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
label
stringclasses
2 values
prob
float64
0.5
0.73
1406.3837
Thomas Laurent
Xavier Bresson, Huiyi Hu, Thomas Laurent, Arthur Szlam, and James von Brecht
An Incremental Reseeding Strategy for Clustering
null
null
null
null
stat.ML cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
In this work we propose a simple and easily parallelizable algorithm for multiway graph partitioning. The algorithm alternates between three basic components: diffusing seed vertices over the graph, thresholding the diffused seeds, and then randomly reseeding the thresholded clusters. We demonstrate experimentally that the proper combination of these ingredients leads to an algorithm that achieves state-of-the-art performance in terms of cluster purity on standard benchmarks datasets. Moreover, the algorithm runs an order of magnitude faster than the other algorithms that achieve comparable results in terms of accuracy. We also describe a coarsen, cluster and refine approach similar to GRACLUS and METIS that removes an additional order of magnitude from the runtime of our algorithm while still maintaining competitive accuracy.
[ { "version": "v1", "created": "Sun, 15 Jun 2014 18:30:51 GMT" } ]
2014-06-17T00:00:00
[ [ "Bresson", "Xavier", "" ], [ "Hu", "Huiyi", "" ], [ "Laurent", "Thomas", "" ], [ "Szlam", "Arthur", "" ], [ "von Brecht", "James", "" ] ]
TITLE: An Incremental Reseeding Strategy for Clustering ABSTRACT: In this work we propose a simple and easily parallelizable algorithm for multiway graph partitioning. The algorithm alternates between three basic components: diffusing seed vertices over the graph, thresholding the diffused seeds, and then randomly reseeding the thresholded clusters. We demonstrate experimentally that the proper combination of these ingredients leads to an algorithm that achieves state-of-the-art performance in terms of cluster purity on standard benchmarks datasets. Moreover, the algorithm runs an order of magnitude faster than the other algorithms that achieve comparable results in terms of accuracy. We also describe a coarsen, cluster and refine approach similar to GRACLUS and METIS that removes an additional order of magnitude from the runtime of our algorithm while still maintaining competitive accuracy.
no_new_dataset
0.71202
1406.3901
Liya Fan
Liya Fan, Bo Gao, Fa Zhang, Zhiyong Liu
OS4M: Achieving Global Load Balance of MapReduce Workload by Scheduling at the Operation Level
arXiv admin note: substantial text overlap with arXiv:1401.0355
null
null
null
cs.DC
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The efficiency of MapReduce is closely related to its load balance. Existing works on MapReduce load balance focus on coarse-grained scheduling. This study concerns fine-grained scheduling on MapReduce operations, with each operation representing one invocation of the Map or Reduce function. By default, MapReduce adopts the hash-based method to schedule Reduce operations, which often leads to poor load balance. In addition, the copy phase of Reduce tasks overlaps with Map tasks, which significantly hinders the progress of Map tasks due to I/O contention. Moreover, the three phases of Reduce tasks run in sequence, while consuming different resources, thereby under-utilizing resources. To overcome these problems, we introduce a set of mechanisms named OS4M (Operation Scheduling for MapReduce) to improve MapReduce's performance. OS4M achieves load balance by collecting statistics of all Map operations, and calculates a globally optimal schedule to distribute Reduce operations. With OS4M, the copy phase of Reduce tasks no longer overlaps with Map tasks, and the three phases of Reduce tasks are pipelined based on their operation loads. OS4M has been transparently incorporated into MapReduce. Evaluations on standard benchmarks show that OS4M's job duration can be shortened by up to 42%, compared with a baseline of Hadoop.
[ { "version": "v1", "created": "Mon, 16 Jun 2014 04:16:41 GMT" } ]
2014-06-17T00:00:00
[ [ "Fan", "Liya", "" ], [ "Gao", "Bo", "" ], [ "Zhang", "Fa", "" ], [ "Liu", "Zhiyong", "" ] ]
TITLE: OS4M: Achieving Global Load Balance of MapReduce Workload by Scheduling at the Operation Level ABSTRACT: The efficiency of MapReduce is closely related to its load balance. Existing works on MapReduce load balance focus on coarse-grained scheduling. This study concerns fine-grained scheduling on MapReduce operations, with each operation representing one invocation of the Map or Reduce function. By default, MapReduce adopts the hash-based method to schedule Reduce operations, which often leads to poor load balance. In addition, the copy phase of Reduce tasks overlaps with Map tasks, which significantly hinders the progress of Map tasks due to I/O contention. Moreover, the three phases of Reduce tasks run in sequence, while consuming different resources, thereby under-utilizing resources. To overcome these problems, we introduce a set of mechanisms named OS4M (Operation Scheduling for MapReduce) to improve MapReduce's performance. OS4M achieves load balance by collecting statistics of all Map operations, and calculates a globally optimal schedule to distribute Reduce operations. With OS4M, the copy phase of Reduce tasks no longer overlaps with Map tasks, and the three phases of Reduce tasks are pipelined based on their operation loads. OS4M has been transparently incorporated into MapReduce. Evaluations on standard benchmarks show that OS4M's job duration can be shortened by up to 42%, compared with a baseline of Hadoop.
no_new_dataset
0.704935
1406.3949
Jamil Ahmad
Jamil Ahmad, Zahoor Jan, Zia-ud-Din and Shoaib Muhammad Khan
A Fusion of Labeled-Grid Shape Descriptors with Weighted Ranking Algorithm for Shapes Recognition
null
World Applied Sciences Journal, vol. 31(6), pp. 1207-1213, 2014
10.5829/idosi.wasj.2014.31.06.353
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Retrieving similar images from a large dataset based on the image content has been a very active research area and is a very challenging task. Studies have shown that retrieving similar images based on their shape is a very effective method. For this purpose a large number of methods exist in literature. The combination of more than one feature has also been investigated for this purpose and has shown promising results. In this paper a fusion based shapes recognition method has been proposed. A set of local boundary based and region based features are derived from the labeled grid based representation of the shape and are combined with a few global shape features to produce a composite shape descriptor. This composite shape descriptor is then used in a weighted ranking algorithm to find similarities among shapes from a large dataset. The experimental analysis has shown that the proposed method is powerful enough to discriminate the geometrically similar shapes from the non-similar ones.
[ { "version": "v1", "created": "Mon, 16 Jun 2014 09:50:04 GMT" } ]
2014-06-17T00:00:00
[ [ "Ahmad", "Jamil", "" ], [ "Jan", "Zahoor", "" ], [ "Zia-ud-Din", "", "" ], [ "Khan", "Shoaib Muhammad", "" ] ]
TITLE: A Fusion of Labeled-Grid Shape Descriptors with Weighted Ranking Algorithm for Shapes Recognition ABSTRACT: Retrieving similar images from a large dataset based on the image content has been a very active research area and is a very challenging task. Studies have shown that retrieving similar images based on their shape is a very effective method. For this purpose a large number of methods exist in literature. The combination of more than one feature has also been investigated for this purpose and has shown promising results. In this paper a fusion based shapes recognition method has been proposed. A set of local boundary based and region based features are derived from the labeled grid based representation of the shape and are combined with a few global shape features to produce a composite shape descriptor. This composite shape descriptor is then used in a weighted ranking algorithm to find similarities among shapes from a large dataset. The experimental analysis has shown that the proposed method is powerful enough to discriminate the geometrically similar shapes from the non-similar ones.
no_new_dataset
0.711994
1312.1743
Deva Ramanan
Deva Ramanan
Dual coordinate solvers for large-scale structural SVMs
null
null
null
null
cs.LG cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
This manuscript describes a method for training linear SVMs (including binary SVMs, SVM regression, and structural SVMs) from large, out-of-core training datasets. Current strategies for large-scale learning fall into one of two camps; batch algorithms which solve the learning problem given a finite datasets, and online algorithms which can process out-of-core datasets. The former typically requires datasets small enough to fit in memory. The latter is often phrased as a stochastic optimization problem; such algorithms enjoy strong theoretical properties but often require manual tuned annealing schedules, and may converge slowly for problems with large output spaces (e.g., structural SVMs). We discuss an algorithm for an "intermediate" regime in which the data is too large to fit in memory, but the active constraints (support vectors) are small enough to remain in memory. In this case, one can design rather efficient learning algorithms that are as stable as batch algorithms, but capable of processing out-of-core datasets. We have developed such a MATLAB-based solver and used it to train a collection of recognition systems for articulated pose estimation, facial analysis, 3D object recognition, and action classification, all with publicly-available code. This writeup describes the solver in detail.
[ { "version": "v1", "created": "Fri, 6 Dec 2013 00:55:51 GMT" }, { "version": "v2", "created": "Fri, 13 Jun 2014 04:10:06 GMT" } ]
2014-06-16T00:00:00
[ [ "Ramanan", "Deva", "" ] ]
TITLE: Dual coordinate solvers for large-scale structural SVMs ABSTRACT: This manuscript describes a method for training linear SVMs (including binary SVMs, SVM regression, and structural SVMs) from large, out-of-core training datasets. Current strategies for large-scale learning fall into one of two camps; batch algorithms which solve the learning problem given a finite datasets, and online algorithms which can process out-of-core datasets. The former typically requires datasets small enough to fit in memory. The latter is often phrased as a stochastic optimization problem; such algorithms enjoy strong theoretical properties but often require manual tuned annealing schedules, and may converge slowly for problems with large output spaces (e.g., structural SVMs). We discuss an algorithm for an "intermediate" regime in which the data is too large to fit in memory, but the active constraints (support vectors) are small enough to remain in memory. In this case, one can design rather efficient learning algorithms that are as stable as batch algorithms, but capable of processing out-of-core datasets. We have developed such a MATLAB-based solver and used it to train a collection of recognition systems for articulated pose estimation, facial analysis, 3D object recognition, and action classification, all with publicly-available code. This writeup describes the solver in detail.
no_new_dataset
0.710829
1406.0455
Cheng Chen
Cheng Chen, Lan Zheng, Venkatesh Srinivasan, Alex Thomo, Kui Wu, Anthony Sukow
Buyer to Seller Recommendation under Constraints
9 pages, 7 figures
null
null
null
cs.SI cs.GT q-fin.GN q-fin.ST
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The majority of recommender systems are designed to recommend items (such as movies and products) to users. We focus on the problem of recommending buyers to sellers which comes with new challenges: (1) constraints on the number of recommendations buyers are part of before they become overwhelmed, (2) constraints on the number of recommendations sellers receive within their budget, and (3) constraints on the set of buyers that sellers want to receive (e.g., no more than two people from the same household). We propose the following critical problems of recommending buyers to sellers: Constrained Recommendation (C-REC) capturing the first two challenges, and Conflict-Aware Constrained Recommendation (CAC-REC) capturing all three challenges at the same time. We show that C-REC can be modeled using linear programming and can be efficiently solved using modern solvers. On the other hand, we show that CAC-REC is NP-hard. We propose two approximate algorithms to solve CAC-REC and show that they achieve close to optimal solutions via comprehensive experiments using real-world datasets.
[ { "version": "v1", "created": "Mon, 2 Jun 2014 17:45:52 GMT" }, { "version": "v2", "created": "Mon, 9 Jun 2014 05:32:29 GMT" }, { "version": "v3", "created": "Fri, 13 Jun 2014 17:34:26 GMT" } ]
2014-06-16T00:00:00
[ [ "Chen", "Cheng", "" ], [ "Zheng", "Lan", "" ], [ "Srinivasan", "Venkatesh", "" ], [ "Thomo", "Alex", "" ], [ "Wu", "Kui", "" ], [ "Sukow", "Anthony", "" ] ]
TITLE: Buyer to Seller Recommendation under Constraints ABSTRACT: The majority of recommender systems are designed to recommend items (such as movies and products) to users. We focus on the problem of recommending buyers to sellers which comes with new challenges: (1) constraints on the number of recommendations buyers are part of before they become overwhelmed, (2) constraints on the number of recommendations sellers receive within their budget, and (3) constraints on the set of buyers that sellers want to receive (e.g., no more than two people from the same household). We propose the following critical problems of recommending buyers to sellers: Constrained Recommendation (C-REC) capturing the first two challenges, and Conflict-Aware Constrained Recommendation (CAC-REC) capturing all three challenges at the same time. We show that C-REC can be modeled using linear programming and can be efficiently solved using modern solvers. On the other hand, we show that CAC-REC is NP-hard. We propose two approximate algorithms to solve CAC-REC and show that they achieve close to optimal solutions via comprehensive experiments using real-world datasets.
no_new_dataset
0.708389
1402.0929
Jasper Snoek
Jasper Snoek, Kevin Swersky, Richard S. Zemel and Ryan P. Adams
Input Warping for Bayesian Optimization of Non-stationary Functions
null
null
null
null
stat.ML cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Bayesian optimization has proven to be a highly effective methodology for the global optimization of unknown, expensive and multimodal functions. The ability to accurately model distributions over functions is critical to the effectiveness of Bayesian optimization. Although Gaussian processes provide a flexible prior over functions which can be queried efficiently, there are various classes of functions that remain difficult to model. One of the most frequently occurring of these is the class of non-stationary functions. The optimization of the hyperparameters of machine learning algorithms is a problem domain in which parameters are often manually transformed a priori, for example by optimizing in "log-space," to mitigate the effects of spatially-varying length scale. We develop a methodology for automatically learning a wide family of bijective transformations or warpings of the input space using the Beta cumulative distribution function. We further extend the warping framework to multi-task Bayesian optimization so that multiple tasks can be warped into a jointly stationary space. On a set of challenging benchmark optimization tasks, we observe that the inclusion of warping greatly improves on the state-of-the-art, producing better results faster and more reliably.
[ { "version": "v1", "created": "Wed, 5 Feb 2014 03:55:39 GMT" }, { "version": "v2", "created": "Thu, 20 Feb 2014 22:00:38 GMT" }, { "version": "v3", "created": "Wed, 11 Jun 2014 20:32:11 GMT" } ]
2014-06-13T00:00:00
[ [ "Snoek", "Jasper", "" ], [ "Swersky", "Kevin", "" ], [ "Zemel", "Richard S.", "" ], [ "Adams", "Ryan P.", "" ] ]
TITLE: Input Warping for Bayesian Optimization of Non-stationary Functions ABSTRACT: Bayesian optimization has proven to be a highly effective methodology for the global optimization of unknown, expensive and multimodal functions. The ability to accurately model distributions over functions is critical to the effectiveness of Bayesian optimization. Although Gaussian processes provide a flexible prior over functions which can be queried efficiently, there are various classes of functions that remain difficult to model. One of the most frequently occurring of these is the class of non-stationary functions. The optimization of the hyperparameters of machine learning algorithms is a problem domain in which parameters are often manually transformed a priori, for example by optimizing in "log-space," to mitigate the effects of spatially-varying length scale. We develop a methodology for automatically learning a wide family of bijective transformations or warpings of the input space using the Beta cumulative distribution function. We further extend the warping framework to multi-task Bayesian optimization so that multiple tasks can be warped into a jointly stationary space. On a set of challenging benchmark optimization tasks, we observe that the inclusion of warping greatly improves on the state-of-the-art, producing better results faster and more reliably.
no_new_dataset
0.708855
1406.2375
Xiaochen Lian
Wenhao Lu, Xiaochen Lian and Alan Yuille
Parsing Semantic Parts of Cars Using Graphical Models and Segment Appearance Consistency
12 pages, CBMM memo
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
This paper addresses the problem of semantic part parsing (segmentation) of cars, i.e.assigning every pixel within the car to one of the parts (e.g.body, window, lights, license plates and wheels). We formulate this as a landmark identification problem, where a set of landmarks specifies the boundaries of the parts. A novel mixture of graphical models is proposed, which dynamically couples the landmarks to a hierarchy of segments. When modeling pairwise relation between landmarks, this coupling enables our model to exploit the local image contents in addition to spatial deformation, an aspect that most existing graphical models ignore. In particular, our model enforces appearance consistency between segments within the same part. Parsing the car, including finding the optimal coupling between landmarks and segments in the hierarchy, is performed by dynamic programming. We evaluate our method on a subset of PASCAL VOC 2010 car images and on the car subset of 3D Object Category dataset (CAR3D). We show good results and, in particular, quantify the effectiveness of using the segment appearance consistency in terms of accuracy of part localization and segmentation.
[ { "version": "v1", "created": "Mon, 9 Jun 2014 22:16:57 GMT" }, { "version": "v2", "created": "Wed, 11 Jun 2014 23:39:41 GMT" } ]
2014-06-13T00:00:00
[ [ "Lu", "Wenhao", "" ], [ "Lian", "Xiaochen", "" ], [ "Yuille", "Alan", "" ] ]
TITLE: Parsing Semantic Parts of Cars Using Graphical Models and Segment Appearance Consistency ABSTRACT: This paper addresses the problem of semantic part parsing (segmentation) of cars, i.e.assigning every pixel within the car to one of the parts (e.g.body, window, lights, license plates and wheels). We formulate this as a landmark identification problem, where a set of landmarks specifies the boundaries of the parts. A novel mixture of graphical models is proposed, which dynamically couples the landmarks to a hierarchy of segments. When modeling pairwise relation between landmarks, this coupling enables our model to exploit the local image contents in addition to spatial deformation, an aspect that most existing graphical models ignore. In particular, our model enforces appearance consistency between segments within the same part. Parsing the car, including finding the optimal coupling between landmarks and segments in the hierarchy, is performed by dynamic programming. We evaluate our method on a subset of PASCAL VOC 2010 car images and on the car subset of 3D Object Category dataset (CAR3D). We show good results and, in particular, quantify the effectiveness of using the segment appearance consistency in terms of accuracy of part localization and segmentation.
no_new_dataset
0.710396
1406.2807
Yin Li
Yin Li, Xiaodi Hou, Christof Koch, James M. Rehg, Alan L. Yuille
The Secrets of Salient Object Segmentation
15 pages, 8 figures. Conference version was accepted by CVPR 2014
null
null
CBMM Memmo #14
cs.CV
http://creativecommons.org/licenses/by-nc-sa/3.0/
In this paper we provide an extensive evaluation of fixation prediction and salient object segmentation algorithms as well as statistics of major datasets. Our analysis identifies serious design flaws of existing salient object benchmarks, called the dataset design bias, by over emphasizing the stereotypical concepts of saliency. The dataset design bias does not only create the discomforting disconnection between fixations and salient object segmentation, but also misleads the algorithm designing. Based on our analysis, we propose a new high quality dataset that offers both fixation and salient object segmentation ground-truth. With fixations and salient object being presented simultaneously, we are able to bridge the gap between fixations and salient objects, and propose a novel method for salient object segmentation. Finally, we report significant benchmark progress on three existing datasets of segmenting salient objects
[ { "version": "v1", "created": "Wed, 11 Jun 2014 07:46:03 GMT" }, { "version": "v2", "created": "Thu, 12 Jun 2014 17:35:08 GMT" } ]
2014-06-13T00:00:00
[ [ "Li", "Yin", "" ], [ "Hou", "Xiaodi", "" ], [ "Koch", "Christof", "" ], [ "Rehg", "James M.", "" ], [ "Yuille", "Alan L.", "" ] ]
TITLE: The Secrets of Salient Object Segmentation ABSTRACT: In this paper we provide an extensive evaluation of fixation prediction and salient object segmentation algorithms as well as statistics of major datasets. Our analysis identifies serious design flaws of existing salient object benchmarks, called the dataset design bias, by over emphasizing the stereotypical concepts of saliency. The dataset design bias does not only create the discomforting disconnection between fixations and salient object segmentation, but also misleads the algorithm designing. Based on our analysis, we propose a new high quality dataset that offers both fixation and salient object segmentation ground-truth. With fixations and salient object being presented simultaneously, we are able to bridge the gap between fixations and salient objects, and propose a novel method for salient object segmentation. Finally, we report significant benchmark progress on three existing datasets of segmenting salient objects
new_dataset
0.713226
1406.3270
Matthieu Geist
Matthieu Geist, Olivier Pietquin
Kalman Temporal Differences
null
Journal Of Artificial Intelligence Research, Volume 39, pages 483-532, 2010
10.1613/jair.3077
null
cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Because reinforcement learning suffers from a lack of scalability, online value (and Q-) function approximation has received increasing interest this last decade. This contribution introduces a novel approximation scheme, namely the Kalman Temporal Differences (KTD) framework, that exhibits the following features: sample-efficiency, non-linear approximation, non-stationarity handling and uncertainty management. A first KTD-based algorithm is provided for deterministic Markov Decision Processes (MDP) which produces biased estimates in the case of stochastic transitions. Than the eXtended KTD framework (XKTD), solving stochastic MDP, is described. Convergence is analyzed for special cases for both deterministic and stochastic transitions. Related algorithms are experimented on classical benchmarks. They compare favorably to the state of the art while exhibiting the announced features.
[ { "version": "v1", "created": "Thu, 16 Jan 2014 05:02:28 GMT" } ]
2014-06-13T00:00:00
[ [ "Geist", "Matthieu", "" ], [ "Pietquin", "Olivier", "" ] ]
TITLE: Kalman Temporal Differences ABSTRACT: Because reinforcement learning suffers from a lack of scalability, online value (and Q-) function approximation has received increasing interest this last decade. This contribution introduces a novel approximation scheme, namely the Kalman Temporal Differences (KTD) framework, that exhibits the following features: sample-efficiency, non-linear approximation, non-stationarity handling and uncertainty management. A first KTD-based algorithm is provided for deterministic Markov Decision Processes (MDP) which produces biased estimates in the case of stochastic transitions. Than the eXtended KTD framework (XKTD), solving stochastic MDP, is described. Convergence is analyzed for special cases for both deterministic and stochastic transitions. Related algorithms are experimented on classical benchmarks. They compare favorably to the state of the art while exhibiting the announced features.
no_new_dataset
0.707436
1406.3282
Erik Cuevas
Erik Cuevas, Miguel Cienfuegos, Daniel Zaldivar and Marco Perez
A swarm optimization algorithm inspired in the behavior of the social-spider
21 Pages
Expert Systems with Applications, 40 (16), (2013), pp. 6374-6384
null
null
cs.NE
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Swarm intelligence is a research field that models the collective behavior in swarms of insects or animals. Several algorithms arising from such models have been proposed to solve a wide range of complex optimization problems. In this paper, a novel swarm algorithm called the Social Spider Optimization (SSO) is proposed for solving optimization tasks. The SSO algorithm is based on the simulation of cooperative behavior of social-spiders. In the proposed algorithm, individuals emulate a group of spiders which interact to each other based on the biological laws of the cooperative colony. The algorithm considers two different search agents (spiders): males and females. Depending on gender, each individual is conducted by a set of different evolutionary operators which mimic different cooperative behaviors that are typically found in the colony. In order to illustrate the proficiency and robustness of the proposed approach, it is compared to other well-known evolutionary methods. The comparison examines several standard benchmark functions that are commonly considered within the literature of evolutionary algorithms. The outcome shows a high performance of the proposed method for searching a global optimum with several benchmark functions.
