": "v1",
"created": "Sun, 2 Mar 2014 16:58:29 GMT"
}
] | 2014-03-04T00:00:00 | [
[
"Sbalzarini",
"Ivo F.",
""
],
[
"Schneider",
"Sophie",
""
],
[
"Cardinale",
"Janick",
""
]
] | TITLE: Particle methods enable fast and simple approximation of Sobolev
gradients in image segmentation
ABSTRACT: Bio-image analysis is challenging due to inhomogeneous intensity
distributions and high levels of noise in the images. Bayesian inference
provides a principled way for regularizing the problem using prior knowledge. A
fundamental choice is how one measures "distances" between shapes in an image.
It has been shown that the straightforward geometric L2 distance is degenerate
and leads to pathological situations. This is avoided when using Sobolev
gradients, rendering the segmentation problem less ill-posed. The high
computational cost and implementation overhead of Sobolev gradients, however,
have hampered practical applications. We show how particle methods as applied
to image segmentation allow for a simple and computationally efficient
implementation of Sobolev gradients. We show that the evaluation of Sobolev
gradients amounts to particle-particle interactions along the contour in an
image. We extend an existing particle-based segmentation algorithm to using
Sobolev gradients. Using synthetic and real-world images, we benchmark the
results for both 2D and 3D images using piecewise smooth and piecewise constant
region models. The present particle approximation of Sobolev gradients is 2.8
to 10 times faster than the previous reference implementation, but retains the
known favorable properties of Sobolev gradients. This speedup is achieved by
using local particle-particle interactions instead of solving a global Poisson
equation at each iteration. The computational time per iteration is higher for
Sobolev gradients than for L2 gradients. Since Sobolev gradients precondition
the optimization problem, however, a smaller number of overall iterations may
be necessary for the algorithm to converge, which can in some cases amortize
the higher per-iteration cost.
|
1403.0316 | Kang Zhang | Kang Zhang, Yuqiang Fang, Dongbo Min, Lifeng Sun, Shiqiang Yang.
Shuicheng Yan, Qi Tian | Cross-Scale Cost Aggregation for Stereo Matching | To Appear in 2013 IEEE Conference on Computer Vision and Pattern
Recognition (CVPR). 2014 (poster, 29.88%) | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Human beings process stereoscopic correspondence across multiple scales.
However, this bio-inspiration is ignored by state-of-the-art cost aggregation
methods for dense stereo correspondence. In this paper, a generic cross-scale
cost aggregation framework is proposed to allow multi-scale interaction in cost
aggregation. We firstly reformulate cost aggregation from a unified
optimization perspective and show that different cost aggregation methods
essentially differ in the choices of similarity kernels. Then, an inter-scale
regularizer is introduced into optimization and solving this new optimization
problem leads to the proposed framework. Since the regularization term is
independent of the similarity kernel, various cost aggregation methods can be
integrated into the proposed general framework. We show that the cross-scale
framework is important as it effectively and efficiently expands
state-of-the-art cost aggregation methods and leads to significant
improvements, when evaluated on Middlebury, KITTI and New Tsukuba datasets.
| [
{
"version": "v1",
"created": "Mon, 3 Mar 2014 05:20:28 GMT"
}
] | 2014-03-04T00:00:00 | [
[
"Zhang",
"Kang",
""
],
[
"Fang",
"Yuqiang",
""
],
[
"Min",
"Dongbo",
""
],
[
"Sun",
"Lifeng",
""
],
[
"Yan",
"Shiqiang Yang. Shuicheng",
""
],
[
"Tian",
"Qi",
""
]
] | TITLE: Cross-Scale Cost Aggregation for Stereo Matching
ABSTRACT: Human beings process stereoscopic correspondence across multiple scales.
However, this bio-inspiration is ignored by state-of-the-art cost aggregation
methods for dense stereo correspondence. In this paper, a generic cross-scale
cost aggregation framework is proposed to allow multi-scale interaction in cost
aggregation. We firstly reformulate cost aggregation from a unified
optimization perspective and show that different cost aggregation methods
essentially differ in the choices of similarity kernels. Then, an inter-scale
regularizer is introduced into optimization and solving this new optimization
problem leads to the proposed framework. Since the regularization term is
independent of the similarity kernel, various cost aggregation methods can be
integrated into the proposed general framework. We show that the cross-scale
framework is important as it effectively and efficiently expands
state-of-the-art cost aggregation methods and leads to significant
improvements, when evaluated on Middlebury, KITTI and New Tsukuba datasets.
|
1403.0481 | Arindam Chaudhuri AC | Arindam Chaudhuri | Support Vector Machine Model for Currency Crisis Discrimination | Book Chapter Selected Works in Infrastructural Finance, Rudra P.
Pradhan, Indian Institute of Technology Kharagpur, Editor, Macmillan
Publishers, India, pp 249 - 256, 2011 | null | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Support Vector Machine (SVM) is powerful classification technique based on
the idea of structural risk minimization. Use of kernel function enables curse
of dimensionality to be addressed. However, proper kernel function for certain
problem is dependent on specific dataset and as such there is no good method on
choice of kernel function. In this paper, SVM is used to build empirical models
of currency crisis in Argentina. An estimation technique is developed by
training model on real life data set which provides reasonably accurate model
outputs and helps policy makers to identify situations in which currency crisis
may happen. The third and fourth order polynomial kernel is generally best
choice to achieve high generalization of classifier performance. SVM has high
level of maturity with algorithms that are simple, easy to implement, tolerates
curse of dimensionality and good empirical performance. The satisfactory
results show that currency crisis situation is properly emulated using only
small fraction of database and could be used as an evaluation tool as well as
an early warning system. To the best of knowledge this is the first work on SVM
approach for currency crisis evaluation of Argentina.
| [
{
"version": "v1",
"created": "Mon, 3 Mar 2014 16:34:38 GMT"
}
] | 2014-03-04T00:00:00 | [
[
"Chaudhuri",
"Arindam",
""
]
] | TITLE: Support Vector Machine Model for Currency Crisis Discrimination
ABSTRACT: Support Vector Machine (SVM) is powerful classification technique based on
the idea of structural risk minimization. Use of kernel function enables curse
of dimensionality to be addressed. However, proper kernel function for certain
problem is dependent on specific dataset and as such there is no good method on
choice of kernel function. In this paper, SVM is used to build empirical models
of currency crisis in Argentina. An estimation technique is developed by
training model on real life data set which provides reasonably accurate model
outputs and helps policy makers to identify situations in which currency crisis
may happen. The third and fourth order polynomial kernel is generally best
choice to achieve high generalization of classifier performance. SVM has high
level of maturity with algorithms that are simple, easy to implement, tolerates
curse of dimensionality and good empirical performance. The satisfactory
results show that currency crisis situation is properly emulated using only
small fraction of database and could be used as an evaluation tool as well as
an early warning system. To the best of knowledge this is the first work on SVM
approach for currency crisis evaluation of Argentina.
|
1310.2959 | Partha Talukdar | Partha Pratim Talukdar, William Cohen | Scaling Graph-based Semi Supervised Learning to Large Number of Labels
Using Count-Min Sketch | 9 pages | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Graph-based Semi-supervised learning (SSL) algorithms have been successfully
used in a large number of applications. These methods classify initially
unlabeled nodes by propagating label information over the structure of graph
starting from seed nodes. Graph-based SSL algorithms usually scale linearly
with the number of distinct labels (m), and require O(m) space on each node.