[ { "version": "v1", "created": "Thu, 12 Jun 2014 16:29:14 GMT" } ]
2014-06-13T00:00:00
[ [ "Cuevas", "Erik", "" ], [ "Cienfuegos", "Miguel", "" ], [ "Zaldivar", "Daniel", "" ], [ "Perez", "Marco", "" ] ]
TITLE: A swarm optimization algorithm inspired in the behavior of the social-spider ABSTRACT: Swarm intelligence is a research field that models the collective behavior in swarms of insects or animals. Several algorithms arising from such models have been proposed to solve a wide range of complex optimization problems. In this paper, a novel swarm algorithm called the Social Spider Optimization (SSO) is proposed for solving optimization tasks. The SSO algorithm is based on the simulation of cooperative behavior of social-spiders. In the proposed algorithm, individuals emulate a group of spiders which interact to each other based on the biological laws of the cooperative colony. The algorithm considers two different search agents (spiders): males and females. Depending on gender, each individual is conducted by a set of different evolutionary operators which mimic different cooperative behaviors that are typically found in the colony. In order to illustrate the proficiency and robustness of the proposed approach, it is compared to other well-known evolutionary methods. The comparison examines several standard benchmark functions that are commonly considered within the literature of evolutionary algorithms. The outcome shows a high performance of the proposed method for searching a global optimum with several benchmark functions.
no_new_dataset
0.710823
1406.2732
George Papandreou
George Papandreou
Deep Epitomic Convolutional Neural Networks
9 pages
null
null
null
cs.CV cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Deep convolutional neural networks have recently proven extremely competitive in challenging image recognition tasks. This paper proposes the epitomic convolution as a new building block for deep neural networks. An epitomic convolution layer replaces a pair of consecutive convolution and max-pooling layers found in standard deep convolutional neural networks. The main version of the proposed model uses mini-epitomes in place of filters and computes responses invariant to small translations by epitomic search instead of max-pooling over image positions. The topographic version of the proposed model uses large epitomes to learn filter maps organized in translational topographies. We show that error back-propagation can successfully learn multiple epitomic layers in a supervised fashion. The effectiveness of the proposed method is assessed in image classification tasks on standard benchmarks. Our experiments on Imagenet indicate improved recognition performance compared to standard convolutional neural networks of similar architecture. Our models pre-trained on Imagenet perform excellently on Caltech-101. We also obtain competitive image classification results on the small-image MNIST and CIFAR-10 datasets.
[ { "version": "v1", "created": "Tue, 10 Jun 2014 22:07:01 GMT" } ]
2014-06-12T00:00:00
[ [ "Papandreou", "George", "" ] ]
TITLE: Deep Epitomic Convolutional Neural Networks ABSTRACT: Deep convolutional neural networks have recently proven extremely competitive in challenging image recognition tasks. This paper proposes the epitomic convolution as a new building block for deep neural networks. An epitomic convolution layer replaces a pair of consecutive convolution and max-pooling layers found in standard deep convolutional neural networks. The main version of the proposed model uses mini-epitomes in place of filters and computes responses invariant to small translations by epitomic search instead of max-pooling over image positions. The topographic version of the proposed model uses large epitomes to learn filter maps organized in translational topographies. We show that error back-propagation can successfully learn multiple epitomic layers in a supervised fashion. The effectiveness of the proposed method is assessed in image classification tasks on standard benchmarks. Our experiments on Imagenet indicate improved recognition performance compared to standard convolutional neural networks of similar architecture. Our models pre-trained on Imagenet perform excellently on Caltech-101. We also obtain competitive image classification results on the small-image MNIST and CIFAR-10 datasets.
no_new_dataset
0.711794
1406.1833
Kenneth Stanley
Paul A. Szerlip, Gregory Morse, Justin K. Pugh, and Kenneth O. Stanley
Unsupervised Feature Learning through Divergent Discriminative Feature Accumulation
Corrected citation formatting
null
null
null
cs.NE cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Unlike unsupervised approaches such as autoencoders that learn to reconstruct their inputs, this paper introduces an alternative approach to unsupervised feature learning called divergent discriminative feature accumulation (DDFA) that instead continually accumulates features that make novel discriminations among the training set. Thus DDFA features are inherently discriminative from the start even though they are trained without knowledge of the ultimate classification problem. Interestingly, DDFA also continues to add new features indefinitely (so it does not depend on a hidden layer size), is not based on minimizing error, and is inherently divergent instead of convergent, thereby providing a unique direction of research for unsupervised feature learning. In this paper the quality of its learned features is demonstrated on the MNIST dataset, where its performance confirms that indeed DDFA is a viable technique for learning useful features.
[ { "version": "v1", "created": "Fri, 6 Jun 2014 23:45:03 GMT" }, { "version": "v2", "created": "Tue, 10 Jun 2014 03:37:45 GMT" } ]
2014-06-11T00:00:00
[ [ "Szerlip", "Paul A.", "" ], [ "Morse", "Gregory", "" ], [ "Pugh", "Justin K.", "" ], [ "Stanley", "Kenneth O.", "" ] ]
TITLE: Unsupervised Feature Learning through Divergent Discriminative Feature Accumulation ABSTRACT: Unlike unsupervised approaches such as autoencoders that learn to reconstruct their inputs, this paper introduces an alternative approach to unsupervised feature learning called divergent discriminative feature accumulation (DDFA) that instead continually accumulates features that make novel discriminations among the training set. Thus DDFA features are inherently discriminative from the start even though they are trained without knowledge of the ultimate classification problem. Interestingly, DDFA also continues to add new features indefinitely (so it does not depend on a hidden layer size), is not based on minimizing error, and is inherently divergent instead of convergent, thereby providing a unique direction of research for unsupervised feature learning. In this paper the quality of its learned features is demonstrated on the MNIST dataset, where its performance confirms that indeed DDFA is a viable technique for learning useful features.
no_new_dataset
0.709265
1406.2392
Ryan Compton
Ryan Compton, Matthew S. Keegan, Jiejun Xu
Inferring the geographic focus of online documents from social media sharing patterns
6 pages, 10 figures, Computational Approaches to Social Modeling (ChASM) Workshop, WebSci 2014, Bloomington, Indiana-June 24-26 2014
null
null
null
cs.SI physics.soc-ph
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Determining the geographic focus of digital media is an essential first step for modern geographic information retrieval. However, publicly-visible location annotations are remarkably sparse in online data. In this work, we demonstrate a method which infers the geographic focus of an online document by examining the locations of Twitter users who share links to the document. We apply our geotagging technique to multiple datasets built from different content: manually-annotated news articles, GDELT, YouTube, Flickr, Twitter, and Tumblr.
[ { "version": "v1", "created": "Tue, 10 Jun 2014 00:34:55 GMT" } ]
2014-06-11T00:00:00
[ [ "Compton", "Ryan", "" ], [ "Keegan", "Matthew S.", "" ], [ "Xu", "Jiejun", "" ] ]
TITLE: Inferring the geographic focus of online documents from social media sharing patterns ABSTRACT: Determining the geographic focus of digital media is an essential first step for modern geographic information retrieval. However, publicly-visible location annotations are remarkably sparse in online data. In this work, we demonstrate a method which infers the geographic focus of an online document by examining the locations of Twitter users who share links to the document. We apply our geotagging technique to multiple datasets built from different content: manually-annotated news articles, GDELT, YouTube, Flickr, Twitter, and Tumblr.
no_new_dataset
0.710019
1406.2545
Fabricio de Franca Olivetti
Fabricio Olivetti de Franca and Guilherme Palermo Coelho
A Flexible Fitness Function for Community Detection in Complex Networks
null
null
null
null
cs.NE cs.SI physics.soc-ph
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Most community detection algorithms from the literature work as optimization tools that minimize a given \textit{fitness function}, while assuming that each node belongs to a single community. Since there is no hard concept of what a community is, most proposed fitness functions focus on a particular definition. As such, these functions do not always lead to partitions that correspond to those observed in practice. This paper proposes a new flexible fitness function that allows the identification of communities with distinct characteristics. Such flexibility was evaluated through the adoption of an immune-inspired optimization algorithm, named cob-aiNet[C], to identify both disjoint and overlapping communities in a set of benchmark networks. The results have shown that the obtained partitions are much closer to the ground-truth than those obtained by the optimization of the modularity function.
[ { "version": "v1", "created": "Tue, 10 Jun 2014 13:39:06 GMT" } ]
2014-06-11T00:00:00
[ [ "de Franca", "Fabricio Olivetti", "" ], [ "Coelho", "Guilherme Palermo", "" ] ]
TITLE: A Flexible Fitness Function for Community Detection in Complex Networks ABSTRACT: Most community detection algorithms from the literature work as optimization tools that minimize a given \textit{fitness function}, while assuming that each node belongs to a single community. Since there is no hard concept of what a community is, most proposed fitness functions focus on a particular definition. As such, these functions do not always lead to partitions that correspond to those observed in practice. This paper proposes a new flexible fitness function that allows the identification of communities with distinct characteristics. Such flexibility was evaluated through the adoption of an immune-inspired optimization algorithm, named cob-aiNet[C], to identify both disjoint and overlapping communities in a set of benchmark networks. The results have shown that the obtained partitions are much closer to the ground-truth than those obtained by the optimization of the modularity function.
no_new_dataset
0.711186
1406.2644
Ali Elouafiq
Ali Elouafiq and Redouan Abid
Geographical Asynchronous Information Access for Distributed Systems
7 pages
null
null
null
cs.DC cs.DB
http://creativecommons.org/licenses/by/3.0/
Non-relational databases are the common means of data storage in the Cloud, and optimizing the data access is of paramount importance into determining the overall Cloud system performance. In this paper, we present GAIA, a novel model for retrieving and managing correlated geo-localized data in the cloud environment. We survey and compare the existing models used mostly in Geographical Information Systems (GIS), mainly the Grid model and the Coordinates Projection model. Besides, we present a benchmark comparing the efficiency of the models. Using extensive experimentation, we show that GAIA outperforms the existing models by its high efficiency which is of O(log(n)), and this mainly thanks to its combination of projection with cell decomposition. The other models have a linear efficiency of O(n). The presented model is designed from the ground up to support GIS and is designed to suit both cloud and parallel computing.
[ { "version": "v1", "created": "Tue, 10 Jun 2014 17:33:56 GMT" } ]
2014-06-11T00:00:00
[ [ "Elouafiq", "Ali", "" ], [ "Abid", "Redouan", "" ] ]
TITLE: Geographical Asynchronous Information Access for Distributed Systems ABSTRACT: Non-relational databases are the common means of data storage in the Cloud, and optimizing the data access is of paramount importance into determining the overall Cloud system performance. In this paper, we present GAIA, a novel model for retrieving and managing correlated geo-localized data in the cloud environment. We survey and compare the existing models used mostly in Geographical Information Systems (GIS), mainly the Grid model and the Coordinates Projection model. Besides, we present a benchmark comparing the efficiency of the models. Using extensive experimentation, we show that GAIA outperforms the existing models by its high efficiency which is of O(log(n)), and this mainly thanks to its combination of projection with cell decomposition. The other models have a linear efficiency of O(n). The presented model is designed from the ground up to support GIS and is designed to suit both cloud and parallel computing.
no_new_dataset
0.711005
1201.4342
Tobias Buer
Tobias Buer and Herbert Kopfer
A Pareto-metaheuristic for a bi-objective winner determination problem in a combinatorial reverse auction
Accepted for publication in Computers & Operations Research, available online, Computers & Operations Research, 2013
Computers & Operations Research 41 (2014), 208-220
10.1016/j.cor.2013.04.004
null
cs.GT cs.AI math.OC
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The bi-objective winner determination problem (2WDP-SC) of a combinatorial procurement auction for transport contracts is characterized by a set B of bundle bids, with each bundle bid b in B consisting of a bidding carrier c_b, a bid price p_b, and a set tau_b transport contracts which is a subset of the set T of tendered transport contracts. Additionally, the transport quality q_{t,c_b} is given which is expected to be realized when a transport contract t is executed by a carrier c_b. The task of the auctioneer is to find a set X of winning bids (X subset B), such that each transport contract is part of at least one winning bid, the total procurement costs are minimized, and the total transport quality is maximized. This article presents a metaheuristic approach for the 2WDP-SC which integrates the greedy randomized adaptive search procedure with a two-stage candidate component selection procedure, large neighborhood search, and self-adaptive parameter setting in order to find a competitive set of non-dominated solutions. The heuristic outperforms all existing approaches. For seven small benchmark instances, the heuristic is the sole approach that finds all Pareto-optimal solutions. For 28 out of 30 large instances, none of the existing approaches is able to compute a solution that dominates a solution found by the proposed heuristic.
[ { "version": "v1", "created": "Fri, 20 Jan 2012 17:09:22 GMT" }, { "version": "v2", "created": "Mon, 22 Apr 2013 12:25:42 GMT" } ]
2014-06-10T00:00:00
[ [ "Buer", "Tobias", "" ], [ "Kopfer", "Herbert", "" ] ]
TITLE: A Pareto-metaheuristic for a bi-objective winner determination problem in a combinatorial reverse auction ABSTRACT: The bi-objective winner determination problem (2WDP-SC) of a combinatorial procurement auction for transport contracts is characterized by a set B of bundle bids, with each bundle bid b in B consisting of a bidding carrier c_b, a bid price p_b, and a set tau_b transport contracts which is a subset of the set T of tendered transport contracts. Additionally, the transport quality q_{t,c_b} is given which is expected to be realized when a transport contract t is executed by a carrier c_b. The task of the auctioneer is to find a set X of winning bids (X subset B), such that each transport contract is part of at least one winning bid, the total procurement costs are minimized, and the total transport quality is maximized. This article presents a metaheuristic approach for the 2WDP-SC which integrates the greedy randomized adaptive search procedure with a two-stage candidate component selection procedure, large neighborhood search, and self-adaptive parameter setting in order to find a competitive set of non-dominated solutions. The heuristic outperforms all existing approaches. For seven small benchmark instances, the heuristic is the sole approach that finds all Pareto-optimal solutions. For 28 out of 30 large instances, none of the existing approaches is able to compute a solution that dominates a solution found by the proposed heuristic.
no_new_dataset
0.7023
1204.0479
Tobias Buer
Tobias Buer and J\"org Homberger and Hermann Gehring
A collaborative ant colony metaheuristic for distributed multi-level lot-sizing
null
International Journal of Production Research 51 (17) 2013, 5253-5270
10.1080/00207543.2013.802822
null
cs.AI cs.DC
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The paper presents an ant colony optimization metaheuristic for collaborative planning. Collaborative planning is used to coordinate individual plans of self-interested decision makers with private information in order to increase the overall benefit of the coalition. The method consists of a new search graph based on encoded solutions. Distributed and private information is integrated via voting mechanisms and via a simple but effective collaborative local search procedure. The approach is applied to a distributed variant of the multi-level lot-sizing problem and evaluated by means of 352 benchmark instances from the literature. The proposed approach clearly outperforms existing approaches on the sets of medium and large sized instances. While the best method in the literature so far achieves an average deviation from the best known non-distributed solutions of 46 percent for the set of the largest instances, for example, the presented approach reduces the average deviation to only 5 percent.
[ { "version": "v1", "created": "Mon, 2 Apr 2012 17:44:45 GMT" } ]
2014-06-10T00:00:00
[ [ "Buer", "Tobias", "" ], [ "Homberger", "Jörg", "" ], [ "Gehring", "Hermann", "" ] ]
TITLE: A collaborative ant colony metaheuristic for distributed multi-level lot-sizing ABSTRACT: The paper presents an ant colony optimization metaheuristic for collaborative planning. Collaborative planning is used to coordinate individual plans of self-interested decision makers with private information in order to increase the overall benefit of the coalition. The method consists of a new search graph based on encoded solutions. Distributed and private information is integrated via voting mechanisms and via a simple but effective collaborative local search procedure. The approach is applied to a distributed variant of the multi-level lot-sizing problem and evaluated by means of 352 benchmark instances from the literature. The proposed approach clearly outperforms existing approaches on the sets of medium and large sized instances. While the best method in the literature so far achieves an average deviation from the best known non-distributed solutions of 46 percent for the set of the largest instances, for example, the presented approach reduces the average deviation to only 5 percent.
no_new_dataset
0.709868
1312.4564
Peilin Zhao
Peilin Zhao, Jinwei Yang, Tong Zhang, Ping Li
Adaptive Stochastic Alternating Direction Method of Multipliers
13 pages
null
null
null
stat.ML cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The Alternating Direction Method of Multipliers (ADMM) has been studied for years. The traditional ADMM algorithm needs to compute, at each iteration, an (empirical) expected loss function on all training examples, resulting in a computational complexity proportional to the number of training examples. To reduce the time complexity, stochastic ADMM algorithms were proposed to replace the expected function with a random loss function associated with one uniformly drawn example plus a Bregman divergence. The Bregman divergence, however, is derived from a simple second order proximal function, the half squared norm, which could be a suboptimal choice. In this paper, we present a new family of stochastic ADMM algorithms with optimal second order proximal functions, which produce a new family of adaptive subgradient methods. We theoretically prove that their regret bounds are as good as the bounds which could be achieved by the best proximal function that can be chosen in hindsight. Encouraging empirical results on a variety of real-world datasets confirm the effectiveness and efficiency of the proposed algorithms.
[ { "version": "v1", "created": "Mon, 16 Dec 2013 21:22:46 GMT" }, { "version": "v2", "created": "Sun, 22 Dec 2013 01:59:05 GMT" }, { "version": "v3", "created": "Thu, 5 Jun 2014 07:03:48 GMT" }, { "version": "v4", "created": "Mon, 9 Jun 2014 09:31:13 GMT" } ]
2014-06-10T00:00:00
[ [ "Zhao", "Peilin", "" ], [ "Yang", "Jinwei", "" ], [ "Zhang", "Tong", "" ], [ "Li", "Ping", "" ] ]
TITLE: Adaptive Stochastic Alternating Direction Method of Multipliers ABSTRACT: The Alternating Direction Method of Multipliers (ADMM) has been studied for years. The traditional ADMM algorithm needs to compute, at each iteration, an (empirical) expected loss function on all training examples, resulting in a computational complexity proportional to the number of training examples. To reduce the time complexity, stochastic ADMM algorithms were proposed to replace the expected function with a random loss function associated with one uniformly drawn example plus a Bregman divergence. The Bregman divergence, however, is derived from a simple second order proximal function, the half squared norm, which could be a suboptimal choice. In this paper, we present a new family of stochastic ADMM algorithms with optimal second order proximal functions, which produce a new family of adaptive subgradient methods. We theoretically prove that their regret bounds are as good as the bounds which could be achieved by the best proximal function that can be chosen in hindsight. Encouraging empirical results on a variety of real-world datasets confirm the effectiveness and efficiency of the proposed algorithms.
no_new_dataset
0.708036
1406.1928
Tobias Buer
Tobias Buer
An exact and two heuristic strategies for truthful bidding in combinatorial transport auctions
null
null
null
null
cs.GT cs.AI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
To support a freight carrier in a combinatorial transport auction, we proposes an exact and two heuristic strategies for bidding on subsets of requests. The exact bidding strategy is based on the concept of elementary request combinations. We show that it is sufficient and necessary for a carrier to bid on each elementary request combination in order to guarantee the same result as bidding on each element of the powerset of the set of tendered requests. Both heuristic bidding strategies identify promising request combinations. For this, pairwise synergies based on saving values as well as the capacitated p-median problem are used. The bidding strategies are evaluated by a computational study that simulates an auction. It is based on 174 benchmark instances and therefore easily extendable by other researchers. On average, the two heuristic strategies achieve 91 percent and 81 percent of the available sales potential while generating 36 and only 4 percent of the bundle bids of the exact strategy. Therefore, the proposed bidding strategies help a carrier to increase her chance to win and at the same time reduce the computational burden to participate in a combinatorial transport auction.
[ { "version": "v1", "created": "Sat, 7 Jun 2014 20:57:26 GMT" } ]
2014-06-10T00:00:00
[ [ "Buer", "Tobias", "" ] ]
TITLE: An exact and two heuristic strategies for truthful bidding in combinatorial transport auctions ABSTRACT: To support a freight carrier in a combinatorial transport auction, we proposes an exact and two heuristic strategies for bidding on subsets of requests. The exact bidding strategy is based on the concept of elementary request combinations. We show that it is sufficient and necessary for a carrier to bid on each elementary request combination in order to guarantee the same result as bidding on each element of the powerset of the set of tendered requests. Both heuristic bidding strategies identify promising request combinations. For this, pairwise synergies based on saving values as well as the capacitated p-median problem are used. The bidding strategies are evaluated by a computational study that simulates an auction. It is based on 174 benchmark instances and therefore easily extendable by other researchers. On average, the two heuristic strategies achieve 91 percent and 81 percent of the available sales potential while generating 36 and only 4 percent of the bundle bids of the exact strategy. Therefore, the proposed bidding strategies help a carrier to increase her chance to win and at the same time reduce the computational burden to participate in a combinatorial transport auction.
no_new_dataset
0.712595
1406.2031
Xianjie Chen
Xianjie Chen, Roozbeh Mottaghi, Xiaobai Liu, Sanja Fidler, Raquel Urtasun, Alan Yuille
Detect What You Can: Detecting and Representing Objects using Holistic Models and Body Parts
CBMM memo
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Detecting objects becomes difficult when we need to deal with large shape deformation, occlusion and low resolution. We propose a novel approach to i) handle large deformations and partial occlusions in animals (as examples of highly deformable objects), ii) describe them in terms of body parts, and iii) detect them when their body parts are hard to detect (e.g., animals depicted at low resolution). We represent the holistic object and body parts separately and use a fully connected model to arrange templates for the holistic object and body parts. Our model automatically decouples the holistic object or body parts from the model when they are hard to detect. This enables us to represent a large number of holistic object and body part combinations to better deal with different "detectability" patterns caused by deformations, occlusion and/or low resolution. We apply our method to the six animal categories in the PASCAL VOC dataset and show that our method significantly improves state-of-the-art (by 4.1% AP) and provides a richer representation for objects. During training we use annotations for body parts (e.g., head, torso, etc), making use of a new dataset of fully annotated object parts for PASCAL VOC 2010, which provides a mask for each part.
[ { "version": "v1", "created": "Sun, 8 Jun 2014 21:44:18 GMT" } ]
2014-06-10T00:00:00
[ [ "Chen", "Xianjie", "" ], [ "Mottaghi", "Roozbeh", "" ], [ "Liu", "Xiaobai", "" ], [ "Fidler", "Sanja", "" ], [ "Urtasun", "Raquel", "" ], [ "Yuille", "Alan", "" ] ]
TITLE: Detect What You Can: Detecting and Representing Objects using Holistic Models and Body Parts ABSTRACT: Detecting objects becomes difficult when we need to deal with large shape deformation, occlusion and low resolution. We propose a novel approach to i) handle large deformations and partial occlusions in animals (as examples of highly deformable objects), ii) describe them in terms of body parts, and iii) detect them when their body parts are hard to detect (e.g., animals depicted at low resolution). We represent the holistic object and body parts separately and use a fully connected model to arrange templates for the holistic object and body parts. Our model automatically decouples the holistic object or body parts from the model when they are hard to detect. This enables us to represent a large number of holistic object and body part combinations to better deal with different "detectability" patterns caused by deformations, occlusion and/or low resolution. We apply our method to the six animal categories in the PASCAL VOC dataset and show that our method significantly improves state-of-the-art (by 4.1% AP) and provides a richer representation for objects. During training we use annotations for body parts (e.g., head, torso, etc), making use of a new dataset of fully annotated object parts for PASCAL VOC 2010, which provides a mask for each part.
new_dataset
0.674024
1406.2049
Xue Li
Xue Li, Yu-Jin Zhang, Bin Shen, Bao-Di Liu
Image Tag Completion by Low-rank Factorization with Dual Reconstruction Structure Preserved
null
null
null
null
cs.CV cs.IR
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
A novel tag completion algorithm is proposed in this paper, which is designed with the following features: 1) Low-rank and error s-parsity: the incomplete initial tagging matrix D is decomposed into the complete tagging matrix A and a sparse error matrix E. However, instead of minimizing its nuclear norm, A is further factor-ized into a basis matrix U and a sparse coefficient matrix V, i.e. D=UV+E. This low-rank formulation encapsulating sparse coding enables our algorithm to recover latent structures from noisy initial data and avoid performing too much denoising; 2) Local reconstruction structure consistency: to steer the completion of D, the local linear reconstruction structures in feature space and tag space are obtained and preserved by U and V respectively. Such a scheme could alleviate the negative effect of distances measured by low-level features and incomplete tags. Thus, we can seek a balance between exploiting as much information and not being mislead to suboptimal performance. Experiments conducted on Corel5k dataset and the newly issued Flickr30Concepts dataset demonstrate the effectiveness and efficiency of the proposed method.