Unfortunately, there exist many applications of practical significance with
very large m over large graphs, demanding better space and time complexity. In
this paper, we propose MAD-SKETCH, a novel graph-based SSL algorithm which
compactly stores label distribution on each node using Count-min Sketch, a
randomized data structure. We present theoretical analysis showing that under
mild conditions, MAD-SKETCH can reduce space complexity at each node from O(m)
to O(log m), and achieve similar savings in time complexity as well. We support
our analysis through experiments on multiple real world datasets. We observe
that MAD-SKETCH achieves similar performance as existing state-of-the-art
graph- based SSL algorithms, while requiring smaller memory footprint and at
the same time achieving up to 10x speedup. We find that MAD-SKETCH is able to
scale to datasets with one million labels, which is beyond the scope of
existing graph- based SSL algorithms.
| [
{
"version": "v1",
"created": "Thu, 10 Oct 2013 20:30:06 GMT"
},
{
"version": "v2",
"created": "Thu, 27 Feb 2014 21:19:41 GMT"
}
] | 2014-03-03T00:00:00 | [
[
"Talukdar",
"Partha Pratim",
""
],
[
"Cohen",
"William",
""
]
] | TITLE: Scaling Graph-based Semi Supervised Learning to Large Number of Labels
Using Count-Min Sketch
ABSTRACT: Graph-based Semi-supervised learning (SSL) algorithms have been successfully
used in a large number of applications. These methods classify initially
unlabeled nodes by propagating label information over the structure of graph
starting from seed nodes. Graph-based SSL algorithms usually scale linearly
with the number of distinct labels (m), and require O(m) space on each node.
Unfortunately, there exist many applications of practical significance with
very large m over large graphs, demanding better space and time complexity. In
this paper, we propose MAD-SKETCH, a novel graph-based SSL algorithm which
compactly stores label distribution on each node using Count-min Sketch, a
randomized data structure. We present theoretical analysis showing that under
mild conditions, MAD-SKETCH can reduce space complexity at each node from O(m)
to O(log m), and achieve similar savings in time complexity as well. We support
our analysis through experiments on multiple real world datasets. We observe
that MAD-SKETCH achieves similar performance as existing state-of-the-art
graph- based SSL algorithms, while requiring smaller memory footprint and at
the same time achieving up to 10x speedup. We find that MAD-SKETCH is able to
scale to datasets with one million labels, which is beyond the scope of
existing graph- based SSL algorithms.
|
1311.0680 | Bartosz Hawelka | Bartosz Hawelka, Izabela Sitko, Euro Beinat, Stanislav Sobolevsky,
Pavlos Kazakopoulos and Carlo Ratti | Geo-located Twitter as the proxy for global mobility patterns | 17 pages, 13 figures | null | 10.1080/15230406.2014.890072 | null | cs.SI physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In the advent of a pervasive presence of location sharing services
researchers gained an unprecedented access to the direct records of human
activity in space and time. This paper analyses geo-located Twitter messages in
order to uncover global patterns of human mobility. Based on a dataset of
almost a billion tweets recorded in 2012 we estimate volumes of international
travelers in respect to their country of residence. We examine mobility
profiles of different nations looking at the characteristics such as mobility
rate, radius of gyration, diversity of destinations and a balance of the
inflows and outflows. The temporal patterns disclose the universal seasons of
increased international mobility and the peculiar national nature of overseen
travels. Our analysis of the community structure of the Twitter mobility
network, obtained with the iterative network partitioning, reveals spatially
cohesive regions that follow the regional division of the world. Finally, we
validate our result with the global tourism statistics and mobility models
provided by other authors, and argue that Twitter is a viable source to
understand and quantify global mobility patterns.
| [
{
"version": "v1",
"created": "Mon, 4 Nov 2013 12:46:08 GMT"
},
{
"version": "v2",
"created": "Sat, 28 Dec 2013 13:40:30 GMT"
}
] | 2014-03-03T00:00:00 | [
[
"Hawelka",
"Bartosz",
""
],
[
"Sitko",
"Izabela",
""
],
[
"Beinat",
"Euro",
""
],
[
"Sobolevsky",
"Stanislav",
""
],
[
"Kazakopoulos",
"Pavlos",
""
],
[
"Ratti",
"Carlo",
""
]
] | TITLE: Geo-located Twitter as the proxy for global mobility patterns
ABSTRACT: In the advent of a pervasive presence of location sharing services
researchers gained an unprecedented access to the direct records of human
activity in space and time. This paper analyses geo-located Twitter messages in
order to uncover global patterns of human mobility. Based on a dataset of
almost a billion tweets recorded in 2012 we estimate volumes of international
travelers in respect to their country of residence. We examine mobility
profiles of different nations looking at the characteristics such as mobility
rate, radius of gyration, diversity of destinations and a balance of the
inflows and outflows. The temporal patterns disclose the universal seasons of
increased international mobility and the peculiar national nature of overseen
travels. Our analysis of the community structure of the Twitter mobility
network, obtained with the iterative network partitioning, reveals spatially
cohesive regions that follow the regional division of the world. Finally, we
validate our result with the global tourism statistics and mobility models
provided by other authors, and argue that Twitter is a viable source to
understand and quantify global mobility patterns.
|
1402.5596 | Jason Lee | Jason D Lee and Jonathan E Taylor | Exact Post Model Selection Inference for Marginal Screening | null | null | null | null | stat.ME cs.LG math.ST stat.ML stat.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We develop a framework for post model selection inference, via marginal
screening, in linear regression. At the core of this framework is a result that
characterizes the exact distribution of linear functions of the response $y$,
conditional on the model being selected (``condition on selection" framework).