[ { "version": "v1", "created": "Mon, 9 Jun 2014 01:22:43 GMT" } ]
2014-06-10T00:00:00
[ [ "Li", "Xue", "" ], [ "Zhang", "Yu-Jin", "" ], [ "Shen", "Bin", "" ], [ "Liu", "Bao-Di", "" ] ]
TITLE: Image Tag Completion by Low-rank Factorization with Dual Reconstruction Structure Preserved ABSTRACT: A novel tag completion algorithm is proposed in this paper, which is designed with the following features: 1) Low-rank and error s-parsity: the incomplete initial tagging matrix D is decomposed into the complete tagging matrix A and a sparse error matrix E. However, instead of minimizing its nuclear norm, A is further factor-ized into a basis matrix U and a sparse coefficient matrix V, i.e. D=UV+E. This low-rank formulation encapsulating sparse coding enables our algorithm to recover latent structures from noisy initial data and avoid performing too much denoising; 2) Local reconstruction structure consistency: to steer the completion of D, the local linear reconstruction structures in feature space and tag space are obtained and preserved by U and V respectively. Such a scheme could alleviate the negative effect of distances measured by low-level features and incomplete tags. Thus, we can seek a balance between exploiting as much information and not being mislead to suboptimal performance. Experiments conducted on Corel5k dataset and the newly issued Flickr30Concepts dataset demonstrate the effectiveness and efficiency of the proposed method.
no_new_dataset
0.708175
1406.2099
Zahid Halim
Tufail Muhammad, Zahid Halim and Majid Ali Khan
ClassSpy: Java Object Pattern Visualization Tool
ICOMS-2013. International Conference on Modeling and Simulation, 25-27 November, Islamabad
null
null
null
cs.PL cs.SE
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Modern java programs consist of large number of classes as well as vast amount of objects instantiated during program execution. Software developers are always keen to know the number of objects created for each class. This information is helpful for a developer in understanding the packages/classes of a program and optimizing their code. However, understanding such a vast amount of information is not a trivial task. Visualization helps to depict this information on a single screen and to comprehend it efficiently. This paper presents a visualization approach that depicts information about all the objects instantiated during the program execution. The proposed technique is more space efficient and scalable to handle vast datasets, at the same time helpful to identify the key program components. This easy to use interface provides user an environment to glimpse the entire objects on a single screen. The proposed approach allows sorting objects at class, thread and method levels. Effectiveness and usability of the proposed approach is shown through case studies.
[ { "version": "v1", "created": "Mon, 9 Jun 2014 07:44:56 GMT" } ]
2014-06-10T00:00:00
[ [ "Muhammad", "Tufail", "" ], [ "Halim", "Zahid", "" ], [ "Khan", "Majid Ali", "" ] ]
TITLE: ClassSpy: Java Object Pattern Visualization Tool ABSTRACT: Modern java programs consist of large number of classes as well as vast amount of objects instantiated during program execution. Software developers are always keen to know the number of objects created for each class. This information is helpful for a developer in understanding the packages/classes of a program and optimizing their code. However, understanding such a vast amount of information is not a trivial task. Visualization helps to depict this information on a single screen and to comprehend it efficiently. This paper presents a visualization approach that depicts information about all the objects instantiated during the program execution. The proposed technique is more space efficient and scalable to handle vast datasets, at the same time helpful to identify the key program components. This easy to use interface provides user an environment to glimpse the entire objects on a single screen. The proposed approach allows sorting objects at class, thread and method levels. Effectiveness and usability of the proposed approach is shown through case studies.
no_new_dataset
0.707234
1406.2282
Chunyu Wang
Chunyu Wang, Yizhou Wang, Zhouchen Lin, Alan L. Yuille, Wen Gao
Robust Estimation of 3D Human Poses from a Single Image
null
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Human pose estimation is a key step to action recognition. We propose a method of estimating 3D human poses from a single image, which works in conjunction with an existing 2D pose/joint detector. 3D pose estimation is challenging because multiple 3D poses may correspond to the same 2D pose after projection due to the lack of depth information. Moreover, current 2D pose estimators are usually inaccurate which may cause errors in the 3D estimation. We address the challenges in three ways: (i) We represent a 3D pose as a linear combination of a sparse set of bases learned from 3D human skeletons. (ii) We enforce limb length constraints to eliminate anthropomorphically implausible skeletons. (iii) We estimate a 3D pose by minimizing the $L_1$-norm error between the projection of the 3D pose and the corresponding 2D detection. The $L_1$-norm loss term is robust to inaccurate 2D joint estimations. We use the alternating direction method (ADM) to solve the optimization problem efficiently. Our approach outperforms the state-of-the-arts on three benchmark datasets.
[ { "version": "v1", "created": "Mon, 9 Jun 2014 18:55:31 GMT" } ]
2014-06-10T00:00:00
[ [ "Wang", "Chunyu", "" ], [ "Wang", "Yizhou", "" ], [ "Lin", "Zhouchen", "" ], [ "Yuille", "Alan L.", "" ], [ "Gao", "Wen", "" ] ]
TITLE: Robust Estimation of 3D Human Poses from a Single Image ABSTRACT: Human pose estimation is a key step to action recognition. We propose a method of estimating 3D human poses from a single image, which works in conjunction with an existing 2D pose/joint detector. 3D pose estimation is challenging because multiple 3D poses may correspond to the same 2D pose after projection due to the lack of depth information. Moreover, current 2D pose estimators are usually inaccurate which may cause errors in the 3D estimation. We address the challenges in three ways: (i) We represent a 3D pose as a linear combination of a sparse set of bases learned from 3D human skeletons. (ii) We enforce limb length constraints to eliminate anthropomorphically implausible skeletons. (iii) We estimate a 3D pose by minimizing the $L_1$-norm error between the projection of the 3D pose and the corresponding 2D detection. The $L_1$-norm loss term is robust to inaccurate 2D joint estimations. We use the alternating direction method (ADM) to solve the optimization problem efficiently. Our approach outperforms the state-of-the-arts on three benchmark datasets.
no_new_dataset
0.708572
1406.2283
David Eigen
David Eigen and Christian Puhrsch and Rob Fergus
Depth Map Prediction from a Single Image using a Multi-Scale Deep Network
null
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Predicting depth is an essential component in understanding the 3D geometry of a scene. While for stereo images local correspondence suffices for estimation, finding depth relations from a single image is less straightforward, requiring integration of both global and local information from various cues. Moreover, the task is inherently ambiguous, with a large source of uncertainty coming from the overall scale. In this paper, we present a new method that addresses this task by employing two deep network stacks: one that makes a coarse global prediction based on the entire image, and another that refines this prediction locally. We also apply a scale-invariant error to help measure depth relations rather than scale. By leveraging the raw datasets as large sources of training data, our method achieves state-of-the-art results on both NYU Depth and KITTI, and matches detailed depth boundaries without the need for superpixelation.
[ { "version": "v1", "created": "Mon, 9 Jun 2014 19:01:18 GMT" } ]
2014-06-10T00:00:00
[ [ "Eigen", "David", "" ], [ "Puhrsch", "Christian", "" ], [ "Fergus", "Rob", "" ] ]
TITLE: Depth Map Prediction from a Single Image using a Multi-Scale Deep Network ABSTRACT: Predicting depth is an essential component in understanding the 3D geometry of a scene. While for stereo images local correspondence suffices for estimation, finding depth relations from a single image is less straightforward, requiring integration of both global and local information from various cues. Moreover, the task is inherently ambiguous, with a large source of uncertainty coming from the overall scale. In this paper, we present a new method that addresses this task by employing two deep network stacks: one that makes a coarse global prediction based on the entire image, and another that refines this prediction locally. We also apply a scale-invariant error to help measure depth relations rather than scale. By leveraging the raw datasets as large sources of training data, our method achieves state-of-the-art results on both NYU Depth and KITTI, and matches detailed depth boundaries without the need for superpixelation.
no_new_dataset
0.71222
1401.8257
Shuai Li
Claudio Gentile, Shuai Li, Giovanni Zappella
Online Clustering of Bandits
In E. Xing and T. Jebara (Eds.), Proceedings of 31st International Conference on Machine Learning, Journal of Machine Learning Research Workshop and Conference Proceedings, Vol.32 (JMLR W&CP-32), Beijing, China, Jun. 21-26, 2014 (ICML 2014), Submitted by Shuai Li (https://sites.google.com/site/shuailidotsli)
null
null
null
cs.LG stat.ML
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We introduce a novel algorithmic approach to content recommendation based on adaptive clustering of exploration-exploitation ("bandit") strategies. We provide a sharp regret analysis of this algorithm in a standard stochastic noise setting, demonstrate its scalability properties, and prove its effectiveness on a number of artificial and real-world datasets. Our experiments show a significant increase in prediction performance over state-of-the-art methods for bandit problems.
[ { "version": "v1", "created": "Fri, 31 Jan 2014 18:49:42 GMT" }, { "version": "v2", "created": "Tue, 13 May 2014 07:13:06 GMT" }, { "version": "v3", "created": "Fri, 6 Jun 2014 13:59:04 GMT" } ]
2014-06-09T00:00:00
[ [ "Gentile", "Claudio", "" ], [ "Li", "Shuai", "" ], [ "Zappella", "Giovanni", "" ] ]
TITLE: Online Clustering of Bandits ABSTRACT: We introduce a novel algorithmic approach to content recommendation based on adaptive clustering of exploration-exploitation ("bandit") strategies. We provide a sharp regret analysis of this algorithm in a standard stochastic noise setting, demonstrate its scalability properties, and prove its effectiveness on a number of artificial and real-world datasets. Our experiments show a significant increase in prediction performance over state-of-the-art methods for bandit problems.
no_new_dataset
0.707556
1406.0588
Shasha Bu
Shasha Bu and Yu-Jin Zhang
Image retrieval with hierarchical matching pursuit
5 pages, 6 figures, conference
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
A novel representation of images for image retrieval is introduced in this paper, by using a new type of feature with remarkable discriminative power. Despite the multi-scale nature of objects, most existing models perform feature extraction on a fixed scale, which will inevitably degrade the performance of the whole system. Motivated by this, we introduce a hierarchical sparse coding architecture for image retrieval to explore multi-scale cues. Sparse codes extracted on lower layers are transmitted to higher layers recursively. With this mechanism, cues from different scales are fused. Experiments on the Holidays dataset show that the proposed method achieves an excellent retrieval performance with a small code length.
[ { "version": "v1", "created": "Tue, 3 Jun 2014 06:32:24 GMT" }, { "version": "v2", "created": "Thu, 5 Jun 2014 02:23:21 GMT" } ]
2014-06-06T00:00:00
[ [ "Bu", "Shasha", "" ], [ "Zhang", "Yu-Jin", "" ] ]
TITLE: Image retrieval with hierarchical matching pursuit ABSTRACT: A novel representation of images for image retrieval is introduced in this paper, by using a new type of feature with remarkable discriminative power. Despite the multi-scale nature of objects, most existing models perform feature extraction on a fixed scale, which will inevitably degrade the performance of the whole system. Motivated by this, we introduce a hierarchical sparse coding architecture for image retrieval to explore multi-scale cues. Sparse codes extracted on lower layers are transmitted to higher layers recursively. With this mechanism, cues from different scales are fused. Experiments on the Holidays dataset show that the proposed method achieves an excellent retrieval performance with a small code length.
no_new_dataset
0.711443
1406.1167
Xu-Cheng Yin
Xu-Cheng Yin and Chun Yang and Hong-Wei Hao
Learning to Diversify via Weighted Kernels for Classifier Ensemble
Submitted to IEEE Trans. Pattern Analysis and Machine Intelligence (TPAMI)
null
null
null
cs.LG cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Classifier ensemble generally should combine diverse component classifiers. However, it is difficult to give a definitive connection between diversity measure and ensemble accuracy. Given a list of available component classifiers, how to adaptively and diversely ensemble classifiers becomes a big challenge in the literature. In this paper, we argue that diversity, not direct diversity on samples but adaptive diversity with data, is highly correlated to ensemble accuracy, and we propose a novel technology for classifier ensemble, learning to diversify, which learns to adaptively combine classifiers by considering both accuracy and diversity. Specifically, our approach, Learning TO Diversify via Weighted Kernels (L2DWK), performs classifier combination by optimizing a direct but simple criterion: maximizing ensemble accuracy and adaptive diversity simultaneously by minimizing a convex loss function. Given a measure formulation, the diversity is calculated with weighted kernels (i.e., the diversity is measured on the component classifiers' outputs which are kernelled and weighted), and the kernel weights are automatically learned. We minimize this loss function by estimating the kernel weights in conjunction with the classifier weights, and propose a self-training algorithm for conducting this convex optimization procedure iteratively. Extensive experiments on a variety of 32 UCI classification benchmark datasets show that the proposed approach consistently outperforms state-of-the-art ensembles such as Bagging, AdaBoost, Random Forests, Gasen, Regularized Selective Ensemble, and Ensemble Pruning via Semi-Definite Programming.
[ { "version": "v1", "created": "Wed, 4 Jun 2014 09:16:42 GMT" } ]
2014-06-06T00:00:00
[ [ "Yin", "Xu-Cheng", "" ], [ "Yang", "Chun", "" ], [ "Hao", "Hong-Wei", "" ] ]
TITLE: Learning to Diversify via Weighted Kernels for Classifier Ensemble ABSTRACT: Classifier ensemble generally should combine diverse component classifiers. However, it is difficult to give a definitive connection between diversity measure and ensemble accuracy. Given a list of available component classifiers, how to adaptively and diversely ensemble classifiers becomes a big challenge in the literature. In this paper, we argue that diversity, not direct diversity on samples but adaptive diversity with data, is highly correlated to ensemble accuracy, and we propose a novel technology for classifier ensemble, learning to diversify, which learns to adaptively combine classifiers by considering both accuracy and diversity. Specifically, our approach, Learning TO Diversify via Weighted Kernels (L2DWK), performs classifier combination by optimizing a direct but simple criterion: maximizing ensemble accuracy and adaptive diversity simultaneously by minimizing a convex loss function. Given a measure formulation, the diversity is calculated with weighted kernels (i.e., the diversity is measured on the component classifiers' outputs which are kernelled and weighted), and the kernel weights are automatically learned. We minimize this loss function by estimating the kernel weights in conjunction with the classifier weights, and propose a self-training algorithm for conducting this convex optimization procedure iteratively. Extensive experiments on a variety of 32 UCI classification benchmark datasets show that the proposed approach consistently outperforms state-of-the-art ensembles such as Bagging, AdaBoost, Random Forests, Gasen, Regularized Selective Ensemble, and Ensemble Pruning via Semi-Definite Programming.
no_new_dataset
0.710842
1406.1256
Ian Manchester
Ian R. Manchester and Jean-Jacques E. Slotine
Output-Feedback Control of Nonlinear Systems using Control Contraction Metrics and Convex Optimization
Conference submission
null
null
null
math.OC cs.SY
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Control contraction metrics (CCMs) are a new approach to nonlinear control design based on contraction theory. The resulting design problems are expressed as pointwise linear matrix inequalities and are and well-suited to solution via convex optimization. In this paper, we extend the theory on CCMs by showing that a pair of "dual" observer and controller problems can be solved using pointwise linear matrix inequalities, and that when a solution exists a separation principle holds. That is, a stabilizing output-feedback controller can be found. The procedure is demonstrated using a benchmark problem of nonlinear control: the Moore-Greitzer jet engine compressor model.
[ { "version": "v1", "created": "Thu, 5 Jun 2014 02:27:03 GMT" } ]
2014-06-06T00:00:00
[ [ "Manchester", "Ian R.", "" ], [ "Slotine", "Jean-Jacques E.", "" ] ]
TITLE: Output-Feedback Control of Nonlinear Systems using Control Contraction Metrics and Convex Optimization ABSTRACT: Control contraction metrics (CCMs) are a new approach to nonlinear control design based on contraction theory. The resulting design problems are expressed as pointwise linear matrix inequalities and are and well-suited to solution via convex optimization. In this paper, we extend the theory on CCMs by showing that a pair of "dual" observer and controller problems can be solved using pointwise linear matrix inequalities, and that when a solution exists a separation principle holds. That is, a stabilizing output-feedback controller can be found. The procedure is demonstrated using a benchmark problem of nonlinear control: the Moore-Greitzer jet engine compressor model.
no_new_dataset
0.708591
1207.6430
Christoph Brune
Braxton Osting and Christoph Brune and Stanley J. Osher
Optimal Data Collection For Informative Rankings Expose Well-Connected Graphs
31 pages, 10 figures, 3 tables
null
null
UCLA CAM report 12-32
stat.ML cs.LG stat.AP
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Given a graph where vertices represent alternatives and arcs represent pairwise comparison data, the statistical ranking problem is to find a potential function, defined on the vertices, such that the gradient of the potential function agrees with the pairwise comparisons. Our goal in this paper is to develop a method for collecting data for which the least squares estimator for the ranking problem has maximal Fisher information. Our approach, based on experimental design, is to view data collection as a bi-level optimization problem where the inner problem is the ranking problem and the outer problem is to identify data which maximizes the informativeness of the ranking. Under certain assumptions, the data collection problem decouples, reducing to a problem of finding multigraphs with large algebraic connectivity. This reduction of the data collection problem to graph-theoretic questions is one of the primary contributions of this work. As an application, we study the Yahoo! Movie user rating dataset and demonstrate that the addition of a small number of well-chosen pairwise comparisons can significantly increase the Fisher informativeness of the ranking. As another application, we study the 2011-12 NCAA football schedule and propose schedules with the same number of games which are significantly more informative. Using spectral clustering methods to identify highly-connected communities within the division, we argue that the NCAA could improve its notoriously poor rankings by simply scheduling more out-of-conference games.
[ { "version": "v1", "created": "Thu, 26 Jul 2012 23:14:34 GMT" }, { "version": "v2", "created": "Wed, 4 Jun 2014 08:31:57 GMT" } ]
2014-06-05T00:00:00
[ [ "Osting", "Braxton", "" ], [ "Brune", "Christoph", "" ], [ "Osher", "Stanley J.", "" ] ]
TITLE: Optimal Data Collection For Informative Rankings Expose Well-Connected Graphs ABSTRACT: Given a graph where vertices represent alternatives and arcs represent pairwise comparison data, the statistical ranking problem is to find a potential function, defined on the vertices, such that the gradient of the potential function agrees with the pairwise comparisons. Our goal in this paper is to develop a method for collecting data for which the least squares estimator for the ranking problem has maximal Fisher information. Our approach, based on experimental design, is to view data collection as a bi-level optimization problem where the inner problem is the ranking problem and the outer problem is to identify data which maximizes the informativeness of the ranking. Under certain assumptions, the data collection problem decouples, reducing to a problem of finding multigraphs with large algebraic connectivity. This reduction of the data collection problem to graph-theoretic questions is one of the primary contributions of this work. As an application, we study the Yahoo! Movie user rating dataset and demonstrate that the addition of a small number of well-chosen pairwise comparisons can significantly increase the Fisher informativeness of the ranking. As another application, we study the 2011-12 NCAA football schedule and propose schedules with the same number of games which are significantly more informative. Using spectral clustering methods to identify highly-connected communities within the division, we argue that the NCAA could improve its notoriously poor rankings by simply scheduling more out-of-conference games.
no_new_dataset
0.709069
1406.1061
Vit Novacek
Vit Novacek
A Methodology for Empirical Analysis of LOD Datasets
A current working draft of the paper submitted to the ISWC'14 conference (track information available here: http://iswc2014.semanticweb.org/call-replication-benchmark-data-software-papers)
null
null
null
cs.AI cs.SI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
CoCoE stands for Complexity, Coherence and Entropy, and presents an extensible methodology for empirical analysis of Linked Open Data (i.e., RDF graphs). CoCoE can offer answers to questions like: Is dataset A better than B for knowledge discovery since it is more complex and informative?, Is dataset X better than Y for simple value lookups due its flatter structure?, etc. In order to address such questions, we introduce a set of well-founded measures based on complementary notions from distributional semantics, network analysis and information theory. These measures are part of a specific implementation of the CoCoE methodology that is available for download. Last but not least, we illustrate CoCoE by its application to selected biomedical RDF datasets.
[ { "version": "v1", "created": "Wed, 4 Jun 2014 14:45:43 GMT" } ]
2014-06-05T00:00:00
[ [ "Novacek", "Vit", "" ] ]
TITLE: A Methodology for Empirical Analysis of LOD Datasets ABSTRACT: CoCoE stands for Complexity, Coherence and Entropy, and presents an extensible methodology for empirical analysis of Linked Open Data (i.e., RDF graphs). CoCoE can offer answers to questions like: Is dataset A better than B for knowledge discovery since it is more complex and informative?, Is dataset X better than Y for simple value lookups due its flatter structure?, etc. In order to address such questions, we introduce a set of well-founded measures based on complementary notions from distributional semantics, network analysis and information theory. These measures are part of a specific implementation of the CoCoE methodology that is available for download. Last but not least, we illustrate CoCoE by its application to selected biomedical RDF datasets.
no_new_dataset
0.708862
1406.1137
Gang Wang
Gang Wang, Tianyi Wang, Bolun Wang, Divya Sambasivan, Zengbin Zhang, Haitao Zheng, Ben Y. Zhao
Crowds on Wall Street: Extracting Value from Social Investing Platforms
null
null
null
null
cs.SI physics.soc-ph
http://creativecommons.org/licenses/by/3.0/
For decades, the world of financial advisors has been dominated by large investment banks such as Goldman Sachs. In recent years, user-contributed investment services such as SeekingAlpha and StockTwits have grown to millions of users. In this paper, we seek to understand the quality and impact of content on social investment platforms, by empirically analyzing complete datasets of SeekingAlpha articles (9 years) and StockTwits messages (4 years). We develop sentiment analysis tools and correlate contributed content to the historical performance of relevant stocks. While SeekingAlpha articles and StockTwits messages provide minimal correlation to stock performance in aggregate, a subset of authors contribute more valuable (predictive) content. We show that these authors can be identified via both empirical methods or by user interactions, and investments using their analysis significantly outperform broader markets. Finally, we conduct a user survey that sheds light on users views of SeekingAlpha content and stock manipulation.
[ { "version": "v1", "created": "Wed, 4 Jun 2014 18:34:32 GMT" } ]
2014-06-05T00:00:00
[ [ "Wang", "Gang", "" ], [ "Wang", "Tianyi", "" ], [ "Wang", "Bolun", "" ], [ "Sambasivan", "Divya", "" ], [ "Zhang", "Zengbin", "" ], [ "Zheng", "Haitao", "" ], [ "Zhao", "Ben Y.", "" ] ]
TITLE: Crowds on Wall Street: Extracting Value from Social Investing Platforms ABSTRACT: For decades, the world of financial advisors has been dominated by large investment banks such as Goldman Sachs. In recent years, user-contributed investment services such as SeekingAlpha and StockTwits have grown to millions of users. In this paper, we seek to understand the quality and impact of content on social investment platforms, by empirically analyzing complete datasets of SeekingAlpha articles (9 years) and StockTwits messages (4 years). We develop sentiment analysis tools and correlate contributed content to the historical performance of relevant stocks. While SeekingAlpha articles and StockTwits messages provide minimal correlation to stock performance in aggregate, a subset of authors contribute more valuable (predictive) content. We show that these authors can be identified via both empirical methods or by user interactions, and investments using their analysis significantly outperform broader markets. Finally, we conduct a user survey that sheds light on users views of SeekingAlpha content and stock manipulation.
no_new_dataset
0.71081
1401.5842
Florian Zuleger
Moritz Sinn, Florian Zuleger, Helmut Veith
A Simple and Scalable Static Analysis for Bound Analysis and Amortized Complexity Analysis
null
null
null
null
cs.PL cs.DS
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We present the first scalable bound analysis that achieves amortized complexity analysis. In contrast to earlier work, our bound analysis is not based on general purpose reasoners such as abstract interpreters, software model checkers or computer algebra tools. Rather, we derive bounds directly from abstract program models, which we obtain from programs by comparatively simple invariant generation and symbolic execution techniques. As a result, we obtain an analysis that is more predictable and more scalable than earlier approaches. Our experiments demonstrate that our analysis is fast and at the same time able to compute bounds for challenging loops in a large real-world benchmark. Technically, our approach is based on lossy vector addition systems (VASS). Our bound analysis first computes a lexicographic ranking function that proves the termination of a VASS, and then derives a bound from this ranking function. Our methodology achieves amortized analysis based on a new insight how lexicographic ranking functions can be used for bound analysis.