This allows us to construct valid confidence intervals and hypothesis tests for
regression coefficients that account for the selection procedure. In contrast
to recent work in high-dimensional statistics, our results are exact
(non-asymptotic) and require no eigenvalue-like assumptions on the design
matrix $X$. Furthermore, the computational cost of marginal regression,
constructing confidence intervals and hypothesis testing is negligible compared
to the cost of linear regression, thus making our methods particularly suitable
for extremely large datasets. Although we focus on marginal screening to
illustrate the applicability of the condition on selection framework, this
framework is much more broadly applicable. We show how to apply the proposed
framework to several other selection procedures including orthogonal matching
pursuit, non-negative least squares, and marginal screening+Lasso.
| [
{
"version": "v1",
"created": "Sun, 23 Feb 2014 10:30:21 GMT"
},
{
"version": "v2",
"created": "Fri, 28 Feb 2014 00:28:21 GMT"
}
] | 2014-03-03T00:00:00 | [
[
"Lee",
"Jason D",
""
],
[
"Taylor",
"Jonathan E",
""
]
] | TITLE: Exact Post Model Selection Inference for Marginal Screening
ABSTRACT: We develop a framework for post model selection inference, via marginal
screening, in linear regression. At the core of this framework is a result that
characterizes the exact distribution of linear functions of the response $y$,
conditional on the model being selected (``condition on selection" framework).
This allows us to construct valid confidence intervals and hypothesis tests for
regression coefficients that account for the selection procedure. In contrast
to recent work in high-dimensional statistics, our results are exact
(non-asymptotic) and require no eigenvalue-like assumptions on the design
matrix $X$. Furthermore, the computational cost of marginal regression,
constructing confidence intervals and hypothesis testing is negligible compared
to the cost of linear regression, thus making our methods particularly suitable
for extremely large datasets. Although we focus on marginal screening to
illustrate the applicability of the condition on selection framework, this
framework is much more broadly applicable. We show how to apply the proposed
framework to several other selection procedures including orthogonal matching
pursuit, non-negative least squares, and marginal screening+Lasso.
|
1312.3245 | Hien Thi Thu Truong | Hien Thi Thu Truong, Eemil Lagerspetz, Petteri Nurmi, Adam J. Oliner,
Sasu Tarkoma, N. Asokan, Sourav Bhattacharya | The Company You Keep: Mobile Malware Infection Rates and Inexpensive
Risk Indicators | null | null | 10.1145/2566486.2568046 | null | cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | There is little information from independent sources in the public domain
about mobile malware infection rates. The only previous independent estimate
(0.0009%) [12], was based on indirect measurements obtained from domain name
resolution traces. In this paper, we present the first independent study of
malware infection rates and associated risk factors using data collected
directly from over 55,000 Android devices. We find that the malware infection
rates in Android devices estimated using two malware datasets (0.28% and
0.26%), though small, are significantly higher than the previous independent
estimate. Using our datasets, we investigate how indicators extracted
inexpensively from the devices correlate with malware infection. Based on the
hypothesis that some application stores have a greater density of malicious
applications and that advertising within applications and cross-promotional
deals may act as infection vectors, we investigate whether the set of
applications used on a device can serve as an indicator for infection of that
device. Our analysis indicates that this alone is not an accurate indicator for
pinpointing infection. However, it is a very inexpensive but surprisingly
useful way for significantly narrowing down the pool of devices on which
expensive monitoring and analysis mechanisms must be deployed. Using our two
malware datasets we show that this indicator performs 4.8 and 4.6 times
(respectively) better at identifying infected devices than the baseline of
random checks. Such indicators can be used, for example, in the search for new
or previously undetected malware. It is therefore a technique that can
complement standard malware scanning by anti-malware tools. Our analysis also
demonstrates a marginally significant difference in battery use between
infected and clean devices.
| [
{
"version": "v1",
"created": "Wed, 11 Dec 2013 17:06:16 GMT"
},
{
"version": "v2",
"created": "Thu, 27 Feb 2014 16:58:12 GMT"
}
] | 2014-02-28T00:00:00 | [
[
"Truong",
"Hien Thi Thu",
""
],
[
"Lagerspetz",
"Eemil",
""
],
[
"Nurmi",
"Petteri",
""
],
[
"Oliner",
"Adam J.",
""
],
[
"Tarkoma",
"Sasu",
""
],
[
"Asokan",
"N.",
""
],
[
"Bhattacharya",
"Sourav",
""
]
] | TITLE: The Company You Keep: Mobile Malware Infection Rates and Inexpensive
Risk Indicators
ABSTRACT: There is little information from independent sources in the public domain
about mobile malware infection rates. The only previous independent estimate
(0.0009%) [12], was based on indirect measurements obtained from domain name
resolution traces. In this paper, we present the first independent study of
malware infection rates and associated risk factors using data collected
directly from over 55,000 Android devices. We find that the malware infection
rates in Android devices estimated using two malware datasets (0.28% and
0.26%), though small, are significantly higher than the previous independent
estimate. Using our datasets, we investigate how indicators extracted
inexpensively from the devices correlate with malware infection. Based on the
hypothesis that some application stores have a greater density of malicious
applications and that advertising within applications and cross-promotional
deals may act as infection vectors, we investigate whether the set of
applications used on a device can serve as an indicator for infection of that
device. Our analysis indicates that this alone is not an accurate indicator for
pinpointing infection. However, it is a very inexpensive but surprisingly
useful way for significantly narrowing down the pool of devices on which
expensive monitoring and analysis mechanisms must be deployed. Using our two
malware datasets we show that this indicator performs 4.8 and 4.6 times
(respectively) better at identifying infected devices than the baseline of
random checks. Such indicators can be used, for example, in the search for new
or previously undetected malware. It is therefore a technique that can
complement standard malware scanning by anti-malware tools. Our analysis also
demonstrates a marginally significant difference in battery use between
infected and clean devices.
|
1401.5465 | Wanling Gao | Zijian Ming, Chunjie Luo, Wanling Gao, Rui Han, Qiang Yang, Lei Wang,
and Jianfeng Zhan | BDGS: A Scalable Big Data Generator Suite in Big Data Benchmarking | null | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Data generation is a key issue in big data benchmarking that aims to generate
application-specific data sets to meet the 4V requirements of big data.