[ { "version": "v1", "created": "Thu, 23 Jan 2014 01:40:43 GMT" }, { "version": "v2", "created": "Mon, 2 Jun 2014 20:44:29 GMT" } ]
2014-06-04T00:00:00
[ [ "Sinn", "Moritz", "" ], [ "Zuleger", "Florian", "" ], [ "Veith", "Helmut", "" ] ]
TITLE: A Simple and Scalable Static Analysis for Bound Analysis and Amortized Complexity Analysis ABSTRACT: We present the first scalable bound analysis that achieves amortized complexity analysis. In contrast to earlier work, our bound analysis is not based on general purpose reasoners such as abstract interpreters, software model checkers or computer algebra tools. Rather, we derive bounds directly from abstract program models, which we obtain from programs by comparatively simple invariant generation and symbolic execution techniques. As a result, we obtain an analysis that is more predictable and more scalable than earlier approaches. Our experiments demonstrate that our analysis is fast and at the same time able to compute bounds for challenging loops in a large real-world benchmark. Technically, our approach is based on lossy vector addition systems (VASS). Our bound analysis first computes a lexicographic ranking function that proves the termination of a VASS, and then derives a bound from this ranking function. Our methodology achieves amortized analysis based on a new insight how lexicographic ranking functions can be used for bound analysis.
no_new_dataset
0.710057
1405.4138
Reza Azizi
Reza Azizi
Empirical Study of Artificial Fish Swarm Algorithm
null
International Journal of Computing, Communications and Networking (IJCCN) , Volume 3, No.1, Pages 01-07, March 2014
null
null
cs.AI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Artificial fish swarm algorithm (AFSA) is one of the swarm intelligence optimization algorithms that works based on population and stochastic search. In order to achieve acceptable result, there are many parameters needs to be adjusted in AFSA. Among these parameters, visual and step are very significant in view of the fact that artificial fish basically move based on these parameters. In standard AFSA, these two parameters remain constant until the algorithm termination. Large values of these parameters increase the capability of algorithm in global search, while small values improve the local search ability of the algorithm. In this paper, we empirically study the performance of the AFSA and different approaches to balance between local and global exploration have been tested based on the adaptive modification of visual and step during algorithm execution. The proposed approaches have been evaluated based on the four well-known benchmark functions. Experimental results show considerable positive impact on the performance of AFSA.
[ { "version": "v1", "created": "Fri, 16 May 2014 12:02:42 GMT" } ]
2014-06-04T00:00:00
[ [ "Azizi", "Reza", "" ] ]
TITLE: Empirical Study of Artificial Fish Swarm Algorithm ABSTRACT: Artificial fish swarm algorithm (AFSA) is one of the swarm intelligence optimization algorithms that works based on population and stochastic search. In order to achieve acceptable result, there are many parameters needs to be adjusted in AFSA. Among these parameters, visual and step are very significant in view of the fact that artificial fish basically move based on these parameters. In standard AFSA, these two parameters remain constant until the algorithm termination. Large values of these parameters increase the capability of algorithm in global search, while small values improve the local search ability of the algorithm. In this paper, we empirically study the performance of the AFSA and different approaches to balance between local and global exploration have been tested based on the adaptive modification of visual and step during algorithm execution. The proposed approaches have been evaluated based on the four well-known benchmark functions. Experimental results show considerable positive impact on the performance of AFSA.
no_new_dataset
0.709403
1406.0582
Gcg Equipe
Lukasz Domagala (INRIA Grenoble Rh\^one-Alpes), Fabrice Rastello (INRIA Grenoble Rh\^one-Alpes), Sadayappan Ponnuswany, Duco Van Amstel (INRIA Grenoble Rh\^one-Alpes)
A Tiling Perspective for Register Optimization
null
N° RR-8541 (2014)
null
RR-8541
cs.PL
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Register allocation is a much studied problem. A particularly important context for optimizing register allocation is within loops, since a significant fraction of the execution time of programs is often inside loop code. A variety of algorithms have been proposed in the past for register allocation, but the complexity of the problem has resulted in a decoupling of several important aspects, including loop unrolling, register promotion, and instruction reordering. In this paper, we develop an approach to register allocation and promotion in a unified optimization framework that simultaneously considers the impact of loop unrolling and instruction scheduling. This is done via a novel instruction tiling approach where instructions within a loop are represented along one dimension and innermost loop iterations along the other dimension. By exploiting the regularity along the loop dimension, and imposing essential dependence based constraints on intra-tile execution order, the problem of optimizing register pressure is cast in a constraint programming formalism. Experimental results are provided from thousands of innermost loops extracted from the SPEC benchmarks, demonstrating improvements over the current state-of-the-art.
[ { "version": "v1", "created": "Tue, 3 Jun 2014 06:04:52 GMT" } ]
2014-06-04T00:00:00
[ [ "Domagala", "Lukasz", "", "INRIA Grenoble Rhône-Alpes" ], [ "Rastello", "Fabrice", "", "INRIA Grenoble Rhône-Alpes" ], [ "Ponnuswany", "Sadayappan", "", "INRIA\n Grenoble Rhône-Alpes" ], [ "Van Amstel", "Duco", "", "INRIA\n Grenoble Rhône-Alpes" ] ]
TITLE: A Tiling Perspective for Register Optimization ABSTRACT: Register allocation is a much studied problem. A particularly important context for optimizing register allocation is within loops, since a significant fraction of the execution time of programs is often inside loop code. A variety of algorithms have been proposed in the past for register allocation, but the complexity of the problem has resulted in a decoupling of several important aspects, including loop unrolling, register promotion, and instruction reordering. In this paper, we develop an approach to register allocation and promotion in a unified optimization framework that simultaneously considers the impact of loop unrolling and instruction scheduling. This is done via a novel instruction tiling approach where instructions within a loop are represented along one dimension and innermost loop iterations along the other dimension. By exploiting the regularity along the loop dimension, and imposing essential dependence based constraints on intra-tile execution order, the problem of optimizing register pressure is cast in a constraint programming formalism. Experimental results are provided from thousands of innermost loops extracted from the SPEC benchmarks, demonstrating improvements over the current state-of-the-art.
no_new_dataset
0.708767
1406.0599
Zhenghong Chen
Zhenghong Chen, Xiaoxian Tang, Bican Xia
Hierarchical Comprehensive Triangular Decomposition
null
null
null
null
cs.SC
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The concept of comprehensive triangular decomposition (CTD) was first introduced by Chen et al. in their CASC'2007 paper and could be viewed as an analogue of comprehensive Grobner systems for parametric polynomial systems. The first complete algorithm for computing CTD was also proposed in that paper and implemented in the RegularChains library in Maple. Following our previous work on generic regular decomposition for parametric polynomial systems, we introduce in this paper a so-called hierarchical strategy for computing CTDs. Roughly speaking, for a given parametric system, the parametric space is divided into several sub-spaces of different dimensions and we compute CTDs over those sub-spaces one by one. So, it is possible that, for some benchmarks, it is difficult to compute CTDs in reasonable time while this strategy can obtain some "partial" solutions over some parametric sub-spaces. The program based on this strategy has been tested on a number of benchmarks from the literature. Experimental results on these benchmarks with comparison to RegularChains are reported and may be valuable for developing more efficient triangularization tools.
[ { "version": "v1", "created": "Tue, 3 Jun 2014 07:13:17 GMT" } ]
2014-06-04T00:00:00
[ [ "Chen", "Zhenghong", "" ], [ "Tang", "Xiaoxian", "" ], [ "Xia", "Bican", "" ] ]
TITLE: Hierarchical Comprehensive Triangular Decomposition ABSTRACT: The concept of comprehensive triangular decomposition (CTD) was first introduced by Chen et al. in their CASC'2007 paper and could be viewed as an analogue of comprehensive Grobner systems for parametric polynomial systems. The first complete algorithm for computing CTD was also proposed in that paper and implemented in the RegularChains library in Maple. Following our previous work on generic regular decomposition for parametric polynomial systems, we introduce in this paper a so-called hierarchical strategy for computing CTDs. Roughly speaking, for a given parametric system, the parametric space is divided into several sub-spaces of different dimensions and we compute CTDs over those sub-spaces one by one. So, it is possible that, for some benchmarks, it is difficult to compute CTDs in reasonable time while this strategy can obtain some "partial" solutions over some parametric sub-spaces. The program based on this strategy has been tested on a number of benchmarks from the literature. Experimental results on these benchmarks with comparison to RegularChains are reported and may be valuable for developing more efficient triangularization tools.
no_new_dataset
0.709629
1406.0680
Ziqiong Liu
Ziqiong Liu, Shengjin Wang, Liang Zheng, Qi Tian
Visual Reranking with Improved Image Graph
null
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
This paper introduces an improved reranking method for the Bag-of-Words (BoW) based image search. Built on [1], a directed image graph robust to outlier distraction is proposed. In our approach, the relevance among images is encoded in the image graph, based on which the initial rank list is refined. Moreover, we show that the rank-level feature fusion can be adopted in this reranking method as well. Taking advantage of the complementary nature of various features, the reranking performance is further enhanced. Particularly, we exploit the reranking method combining the BoW and color information. Experiments on two benchmark datasets demonstrate that ourmethod yields significant improvements and the reranking results are competitive to the state-of-the-art methods.
[ { "version": "v1", "created": "Tue, 3 Jun 2014 12:07:12 GMT" } ]
2014-06-04T00:00:00
[ [ "Liu", "Ziqiong", "" ], [ "Wang", "Shengjin", "" ], [ "Zheng", "Liang", "" ], [ "Tian", "Qi", "" ] ]
TITLE: Visual Reranking with Improved Image Graph ABSTRACT: This paper introduces an improved reranking method for the Bag-of-Words (BoW) based image search. Built on [1], a directed image graph robust to outlier distraction is proposed. In our approach, the relevance among images is encoded in the image graph, based on which the initial rank list is refined. Moreover, we show that the rank-level feature fusion can be adopted in this reranking method as well. Taking advantage of the complementary nature of various features, the reranking performance is further enhanced. Particularly, we exploit the reranking method combining the BoW and color information. Experiments on two benchmark datasets demonstrate that ourmethod yields significant improvements and the reranking results are competitive to the state-of-the-art methods.
no_new_dataset
0.710622
1406.0132
Liang Zheng
Liang Zheng, Shengjin Wang, Fei He, Qi Tian
Seeing the Big Picture: Deep Embedding with Contextual Evidences
10 pages, 13 figures, 7 tables, submitted to ACM Multimedia 2014
null
null
null
cs.CV
http://creativecommons.org/licenses/by/3.0/
In the Bag-of-Words (BoW) model based image retrieval task, the precision of visual matching plays a critical role in improving retrieval performance. Conventionally, local cues of a keypoint are employed. However, such strategy does not consider the contextual evidences of a keypoint, a problem which would lead to the prevalence of false matches. To address this problem, this paper defines "true match" as a pair of keypoints which are similar on three levels, i.e., local, regional, and global. Then, a principled probabilistic framework is established, which is capable of implicitly integrating discriminative cues from all these feature levels. Specifically, the Convolutional Neural Network (CNN) is employed to extract features from regional and global patches, leading to the so-called "Deep Embedding" framework. CNN has been shown to produce excellent performance on a dozen computer vision tasks such as image classification and detection, but few works have been done on BoW based image retrieval. In this paper, firstly we show that proper pre-processing techniques are necessary for effective usage of CNN feature. Then, in the attempt to fit it into our model, a novel indexing structure called "Deep Indexing" is introduced, which dramatically reduces memory usage. Extensive experiments on three benchmark datasets demonstrate that, the proposed Deep Embedding method greatly promotes the retrieval accuracy when CNN feature is integrated. We show that our method is efficient in terms of both memory and time cost, and compares favorably with the state-of-the-art methods.
[ { "version": "v1", "created": "Sun, 1 Jun 2014 05:04:28 GMT" } ]
2014-06-03T00:00:00
[ [ "Zheng", "Liang", "" ], [ "Wang", "Shengjin", "" ], [ "He", "Fei", "" ], [ "Tian", "Qi", "" ] ]
TITLE: Seeing the Big Picture: Deep Embedding with Contextual Evidences ABSTRACT: In the Bag-of-Words (BoW) model based image retrieval task, the precision of visual matching plays a critical role in improving retrieval performance. Conventionally, local cues of a keypoint are employed. However, such strategy does not consider the contextual evidences of a keypoint, a problem which would lead to the prevalence of false matches. To address this problem, this paper defines "true match" as a pair of keypoints which are similar on three levels, i.e., local, regional, and global. Then, a principled probabilistic framework is established, which is capable of implicitly integrating discriminative cues from all these feature levels. Specifically, the Convolutional Neural Network (CNN) is employed to extract features from regional and global patches, leading to the so-called "Deep Embedding" framework. CNN has been shown to produce excellent performance on a dozen computer vision tasks such as image classification and detection, but few works have been done on BoW based image retrieval. In this paper, firstly we show that proper pre-processing techniques are necessary for effective usage of CNN feature. Then, in the attempt to fit it into our model, a novel indexing structure called "Deep Indexing" is introduced, which dramatically reduces memory usage. Extensive experiments on three benchmark datasets demonstrate that, the proposed Deep Embedding method greatly promotes the retrieval accuracy when CNN feature is integrated. We show that our method is efficient in terms of both memory and time cost, and compares favorably with the state-of-the-art methods.
no_new_dataset
0.708811
1406.0304
Markus Schneider
Markus Schneider and Fabio Ramos
Transductive Learning for Multi-Task Copula Processes
null
null
null
null
cs.LG stat.ML
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We tackle the problem of multi-task learning with copula process. Multivariable prediction in spatial and spatial-temporal processes such as natural resource estimation and pollution monitoring have been typically addressed using techniques based on Gaussian processes and co-Kriging. While the Gaussian prior assumption is convenient from analytical and computational perspectives, nature is dominated by non-Gaussian likelihoods. Copula processes are an elegant and flexible solution to handle various non-Gaussian likelihoods by capturing the dependence structure of random variables with cumulative distribution functions rather than their marginals. We show how multi-task learning for copula processes can be used to improve multivariable prediction for problems where the simple Gaussianity prior assumption does not hold. Then, we present a transductive approximation for multi-task learning and derive analytical expressions for the copula process model. The approach is evaluated and compared to other techniques in one artificial dataset and two publicly available datasets for natural resource estimation and concrete slump prediction.
[ { "version": "v1", "created": "Mon, 2 Jun 2014 09:22:49 GMT" } ]
2014-06-03T00:00:00
[ [ "Schneider", "Markus", "" ], [ "Ramos", "Fabio", "" ] ]
TITLE: Transductive Learning for Multi-Task Copula Processes ABSTRACT: We tackle the problem of multi-task learning with copula process. Multivariable prediction in spatial and spatial-temporal processes such as natural resource estimation and pollution monitoring have been typically addressed using techniques based on Gaussian processes and co-Kriging. While the Gaussian prior assumption is convenient from analytical and computational perspectives, nature is dominated by non-Gaussian likelihoods. Copula processes are an elegant and flexible solution to handle various non-Gaussian likelihoods by capturing the dependence structure of random variables with cumulative distribution functions rather than their marginals. We show how multi-task learning for copula processes can be used to improve multivariable prediction for problems where the simple Gaussianity prior assumption does not hold. Then, we present a transductive approximation for multi-task learning and derive analytical expressions for the copula process model. The approach is evaluated and compared to other techniques in one artificial dataset and two publicly available datasets for natural resource estimation and concrete slump prediction.
no_new_dataset
0.708452
1406.0312
Naila Murray
Naila Murray and Florent Perronnin
Generalized Max Pooling
(to appear) CVPR 2014 - IEEE Conference on Computer Vision & Pattern Recognition (2014)
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
State-of-the-art patch-based image representations involve a pooling operation that aggregates statistics computed from local descriptors. Standard pooling operations include sum- and max-pooling. Sum-pooling lacks discriminability because the resulting representation is strongly influenced by frequent yet often uninformative descriptors, but only weakly influenced by rare yet potentially highly-informative ones. Max-pooling equalizes the influence of frequent and rare descriptors but is only applicable to representations that rely on count statistics, such as the bag-of-visual-words (BOV) and its soft- and sparse-coding extensions. We propose a novel pooling mechanism that achieves the same effect as max-pooling but is applicable beyond the BOV and especially to the state-of-the-art Fisher Vector -- hence the name Generalized Max Pooling (GMP). It involves equalizing the similarity between each patch and the pooled representation, which is shown to be equivalent to re-weighting the per-patch statistics. We show on five public image classification benchmarks that the proposed GMP can lead to significant performance gains with respect to heuristic alternatives.
[ { "version": "v1", "created": "Mon, 2 Jun 2014 10:17:03 GMT" } ]
2014-06-03T00:00:00
[ [ "Murray", "Naila", "" ], [ "Perronnin", "Florent", "" ] ]
TITLE: Generalized Max Pooling ABSTRACT: State-of-the-art patch-based image representations involve a pooling operation that aggregates statistics computed from local descriptors. Standard pooling operations include sum- and max-pooling. Sum-pooling lacks discriminability because the resulting representation is strongly influenced by frequent yet often uninformative descriptors, but only weakly influenced by rare yet potentially highly-informative ones. Max-pooling equalizes the influence of frequent and rare descriptors but is only applicable to representations that rely on count statistics, such as the bag-of-visual-words (BOV) and its soft- and sparse-coding extensions. We propose a novel pooling mechanism that achieves the same effect as max-pooling but is applicable beyond the BOV and especially to the state-of-the-art Fisher Vector -- hence the name Generalized Max Pooling (GMP). It involves equalizing the similarity between each patch and the pooled representation, which is shown to be equivalent to re-weighting the per-patch statistics. We show on five public image classification benchmarks that the proposed GMP can lead to significant performance gains with respect to heuristic alternatives.
no_new_dataset
0.71145
1405.7958
George Teodoro
George Teodoro, Tony Pan, Tahsin Kurc, Jun Kong, Lee Cooper, Scott Klasky, Joel Saltz
Region Templates: Data Representation and Management for Large-Scale Image Analysis
43 pages, 17 figures
null
null
null
cs.DC
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Distributed memory machines equipped with CPUs and GPUs (hybrid computing nodes) are hard to program because of the multiple layers of memory and heterogeneous computing configurations. In this paper, we introduce a region template abstraction for the efficient management of common data types used in analysis of large datasets of high resolution images on clusters of hybrid computing nodes. The region template provides a generic container template for common data structures, such as points, arrays, regions, and object sets, within a spatial and temporal bounding box. The region template abstraction enables different data management strategies and data I/O implementations, while providing a homogeneous, unified interface to the application for data storage and retrieval. The execution of region templates applications is coordinated by a runtime system that supports efficient execution in hybrid machines. Region templates applications are represented as hierarchical dataflow in which each computing stage may be represented as another dataflow of finer-grain tasks. A number of optimizations for hybrid machines are available in our runtime system, including performance-aware scheduling for maximizing utilization of computing devices and techniques to reduce impact of data transfers between CPUs and GPUs. An experimental evaluation on a state-of-the-art hybrid cluster using a microscopy imaging study shows that this abstraction adds negligible overhead (about 3%) and achieves good scalability.
[ { "version": "v1", "created": "Fri, 30 May 2014 19:22:46 GMT" } ]
2014-06-02T00:00:00
[ [ "Teodoro", "George", "" ], [ "Pan", "Tony", "" ], [ "Kurc", "Tahsin", "" ], [ "Kong", "Jun", "" ], [ "Cooper", "Lee", "" ], [ "Klasky", "Scott", "" ], [ "Saltz", "Joel", "" ] ]
TITLE: Region Templates: Data Representation and Management for Large-Scale Image Analysis ABSTRACT: Distributed memory machines equipped with CPUs and GPUs (hybrid computing nodes) are hard to program because of the multiple layers of memory and heterogeneous computing configurations. In this paper, we introduce a region template abstraction for the efficient management of common data types used in analysis of large datasets of high resolution images on clusters of hybrid computing nodes. The region template provides a generic container template for common data structures, such as points, arrays, regions, and object sets, within a spatial and temporal bounding box. The region template abstraction enables different data management strategies and data I/O implementations, while providing a homogeneous, unified interface to the application for data storage and retrieval. The execution of region templates applications is coordinated by a runtime system that supports efficient execution in hybrid machines. Region templates applications are represented as hierarchical dataflow in which each computing stage may be represented as another dataflow of finer-grain tasks. A number of optimizations for hybrid machines are available in our runtime system, including performance-aware scheduling for maximizing utilization of computing devices and techniques to reduce impact of data transfers between CPUs and GPUs. An experimental evaluation on a state-of-the-art hybrid cluster using a microscopy imaging study shows that this abstraction adds negligible overhead (about 3%) and achieves good scalability.
no_new_dataset
0.710201
1405.7962
David Monniaux
Julien Henry (VERIMAG - IMAG), Mihail Asavoae (VERIMAG - IMAG), David Monniaux (VERIMAG - IMAG), Claire Ma\"iza (VERIMAG - IMAG)
How to Compute Worst-Case Execution Time by Optimization Modulo Theory and a Clever Encoding of Program Semantics
ACM SIGPLAN/SIGBED Conference on Languages, Compilers and Tools for Embedded Systems 2014, Edimbourg : United Kingdom (2014)
null
null
null
cs.PL cs.LO
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
In systems with hard real-time constraints, it is necessary to compute upper bounds on the worst-case execution time (WCET) of programs; the closer the bound to the real WCET, the better. This is especially the case of synchronous reactive control loops with a fixed clock; the WCET of the loop body must not exceed the clock period. We compute the WCET (or at least a close upper bound thereof) as the solution of an optimization modulo theory problem that takes into account the semantics of the program, in contrast to other methods that compute the longest path whether or not it is feasible according to these semantics. Optimization modulo theory extends satisfiability modulo theory (SMT) to maximization problems. Immediate encodings of WCET problems into SMT yield formulas intractable for all current production-grade solvers; this is inherent to the DPLL(T) approach to SMT implemented in these solvers. By conjoining some appropriate "cuts" to these formulas, we considerably reduce the computation time of the SMT-solver. We experimented our approach on a variety of control programs, using the OTAWA analyzer both as baseline and as underlying microarchitectural analysis for our analysis, and show notable improvement on the WCET bound on a variety of benchmarks and control programs.
[ { "version": "v1", "created": "Fri, 30 May 2014 19:29:16 GMT" } ]
2014-06-02T00:00:00
[ [ "Henry", "Julien", "", "VERIMAG - IMAG" ], [ "Asavoae", "Mihail", "", "VERIMAG - IMAG" ], [ "Monniaux", "David", "", "VERIMAG - IMAG" ], [ "Maïza", "Claire", "", "VERIMAG - IMAG" ] ]
TITLE: How to Compute Worst-Case Execution Time by Optimization Modulo Theory and a Clever Encoding of Program Semantics ABSTRACT: In systems with hard real-time constraints, it is necessary to compute upper bounds on the worst-case execution time (WCET) of programs; the closer the bound to the real WCET, the better. This is especially the case of synchronous reactive control loops with a fixed clock; the WCET of the loop body must not exceed the clock period. We compute the WCET (or at least a close upper bound thereof) as the solution of an optimization modulo theory problem that takes into account the semantics of the program, in contrast to other methods that compute the longest path whether or not it is feasible according to these semantics. Optimization modulo theory extends satisfiability modulo theory (SMT) to maximization problems. Immediate encodings of WCET problems into SMT yield formulas intractable for all current production-grade solvers; this is inherent to the DPLL(T) approach to SMT implemented in these solvers. By conjoining some appropriate "cuts" to these formulas, we considerably reduce the computation time of the SMT-solver. We experimented our approach on a variety of control programs, using the OTAWA analyzer both as baseline and as underlying microarchitectural analysis for our analysis, and show notable improvement on the WCET bound on a variety of benchmarks and control programs.
no_new_dataset
0.70676
1405.7397
Kamal Sarkar
Vivekananda Gayen, Kamal Sarkar
An HMM Based Named Entity Recognition System for Indian Languages: The JU System at ICON 2013
The ICON 2013 tools contest on Named Entity Recognition in Indian languages (IL) co-located with the 10th International Conference on Natural Language Processing(ICON), CDAC Noida, India,18-20 December, 2013
null
null
null
cs.CL
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
This paper reports about our work in the ICON 2013 NLP TOOLS CONTEST on Named Entity Recognition. We submitted runs for Bengali, English, Hindi, Marathi, Punjabi, Tamil and Telugu. A statistical HMM (Hidden Markov Models) based model has been used to implement our system. The system has been trained and tested on the NLP TOOLS CONTEST: ICON 2013 datasets. Our system obtains F-measures of 0.8599, 0.7704, 0.7520, 0.4289, 0.5455, 0.4466, and 0.4003 for Bengali, English, Hindi, Marathi, Punjabi, Tamil and Telugu respectively.
[ { "version": "v1", "created": "Wed, 28 May 2014 21:05:00 GMT" } ]
2014-05-30T00:00:00
[ [ "Gayen", "Vivekananda", "" ], [ "Sarkar", "Kamal", "" ] ]
TITLE: An HMM Based Named Entity Recognition System for Indian Languages: The JU System at ICON 2013 ABSTRACT: This paper reports about our work in the ICON 2013 NLP TOOLS CONTEST on Named Entity Recognition. We submitted runs for Bengali, English, Hindi, Marathi, Punjabi, Tamil and Telugu. A statistical HMM (Hidden Markov Models) based model has been used to implement our system. The system has been trained and tested on the NLP TOOLS CONTEST: ICON 2013 datasets. Our system obtains F-measures of 0.8599, 0.7704, 0.7520, 0.4289, 0.5455, 0.4466, and 0.4003 for Bengali, English, Hindi, Marathi, Punjabi, Tamil and Telugu respectively.
no_new_dataset
0.709038
1405.7545
Michael Sapienza
Michael Sapienza and Fabio Cuzzolin and Philip H.S. Torr
Feature sampling and partitioning for visual vocabulary generation on large action classification datasets
null
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The recent trend in action recognition is towards larger datasets, an increasing number of action classes and larger visual vocabularies. State-of-the-art human action classification in challenging video data is currently based on a bag-of-visual-words pipeline in which space-time features are aggregated globally to form a histogram. The strategies chosen to sample features and construct a visual vocabulary are critical to performance, in fact often dominating performance. In this work we provide a critical evaluation of various approaches to building a vocabulary and show that good practises do have a significant impact. By subsampling and partitioning features strategically, we are able to achieve state-of-the-art results on 5 major action recognition datasets using relatively small visual vocabularies.
[ { "version": "v1", "created": "Thu, 29 May 2014 13:09:52 GMT" } ]
2014-05-30T00:00:00
[ [ "Sapienza", "Michael", "" ], [ "Cuzzolin", "Fabio", "" ], [ "Torr", "Philip H. S.", "" ] ]
TITLE: Feature sampling and partitioning for visual vocabulary generation on large action classification datasets ABSTRACT: The recent trend in action recognition is towards larger datasets, an increasing number of action classes and larger visual vocabularies. State-of-the-art human action classification in challenging video data is currently based on a bag-of-visual-words pipeline in which space-time features are aggregated globally to form a histogram. The strategies chosen to sample features and construct a visual vocabulary are critical to performance, in fact often dominating performance. In this work we provide a critical evaluation of various approaches to building a vocabulary and show that good practises do have a significant impact. By subsampling and partitioning features strategically, we are able to achieve state-of-the-art results on 5 major action recognition datasets using relatively small visual vocabularies.
no_new_dataset
0.712045
1405.7631
Mostafa Salehi
Motahareh Eslami Mehdiabadi, Hamid R. Rabiee, Mostafa Salehi
Diffusion-Aware Sampling and Estimation in Information Diffusion Networks
8 pages, 4 figures, Published in: International Confernece on Social Computing 2012 (SocialCom12)
null
10.1109/SocialCom-PASSAT.2012.98
null
cs.SI physics.soc-ph
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Partially-observed data collected by sampling methods is often being studied to obtain the characteristics of information diffusion networks. However, these methods usually do not consider the behavior of diffusion process. In this paper, we propose a novel two-step (sampling/estimation) measurement framework by utilizing the diffusion process characteristics. To this end, we propose a link-tracing based sampling design which uses the infection times as local information without any knowledge about the latent structure of diffusion network. To correct the bias of sampled data, we introduce three estimators for different categories; link-based, node-based, and cascade-based. To the best of our knowledge, this is the first attempt to introduce a complete measurement framework for diffusion networks. We also show that the estimator plays an important role in correcting the bias of sampling from diffusion networks. Our comprehensive empirical analysis over large synthetic and real datasets demonstrates that in average, the proposed framework outperforms the common BFS and RW sampling methods in terms of link-based characteristics by about 37% and 35%, respectively.
[ { "version": "v1", "created": "Thu, 29 May 2014 17:52:04 GMT" } ]
2014-05-30T00:00:00
[ [ "Mehdiabadi", "Motahareh Eslami", "" ], [ "Rabiee", "Hamid R.", "" ], [ "Salehi", "Mostafa", "" ] ]
TITLE: Diffusion-Aware Sampling and Estimation in Information Diffusion Networks ABSTRACT: Partially-observed data collected by sampling methods is often being studied to obtain the characteristics of information diffusion networks. However, these methods usually do not consider the behavior of diffusion process. In this paper, we propose a novel two-step (sampling/estimation) measurement framework by utilizing the diffusion process characteristics. To this end, we propose a link-tracing based sampling design which uses the infection times as local information without any knowledge about the latent structure of diffusion network. To correct the bias of sampled data, we introduce three estimators for different categories; link-based, node-based, and cascade-based. To the best of our knowledge, this is the first attempt to introduce a complete measurement framework for diffusion networks. We also show that the estimator plays an important role in correcting the bias of sampling from diffusion networks. Our comprehensive empirical analysis over large synthetic and real datasets demonstrates that in average, the proposed framework outperforms the common BFS and RW sampling methods in terms of link-based characteristics by about 37% and 35%, respectively.
no_new_dataset
0.711205
1312.6190
Son Tran
Son N. Tran, Artur d'Avila Garcez
Adaptive Feature Ranking for Unsupervised Transfer Learning
9 pages 7 figures, new experimental results on ranking and transfer have been added, typo fixed
null
null
null
cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Transfer Learning is concerned with the application of knowledge gained from solving a problem to a different but related problem domain. In this paper, we propose a method and efficient algorithm for ranking and selecting representations from a Restricted Boltzmann Machine trained on a source domain to be transferred onto a target domain. Experiments carried out using the MNIST, ICDAR and TiCC image datasets show that the proposed adaptive feature ranking and transfer learning method offers statistically significant improvements on the training of RBMs. Our method is general in that the knowledge chosen by the ranking function does not depend on its relation to any specific target domain, and it works with unsupervised learning and knowledge-based transfer.
[ { "version": "v1", "created": "Sat, 21 Dec 2013 01:50:08 GMT" }, { "version": "v2", "created": "Wed, 28 May 2014 16:35:17 GMT" } ]
2014-05-29T00:00:00
[ [ "Tran", "Son N.", "" ], [ "Garcez", "Artur d'Avila", "" ] ]
TITLE: Adaptive Feature Ranking for Unsupervised Transfer Learning ABSTRACT: Transfer Learning is concerned with the application of knowledge gained from solving a problem to a different but related problem domain. In this paper, we propose a method and efficient algorithm for ranking and selecting representations from a Restricted Boltzmann Machine trained on a source domain to be transferred onto a target domain. Experiments carried out using the MNIST, ICDAR and TiCC image datasets show that the proposed adaptive feature ranking and transfer learning method offers statistically significant improvements on the training of RBMs. Our method is general in that the knowledge chosen by the ranking function does not depend on its relation to any specific target domain, and it works with unsupervised learning and knowledge-based transfer.
no_new_dataset
0.708855
1405.6804
Zhuowen Tu
Zhuowen Tu and Piotr Dollar and Yingnian Wu
Layered Logic Classifiers: Exploring the `And' and `Or' Relations
null
null
null
null
stat.ML cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Designing effective and efficient classifier for pattern analysis is a key problem in machine learning and computer vision. Many the solutions to the problem require to perform logic operations such as `and', `or', and `not'. Classification and regression tree (CART) include these operations explicitly. Other methods such as neural networks, SVM, and boosting learn/compute a weighted sum on features (weak classifiers), which weakly perform the 'and' and 'or' operations. However, it is hard for these classifiers to deal with the 'xor' pattern directly. In this paper, we propose layered logic classifiers for patterns of complicated distributions by combining the `and', `or', and `not' operations. The proposed algorithm is very general and easy to implement. We test the classifiers on several typical datasets from the Irvine repository and two challenging vision applications, object segmentation and pedestrian detection. We observe significant improvements on all the datasets over the widely used decision stump based AdaBoost algorithm. The resulting classifiers have much less training complexity than decision tree based AdaBoost, and can be applied in a wide range of domains.
[ { "version": "v1", "created": "Tue, 27 May 2014 06:29:01 GMT" }, { "version": "v2", "created": "Wed, 28 May 2014 00:51:08 GMT" } ]
2014-05-29T00:00:00
[ [ "Tu", "Zhuowen", "" ], [ "Dollar", "Piotr", "" ], [ "Wu", "Yingnian", "" ] ]
TITLE: Layered Logic Classifiers: Exploring the `And' and `Or' Relations ABSTRACT: Designing effective and efficient classifier for pattern analysis is a key problem in machine learning and computer vision. Many the solutions to the problem require to perform logic operations such as `and', `or', and `not'. Classification and regression tree (CART) include these operations explicitly. Other methods such as neural networks, SVM, and boosting learn/compute a weighted sum on features (weak classifiers), which weakly perform the 'and' and 'or' operations. However, it is hard for these classifiers to deal with the 'xor' pattern directly. In this paper, we propose layered logic classifiers for patterns of complicated distributions by combining the `and', `or', and `not' operations. The proposed algorithm is very general and easy to implement. We test the classifiers on several typical datasets from the Irvine repository and two challenging vision applications, object segmentation and pedestrian detection. We observe significant improvements on all the datasets over the widely used decision stump based AdaBoost algorithm. The resulting classifiers have much less training complexity than decision tree based AdaBoost, and can be applied in a wide range of domains.
no_new_dataset
0.710057
1405.7258
Mostafa Salehi
Motahareh Eslami Mehdiabadi, Hamid R. Rabiee, Mostafa Salehi
Sampling from Diffusion Networks
Published in Proceedings of the 2012 International Conference on Social Informatics, Pages 106-112
null
10.1109/SocialInformatics.2012.79
null
cs.SI physics.soc-ph
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The diffusion phenomenon has a remarkable impact on Online Social Networks (OSNs). Gathering diffusion data over these large networks encounters many challenges which can be alleviated by adopting a suitable sampling approach. The contributions of this paper is twofold. First we study the sampling approaches over diffusion networks, and for the first time, classify these approaches into two categories; (1) Structure-based Sampling (SBS), and (2) Diffusion-based Sampling (DBS). The dependency of the former approach to topological features of the network, and unavailability of real diffusion paths in the latter, converts the problem of choosing an appropriate sampling approach to a trade-off. Second, we formally define the diffusion network sampling problem and propose a number of new diffusion-based characteristics to evaluate introduced sampling approaches. Our experiments on large scale synthetic and real datasets show that although DBS performs much better than SBS in higher sampling rates (16% ~ 29% on average), their performances differ about 7% in lower sampling rates. Therefore, in real large scale systems with low sampling rate requirements, SBS would be a better choice according to its lower time complexity in gathering data compared to DBS. Moreover, we show that the introduced sampling approaches (SBS and DBS) play a more important role than the graph exploration techniques such as Breadth-First Search (BFS) and Random Walk (RW) in the analysis of diffusion processes.
[ { "version": "v1", "created": "Wed, 28 May 2014 14:33:02 GMT" } ]
2014-05-29T00:00:00
[ [ "Mehdiabadi", "Motahareh Eslami", "" ], [ "Rabiee", "Hamid R.", "" ], [ "Salehi", "Mostafa", "" ] ]
TITLE: Sampling from Diffusion Networks ABSTRACT: The diffusion phenomenon has a remarkable impact on Online Social Networks (OSNs). Gathering diffusion data over these large networks encounters many challenges which can be alleviated by adopting a suitable sampling approach. The contributions of this paper is twofold. First we study the sampling approaches over diffusion networks, and for the first time, classify these approaches into two categories; (1) Structure-based Sampling (SBS), and (2) Diffusion-based Sampling (DBS). The dependency of the former approach to topological features of the network, and unavailability of real diffusion paths in the latter, converts the problem of choosing an appropriate sampling approach to a trade-off. Second, we formally define the diffusion network sampling problem and propose a number of new diffusion-based characteristics to evaluate introduced sampling approaches. Our experiments on large scale synthetic and real datasets show that although DBS performs much better than SBS in higher sampling rates (16% ~ 29% on average), their performances differ about 7% in lower sampling rates. Therefore, in real large scale systems with low sampling rate requirements, SBS would be a better choice according to its lower time complexity in gathering data compared to DBS. Moreover, we show that the introduced sampling approaches (SBS and DBS) play a more important role than the graph exploration techniques such as Breadth-First Search (BFS) and Random Walk (RW) in the analysis of diffusion processes.
no_new_dataset
0.711857
1405.1213
Oscar Danielsson
Oscar Danielsson and Omid Aghazadeh
Human Pose Estimation from RGB Input Using Synthetic Training Data
6 pages
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We address the problem of estimating the pose of humans using RGB image input. More specifically, we are using a random forest classifier to classify pixels into joint-based body part categories, much similar to the famous Kinect pose estimator [11], [12]. However, we are using pure RGB input, i.e. no depth. Since the random forest requires a large number of training examples, we are using computer graphics generated, synthetic training data. In addition, we assume that we have access to a large number of real images with bounding box labels, extracted for example by a pedestrian detector or a tracking system. We propose a new objective function for random forest training that uses the weakly labeled data from the target domain to encourage the learner to select features that generalize from the synthetic source domain to the real target domain. We demonstrate on a publicly available dataset [6] that the proposed objective function yields a classifier that significantly outperforms a baseline classifier trained using the standard entropy objective [10].
[ { "version": "v1", "created": "Tue, 6 May 2014 10:13:08 GMT" }, { "version": "v2", "created": "Tue, 27 May 2014 12:23:54 GMT" } ]
2014-05-28T00:00:00
[ [ "Danielsson", "Oscar", "" ], [ "Aghazadeh", "Omid", "" ] ]
TITLE: Human Pose Estimation from RGB Input Using Synthetic Training Data ABSTRACT: We address the problem of estimating the pose of humans using RGB image input. More specifically, we are using a random forest classifier to classify pixels into joint-based body part categories, much similar to the famous Kinect pose estimator [11], [12]. However, we are using pure RGB input, i.e. no depth. Since the random forest requires a large number of training examples, we are using computer graphics generated, synthetic training data. In addition, we assume that we have access to a large number of real images with bounding box labels, extracted for example by a pedestrian detector or a tracking system. We propose a new objective function for random forest training that uses the weakly labeled data from the target domain to encourage the learner to select features that generalize from the synthetic source domain to the real target domain. We demonstrate on a publicly available dataset [6] that the proposed objective function yields a classifier that significantly outperforms a baseline classifier trained using the standard entropy objective [10].
no_new_dataset
0.710007
1405.6886
Rasmus Troelsg{\aa}rd
Rasmus Troelsg{\aa}rd, Bj{\o}rn Sand Jensen, Lars Kai Hansen
A Topic Model Approach to Multi-Modal Similarity
topic modelling workshop at NIPS 2013
null
null
null
cs.IR stat.ML
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Calculating similarities between objects defined by many heterogeneous data modalities is an important challenge in many multimedia applications. We use a multi-modal topic model as a basis for defining such a similarity between objects. We propose to compare the resulting similarities from different model realizations using the non-parametric Mantel test. The approach is evaluated on a music dataset.
[ { "version": "v1", "created": "Tue, 27 May 2014 12:34:24 GMT" } ]
2014-05-28T00:00:00
[ [ "Troelsgård", "Rasmus", "" ], [ "Jensen", "Bjørn Sand", "" ], [ "Hansen", "Lars Kai", "" ] ]
TITLE: A Topic Model Approach to Multi-Modal Similarity ABSTRACT: Calculating similarities between objects defined by many heterogeneous data modalities is an important challenge in many multimedia applications. We use a multi-modal topic model as a basis for defining such a similarity between objects. We propose to compare the resulting similarities from different model realizations using the non-parametric Mantel test. The approach is evaluated on a music dataset.
no_new_dataset
0.710842
1405.6922
Omid Aghazadeh
Omid Aghazadeh and Stefan Carlsson
Large Scale, Large Margin Classification using Indefinite Similarity Measures
null
null
null
null
cs.LG cs.CV stat.ML
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Despite the success of the popular kernelized support vector machines, they have two major limitations: they are restricted to Positive Semi-Definite (PSD) kernels, and their training complexity scales at least quadratically with the size of the data. Many natural measures of similarity between pairs of samples are not PSD e.g. invariant kernels, and those that are implicitly or explicitly defined by latent variable models. In this paper, we investigate scalable approaches for using indefinite similarity measures in large margin frameworks. In particular we show that a normalization of similarity to a subset of the data points constitutes a representation suitable for linear classifiers. The result is a classifier which is competitive to kernelized SVM in terms of accuracy, despite having better training and test time complexities. Experimental results demonstrate that on CIFAR-10 dataset, the model equipped with similarity measures invariant to rigid and non-rigid deformations, can be made more than 5 times sparser while being more accurate than kernelized SVM using RBF kernels.
[ { "version": "v1", "created": "Tue, 27 May 2014 14:18:26 GMT" } ]
2014-05-28T00:00:00
[ [ "Aghazadeh", "Omid", "" ], [ "Carlsson", "Stefan", "" ] ]
TITLE: Large Scale, Large Margin Classification using Indefinite Similarity Measures ABSTRACT: Despite the success of the popular kernelized support vector machines, they have two major limitations: they are restricted to Positive Semi-Definite (PSD) kernels, and their training complexity scales at least quadratically with the size of the data. Many natural measures of similarity between pairs of samples are not PSD e.g. invariant kernels, and those that are implicitly or explicitly defined by latent variable models. In this paper, we investigate scalable approaches for using indefinite similarity measures in large margin frameworks. In particular we show that a normalization of similarity to a subset of the data points constitutes a representation suitable for linear classifiers. The result is a classifier which is competitive to kernelized SVM in terms of accuracy, despite having better training and test time complexities. Experimental results demonstrate that on CIFAR-10 dataset, the model equipped with similarity measures invariant to rigid and non-rigid deformations, can be made more than 5 times sparser while being more accurate than kernelized SVM using RBF kernels.
no_new_dataset
0.710459
1206.5333
Leon Derczynski
Naushad UzZaman, Hector Llorens, James Allen, Leon Derczynski, Marc Verhagen and James Pustejovsky
TempEval-3: Evaluating Events, Time Expressions, and Temporal Relations
null
null
null
null
cs.CL
http://creativecommons.org/licenses/by/3.0/
We describe the TempEval-3 task which is currently in preparation for the SemEval-2013 evaluation exercise. The aim of TempEval is to advance research on temporal information processing. TempEval-3 follows on from previous TempEval events, incorporating: a three-part task structure covering event, temporal expression and temporal relation extraction; a larger dataset; and single overall task quality scores.
[ { "version": "v1", "created": "Fri, 22 Jun 2012 22:30:44 GMT" }, { "version": "v2", "created": "Sun, 25 May 2014 19:10:12 GMT" } ]
2014-05-27T00:00:00
[ [ "UzZaman", "Naushad", "" ], [ "Llorens", "Hector", "" ], [ "Allen", "James", "" ], [ "Derczynski", "Leon", "" ], [ "Verhagen", "Marc", "" ], [ "Pustejovsky", "James", "" ] ]
TITLE: TempEval-3: Evaluating Events, Time Expressions, and Temporal Relations ABSTRACT: We describe the TempEval-3 task which is currently in preparation for the SemEval-2013 evaluation exercise. The aim of TempEval is to advance research on temporal information processing. TempEval-3 follows on from previous TempEval events, incorporating: a three-part task structure covering event, temporal expression and temporal relation extraction; a larger dataset; and single overall task quality scores.
no_new_dataset
0.706792
1306.1091
Yoshua Bengio
Yoshua Bengio, \'Eric Thibodeau-Laufer, Guillaume Alain and Jason Yosinski
Deep Generative Stochastic Networks Trainable by Backprop
arXiv admin note: text overlap with arXiv:1305.0445, Also published in ICML'2014
null
null
null
cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We introduce a novel training principle for probabilistic models that is an alternative to maximum likelihood. The proposed Generative Stochastic Networks (GSN) framework is based on learning the transition operator of a Markov chain whose stationary distribution estimates the data distribution. The transition distribution of the Markov chain is conditional on the previous state, generally involving a small move, so this conditional distribution has fewer dominant modes, being unimodal in the limit of small moves. Thus, it is easier to learn because it is easier to approximate its partition function, more like learning to perform supervised function approximation, with gradients that can be obtained by backprop. We provide theorems that generalize recent work on the probabilistic interpretation of denoising autoencoders and obtain along the way an interesting justification for dependency networks and generalized pseudolikelihood, along with a definition of an appropriate joint distribution and sampling mechanism even when the conditionals are not consistent. GSNs can be used with missing inputs and can be used to sample subsets of variables given the rest. We validate these theoretical results with experiments on two image datasets using an architecture that mimics the Deep Boltzmann Machine Gibbs sampler but allows training to proceed with simple backprop, without the need for layerwise pretraining.
[ { "version": "v1", "created": "Wed, 5 Jun 2013 13:01:14 GMT" }, { "version": "v2", "created": "Fri, 7 Jun 2013 16:55:38 GMT" }, { "version": "v3", "created": "Fri, 25 Oct 2013 07:04:58 GMT" }, { "version": "v4", "created": "Wed, 18 Dec 2013 19:46:07 GMT" }, { "version": "v5", "created": "Sat, 24 May 2014 00:05:18 GMT" } ]
2014-05-27T00:00:00
[ [ "Bengio", "Yoshua", "" ], [ "Thibodeau-Laufer", "Éric", "" ], [ "Alain", "Guillaume", "" ], [ "Yosinski", "Jason", "" ] ]
TITLE: Deep Generative Stochastic Networks Trainable by Backprop ABSTRACT: We introduce a novel training principle for probabilistic models that is an alternative to maximum likelihood. The proposed Generative Stochastic Networks (GSN) framework is based on learning the transition operator of a Markov chain whose stationary distribution estimates the data distribution. The transition distribution of the Markov chain is conditional on the previous state, generally involving a small move, so this conditional distribution has fewer dominant modes, being unimodal in the limit of small moves. Thus, it is easier to learn because it is easier to approximate its partition function, more like learning to perform supervised function approximation, with gradients that can be obtained by backprop. We provide theorems that generalize recent work on the probabilistic interpretation of denoising autoencoders and obtain along the way an interesting justification for dependency networks and generalized pseudolikelihood, along with a definition of an appropriate joint distribution and sampling mechanism even when the conditionals are not consistent. GSNs can be used with missing inputs and can be used to sample subsets of variables given the rest. We validate these theoretical results with experiments on two image datasets using an architecture that mimics the Deep Boltzmann Machine Gibbs sampler but allows training to proceed with simple backprop, without the need for layerwise pretraining.
no_new_dataset
0.710201
1405.4028
Anvesh Komuravelli
Anvesh Komuravelli, Arie Gurfinkel, Sagar Chaki
SMT-based Model Checking for Recursive Programs
originally published as part of the proceedings of CAV 2014; fixed typos, better wording at some places
null
null
null
cs.LO
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We present an SMT-based symbolic model checking algorithm for safety verification of recursive programs. The algorithm is modular and analyzes procedures individually. Unlike other SMT-based approaches, it maintains both "over-" and "under-approximations" of procedure summaries. Under-approximations are used to analyze procedure calls without inlining. Over-approximations are used to block infeasible counterexamples and detect convergence to a proof. We show that for programs and properties over a decidable theory, the algorithm is guaranteed to find a counterexample, if one exists. However, efficiency depends on an oracle for quantifier elimination (QE). For Boolean Programs, the algorithm is a polynomial decision procedure, matching the worst-case bounds of the best BDD-based algorithms. For Linear Arithmetic (integers and rationals), we give an efficient instantiation of the algorithm by applying QE "lazily". We use existing interpolation techniques to over-approximate QE and introduce "Model Based Projection" to under-approximate QE. Empirical evaluation on SV-COMP benchmarks shows that our algorithm improves significantly on the state-of-the-art.