Specifically, big data generators need to generate scalable data (Volume) of
different types (Variety) under controllable generation rates (Velocity) while
keeping the important characteristics of raw data (Veracity). This gives rise
to various new challenges about how we design generators efficiently and
successfully. To date, most existing techniques can only generate limited types
of data and support specific big data systems such as Hadoop. Hence we develop
a tool, called Big Data Generator Suite (BDGS), to efficiently generate
scalable big data while employing data models derived from real data to
preserve data veracity. The effectiveness of BDGS is demonstrated by developing
six data generators covering three representative data types (structured,
semi-structured and unstructured) and three data sources (text, graph, and
table data).
| [
{
"version": "v1",
"created": "Wed, 22 Jan 2014 02:17:52 GMT"
},
{
"version": "v2",
"created": "Thu, 20 Feb 2014 08:13:48 GMT"
},
{
"version": "v3",
"created": "Thu, 27 Feb 2014 03:40:26 GMT"
}
] | 2014-02-28T00:00:00 | [
[
"Ming",
"Zijian",
""
],
[
"Luo",
"Chunjie",
""
],
[
"Gao",
"Wanling",
""
],
[
"Han",
"Rui",
""
],
[
"Yang",
"Qiang",
""
],
[
"Wang",
"Lei",
""
],
[
"Zhan",
"Jianfeng",
""
]
] | TITLE: BDGS: A Scalable Big Data Generator Suite in Big Data Benchmarking
ABSTRACT: Data generation is a key issue in big data benchmarking that aims to generate
application-specific data sets to meet the 4V requirements of big data.
Specifically, big data generators need to generate scalable data (Volume) of
different types (Variety) under controllable generation rates (Velocity) while
keeping the important characteristics of raw data (Veracity). This gives rise
to various new challenges about how we design generators efficiently and
successfully. To date, most existing techniques can only generate limited types
of data and support specific big data systems such as Hadoop. Hence we develop
a tool, called Big Data Generator Suite (BDGS), to efficiently generate
scalable big data while employing data models derived from real data to
preserve data veracity. The effectiveness of BDGS is demonstrated by developing
six data generators covering three representative data types (structured,
semi-structured and unstructured) and three data sources (text, graph, and
table data).
|
1402.6366 | Mustafa Abdul Salam | Osman Hegazy, Omar S. Soliman and Mustafa Abdul Salam | LSSVM-ABC Algorithm for Stock Price prediction | 12 pages. International Journal of Computer Trends and Technology
(IJCTT)2014 | null | null | null | cs.CE cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, Artificial Bee Colony (ABC) algorithm which inspired from the
behavior of honey bees swarm is presented. ABC is a stochastic population-based
evolutionary algorithm for problem solving. ABC algorithm, which is considered
one of the most recently swarm intelligent techniques, is proposed to optimize
least square support vector machine (LSSVM) to predict the daily stock prices.
The proposed model is based on the study of stocks historical data, technical
indicators and optimizing LSSVM with ABC algorithm. ABC selects best free
parameters combination for LSSVM to avoid over-fitting and local minima
problems and improve prediction accuracy. LSSVM optimized by Particle swarm
optimization (PSO) algorithm, LSSVM, and ANN techniques are used for comparison
with proposed model. Proposed model tested with twenty datasets representing
different sectors in S&P 500 stock market. Results presented in this paper show
that the proposed model has fast convergence speed, and it also achieves better
accuracy than compared techniques in most cases.
| [
{
"version": "v1",
"created": "Tue, 25 Feb 2014 23:02:08 GMT"
}
] | 2014-02-28T00:00:00 | [
[
"Hegazy",
"Osman",
""
],
[
"Soliman",
"Omar S.",
""
],
[
"Salam",
"Mustafa Abdul",
""
]
] | TITLE: LSSVM-ABC Algorithm for Stock Price prediction
ABSTRACT: In this paper, Artificial Bee Colony (ABC) algorithm which inspired from the
behavior of honey bees swarm is presented. ABC is a stochastic population-based
evolutionary algorithm for problem solving. ABC algorithm, which is considered
one of the most recently swarm intelligent techniques, is proposed to optimize
least square support vector machine (LSSVM) to predict the daily stock prices.
The proposed model is based on the study of stocks historical data, technical
indicators and optimizing LSSVM with ABC algorithm. ABC selects best free
parameters combination for LSSVM to avoid over-fitting and local minima
problems and improve prediction accuracy. LSSVM optimized by Particle swarm
optimization (PSO) algorithm, LSSVM, and ANN techniques are used for comparison
with proposed model. Proposed model tested with twenty datasets representing
different sectors in S&P 500 stock market. Results presented in this paper show
that the proposed model has fast convergence speed, and it also achieves better
accuracy than compared techniques in most cases.
|
1402.6865 | J\'er\^ome Kunegis | J\'er\^ome Kunegis | Applications of Structural Balance in Signed Social Networks | 37 pages | null | null | null | cs.SI physics.soc-ph | http://creativecommons.org/licenses/by/3.0/ | We present measures, models and link prediction algorithms based on the
structural balance in signed social networks. Certain social networks contain,
in addition to the usual 'friend' links, 'enemy' links. These networks are
called signed social networks. A classical and major concept for signed social
networks is that of structural balance, i.e., the tendency of triangles to be
'balanced' towards including an even number of negative edges, such as
friend-friend-friend and friend-enemy-enemy triangles. In this article, we
introduce several new signed network analysis methods that exploit structural
balance for measuring partial balance, for finding communities of people based
on balance, for drawing signed social networks, and for solving the problem of
link prediction. Notably, the introduced methods are based on the signed graph
Laplacian and on the concept of signed resistance distances. We evaluate our
methods on a collection of four signed social network datasets.
| [
{
"version": "v1",
"created": "Thu, 27 Feb 2014 11:32:50 GMT"
}
] | 2014-02-28T00:00:00 | [
[
"Kunegis",
"Jérôme",
""
]
] | TITLE: Applications of Structural Balance in Signed Social Networks
ABSTRACT: We present measures, models and link prediction algorithms based on the
structural balance in signed social networks. Certain social networks contain,
in addition to the usual 'friend' links, 'enemy' links. These networks are
called signed social networks. A classical and major concept for signed social
networks is that of structural balance, i.e., the tendency of triangles to be
'balanced' towards including an even number of negative edges, such as
friend-friend-friend and friend-enemy-enemy triangles. In this article, we
introduce several new signed network analysis methods that exploit structural
balance for measuring partial balance, for finding communities of people based
on balance, for drawing signed social networks, and for solving the problem of
link prediction. Notably, the introduced methods are based on the signed graph
Laplacian and on the concept of signed resistance distances. We evaluate our
methods on a collection of four signed social network datasets.
|
1402.6942 | Jakub Nalepa | Jakub Nalepa and Zbigniew J. Czech | A Parallel Memetic Algorithm to Solve the Vehicle Routing Problem with
Time Windows | 15 pages | Studia Informatica 33 (1), pp 91-106 (2012) | null | null | cs.DC cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents a parallel memetic algorithm for solving the vehicle
routing problem with time windows (VRPTW). The VRPTW is a well-known NP-hard
discrete optimization problem with two objectives. The main objective is to
minimize the number of vehicles serving customers scattered on the map, and the
second one is to minimize the total distance traveled by the vehicles. Here,
the fleet size is minimized in the first phase of the proposed method using the
parallel heuristic algorithm (PHA), and the traveled distance is minimized in
the second phase by the parallel memetic algorithm (PMA). In both parallel
algorithms, the parallel components co-operate periodically in order to
exchange the best solutions found so far. An extensive experimental study
performed on the Gehring and Homberger's benchmark proves the high convergence
capabilities and robustness of both PHA and PMA. Also, we present the speedup
analysis of the PMA.
| [
{
"version": "v1",
"created": "Thu, 27 Feb 2014 15:41:05 GMT"
}
] | 2014-02-28T00:00:00 | [
[
"Nalepa",
"Jakub",
""
],
[
"Czech",
"Zbigniew J.",
""
]
] | TITLE: A Parallel Memetic Algorithm to Solve the Vehicle Routing Problem with
Time Windows
ABSTRACT: This paper presents a parallel memetic algorithm for solving the vehicle
routing problem with time windows (VRPTW). The VRPTW is a well-known NP-hard
discrete optimization problem with two objectives. The main objective is to
minimize the number of vehicles serving customers scattered on the map, and the
second one is to minimize the total distance traveled by the vehicles. Here,
the fleet size is minimized in the first phase of the proposed method using the
parallel heuristic algorithm (PHA), and the traveled distance is minimized in
the second phase by the parallel memetic algorithm (PMA). In both parallel
algorithms, the parallel components co-operate periodically in order to
exchange the best solutions found so far. An extensive experimental study
performed on the Gehring and Homberger's benchmark proves the high convergence
capabilities and robustness of both PHA and PMA. Also, we present the speedup
analysis of the PMA.
|
1402.7005 | Ziyu Wang | Ziyu Wang, Babak Shakibi, Lin Jin, Nando de Freitas | Bayesian Multi-Scale Optimistic Optimization | 15 pages | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Bayesian optimization is a powerful global optimization technique for
expensive black-box functions. One of its shortcomings is that it requires
auxiliary optimization of an acquisition function at each iteration. This
auxiliary optimization can be costly and very hard to carry out in practice.
Moreover, it creates serious theoretical concerns, as most of the convergence
results assume that the exact optimum of the acquisition function can be found.
In this paper, we introduce a new technique for efficient global optimization
that combines Gaussian process confidence bounds and treed simultaneous
optimistic optimization to eliminate the need for auxiliary optimization of
acquisition functions. The experiments with global optimization benchmarks and
a novel application to automatic information extraction demonstrate that the
resulting technique is more efficient than the two approaches from which it
draws inspiration. Unlike most theoretical analyses of Bayesian optimization
with Gaussian processes, our finite-time convergence rate proofs do not require
exact optimization of an acquisition function. That is, our approach eliminates
the unsatisfactory assumption that a difficult, potentially NP-hard, problem
has to be solved in order to obtain vanishing regret rates.
| [
{
"version": "v1",
"created": "Thu, 27 Feb 2014 18:38:02 GMT"
}
] | 2014-02-28T00:00:00 | [
[
"Wang",
"Ziyu",
""
],
[
"Shakibi",
"Babak",
""
],
[
"Jin",
"Lin",
""
],
[
"de Freitas",
"Nando",
""
]
] | TITLE: Bayesian Multi-Scale Optimistic Optimization
ABSTRACT: Bayesian optimization is a powerful global optimization technique for
expensive black-box functions. One of its shortcomings is that it requires
auxiliary optimization of an acquisition function at each iteration. This
auxiliary optimization can be costly and very hard to carry out in practice.
Moreover, it creates serious theoretical concerns, as most of the convergence
results assume that the exact optimum of the acquisition function can be found.
In this paper, we introduce a new technique for efficient global optimization
that combines Gaussian process confidence bounds and treed simultaneous
optimistic optimization to eliminate the need for auxiliary optimization of
acquisition functions. The experiments with global optimization benchmarks and
a novel application to automatic information extraction demonstrate that the
resulting technique is more efficient than the two approaches from which it
draws inspiration. Unlike most theoretical analyses of Bayesian optimization
with Gaussian processes, our finite-time convergence rate proofs do not require
exact optimization of an acquisition function. That is, our approach eliminates
the unsatisfactory assumption that a difficult, potentially NP-hard, problem
has to be solved in order to obtain vanishing regret rates.
|
1402.7063 | Spyros Sioutas SS | Nikolaos Nodarakis, Spyros Sioutas, Dimitrios Tsoumakos, Giannis
Tzimas and Evaggelia Pitoura | Rapid AkNN Query Processing for Fast Classification of Multidimensional
Data in the Cloud | 12 pages, 14 figures, 4 tables (it will be submitted to DEXA 2014) | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A $k$-nearest neighbor ($k$NN) query determines the $k$ nearest points, using
distance metrics, from a specific location. An all $k$-nearest neighbor
(A$k$NN) query constitutes a variation of a $k$NN query and retrieves the $k$
nearest points for each point inside a database. Their main usage resonates in
spatial databases and they consist the backbone of many location-based
applications and not only (i.e. $k$NN joins in databases, classification in
data mining). So, it is very crucial to develop methods that answer them
efficiently. In this work, we propose a novel method for classifying
multidimensional data using an A$k$NN algorithm in the MapReduce framework. Our
approach exploits space decomposition techniques for processing the
classification procedure in a parallel and distributed manner. To our
knowledge, we are the first to study the classification of multidimensional
objects under this perspective. Through an extensive experimental evaluation we
prove that our solution is efficient and scalable in processing the given
queries. We investigate many different perspectives that can affect the total
computational cost, such as different dataset distributions, number of
dimensions, growth of $k$ value and granularity of space decomposition and
prove that our system is efficient, robust and scalable.