[ { "version": "v1", "created": "Thu, 15 May 2014 22:31:03 GMT" }, { "version": "v2", "created": "Sun, 25 May 2014 17:52:17 GMT" } ]
2014-05-27T00:00:00
[ [ "Komuravelli", "Anvesh", "" ], [ "Gurfinkel", "Arie", "" ], [ "Chaki", "Sagar", "" ] ]
TITLE: SMT-based Model Checking for Recursive Programs ABSTRACT: We present an SMT-based symbolic model checking algorithm for safety verification of recursive programs. The algorithm is modular and analyzes procedures individually. Unlike other SMT-based approaches, it maintains both "over-" and "under-approximations" of procedure summaries. Under-approximations are used to analyze procedure calls without inlining. Over-approximations are used to block infeasible counterexamples and detect convergence to a proof. We show that for programs and properties over a decidable theory, the algorithm is guaranteed to find a counterexample, if one exists. However, efficiency depends on an oracle for quantifier elimination (QE). For Boolean Programs, the algorithm is a polynomial decision procedure, matching the worst-case bounds of the best BDD-based algorithms. For Linear Arithmetic (integers and rationals), we give an efficient instantiation of the algorithm by applying QE "lazily". We use existing interpolation techniques to over-approximate QE and introduce "Model Based Projection" to under-approximate QE. Empirical evaluation on SV-COMP benchmarks shows that our algorithm improves significantly on the state-of-the-art.
no_new_dataset
0.70614
1405.6223
Fangfang Li
Fangfang Li, Guandong Xu, Longbing Cao
Coupled Item-based Matrix Factorization
7 pages submitted to AAAI2014. arXiv admin note: substantial text overlap with arXiv:1404.7467
null
null
null
cs.LG cs.IR
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The essence of the challenges cold start and sparsity in Recommender Systems (RS) is that the extant techniques, such as Collaborative Filtering (CF) and Matrix Factorization (MF), mainly rely on the user-item rating matrix, which sometimes is not informative enough for predicting recommendations. To solve these challenges, the objective item attributes are incorporated as complementary information. However, most of the existing methods for inferring the relationships between items assume that the attributes are "independently and identically distributed (iid)", which does not always hold in reality. In fact, the attributes are more or less coupled with each other by some implicit relationships. Therefore, in this pa-per we propose an attribute-based coupled similarity measure to capture the implicit relationships between items. We then integrate the implicit item coupling into MF to form the Coupled Item-based Matrix Factorization (CIMF) model. Experimental results on two open data sets demonstrate that CIMF outperforms the benchmark methods.
[ { "version": "v1", "created": "Tue, 8 Apr 2014 00:42:16 GMT" } ]
2014-05-27T00:00:00
[ [ "Li", "Fangfang", "" ], [ "Xu", "Guandong", "" ], [ "Cao", "Longbing", "" ] ]
TITLE: Coupled Item-based Matrix Factorization ABSTRACT: The essence of the challenges cold start and sparsity in Recommender Systems (RS) is that the extant techniques, such as Collaborative Filtering (CF) and Matrix Factorization (MF), mainly rely on the user-item rating matrix, which sometimes is not informative enough for predicting recommendations. To solve these challenges, the objective item attributes are incorporated as complementary information. However, most of the existing methods for inferring the relationships between items assume that the attributes are "independently and identically distributed (iid)", which does not always hold in reality. In fact, the attributes are more or less coupled with each other by some implicit relationships. Therefore, in this pa-per we propose an attribute-based coupled similarity measure to capture the implicit relationships between items. We then integrate the implicit item coupling into MF to form the Coupled Item-based Matrix Factorization (CIMF) model. Experimental results on two open data sets demonstrate that CIMF outperforms the benchmark methods.
no_new_dataset
0.707569
1405.6490
Bohar Singh Singh
Bohar Singh, Pawan Luthra
Review of Linpack and Cloudsim on VMM
5 Pages
null
10.14445/22315381/IJETT-V11P251
null
cs.DC
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Virtualization is a framework of dividing the resources of a computer into multiple execution environments which offers a lot of benefits including flexibility, security, ease to configuration and reduction of cost but at the same time it also brings a certain degree of performance overhead. Furthermore, Virtual Machine Monitor (VMM) is the core component of virtual machine (VM) system and its effectiveness greatly impacts the performance of the whole system. This review paper will try to describe the basic knowledge about various virtual machine monitors such as VMware and VirtualBox. It also discussed and explores the benchmark LINPACK and CloudSim available for cloud computing. This benchmark and CloudSim can be used to measure the performance of two different virtual machine monitors in terms of processing speed, time, bandwidth, quality and response of the cloud computing network.
[ { "version": "v1", "created": "Mon, 26 May 2014 07:50:16 GMT" } ]
2014-05-27T00:00:00
[ [ "Singh", "Bohar", "" ], [ "Luthra", "Pawan", "" ] ]
TITLE: Review of Linpack and Cloudsim on VMM ABSTRACT: Virtualization is a framework of dividing the resources of a computer into multiple execution environments which offers a lot of benefits including flexibility, security, ease to configuration and reduction of cost but at the same time it also brings a certain degree of performance overhead. Furthermore, Virtual Machine Monitor (VMM) is the core component of virtual machine (VM) system and its effectiveness greatly impacts the performance of the whole system. This review paper will try to describe the basic knowledge about various virtual machine monitors such as VMware and VirtualBox. It also discussed and explores the benchmark LINPACK and CloudSim available for cloud computing. This benchmark and CloudSim can be used to measure the performance of two different virtual machine monitors in terms of processing speed, time, bandwidth, quality and response of the cloud computing network.
no_new_dataset
0.712188
1405.6133
Arun P V
S.K. Katiyar, P.V. Arun
A review over the applicability of image entropy in analyses of remote sensing datasets
arXiv admin note: substantial text overlap with arXiv:1303.6926
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Entropy is the measure of uncertainty in any data and is adopted for maximisation of mutual information in many remote sensing operations. The availability of wide entropy variations motivated us for an investigation over the suitability preference of these versions to specific operations.
[ { "version": "v1", "created": "Wed, 5 Feb 2014 14:49:36 GMT" } ]
2014-05-26T00:00:00
[ [ "Katiyar", "S. K.", "" ], [ "Arun", "P. V.", "" ] ]
TITLE: A review over the applicability of image entropy in analyses of remote sensing datasets ABSTRACT: Entropy is the measure of uncertainty in any data and is adopted for maximisation of mutual information in many remote sensing operations. The availability of wide entropy variations motivated us for an investigation over the suitability preference of these versions to specific operations.
no_new_dataset
0.711675
1405.6173
Ahmed Ibrahim Taloba
M. H. Marghny, Rasha M. Abd El-Aziz, Ahmed I. Taloba
An Effective Evolutionary Clustering Algorithm: Hepatitis C Case Study
null
null
null
null
cs.NE cs.CE
http://creativecommons.org/licenses/by-nc-sa/3.0/
Clustering analysis plays an important role in scientific research and commercial application. K-means algorithm is a widely used partition method in clustering. However, it is known that the K-means algorithm may get stuck at suboptimal solutions, depending on the choice of the initial cluster centers. In this article, we propose a technique to handle large scale data, which can select initial clustering center purposefully using Genetic algorithms (GAs), reduce the sensitivity to isolated point, avoid dissevering big cluster, and overcome deflexion of data in some degree that caused by the disproportion in data partitioning owing to adoption of multi-sampling. We applied our method to some public datasets these show the advantages of the proposed approach for example Hepatitis C dataset that has been taken from the machine learning warehouse of University of California. Our aim is to evaluate hepatitis dataset. In order to evaluate this dataset we did some preprocessing operation, the reason to preprocessing is to summarize the data in the best and suitable way for our algorithm. Missing values of the instances are adjusted using local mean method.
[ { "version": "v1", "created": "Thu, 27 Feb 2014 11:03:28 GMT" } ]
2014-05-26T00:00:00
[ [ "Marghny", "M. H.", "" ], [ "El-Aziz", "Rasha M. Abd", "" ], [ "Taloba", "Ahmed I.", "" ] ]
TITLE: An Effective Evolutionary Clustering Algorithm: Hepatitis C Case Study ABSTRACT: Clustering analysis plays an important role in scientific research and commercial application. K-means algorithm is a widely used partition method in clustering. However, it is known that the K-means algorithm may get stuck at suboptimal solutions, depending on the choice of the initial cluster centers. In this article, we propose a technique to handle large scale data, which can select initial clustering center purposefully using Genetic algorithms (GAs), reduce the sensitivity to isolated point, avoid dissevering big cluster, and overcome deflexion of data in some degree that caused by the disproportion in data partitioning owing to adoption of multi-sampling. We applied our method to some public datasets these show the advantages of the proposed approach for example Hepatitis C dataset that has been taken from the machine learning warehouse of University of California. Our aim is to evaluate hepatitis dataset. In order to evaluate this dataset we did some preprocessing operation, the reason to preprocessing is to summarize the data in the best and suitable way for our algorithm. Missing values of the instances are adjusted using local mean method.
no_new_dataset
0.7116
1405.3100
Andrea Monacchi
Andrea Monacchi, Dominik Egarter, Wilfried Elmenreich, Salvatore D'Alessandro, Andrea M. Tonello
GREEND: An Energy Consumption Dataset of Households in Italy and Austria
null
null
null
null
cs.OH
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Home energy management systems can be used to monitor and optimize consumption and local production from renewable energy. To assess solutions before their deployment, researchers and designers of those systems demand for energy consumption datasets. In this paper, we present the GREEND dataset, containing detailed power usage information obtained through a measurement campaign in households in Austria and Italy. We provide a description of consumption scenarios and discuss design choices for the sensing infrastructure. Finally, we benchmark the dataset with state-of-the-art techniques in load disaggregation, occupancy detection and appliance usage mining.
[ { "version": "v1", "created": "Tue, 13 May 2014 10:51:32 GMT" }, { "version": "v2", "created": "Thu, 22 May 2014 13:57:03 GMT" } ]
2014-05-23T00:00:00
[ [ "Monacchi", "Andrea", "" ], [ "Egarter", "Dominik", "" ], [ "Elmenreich", "Wilfried", "" ], [ "D'Alessandro", "Salvatore", "" ], [ "Tonello", "Andrea M.", "" ] ]
TITLE: GREEND: An Energy Consumption Dataset of Households in Italy and Austria ABSTRACT: Home energy management systems can be used to monitor and optimize consumption and local production from renewable energy. To assess solutions before their deployment, researchers and designers of those systems demand for energy consumption datasets. In this paper, we present the GREEND dataset, containing detailed power usage information obtained through a measurement campaign in households in Austria and Italy. We provide a description of consumption scenarios and discuss design choices for the sensing infrastructure. Finally, we benchmark the dataset with state-of-the-art techniques in load disaggregation, occupancy detection and appliance usage mining.
new_dataset
0.713444
1405.5245
Luca Bonaventura
G. Tumolo and L. Bonaventura
An accurate and efficient numerical framework for adaptive numerical weather prediction
null
null
null
null
math.NA cs.CE cs.NA physics.ao-ph physics.comp-ph physics.flu-dyn
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We present an accurate and efficient discretization approach for the adaptive discretization of typical model equations employed in numerical weather prediction. A semi-Lagrangian approach is combined with the TR-BDF2 semi-implicit time discretization method and with a spatial discretization based on adaptive discontinuous finite elements. The resulting method has full second order accuracy in time and can employ polynomial bases of arbitrarily high degree in space, is unconditionally stable and can effectively adapt the number of degrees of freedom employed in each element, in order to balance accuracy and computational cost. The p-adaptivity approach employed does not require remeshing, therefore it is especially suitable for applications, such as numerical weather prediction, in which a large number of physical quantities are associated with a given mesh. Furthermore, although the proposed method can be implemented on arbitrary unstructured and nonconforming meshes, even its application on simple Cartesian meshes in spherical coordinates can cure effectively the pole problem by reducing the polynomial degree used in the polar elements. Numerical simulations of classical benchmarks for the shallow water and for the fully compressible Euler equations validate the method and demonstrate its capability to achieve accurate results also at large Courant numbers, with time steps up to 100 times larger than those of typical explicit discretizations of the same problems, while reducing the computational cost thanks to the adaptivity algorithm.
[ { "version": "v1", "created": "Tue, 20 May 2014 21:08:44 GMT" } ]
2014-05-23T00:00:00
[ [ "Tumolo", "G.", "" ], [ "Bonaventura", "L.", "" ] ]
TITLE: An accurate and efficient numerical framework for adaptive numerical weather prediction ABSTRACT: We present an accurate and efficient discretization approach for the adaptive discretization of typical model equations employed in numerical weather prediction. A semi-Lagrangian approach is combined with the TR-BDF2 semi-implicit time discretization method and with a spatial discretization based on adaptive discontinuous finite elements. The resulting method has full second order accuracy in time and can employ polynomial bases of arbitrarily high degree in space, is unconditionally stable and can effectively adapt the number of degrees of freedom employed in each element, in order to balance accuracy and computational cost. The p-adaptivity approach employed does not require remeshing, therefore it is especially suitable for applications, such as numerical weather prediction, in which a large number of physical quantities are associated with a given mesh. Furthermore, although the proposed method can be implemented on arbitrary unstructured and nonconforming meshes, even its application on simple Cartesian meshes in spherical coordinates can cure effectively the pole problem by reducing the polynomial degree used in the polar elements. Numerical simulations of classical benchmarks for the shallow water and for the fully compressible Euler equations validate the method and demonstrate its capability to achieve accurate results also at large Courant numbers, with time steps up to 100 times larger than those of typical explicit discretizations of the same problems, while reducing the computational cost thanks to the adaptivity algorithm.
no_new_dataset
0.710208
1405.5643
Martin Josef Geiger
Sandra Huber, Martin Josef Geiger, Marc Sevaux
Interactive Reference Point-Based Guided Local Search for the Bi-objective Inventory Routing Problem
null
Proceedings of the 10th Metaheuristics International Conference MIC 2013, August 5-8, 2013, Singapore, Pages 152-161
null
null
cs.AI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Eliciting preferences of a decision maker is a key factor to successfully combine search and decision making in an interactive method. Therefore, the progressively integration and simulation of the decision maker is a main concern in an application. We contribute in this direction by proposing an interactive method based on a reference point-based guided local search to the bi-objective Inventory Routing Problem. A local search metaheuristic, working on the delivery intervals, and the Clarke & Wright savings heuristic is employed for the subsequently obtained Vehicle Routing Problem. To elicit preferences, the decision maker selects a reference point to guide the search in interesting subregions. Additionally, the reference point is used as a reservation point to discard solutions outside the cone, introduced as a convergence criterion. Computational results of the reference point-based guided local search are reported and analyzed on benchmark data in order to show the applicability of the approach.
[ { "version": "v1", "created": "Thu, 22 May 2014 07:23:38 GMT" } ]
2014-05-23T00:00:00
[ [ "Huber", "Sandra", "" ], [ "Geiger", "Martin Josef", "" ], [ "Sevaux", "Marc", "" ] ]
TITLE: Interactive Reference Point-Based Guided Local Search for the Bi-objective Inventory Routing Problem ABSTRACT: Eliciting preferences of a decision maker is a key factor to successfully combine search and decision making in an interactive method. Therefore, the progressively integration and simulation of the decision maker is a main concern in an application. We contribute in this direction by proposing an interactive method based on a reference point-based guided local search to the bi-objective Inventory Routing Problem. A local search metaheuristic, working on the delivery intervals, and the Clarke & Wright savings heuristic is employed for the subsequently obtained Vehicle Routing Problem. To elicit preferences, the decision maker selects a reference point to guide the search in interesting subregions. Additionally, the reference point is used as a reservation point to discard solutions outside the cone, introduced as a convergence criterion. Computational results of the reference point-based guided local search are reported and analyzed on benchmark data in order to show the applicability of the approach.
no_new_dataset
0.710427
1405.5661
Yan Kit Li
Yan Kit Li, Min Xu, Chun Ho Ng, Patrick P. C. Lee
Efficient Hybrid Inline and Out-of-line Deduplication for Backup Storage
null
null
null
null
cs.DC cs.DB
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Backup storage systems often remove redundancy across backups via inline deduplication, which works by referring duplicate chunks of the latest backup to those of existing backups. However, inline deduplication degrades restore performance of the latest backup due to fragmentation, and complicates deletion of ex- pired backups due to the sharing of data chunks. While out-of-line deduplication addresses the problems by forward-pointing existing duplicate chunks to those of the latest backup, it introduces additional I/Os of writing and removing duplicate chunks. We design and implement RevDedup, an efficient hybrid inline and out-of-line deduplication system for backup storage. It applies coarse-grained inline deduplication to remove duplicates of the latest backup, and then fine-grained out-of-line reverse deduplication to remove duplicates from older backups. Our reverse deduplication design limits the I/O overhead and prepares for efficient deletion of expired backups. Through extensive testbed experiments using synthetic and real-world datasets, we show that RevDedup can bring high performance to the backup, restore, and deletion operations, while maintaining high storage efficiency comparable to conventional inline deduplication.
[ { "version": "v1", "created": "Thu, 22 May 2014 08:13:18 GMT" } ]
2014-05-23T00:00:00
[ [ "Li", "Yan Kit", "" ], [ "Xu", "Min", "" ], [ "Ng", "Chun Ho", "" ], [ "Lee", "Patrick P. C.", "" ] ]
TITLE: Efficient Hybrid Inline and Out-of-line Deduplication for Backup Storage ABSTRACT: Backup storage systems often remove redundancy across backups via inline deduplication, which works by referring duplicate chunks of the latest backup to those of existing backups. However, inline deduplication degrades restore performance of the latest backup due to fragmentation, and complicates deletion of ex- pired backups due to the sharing of data chunks. While out-of-line deduplication addresses the problems by forward-pointing existing duplicate chunks to those of the latest backup, it introduces additional I/Os of writing and removing duplicate chunks. We design and implement RevDedup, an efficient hybrid inline and out-of-line deduplication system for backup storage. It applies coarse-grained inline deduplication to remove duplicates of the latest backup, and then fine-grained out-of-line reverse deduplication to remove duplicates from older backups. Our reverse deduplication design limits the I/O overhead and prepares for efficient deletion of expired backups. Through extensive testbed experiments using synthetic and real-world datasets, we show that RevDedup can bring high performance to the backup, restore, and deletion operations, while maintaining high storage efficiency comparable to conventional inline deduplication.
no_new_dataset
0.707556
1405.5845
Ben Pringle
Ben Pringle, Mukkai Krishnamoorthy, Kenneth Simons
Case study to approaches to finding patterns in citation networks
16 pages, 6 figures
null
null
null
cs.DL cs.SI physics.soc-ph
http://creativecommons.org/licenses/by/3.0/
Analysis of a dataset including a network of LED patents and their metadata is carried out using several methods in order to answer questions about the domain. We are interested in finding the relationship between the metadata and the network structure; for example, are central patents in the network produced by larger or smaller companies? We begin by exploring the structure of the network without any metadata, applying known techniques in citation analysis and a simple clustering scheme. These techinques are then combined with metadata analysis to draw preliminary conclusions about the dataset.
[ { "version": "v1", "created": "Thu, 22 May 2014 18:21:51 GMT" } ]
2014-05-23T00:00:00
[ [ "Pringle", "Ben", "" ], [ "Krishnamoorthy", "Mukkai", "" ], [ "Simons", "Kenneth", "" ] ]
TITLE: Case study to approaches to finding patterns in citation networks ABSTRACT: Analysis of a dataset including a network of LED patents and their metadata is carried out using several methods in order to answer questions about the domain. We are interested in finding the relationship between the metadata and the network structure; for example, are central patents in the network produced by larger or smaller companies? We begin by exploring the structure of the network without any metadata, applying known techniques in citation analysis and a simple clustering scheme. These techinques are then combined with metadata analysis to draw preliminary conclusions about the dataset.
no_new_dataset
0.710245
1405.5869
Ping Li
Anshumali Shrivastava and Ping Li
Asymmetric LSH (ALSH) for Sublinear Time Maximum Inner Product Search (MIPS)
null
null
null
null
stat.ML cs.DS cs.IR cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We present the first provably sublinear time algorithm for approximate \emph{Maximum Inner Product Search} (MIPS). Our proposal is also the first hashing algorithm for searching with (un-normalized) inner product as the underlying similarity measure. Finding hashing schemes for MIPS was considered hard. We formally show that the existing Locality Sensitive Hashing (LSH) framework is insufficient for solving MIPS, and then we extend the existing LSH framework to allow asymmetric hashing schemes. Our proposal is based on an interesting mathematical phenomenon in which inner products, after independent asymmetric transformations, can be converted into the problem of approximate near neighbor search. This key observation makes efficient sublinear hashing scheme for MIPS possible. In the extended asymmetric LSH (ALSH) framework, we provide an explicit construction of provably fast hashing scheme for MIPS. The proposed construction and the extended LSH framework could be of independent theoretical interest. Our proposed algorithm is simple and easy to implement. We evaluate the method, for retrieving inner products, in the collaborative filtering task of item recommendations on Netflix and Movielens datasets.
[ { "version": "v1", "created": "Thu, 22 May 2014 19:42:57 GMT" } ]
2014-05-23T00:00:00
[ [ "Shrivastava", "Anshumali", "" ], [ "Li", "Ping", "" ] ]
TITLE: Asymmetric LSH (ALSH) for Sublinear Time Maximum Inner Product Search (MIPS) ABSTRACT: We present the first provably sublinear time algorithm for approximate \emph{Maximum Inner Product Search} (MIPS). Our proposal is also the first hashing algorithm for searching with (un-normalized) inner product as the underlying similarity measure. Finding hashing schemes for MIPS was considered hard. We formally show that the existing Locality Sensitive Hashing (LSH) framework is insufficient for solving MIPS, and then we extend the existing LSH framework to allow asymmetric hashing schemes. Our proposal is based on an interesting mathematical phenomenon in which inner products, after independent asymmetric transformations, can be converted into the problem of approximate near neighbor search. This key observation makes efficient sublinear hashing scheme for MIPS possible. In the extended asymmetric LSH (ALSH) framework, we provide an explicit construction of provably fast hashing scheme for MIPS. The proposed construction and the extended LSH framework could be of independent theoretical interest. Our proposed algorithm is simple and easy to implement. We evaluate the method, for retrieving inner products, in the collaborative filtering task of item recommendations on Netflix and Movielens datasets.
no_new_dataset
0.710377
1405.2641
Jyothi K
Jyothi K and Prabhakar C.J
Multi Modal Face Recognition Using Block Based Curvelet Features
17 pages, 5 Figures
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
In this paper, we present multimodal 2D +3D face recognition method using block based curvelet features. The 3D surface of face (Depth Map) is computed from the stereo face images using stereo vision technique. The statistical measures such as mean, standard deviation, variance and entropy are extracted from each block of curvelet subband for both depth and intensity images independently.In order to compute the decision score, the KNN classifier is employed independently for both intensity and depth map. Further, computed decision scoresof intensity and depth map are combined at decision level to improve the face recognition rate. The combination of intensity and depth map is verified experimentally using benchmark face database. The experimental results show that the proposed multimodal method is better than individual modality.