| [
{
"version": "v1",
"created": "Thu, 27 Feb 2014 20:46:09 GMT"
}
] | 2014-02-28T00:00:00 | [
[
"Nodarakis",
"Nikolaos",
""
],
[
"Sioutas",
"Spyros",
""
],
[
"Tsoumakos",
"Dimitrios",
""
],
[
"Tzimas",
"Giannis",
""
],
[
"Pitoura",
"Evaggelia",
""
]
] | TITLE: Rapid AkNN Query Processing for Fast Classification of Multidimensional
Data in the Cloud
ABSTRACT: A $k$-nearest neighbor ($k$NN) query determines the $k$ nearest points, using
distance metrics, from a specific location. An all $k$-nearest neighbor
(A$k$NN) query constitutes a variation of a $k$NN query and retrieves the $k$
nearest points for each point inside a database. Their main usage resonates in
spatial databases and they consist the backbone of many location-based
applications and not only (i.e. $k$NN joins in databases, classification in
data mining). So, it is very crucial to develop methods that answer them
efficiently. In this work, we propose a novel method for classifying
multidimensional data using an A$k$NN algorithm in the MapReduce framework. Our
approach exploits space decomposition techniques for processing the
classification procedure in a parallel and distributed manner. To our
knowledge, we are the first to study the classification of multidimensional
objects under this perspective. Through an extensive experimental evaluation we
prove that our solution is efficient and scalable in processing the given
queries. We investigate many different perspectives that can affect the total
computational cost, such as different dataset distributions, number of
dimensions, growth of $k$ value and granularity of space decomposition and
prove that our system is efficient, robust and scalable.
|
1402.6428 | Vishakha Metre VAM | Jayshree Ghorpade-Aher and Vishakha A. Metre | Clustering Multidimensional Data with PSO based Algorithm | 6 pages,6 figures,3 tables, conference paper | null | null | null | cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Data clustering is a recognized data analysis method in data mining whereas
K-Means is the well known partitional clustering method, possessing pleasant
features. We observed that, K-Means and other partitional clustering techniques
suffer from several limitations such as initial cluster centre selection,
preknowledge of number of clusters, dead unit problem, multiple cluster
membership and premature convergence to local optima. Several optimization
methods are proposed in the literature in order to solve clustering
limitations, but Swarm Intelligence (SI) has achieved its remarkable position
in the concerned area. Particle Swarm Optimization (PSO) is the most popular SI
technique and one of the favorite areas of researchers. In this paper, we
present a brief overview of PSO and applicability of its variants to solve
clustering challenges. Also, we propose an advanced PSO algorithm named as
Subtractive Clustering based Boundary Restricted Adaptive Particle Swarm
Optimization (SC-BR-APSO) algorithm for clustering multidimensional data. For
comparison purpose, we have studied and analyzed various algorithms such as
K-Means, PSO, K-Means-PSO, Hybrid Subtractive + PSO, BRAPSO, and proposed
algorithm on nine different datasets. The motivation behind proposing
SC-BR-APSO algorithm is to deal with multidimensional data clustering, with
minimum error rate and maximum convergence rate.
| [
{
"version": "v1",
"created": "Wed, 26 Feb 2014 06:08:27 GMT"
}
] | 2014-02-27T00:00:00 | [
[
"Ghorpade-Aher",
"Jayshree",
""
],
[
"Metre",
"Vishakha A.",
""
]
] | TITLE: Clustering Multidimensional Data with PSO based Algorithm
ABSTRACT: Data clustering is a recognized data analysis method in data mining whereas
K-Means is the well known partitional clustering method, possessing pleasant
features. We observed that, K-Means and other partitional clustering techniques
suffer from several limitations such as initial cluster centre selection,
preknowledge of number of clusters, dead unit problem, multiple cluster
membership and premature convergence to local optima. Several optimization
methods are proposed in the literature in order to solve clustering
limitations, but Swarm Intelligence (SI) has achieved its remarkable position
in the concerned area. Particle Swarm Optimization (PSO) is the most popular SI
technique and one of the favorite areas of researchers. In this paper, we
present a brief overview of PSO and applicability of its variants to solve
clustering challenges. Also, we propose an advanced PSO algorithm named as
Subtractive Clustering based Boundary Restricted Adaptive Particle Swarm
Optimization (SC-BR-APSO) algorithm for clustering multidimensional data. For
comparison purpose, we have studied and analyzed various algorithms such as
K-Means, PSO, K-Means-PSO, Hybrid Subtractive + PSO, BRAPSO, and proposed
algorithm on nine different datasets. The motivation behind proposing
SC-BR-APSO algorithm is to deal with multidimensional data clustering, with
minimum error rate and maximum convergence rate.
|
1402.6636 | Iain Rice Mr | Iain Rice, Roger Benton, Les Hart and David Lowe | Analysis of Multibeam SONAR Data using Dissimilarity Representations | Presented at IMA Mathematics in Defence 2013 | null | null | null | cs.CE stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper considers the problem of low-dimensional visualisation of very
high dimensional information sources for the purpose of situation awareness in
the maritime environment. In response to the requirement for human decision
support aids to reduce information overload (and specifically, data amenable to
inter-point relative similarity measures) appropriate to the below-water
maritime domain, we are investigating a preliminary prototype topographic
visualisation model. The focus of the current paper is on the mathematical
problem of exploiting a relative dissimilarity representation of signals in a
visual informatics mapping model, driven by real-world sonar systems. An
independent source model is used to analyse the sonar beams from which a simple
probabilistic input model to represent uncertainty is mapped to a latent
visualisation space where data uncertainty can be accommodated. The use of
euclidean and non-euclidean measures are used and the motivation for future use
of non-euclidean measures is made. Concepts are illustrated using a simulated
64 beam weak SNR dataset with realistic sonar targets.
| [
{
"version": "v1",
"created": "Wed, 19 Feb 2014 10:21:34 GMT"
}
] | 2014-02-27T00:00:00 | [
[
"Rice",
"Iain",
""
],
[
"Benton",
"Roger",
""
],
[
"Hart",
"Les",
""
],
[
"Lowe",
"David",
""
]
] | TITLE: Analysis of Multibeam SONAR Data using Dissimilarity Representations
ABSTRACT: This paper considers the problem of low-dimensional visualisation of very
high dimensional information sources for the purpose of situation awareness in
the maritime environment. In response to the requirement for human decision
support aids to reduce information overload (and specifically, data amenable to
inter-point relative similarity measures) appropriate to the below-water
maritime domain, we are investigating a preliminary prototype topographic
visualisation model. The focus of the current paper is on the mathematical
problem of exploiting a relative dissimilarity representation of signals in a
visual informatics mapping model, driven by real-world sonar systems. An
independent source model is used to analyse the sonar beams from which a simple
probabilistic input model to represent uncertainty is mapped to a latent
visualisation space where data uncertainty can be accommodated. The use of
euclidean and non-euclidean measures are used and the motivation for future use
of non-euclidean measures is made. Concepts are illustrated using a simulated
64 beam weak SNR dataset with realistic sonar targets.
|
1402.6650 | Ahmed Sahlol | Ahmed Sahlol and Cheng Suen | A Novel Method for the Recognition of Isolated Handwritten Arabic
Characters | Indicate 13 pages, 5 figures | null | null | null | cs.CV | http://creativecommons.org/licenses/by/3.0/ | There are many difficulties facing a handwritten Arabic recognition system
such as unlimited variation in human handwriting, similarities of distinct
character shapes, interconnections of neighbouring characters and their
position in the word. The typical Optical Character Recognition (OCR) systems
are based mainly on three stages, preprocessing, features extraction and
recognition. This paper proposes new methods for handwritten Arabic character
recognition which is based on novel preprocessing operations including
different kinds of noise removal also different kind of features like
structural, Statistical and Morphological features from the main body of the
character and also from the secondary components. Evaluation of the accuracy of
the selected features is made. The system was trained and tested by back
propagation neural network with CENPRMI dataset. The proposed algorithm
obtained promising results as it is able to recognize 88% of our test set
accurately. In Comparable with other related works we find that our result is
the highest among other published works.
| [
{
"version": "v1",
"created": "Wed, 26 Feb 2014 19:09:09 GMT"
}
] | 2014-02-27T00:00:00 | [
[
"Sahlol",
"Ahmed",
""
],
[
"Suen",
"Cheng",
""
]
] | TITLE: A Novel Method for the Recognition of Isolated Handwritten Arabic
Characters
ABSTRACT: There are many difficulties facing a handwritten Arabic recognition system
such as unlimited variation in human handwriting, similarities of distinct
character shapes, interconnections of neighbouring characters and their
position in the word. The typical Optical Character Recognition (OCR) systems
are based mainly on three stages, preprocessing, features extraction and
recognition. This paper proposes new methods for handwritten Arabic character
recognition which is based on novel preprocessing operations including
different kinds of noise removal also different kind of features like
structural, Statistical and Morphological features from the main body of the
character and also from the secondary components. Evaluation of the accuracy of
the selected features is made. The system was trained and tested by back
propagation neural network with CENPRMI dataset. The proposed algorithm
obtained promising results as it is able to recognize 88% of our test set
accurately. In Comparable with other related works we find that our result is
the highest among other published works.
|
1402.6690 | Jalal Mahmud | Jalal Mahmud, Jilin Chen, Jeffrey Nichols | Why Are You More Engaged? Predicting Social Engagement from Word Use | null | null | null | null | cs.SI cs.CL cs.CY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a study to analyze how word use can predict social engagement
behaviors such as replies and retweets in Twitter. We compute psycholinguistic
category scores from word usage, and investigate how people with different
scores exhibited different reply and retweet behaviors on Twitter. We also
found psycholinguistic categories that show significant correlations with such
social engagement behaviors. In addition, we have built predictive models of
replies and retweets from such psycholinguistic category based features. Our
experiments using a real world dataset collected from Twitter validates that
such predictions can be done with reasonable accuracy.
| [
{
"version": "v1",
"created": "Wed, 26 Feb 2014 20:58:00 GMT"
}
] | 2014-02-27T00:00:00 | [
[
"Mahmud",
"Jalal",
""
],
[
"Chen",
"Jilin",
""
],
[
"Nichols",
"Jeffrey",
""
]
] | TITLE: Why Are You More Engaged? Predicting Social Engagement from Word Use
ABSTRACT: We present a study to analyze how word use can predict social engagement
behaviors such as replies and retweets in Twitter. We compute psycholinguistic
category scores from word usage, and investigate how people with different
scores exhibited different reply and retweet behaviors on Twitter. We also
found psycholinguistic categories that show significant correlations with such
social engagement behaviors. In addition, we have built predictive models of
replies and retweets from such psycholinguistic category based features. Our
experiments using a real world dataset collected from Twitter validates that
such predictions can be done with reasonable accuracy.
|
1212.1522 | Vasilis Gkatzelis | Richard Cole, Vasilis Gkatzelis, Gagan Goel | Mechanism Design for Fair Division | arXiv admin note: substantial text overlap with arXiv:1203.4627 | null | null | null | cs.GT cs.DS cs.MA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We revisit the classic problem of fair division from a mechanism design
perspective, using {\em Proportional Fairness} as a benchmark. In particular,
we aim to allocate a collection of divisible items to a set of agents while
incentivizing the agents to be truthful in reporting their valuations. For the
very large class of homogeneous valuations, we design a truthful mechanism that
provides {\em every agent} with at least a $1/e\approx 0.368$ fraction of her
Proportionally Fair valuation. To complement this result, we show that no
truthful mechanism can guarantee more than a $0.5$ fraction, even for the
restricted class of additive linear valuations. We also propose another
mechanism for additive linear valuations that works really well when every item
is highly demanded. To guarantee truthfulness, our mechanisms discard a
carefully chosen fraction of the allocated resources; we conclude by uncovering
interesting connections between our mechanisms and known mechanisms that use
money instead.
| [
{
"version": "v1",
"created": "Fri, 7 Dec 2012 02:31:27 GMT"
},
{
"version": "v2",
"created": "Tue, 12 Feb 2013 05:47:58 GMT"
},
{
"version": "v3",
"created": "Tue, 25 Feb 2014 00:14:37 GMT"
}
] | 2014-02-26T00:00:00 | [
[
"Cole",
"Richard",
""
],
[
"Gkatzelis",
"Vasilis",
""
],
[
"Goel",
"Gagan",
""
]
] | TITLE: Mechanism Design for Fair Division
ABSTRACT: We revisit the classic problem of fair division from a mechanism design
perspective, using {\em Proportional Fairness} as a benchmark. In particular,
we aim to allocate a collection of divisible items to a set of agents while
incentivizing the agents to be truthful in reporting their valuations. For the
very large class of homogeneous valuations, we design a truthful mechanism that
provides {\em every agent} with at least a $1/e\approx 0.368$ fraction of her
Proportionally Fair valuation. To complement this result, we show that no
truthful mechanism can guarantee more than a $0.5$ fraction, even for the
restricted class of additive linear valuations. We also propose another
mechanism for additive linear valuations that works really well when every item
is highly demanded. To guarantee truthfulness, our mechanisms discard a
carefully chosen fraction of the allocated resources; we conclude by uncovering
interesting connections between our mechanisms and known mechanisms that use
money instead.