[ { "version": "v1", "created": "Mon, 12 May 2014 06:51:08 GMT" }, { "version": "v2", "created": "Wed, 21 May 2014 11:30:51 GMT" } ]
2014-05-22T00:00:00
[ [ "K", "Jyothi", "" ], [ "J", "Prabhakar C.", "" ] ]
TITLE: Multi Modal Face Recognition Using Block Based Curvelet Features ABSTRACT: In this paper, we present multimodal 2D +3D face recognition method using block based curvelet features. The 3D surface of face (Depth Map) is computed from the stereo face images using stereo vision technique. The statistical measures such as mean, standard deviation, variance and entropy are extracted from each block of curvelet subband for both depth and intensity images independently.In order to compute the decision score, the KNN classifier is employed independently for both intensity and depth map. Further, computed decision scoresof intensity and depth map are combined at decision level to improve the face recognition rate. The combination of intensity and depth map is verified experimentally using benchmark face database. The experimental results show that the proposed multimodal method is better than individual modality.
no_new_dataset
0.712676
1405.5422
Erik Cuevas E
Erik Cuevas, Daniel Zaldivar, Marco Perez, Edgar Sanchez and Marte Ramirez
Robust Fuzzy corner detector
15 Pages
Intelligent Automation and Soft Computing, 17 (4), (2011), pp. 415-429
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Reliable corner detection is an important task in determining the shape of different regions within an image. Real-life image data are always imprecise due to inherent uncertainties that may arise from the imaging process such as defocusing, illumination changes, noise, etc. Therefore, the localization and detection of corners has become a difficult task to accomplish under such imperfect situations. On the other hand, Fuzzy systems are well known for their efficient handling of impreciseness and incompleteness, which make them inherently suitable for modelling corner properties by means of a rule-based fuzzy system. The paper presents a corner detection algorithm which employs such fuzzy reasoning. The robustness of the proposed algorithm is compared to well-known conventional corner detectors and its performance is also tested over a number of benchmark images to illustrate the efficiency of the algorithm under uncertainty.
[ { "version": "v1", "created": "Wed, 21 May 2014 13:44:50 GMT" } ]
2014-05-22T00:00:00
[ [ "Cuevas", "Erik", "" ], [ "Zaldivar", "Daniel", "" ], [ "Perez", "Marco", "" ], [ "Sanchez", "Edgar", "" ], [ "Ramirez", "Marte", "" ] ]
TITLE: Robust Fuzzy corner detector ABSTRACT: Reliable corner detection is an important task in determining the shape of different regions within an image. Real-life image data are always imprecise due to inherent uncertainties that may arise from the imaging process such as defocusing, illumination changes, noise, etc. Therefore, the localization and detection of corners has become a difficult task to accomplish under such imperfect situations. On the other hand, Fuzzy systems are well known for their efficient handling of impreciseness and incompleteness, which make them inherently suitable for modelling corner properties by means of a rule-based fuzzy system. The paper presents a corner detection algorithm which employs such fuzzy reasoning. The robustness of the proposed algorithm is compared to well-known conventional corner detectors and its performance is also tested over a number of benchmark images to illustrate the efficiency of the algorithm under uncertainty.
no_new_dataset
0.711663
1405.5459
Adi Makmal
Alexey A. Melnikov, Adi Makmal, and Hans J. Briegel
Projective simulation applied to the grid-world and the mountain-car problem
null
null
null
null
cs.AI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We study the model of projective simulation (PS) which is a novel approach to artificial intelligence (AI). Recently it was shown that the PS agent performs well in a number of simple task environments, also when compared to standard models of reinforcement learning (RL). In this paper we study the performance of the PS agent further in more complicated scenarios. To that end we chose two well-studied benchmarking problems, namely the "grid-world" and the "mountain-car" problem, which challenge the model with large and continuous input space. We compare the performance of the PS agent model with those of existing models and show that the PS agent exhibits competitive performance also in such scenarios.
[ { "version": "v1", "created": "Wed, 21 May 2014 15:51:18 GMT" } ]
2014-05-22T00:00:00
[ [ "Melnikov", "Alexey A.", "" ], [ "Makmal", "Adi", "" ], [ "Briegel", "Hans J.", "" ] ]
TITLE: Projective simulation applied to the grid-world and the mountain-car problem ABSTRACT: We study the model of projective simulation (PS) which is a novel approach to artificial intelligence (AI). Recently it was shown that the PS agent performs well in a number of simple task environments, also when compared to standard models of reinforcement learning (RL). In this paper we study the performance of the PS agent further in more complicated scenarios. To that end we chose two well-studied benchmarking problems, namely the "grid-world" and the "mountain-car" problem, which challenge the model with large and continuous input space. We compare the performance of the PS agent model with those of existing models and show that the PS agent exhibits competitive performance also in such scenarios.
no_new_dataset
0.710396
1405.5461
Justin Kopinsky
Dan Alistarh, Justin Kopinsky, Alexander Matveev, Nir Shavit
The LevelArray: A Fast, Practical Long-Lived Renaming Algorithm
ICDCS 2014
null
null
null
cs.DC cs.DS
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The long-lived renaming problem appears in shared-memory systems where a set of threads need to register and deregister frequently from the computation, while concurrent operations scan the set of currently registered threads. Instances of this problem show up in concurrent implementations of transactional memory, flat combining, thread barriers, and memory reclamation schemes for lock-free data structures. In this paper, we analyze a randomized solution for long-lived renaming. The algorithmic technique we consider, called the LevelArray, has previously been used for hashing and one-shot (single-use) renaming. Our main contribu- tion is to prove that, in long-lived executions, where processes may register and deregister polynomially many times, the technique guarantees constant steps on average and O(log log n) steps with high probability for registering, unit cost for deregistering, and O(n) steps for collect queries, where n is an upper bound on the number of processes that may be active at any point in time. We also show that the algorithm has the surprising property that it is self-healing: under reasonable assumptions on the schedule, operations running while the data structure is in a degraded state implicitly help the data structure re-balance itself. This subtle mechanism obviates the need for expensive periodic rebuilding procedures. Our benchmarks validate this approach, showing that, for typical use parameters, the average number of steps a process takes to register is less than two and the worst-case number of steps is bounded by six, even in executions with billions of operations. We contrast this with other randomized implementations, whose worst-case behavior we show to be unreliable, and with deterministic implementations, whose cost is linear in n.
[ { "version": "v1", "created": "Wed, 21 May 2014 15:57:58 GMT" } ]
2014-05-22T00:00:00
[ [ "Alistarh", "Dan", "" ], [ "Kopinsky", "Justin", "" ], [ "Matveev", "Alexander", "" ], [ "Shavit", "Nir", "" ] ]
TITLE: The LevelArray: A Fast, Practical Long-Lived Renaming Algorithm ABSTRACT: The long-lived renaming problem appears in shared-memory systems where a set of threads need to register and deregister frequently from the computation, while concurrent operations scan the set of currently registered threads. Instances of this problem show up in concurrent implementations of transactional memory, flat combining, thread barriers, and memory reclamation schemes for lock-free data structures. In this paper, we analyze a randomized solution for long-lived renaming. The algorithmic technique we consider, called the LevelArray, has previously been used for hashing and one-shot (single-use) renaming. Our main contribu- tion is to prove that, in long-lived executions, where processes may register and deregister polynomially many times, the technique guarantees constant steps on average and O(log log n) steps with high probability for registering, unit cost for deregistering, and O(n) steps for collect queries, where n is an upper bound on the number of processes that may be active at any point in time. We also show that the algorithm has the surprising property that it is self-healing: under reasonable assumptions on the schedule, operations running while the data structure is in a degraded state implicitly help the data structure re-balance itself. This subtle mechanism obviates the need for expensive periodic rebuilding procedures. Our benchmarks validate this approach, showing that, for typical use parameters, the average number of steps a process takes to register is less than two and the worst-case number of steps is bounded by six, even in executions with billions of operations. We contrast this with other randomized implementations, whose worst-case behavior we show to be unreliable, and with deterministic implementations, whose cost is linear in n.
no_new_dataset
0.709982
1405.5488
Marc'Aurelio Ranzato
Marc'Aurelio Ranzato
On Learning Where To Look
deep learning, vision
null
null
null
cs.CV cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Current automatic vision systems face two major challenges: scalability and extreme variability of appearance. First, the computational time required to process an image typically scales linearly with the number of pixels in the image, therefore limiting the resolution of input images to thumbnail size. Second, variability in appearance and pose of the objects constitute a major hurdle for robust recognition and detection. In this work, we propose a model that makes baby steps towards addressing these challenges. We describe a learning based method that recognizes objects through a series of glimpses. This system performs an amount of computation that scales with the complexity of the input rather than its number of pixels. Moreover, the proposed method is potentially more robust to changes in appearance since its parameters are learned in a data driven manner. Preliminary experiments on a handwritten dataset of digits demonstrate the computational advantages of this approach.
[ { "version": "v1", "created": "Thu, 24 Apr 2014 02:29:19 GMT" } ]
2014-05-22T00:00:00
[ [ "Ranzato", "Marc'Aurelio", "" ] ]
TITLE: On Learning Where To Look ABSTRACT: Current automatic vision systems face two major challenges: scalability and extreme variability of appearance. First, the computational time required to process an image typically scales linearly with the number of pixels in the image, therefore limiting the resolution of input images to thumbnail size. Second, variability in appearance and pose of the objects constitute a major hurdle for robust recognition and detection. In this work, we propose a model that makes baby steps towards addressing these challenges. We describe a learning based method that recognizes objects through a series of glimpses. This system performs an amount of computation that scales with the complexity of the input rather than its number of pixels. Moreover, the proposed method is potentially more robust to changes in appearance since its parameters are learned in a data driven manner. Preliminary experiments on a handwritten dataset of digits demonstrate the computational advantages of this approach.
no_new_dataset
0.710208
1301.0967
Xianyi Zeng
Xianyi Zeng
A general approach to enhance slope limiters on non-uniform rectilinear grids
25 pages, 31 figures
null
null
null
math.NA cs.NA
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Most slope limiter functions in high-resolution finite volume methods to solve hyperbolic conservation laws are designed assuming one-dimensional uniform grids, and they are also used to compute slope limiters in computations on non-uniform rectilinear grids. However, this strategy may lead to either loss of total variation diminishing (TVD) stability for 1D linear problems or the loss of formal second-order accuracy if the grid is highly non-uniform. This is especially true when the limiter function is not piecewise linear. Numerical evidences are provided to support this argument for two popular finite volume strategies: MUSCL in space and method of lines in time (MUSCL-MOL), and capacity-form differencing. In order to deal with this issue, this paper presents a general approach to study and enhance the slope limiter functions for highly non-uniform grids in the MUSCL-MOL framework. This approach extends the classical reconstruct-evolve-project procedure to general grids, and it gives sufficient conditions for a slope limiter function leading to a TVD stable, formal second-order accuracy in space, and symmetry preserving numerical scheme on arbitrary grids. Several widely used limiter functions, including the smooth ones by van Leer and van Albada, are enhanced to satisfy these conditions. These properties are confirmed by solving various one-dimensional and two-dimensional benchmark problems using the enhanced limiters on highly non-uniform rectilinear grids.
[ { "version": "v1", "created": "Sun, 6 Jan 2013 03:12:38 GMT" }, { "version": "v2", "created": "Tue, 20 May 2014 03:27:09 GMT" } ]
2014-05-21T00:00:00
[ [ "Zeng", "Xianyi", "" ] ]
TITLE: A general approach to enhance slope limiters on non-uniform rectilinear grids ABSTRACT: Most slope limiter functions in high-resolution finite volume methods to solve hyperbolic conservation laws are designed assuming one-dimensional uniform grids, and they are also used to compute slope limiters in computations on non-uniform rectilinear grids. However, this strategy may lead to either loss of total variation diminishing (TVD) stability for 1D linear problems or the loss of formal second-order accuracy if the grid is highly non-uniform. This is especially true when the limiter function is not piecewise linear. Numerical evidences are provided to support this argument for two popular finite volume strategies: MUSCL in space and method of lines in time (MUSCL-MOL), and capacity-form differencing. In order to deal with this issue, this paper presents a general approach to study and enhance the slope limiter functions for highly non-uniform grids in the MUSCL-MOL framework. This approach extends the classical reconstruct-evolve-project procedure to general grids, and it gives sufficient conditions for a slope limiter function leading to a TVD stable, formal second-order accuracy in space, and symmetry preserving numerical scheme on arbitrary grids. Several widely used limiter functions, including the smooth ones by van Leer and van Albada, are enhanced to satisfy these conditions. These properties are confirmed by solving various one-dimensional and two-dimensional benchmark problems using the enhanced limiters on highly non-uniform rectilinear grids.
no_new_dataset
0.710245
1405.4979
Razen Al-Harbi
Razen Al-Harbi, Yasser Ebrahim, Panos Kalnis
PHD-Store: An Adaptive SPARQL Engine with Dynamic Partitioning for Distributed RDF Repositories
null
null
null
null
cs.DB
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Many repositories utilize the versatile RDF model to publish data. Repositories are typically distributed and geographically remote, but data are interconnected (e.g., the Semantic Web) and queried globally by a language such as SPARQL. Due to the network cost and the nature of the queries, the execution time can be prohibitively high. Current solutions attempt to minimize the network cost by redistributing all data in a preprocessing phase, but here are two drawbacks: (i) redistribution is based on heuristics that may not benefit many of the future queries; and (ii) the preprocessing phase is very expensive even for moderate size datasets. In this paper we propose PHD-Store, a SPARQL engine for distributed RDF repositories. Our system does not assume any particular initial data placement and does not require prepartitioning; hence, it minimizes the startup cost. Initially, PHD-Store answers queries using a potentially slow distributed semi-join algorithm, but adapts dynamically to the query load by incrementally redistributing frequently accessed data. Redistribution is done in a way that future queries can benefit from fast hash-based parallel execution. Our experiments with synthetic and real data verify that PHD-Store scales to very large datasets; many repositories; converges to comparable or better quality of partitioning than existing methods; and executes large query loads 1 to 2 orders of magnitude faster than our competitors.
[ { "version": "v1", "created": "Tue, 20 May 2014 07:44:03 GMT" } ]
2014-05-21T00:00:00
[ [ "Al-Harbi", "Razen", "" ], [ "Ebrahim", "Yasser", "" ], [ "Kalnis", "Panos", "" ] ]
TITLE: PHD-Store: An Adaptive SPARQL Engine with Dynamic Partitioning for Distributed RDF Repositories ABSTRACT: Many repositories utilize the versatile RDF model to publish data. Repositories are typically distributed and geographically remote, but data are interconnected (e.g., the Semantic Web) and queried globally by a language such as SPARQL. Due to the network cost and the nature of the queries, the execution time can be prohibitively high. Current solutions attempt to minimize the network cost by redistributing all data in a preprocessing phase, but here are two drawbacks: (i) redistribution is based on heuristics that may not benefit many of the future queries; and (ii) the preprocessing phase is very expensive even for moderate size datasets. In this paper we propose PHD-Store, a SPARQL engine for distributed RDF repositories. Our system does not assume any particular initial data placement and does not require prepartitioning; hence, it minimizes the startup cost. Initially, PHD-Store answers queries using a potentially slow distributed semi-join algorithm, but adapts dynamically to the query load by incrementally redistributing frequently accessed data. Redistribution is done in a way that future queries can benefit from fast hash-based parallel execution. Our experiments with synthetic and real data verify that PHD-Store scales to very large datasets; many repositories; converges to comparable or better quality of partitioning than existing methods; and executes large query loads 1 to 2 orders of magnitude faster than our competitors.
no_new_dataset
0.706349
1405.5048
Mihai Polceanu M.Sc.
Mihai Polceanu, C\'edric Buche
Towards A Theory-Of-Mind-Inspired Generic Decision-Making Framework
7 pages, 5 figures, IJCAI 2013 Symposium on AI in Angry Birds
null
null
null
cs.AI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Simulation is widely used to make model-based predictions, but few approaches have attempted this technique in dynamic physical environments of medium to high complexity or in general contexts. After an introduction to the cognitive science concepts from which this work is inspired and the current development in the use of simulation as a decision-making technique, we propose a generic framework based on theory of mind, which allows an agent to reason and perform actions using multiple simulations of automatically created or externally inputted models of the perceived environment. A description of a partial implementation is given, which aims to solve a popular game within the IJCAI2013 AIBirds contest. Results of our approach are presented, in comparison with the competition benchmark. Finally, future developments regarding the framework are discussed.
[ { "version": "v1", "created": "Tue, 20 May 2014 11:54:21 GMT" } ]
2014-05-21T00:00:00
[ [ "Polceanu", "Mihai", "" ], [ "Buche", "Cédric", "" ] ]
TITLE: Towards A Theory-Of-Mind-Inspired Generic Decision-Making Framework ABSTRACT: Simulation is widely used to make model-based predictions, but few approaches have attempted this technique in dynamic physical environments of medium to high complexity or in general contexts. After an introduction to the cognitive science concepts from which this work is inspired and the current development in the use of simulation as a decision-making technique, we propose a generic framework based on theory of mind, which allows an agent to reason and perform actions using multiple simulations of automatically created or externally inputted models of the perceived environment. A description of a partial implementation is given, which aims to solve a popular game within the IJCAI2013 AIBirds contest. Results of our approach are presented, in comparison with the competition benchmark. Finally, future developments regarding the framework are discussed.
no_new_dataset
0.710051
1405.5097
Junzhou Zhao
Junzhou Zhao, John C.S. Lui, Don Towsley, Pinghui Wang, and Xiaohong Guan
Design of Efficient Sampling Methods on Hybrid Social-Affiliation Networks
11 pages, 13 figures, technique report
null
null
null
cs.SI physics.soc-ph
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Graph sampling via crawling has become increasingly popular and important in the study of measuring various characteristics of large scale complex networks. While powerful, it is known to be challenging when the graph is loosely connected or disconnected which slows down the convergence of random walks and can cause poor estimation accuracy. In this work, we observe that the graph under study, or called target graph, usually does not exist in isolation. In many situations, the target graph is related to an auxiliary graph and an affiliation graph, and the target graph becomes well connected when we view it from the perspective of these three graphs together, or called a hybrid social-affiliation graph in this paper. When directly sampling the target graph is difficult or inefficient, we can indirectly sample it efficiently with the assistances of the other two graphs. We design three sampling methods on such a hybrid social-affiliation network. Experiments conducted on both synthetic and real datasets demonstrate the effectiveness of our proposed methods.
[ { "version": "v1", "created": "Tue, 20 May 2014 14:17:19 GMT" } ]
2014-05-21T00:00:00
[ [ "Zhao", "Junzhou", "" ], [ "Lui", "John C. S.", "" ], [ "Towsley", "Don", "" ], [ "Wang", "Pinghui", "" ], [ "Guan", "Xiaohong", "" ] ]
TITLE: Design of Efficient Sampling Methods on Hybrid Social-Affiliation Networks ABSTRACT: Graph sampling via crawling has become increasingly popular and important in the study of measuring various characteristics of large scale complex networks. While powerful, it is known to be challenging when the graph is loosely connected or disconnected which slows down the convergence of random walks and can cause poor estimation accuracy. In this work, we observe that the graph under study, or called target graph, usually does not exist in isolation. In many situations, the target graph is related to an auxiliary graph and an affiliation graph, and the target graph becomes well connected when we view it from the perspective of these three graphs together, or called a hybrid social-affiliation graph in this paper. When directly sampling the target graph is difficult or inefficient, we can indirectly sample it efficiently with the assistances of the other two graphs. We design three sampling methods on such a hybrid social-affiliation network. Experiments conducted on both synthetic and real datasets demonstrate the effectiveness of our proposed methods.
no_new_dataset
0.71039
1405.5172
Erik Cuevas E
Erik Cuevas, Diego Oliva, Daniel Zaldivar, Marco Perez and Gonzalo Pajares
Opposition Based ElectromagnetismLike for Global Optimization
27 Pages
International Journal of Innovative Computing, Information and Control, 8 (12) , (2012), pp. 8181-8198
null
null
cs.AI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Electromagnetismlike Optimization (EMO) is a global optimization algorithm, particularly well suited to solve problems featuring nonlinear and multimodal cost functions. EMO employs searcher agents that emulate a population of charged particles which interact to each other according to electromagnetisms laws of attraction and repulsion. However, EMO usually requires a large number of iterations for a local search procedure; any reduction or cancelling over such number, critically perturb other issues such as convergence, exploration, population diversity and accuracy. This paper presents an enhanced EMO algorithm called OBEMO, which employs the Opposition-Based Learning (OBL) approach to accelerate the global convergence speed. OBL is a machine intelligence strategy which considers the current candidate solution and its opposite value at the same time, achieving a faster exploration of the search space. The proposed OBEMO method significantly reduces the required computational effort yet avoiding any detriment to the good search capabilities of the original EMO algorithm. Experiments are conducted over a comprehensive set of benchmark functions, showing that OBEMO obtains promising performance for most of the discussed test problems.
[ { "version": "v1", "created": "Tue, 20 May 2014 17:52:57 GMT" } ]
2014-05-21T00:00:00
[ [ "Cuevas", "Erik", "" ], [ "Oliva", "Diego", "" ], [ "Zaldivar", "Daniel", "" ], [ "Perez", "Marco", "" ], [ "Pajares", "Gonzalo", "" ] ]
TITLE: Opposition Based ElectromagnetismLike for Global Optimization ABSTRACT: Electromagnetismlike Optimization (EMO) is a global optimization algorithm, particularly well suited to solve problems featuring nonlinear and multimodal cost functions. EMO employs searcher agents that emulate a population of charged particles which interact to each other according to electromagnetisms laws of attraction and repulsion. However, EMO usually requires a large number of iterations for a local search procedure; any reduction or cancelling over such number, critically perturb other issues such as convergence, exploration, population diversity and accuracy. This paper presents an enhanced EMO algorithm called OBEMO, which employs the Opposition-Based Learning (OBL) approach to accelerate the global convergence speed. OBL is a machine intelligence strategy which considers the current candidate solution and its opposite value at the same time, achieving a faster exploration of the search space. The proposed OBEMO method significantly reduces the required computational effort yet avoiding any detriment to the good search capabilities of the original EMO algorithm. Experiments are conducted over a comprehensive set of benchmark functions, showing that OBEMO obtains promising performance for most of the discussed test problems.
no_new_dataset
0.71103
1206.6214
Stefan Hennemann
Stefan Hennemann and Ben Derudder
An Alternative Approach to the Calculation and Analysis of Connectivity in the World City Network
18 pages, 9 figures, 2 tables
Environment and Planning B: Planning and Design 41(3) 392-412
10.1068/b39108
null
physics.soc-ph cs.SI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Empirical research on world cities often draws on Taylor's (2001) notion of an 'interlocking network model', in which office networks of globalized service firms are assumed to shape the spatialities of urban networks. In spite of its many merits, this approach is limited because the resultant adjacency matrices are not really fit for network-analytic calculations. We therefore propose a fresh analytical approach using a primary linkage algorithm that produces a one-mode directed graph based on Taylor's two-mode city/firm network data. The procedure has the advantage of creating less dense networks when compared to the interlocking network model, while nonetheless retaining the network structure apparent in the initial dataset. We randomize the empirical network with a bootstrapping simulation approach, and compare the simulated parameters of this null-model with our empirical network parameter (i.e. betweenness centrality). We find that our approach produces results that are comparable to those of the standard interlocking network model. However, because our approach is based on an actual graph representation and network analysis, we are able to assess cities' position in the network at large. For instance, we find that cities such as Tokyo, Sydney, Melbourne, Almaty and Karachi hold more strategic and valuable positions than suggested in the interlocking networks as they play a bridging role in connecting cities across regions. In general, we argue that our graph representation allows for further and deeper analysis of the original data, further extending world city network research into a theory-based empirical research approach.