|
1301.7006 | Michel Plantie | Michel Crampes, Michel Planti\'e | A Unified Community Detection, Visualization and Analysis method | Submitted to Advances in Complex Systems | null | null | null | cs.SI physics.data-an physics.soc-ph stat.OT | http://creativecommons.org/licenses/publicdomain/ | Community detection in social graphs has attracted researchers' interest for
a long time. With the widespread of social networks on the Internet it has
recently become an important research domain. Most contributions focus upon the
definition of algorithms for optimizing the so-called modularity function. In
the first place interest was limited to unipartite graph inputs and partitioned
community outputs. Recently bipartite graphs, directed graphs and overlapping
communities have been investigated. Few contributions embrace at the same time
the three types of nodes. In this paper we present a method which unifies
commmunity detection for the three types of nodes and at the same time merges
partitionned and overlapping communities. Moreover results are visualized in
such a way that they can be analyzed and semantically interpreted. For
validation we experiment this method on well known simple benchmarks. It is
then applied to real data in three cases. In two examples of photos sets with
tagged people we reveal social networks. A second type of application is of
particularly interest. After applying our method to Human Brain Tractography
Data provided by a team of neurologists, we produce clusters of white fibers in
accordance with other well known clustering methods. Moreover our approach for
visualizing overlapping clusters allows better understanding of the results by
the neurologist team. These last results open up the possibility of applying
community detection methods in other domains such as data analysis with
original enhanced performances.
| [
{
"version": "v1",
"created": "Tue, 29 Jan 2013 18:02:44 GMT"
},
{
"version": "v2",
"created": "Wed, 30 Jan 2013 15:13:52 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jan 2013 16:11:59 GMT"
},
{
"version": "v4",
"created": "Fri, 1 Feb 2013 14:59:59 GMT"
},
{
"version": "v5",
"created": "Fri, 15 Feb 2013 17:41:14 GMT"
},
{
"version": "v6",
"created": "Tue, 25 Feb 2014 17:35:36 GMT"
}
] | 2014-02-26T00:00:00 | [
[
"Crampes",
"Michel",
""
],
[
"Plantié",
"Michel",
""
]
] | TITLE: A Unified Community Detection, Visualization and Analysis method
ABSTRACT: Community detection in social graphs has attracted researchers' interest for
a long time. With the widespread of social networks on the Internet it has
recently become an important research domain. Most contributions focus upon the
definition of algorithms for optimizing the so-called modularity function. In
the first place interest was limited to unipartite graph inputs and partitioned
community outputs. Recently bipartite graphs, directed graphs and overlapping
communities have been investigated. Few contributions embrace at the same time
the three types of nodes. In this paper we present a method which unifies
commmunity detection for the three types of nodes and at the same time merges
partitionned and overlapping communities. Moreover results are visualized in
such a way that they can be analyzed and semantically interpreted. For
validation we experiment this method on well known simple benchmarks. It is
then applied to real data in three cases. In two examples of photos sets with
tagged people we reveal social networks. A second type of application is of
particularly interest. After applying our method to Human Brain Tractography
Data provided by a team of neurologists, we produce clusters of white fibers in
accordance with other well known clustering methods. Moreover our approach for
visualizing overlapping clusters allows better understanding of the results by
the neurologist team. These last results open up the possibility of applying
community detection methods in other domains such as data analysis with
original enhanced performances.
|
1306.1704 | Dmytro Karamshuk | Dmytro Karamshuk, Anastasios Noulas, Salvatore Scellato, Vincenzo
Nicosia, Cecilia Mascolo | Geo-Spotting: Mining Online Location-based Services for Optimal Retail
Store Placement | Proceedings of the 19th ACM SIGKDD international conference on
Knowledge discovery and data mining, Chicago, 2013, Pages 793-801 | null | 10.1145/2487575.2487616 | null | cs.SI cs.CE physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The problem of identifying the optimal location for a new retail store has
been the focus of past research, especially in the field of land economy, due
to its importance in the success of a business. Traditional approaches to the
problem have factored in demographics, revenue and aggregated human flow
statistics from nearby or remote areas. However, the acquisition of relevant
data is usually expensive. With the growth of location-based social networks,
fine grained data describing user mobility and popularity of places has
recently become attainable.
In this paper we study the predictive power of various machine learning
features on the popularity of retail stores in the city through the use of a
dataset collected from Foursquare in New York. The features we mine are based
on two general signals: geographic, where features are formulated according to
the types and density of nearby places, and user mobility, which includes
transitions between venues or the incoming flow of mobile users from distant
areas. Our evaluation suggests that the best performing features are common
across the three different commercial chains considered in the analysis,
although variations may exist too, as explained by heterogeneities in the way
retail facilities attract users. We also show that performance improves
significantly when combining multiple features in supervised learning
algorithms, suggesting that the retail success of a business may depend on
multiple factors.
| [
{
"version": "v1",
"created": "Fri, 7 Jun 2013 12:42:06 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Feb 2014 10:48:20 GMT"
}
] | 2014-02-26T00:00:00 | [
[
"Karamshuk",
"Dmytro",
""
],
[
"Noulas",
"Anastasios",
""
],
[
"Scellato",
"Salvatore",
""
],
[
"Nicosia",
"Vincenzo",
""
],
[
"Mascolo",
"Cecilia",
""
]
] | TITLE: Geo-Spotting: Mining Online Location-based Services for Optimal Retail
Store Placement
ABSTRACT: The problem of identifying the optimal location for a new retail store has
been the focus of past research, especially in the field of land economy, due
to its importance in the success of a business. Traditional approaches to the
problem have factored in demographics, revenue and aggregated human flow
statistics from nearby or remote areas. However, the acquisition of relevant
data is usually expensive. With the growth of location-based social networks,
fine grained data describing user mobility and popularity of places has
recently become attainable.
In this paper we study the predictive power of various machine learning
features on the popularity of retail stores in the city through the use of a
dataset collected from Foursquare in New York. The features we mine are based
on two general signals: geographic, where features are formulated according to
the types and density of nearby places, and user mobility, which includes
transitions between venues or the incoming flow of mobile users from distant
areas. Our evaluation suggests that the best performing features are common
across the three different commercial chains considered in the analysis,
although variations may exist too, as explained by heterogeneities in the way
retail facilities attract users. We also show that performance improves
significantly when combining multiple features in supervised learning
algorithms, suggesting that the retail success of a business may depend on
multiple factors.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.