[ { "version": "v1", "created": "Wed, 27 Jun 2012 09:33:33 GMT" } ]
2014-05-20T00:00:00
[ [ "Hennemann", "Stefan", "" ], [ "Derudder", "Ben", "" ] ]
TITLE: An Alternative Approach to the Calculation and Analysis of Connectivity in the World City Network ABSTRACT: Empirical research on world cities often draws on Taylor's (2001) notion of an 'interlocking network model', in which office networks of globalized service firms are assumed to shape the spatialities of urban networks. In spite of its many merits, this approach is limited because the resultant adjacency matrices are not really fit for network-analytic calculations. We therefore propose a fresh analytical approach using a primary linkage algorithm that produces a one-mode directed graph based on Taylor's two-mode city/firm network data. The procedure has the advantage of creating less dense networks when compared to the interlocking network model, while nonetheless retaining the network structure apparent in the initial dataset. We randomize the empirical network with a bootstrapping simulation approach, and compare the simulated parameters of this null-model with our empirical network parameter (i.e. betweenness centrality). We find that our approach produces results that are comparable to those of the standard interlocking network model. However, because our approach is based on an actual graph representation and network analysis, we are able to assess cities' position in the network at large. For instance, we find that cities such as Tokyo, Sydney, Melbourne, Almaty and Karachi hold more strategic and valuable positions than suggested in the interlocking networks as they play a bridging role in connecting cities across regions. In general, we argue that our graph representation allows for further and deeper analysis of the original data, further extending world city network research into a theory-based empirical research approach.
no_new_dataset
0.712026
1402.4963
Julius Hannink
Julius Hannink, Remco Duits and Erik Bekkers
Vesselness via Multiple Scale Orientation Scores
9 pages, 8 figures
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The multi-scale Frangi vesselness filter is an established tool in (retinal) vascular imaging. However, it cannot cope with crossings or bifurcations, since it only looks for elongated structures. Therefore, we disentangle crossing structures in the image via (multiple scale) invertible orientation scores. The described vesselness filter via scale-orientation scores performs considerably better at enhancing vessels throughout crossings and bifurcations than the Frangi version. Both methods are evaluated on a public dataset. Performance is measured by comparing ground truth data to the segmentation results obtained by basic thresholding and morphological component analysis of the filtered images.
[ { "version": "v1", "created": "Thu, 20 Feb 2014 11:06:35 GMT" }, { "version": "v2", "created": "Thu, 27 Feb 2014 18:30:55 GMT" }, { "version": "v3", "created": "Tue, 4 Mar 2014 12:33:37 GMT" }, { "version": "v4", "created": "Mon, 19 May 2014 09:20:06 GMT" } ]
2014-05-20T00:00:00
[ [ "Hannink", "Julius", "" ], [ "Duits", "Remco", "" ], [ "Bekkers", "Erik", "" ] ]
TITLE: Vesselness via Multiple Scale Orientation Scores ABSTRACT: The multi-scale Frangi vesselness filter is an established tool in (retinal) vascular imaging. However, it cannot cope with crossings or bifurcations, since it only looks for elongated structures. Therefore, we disentangle crossing structures in the image via (multiple scale) invertible orientation scores. The described vesselness filter via scale-orientation scores performs considerably better at enhancing vessels throughout crossings and bifurcations than the Frangi version. Both methods are evaluated on a public dataset. Performance is measured by comparing ground truth data to the segmentation results obtained by basic thresholding and morphological component analysis of the filtered images.
no_new_dataset
0.714061
1405.0386
Michael Huth
Michael Huth, Jim Huan-Pu Kuo, Nir Piterman
Fatal Attractors in Parity Games: Building Blocks for Partial Solvers
null
null
null
null
cs.LO
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Attractors in parity games are a technical device for solving "alternating" reachability of given node sets. A well known solver of parity games - Zielonka's algorithm - uses such attractor computations recursively. We here propose new forms of attractors that are monotone in that they are aware of specific static patterns of colors encountered in reaching a given node set in alternating fashion. Then we demonstrate how these new forms of attractors can be embedded within greatest fixed-point computations to design solvers of parity games that run in polynomial time but are partial in that they may not decide the winning status of all nodes in the input game. Experimental results show that our partial solvers completely solve benchmarks that were constructed to challenge existing full solvers. Our partial solvers also have encouraging run times in practice. For one partial solver we prove that its run-time is at most cubic in the number of nodes in the parity game, that its output game is independent of the order in which monotone attractors are computed, and that it solves all Buechi games and weak games. We then define and study a transformation that converts partial solvers into more precise partial solvers, and we prove that this transformation is sound under very reasonable conditions on the input partial solvers. Noting that one of our partial solvers meets these conditions, we apply its transformation on 1.6 million randomly generated games and so experimentally validate that the transformation can be very effective in increasing the precision of partial solvers.
[ { "version": "v1", "created": "Fri, 2 May 2014 11:45:18 GMT" }, { "version": "v2", "created": "Mon, 19 May 2014 16:59:12 GMT" } ]
2014-05-20T00:00:00
[ [ "Huth", "Michael", "" ], [ "Kuo", "Jim Huan-Pu", "" ], [ "Piterman", "Nir", "" ] ]
TITLE: Fatal Attractors in Parity Games: Building Blocks for Partial Solvers ABSTRACT: Attractors in parity games are a technical device for solving "alternating" reachability of given node sets. A well known solver of parity games - Zielonka's algorithm - uses such attractor computations recursively. We here propose new forms of attractors that are monotone in that they are aware of specific static patterns of colors encountered in reaching a given node set in alternating fashion. Then we demonstrate how these new forms of attractors can be embedded within greatest fixed-point computations to design solvers of parity games that run in polynomial time but are partial in that they may not decide the winning status of all nodes in the input game. Experimental results show that our partial solvers completely solve benchmarks that were constructed to challenge existing full solvers. Our partial solvers also have encouraging run times in practice. For one partial solver we prove that its run-time is at most cubic in the number of nodes in the parity game, that its output game is independent of the order in which monotone attractors are computed, and that it solves all Buechi games and weak games. We then define and study a transformation that converts partial solvers into more precise partial solvers, and we prove that this transformation is sound under very reasonable conditions on the input partial solvers. Noting that one of our partial solvers meets these conditions, we apply its transformation on 1.6 million randomly generated games and so experimentally validate that the transformation can be very effective in increasing the precision of partial solvers.
no_new_dataset
0.706551
1405.4301
Stanislav Sobolevsky
Stanislav Sobolevsky, Izabela Sitko, Sebastian Grauwin, Remi Tachet des Combes, Bartosz Hawelka, Juan Murillo Arias, Carlo Ratti
Mining Urban Performance: Scale-Independent Classification of Cities Based on Individual Economic Transactions
10 pages, 7 figures, to be published in the proceedings of ASE BigDataScience 2014 conference
null
null
null
physics.soc-ph cs.SI q-fin.GN
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Intensive development of urban systems creates a number of challenges for urban planners and policy makers in order to maintain sustainable growth. Running efficient urban policies requires meaningful urban metrics, which could quantify important urban characteristics including various aspects of an actual human behavior. Since a city size is known to have a major, yet often nonlinear, impact on the human activity, it also becomes important to develop scale-free metrics that capture qualitative city properties, beyond the effects of scale. Recent availability of extensive datasets created by human activity involving digital technologies creates new opportunities in this area. In this paper we propose a novel approach of city scoring and classification based on quantitative scale-free metrics related to economic activity of city residents, as well as domestic and foreign visitors. It is demonstrated on the example of Spain, but the proposed methodology is of a general character. We employ a new source of large-scale ubiquitous data, which consists of anonymized countrywide records of bank card transactions collected by one of the largest Spanish banks. Different aspects of the classification reveal important properties of Spanish cities, which significantly complement the pattern that might be discovered with the official socioeconomic statistics.
[ { "version": "v1", "created": "Fri, 16 May 2014 20:36:08 GMT" } ]
2014-05-20T00:00:00
[ [ "Sobolevsky", "Stanislav", "" ], [ "Sitko", "Izabela", "" ], [ "Grauwin", "Sebastian", "" ], [ "Combes", "Remi Tachet des", "" ], [ "Hawelka", "Bartosz", "" ], [ "Arias", "Juan Murillo", "" ], [ "Ratti", "Carlo", "" ] ]
TITLE: Mining Urban Performance: Scale-Independent Classification of Cities Based on Individual Economic Transactions ABSTRACT: Intensive development of urban systems creates a number of challenges for urban planners and policy makers in order to maintain sustainable growth. Running efficient urban policies requires meaningful urban metrics, which could quantify important urban characteristics including various aspects of an actual human behavior. Since a city size is known to have a major, yet often nonlinear, impact on the human activity, it also becomes important to develop scale-free metrics that capture qualitative city properties, beyond the effects of scale. Recent availability of extensive datasets created by human activity involving digital technologies creates new opportunities in this area. In this paper we propose a novel approach of city scoring and classification based on quantitative scale-free metrics related to economic activity of city residents, as well as domestic and foreign visitors. It is demonstrated on the example of Spain, but the proposed methodology is of a general character. We employ a new source of large-scale ubiquitous data, which consists of anonymized countrywide records of bank card transactions collected by one of the largest Spanish banks. Different aspects of the classification reveal important properties of Spanish cities, which significantly complement the pattern that might be discovered with the official socioeconomic statistics.
no_new_dataset
0.708042
1405.4308
Le Lu
Meizhu Liu, Le Lu, Xiaojing Ye, Shipeng Yu
Coarse-to-Fine Classification via Parametric and Nonparametric Models for Computer-Aided Diagnosis
null
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Classification is one of the core problems in Computer-Aided Diagnosis (CAD), targeting for early cancer detection using 3D medical imaging interpretation. High detection sensitivity with desirably low false positive (FP) rate is critical for a CAD system to be accepted as a valuable or even indispensable tool in radiologists' workflow. Given various spurious imagery noises which cause observation uncertainties, this remains a very challenging task. In this paper, we propose a novel, two-tiered coarse-to-fine (CTF) classification cascade framework to tackle this problem. We first obtain classification-critical data samples (e.g., samples on the decision boundary) extracted from the holistic data distributions using a robust parametric model (e.g., \cite{Raykar08}); then we build a graph-embedding based nonparametric classifier on sampled data, which can more accurately preserve or formulate the complex classification boundary. These two steps can also be considered as effective "sample pruning" and "feature pursuing + $k$NN/template matching", respectively. Our approach is validated comprehensively in colorectal polyp detection and lung nodule detection CAD systems, as the top two deadly cancers, using hospital scale, multi-site clinical datasets. The results show that our method achieves overall better classification/detection performance than existing state-of-the-art algorithms using single-layer classifiers, such as the support vector machine variants \cite{Wang08}, boosting \cite{Slabaugh10}, logistic regression \cite{Ravesteijn10}, relevance vector machine \cite{Raykar08}, $k$-nearest neighbor \cite{Murphy09} or spectral projections on graph \cite{Cai08}.
[ { "version": "v1", "created": "Fri, 16 May 2014 21:13:01 GMT" } ]
2014-05-20T00:00:00
[ [ "Liu", "Meizhu", "" ], [ "Lu", "Le", "" ], [ "Ye", "Xiaojing", "" ], [ "Yu", "Shipeng", "" ] ]
TITLE: Coarse-to-Fine Classification via Parametric and Nonparametric Models for Computer-Aided Diagnosis ABSTRACT: Classification is one of the core problems in Computer-Aided Diagnosis (CAD), targeting for early cancer detection using 3D medical imaging interpretation. High detection sensitivity with desirably low false positive (FP) rate is critical for a CAD system to be accepted as a valuable or even indispensable tool in radiologists' workflow. Given various spurious imagery noises which cause observation uncertainties, this remains a very challenging task. In this paper, we propose a novel, two-tiered coarse-to-fine (CTF) classification cascade framework to tackle this problem. We first obtain classification-critical data samples (e.g., samples on the decision boundary) extracted from the holistic data distributions using a robust parametric model (e.g., \cite{Raykar08}); then we build a graph-embedding based nonparametric classifier on sampled data, which can more accurately preserve or formulate the complex classification boundary. These two steps can also be considered as effective "sample pruning" and "feature pursuing + $k$NN/template matching", respectively. Our approach is validated comprehensively in colorectal polyp detection and lung nodule detection CAD systems, as the top two deadly cancers, using hospital scale, multi-site clinical datasets. The results show that our method achieves overall better classification/detection performance than existing state-of-the-art algorithms using single-layer classifiers, such as the support vector machine variants \cite{Wang08}, boosting \cite{Slabaugh10}, logistic regression \cite{Ravesteijn10}, relevance vector machine \cite{Raykar08}, $k$-nearest neighbor \cite{Murphy09} or spectral projections on graph \cite{Cai08}.
no_new_dataset
0.710666
1405.4506
Limin Wang
Xiaojiang Peng and Limin Wang and Xingxing Wang and Yu Qiao
Bag of Visual Words and Fusion Methods for Action Recognition: Comprehensive Study and Good Practice
null
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Video based action recognition is one of the important and challenging problems in computer vision research. Bag of Visual Words model (BoVW) with local features has become the most popular method and obtained the state-of-the-art performance on several realistic datasets, such as the HMDB51, UCF50, and UCF101. BoVW is a general pipeline to construct a global representation from a set of local features, which is mainly composed of five steps: (i) feature extraction, (ii) feature pre-processing, (iii) codebook generation, (iv) feature encoding, and (v) pooling and normalization. Many efforts have been made in each step independently in different scenarios and their effect on action recognition is still unknown. Meanwhile, video data exhibits different views of visual pattern, such as static appearance and motion dynamics. Multiple descriptors are usually extracted to represent these different views. Many feature fusion methods have been developed in other areas and their influence on action recognition has never been investigated before. This paper aims to provide a comprehensive study of all steps in BoVW and different fusion methods, and uncover some good practice to produce a state-of-the-art action recognition system. Specifically, we explore two kinds of local features, ten kinds of encoding methods, eight kinds of pooling and normalization strategies, and three kinds of fusion methods. We conclude that every step is crucial for contributing to the final recognition rate. Furthermore, based on our comprehensive study, we propose a simple yet effective representation, called hybrid representation, by exploring the complementarity of different BoVW frameworks and local descriptors. Using this representation, we obtain the state-of-the-art on the three challenging datasets: HMDB51 (61.1%), UCF50 (92.3%), and UCF101 (87.9%).
[ { "version": "v1", "created": "Sun, 18 May 2014 13:56:07 GMT" } ]
2014-05-20T00:00:00
[ [ "Peng", "Xiaojiang", "" ], [ "Wang", "Limin", "" ], [ "Wang", "Xingxing", "" ], [ "Qiao", "Yu", "" ] ]
TITLE: Bag of Visual Words and Fusion Methods for Action Recognition: Comprehensive Study and Good Practice ABSTRACT: Video based action recognition is one of the important and challenging problems in computer vision research. Bag of Visual Words model (BoVW) with local features has become the most popular method and obtained the state-of-the-art performance on several realistic datasets, such as the HMDB51, UCF50, and UCF101. BoVW is a general pipeline to construct a global representation from a set of local features, which is mainly composed of five steps: (i) feature extraction, (ii) feature pre-processing, (iii) codebook generation, (iv) feature encoding, and (v) pooling and normalization. Many efforts have been made in each step independently in different scenarios and their effect on action recognition is still unknown. Meanwhile, video data exhibits different views of visual pattern, such as static appearance and motion dynamics. Multiple descriptors are usually extracted to represent these different views. Many feature fusion methods have been developed in other areas and their influence on action recognition has never been investigated before. This paper aims to provide a comprehensive study of all steps in BoVW and different fusion methods, and uncover some good practice to produce a state-of-the-art action recognition system. Specifically, we explore two kinds of local features, ten kinds of encoding methods, eight kinds of pooling and normalization strategies, and three kinds of fusion methods. We conclude that every step is crucial for contributing to the final recognition rate. Furthermore, based on our comprehensive study, we propose a simple yet effective representation, called hybrid representation, by exploring the complementarity of different BoVW frameworks and local descriptors. Using this representation, we obtain the state-of-the-art on the three challenging datasets: HMDB51 (61.1%), UCF50 (92.3%), and UCF101 (87.9%).
no_new_dataset
0.708042
1405.4510
Tao Ye
Tao Ye and Kan Zhou and Zhipeng Lu and Jin-Kao Hao
A Memetic Algorithm for the Linear Ordering Problem with Cumulative Costs
null
null
null
null
cs.NE
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
This paper introduces an effective memetic algorithm for the linear ordering problem with cumulative costs. The proposed algorithm combines an order-based recombination operator with an improved forward-backward local search procedure and employs a solution quality based replacement criterion for pool updating. Extensive experiments on 118 well-known benchmark instances show that the proposed algorithm achieves competitive results by identifying 46 new upper bounds. Furthermore, some critical ingredients of our algorithm are analyzed to understand the source of its performance.
[ { "version": "v1", "created": "Sun, 18 May 2014 14:25:56 GMT" } ]
2014-05-20T00:00:00
[ [ "Ye", "Tao", "" ], [ "Zhou", "Kan", "" ], [ "Lu", "Zhipeng", "" ], [ "Hao", "Jin-Kao", "" ] ]
TITLE: A Memetic Algorithm for the Linear Ordering Problem with Cumulative Costs ABSTRACT: This paper introduces an effective memetic algorithm for the linear ordering problem with cumulative costs. The proposed algorithm combines an order-based recombination operator with an improved forward-backward local search procedure and employs a solution quality based replacement criterion for pool updating. Extensive experiments on 118 well-known benchmark instances show that the proposed algorithm achieves competitive results by identifying 46 new upper bounds. Furthermore, some critical ingredients of our algorithm are analyzed to understand the source of its performance.
no_new_dataset
0.708629
1405.4543
Dhruv Mahajan
Dhruv Mahajan, S. Sathiya Keerthi, S. Sundararajan
A Distributed Algorithm for Training Nonlinear Kernel Machines
null
null
null
null
cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
This paper concerns the distributed training of nonlinear kernel machines on Map-Reduce. We show that a re-formulation of Nystr\"om approximation based solution which is solved using gradient based techniques is well suited for this, especially when it is necessary to work with a large number of basis points. The main advantages of this approach are: avoidance of computing the pseudo-inverse of the kernel sub-matrix corresponding to the basis points; simplicity and efficiency of the distributed part of the computations; and, friendliness to stage-wise addition of basis points. We implement the method using an AllReduce tree on Hadoop and demonstrate its value on a few large benchmark datasets.
[ { "version": "v1", "created": "Sun, 18 May 2014 19:54:18 GMT" } ]
2014-05-20T00:00:00
[ [ "Mahajan", "Dhruv", "" ], [ "Keerthi", "S. Sathiya", "" ], [ "Sundararajan", "S.", "" ] ]
TITLE: A Distributed Algorithm for Training Nonlinear Kernel Machines ABSTRACT: This paper concerns the distributed training of nonlinear kernel machines on Map-Reduce. We show that a re-formulation of Nystr\"om approximation based solution which is solved using gradient based techniques is well suited for this, especially when it is necessary to work with a large number of basis points. The main advantages of this approach are: avoidance of computing the pseudo-inverse of the kernel sub-matrix corresponding to the basis points; simplicity and efficiency of the distributed part of the computations; and, friendliness to stage-wise addition of basis points. We implement the method using an AllReduce tree on Hadoop and demonstrate its value on a few large benchmark datasets.
no_new_dataset
0.709988
1405.4572
R. Joshua Tobin
R. Joshua Tobin and Conor J. Houghton
A Kernel-Based Calculation of Information on a Metric Space
null
Entropy 2013, 15(10), 4540-4552
10.3390/e15104540
null
cs.IT math.IT
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Kernel density estimation is a technique for approximating probability distributions. Here, it is applied to the calculation of mutual information on a metric space. This is motivated by the problem in neuroscience of calculating the mutual information between stimuli and spiking responses; the space of these responses is a metric space. It is shown that kernel density estimation on a metric space resembles the k-nearest-neighbor approach. This approach is applied to a toy dataset designed to mimic electrophysiological data.
[ { "version": "v1", "created": "Mon, 19 May 2014 01:17:48 GMT" } ]
2014-05-20T00:00:00
[ [ "Tobin", "R. Joshua", "" ], [ "Houghton", "Conor J.", "" ] ]
TITLE: A Kernel-Based Calculation of Information on a Metric Space ABSTRACT: Kernel density estimation is a technique for approximating probability distributions. Here, it is applied to the calculation of mutual information on a metric space. This is motivated by the problem in neuroscience of calculating the mutual information between stimuli and spiking responses; the space of these responses is a metric space. It is shown that kernel density estimation on a metric space resembles the k-nearest-neighbor approach. This approach is applied to a toy dataset designed to mimic electrophysiological data.
new_dataset
0.716374
1405.4699
Thanasis Naskos
Athanasios Naskos, Emmanouela Stachtiari, Anastasios Gounaris, Panagiotis Katsaros, Dimitrios Tsoumakos, Ioannis Konstantinou, Spyros Sioutas
Cloud elasticity using probabilistic model checking
14 pages
null
null
null
cs.DC cs.LO
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Cloud computing has become the leading paradigm for deploying large-scale infrastructures and running big data applications, due to its capacity of achieving economies of scale. In this work, we focus on one of the most prominent advantages of cloud computing, namely the on-demand resource provisioning, which is commonly referred to as elasticity. Although a lot of effort has been invested in developing systems and mechanisms that enable elasticity, the elasticity decision policies tend to be designed without guaranteeing or quantifying the quality of their operation. This work aims to make the development of elasticity policies more formalized and dependable. We make two distinct contributions. First, we propose an extensible approach to enforcing elasticity through the dynamic instantiation and online quantitative verification of Markov Decision Processes (MDP) using probabilistic model checking. Second, we propose concrete elasticity models and related elasticity policies. We evaluate our decision policies using both real and synthetic datasets in clusters of NoSQL databases. According to the experimental results, our approach improves upon the state-of-the-art in significantly increasing user-defined utility values and decreasing user-defined threshold violations.
[ { "version": "v1", "created": "Mon, 19 May 2014 12:47:16 GMT" } ]
2014-05-20T00:00:00
[ [ "Naskos", "Athanasios", "" ], [ "Stachtiari", "Emmanouela", "" ], [ "Gounaris", "Anastasios", "" ], [ "Katsaros", "Panagiotis", "" ], [ "Tsoumakos", "Dimitrios", "" ], [ "Konstantinou", "Ioannis", "" ], [ "Sioutas", "Spyros", "" ] ]
TITLE: Cloud elasticity using probabilistic model checking ABSTRACT: Cloud computing has become the leading paradigm for deploying large-scale infrastructures and running big data applications, due to its capacity of achieving economies of scale. In this work, we focus on one of the most prominent advantages of cloud computing, namely the on-demand resource provisioning, which is commonly referred to as elasticity. Although a lot of effort has been invested in developing systems and mechanisms that enable elasticity, the elasticity decision policies tend to be designed without guaranteeing or quantifying the quality of their operation. This work aims to make the development of elasticity policies more formalized and dependable. We make two distinct contributions. First, we propose an extensible approach to enforcing elasticity through the dynamic instantiation and online quantitative verification of Markov Decision Processes (MDP) using probabilistic model checking. Second, we propose concrete elasticity models and related elasticity policies. We evaluate our decision policies using both real and synthetic datasets in clusters of NoSQL databases. According to the experimental results, our approach improves upon the state-of-the-art in significantly increasing user-defined utility values and decreasing user-defined threshold violations.
no_new_dataset
0.710584
1403.1024
Hyun Oh Song
Hyun Oh Song, Ross Girshick, Stefanie Jegelka, Julien Mairal, Zaid Harchaoui, Trevor Darrell
On learning to localize objects with minimal supervision
null
null
null
null
cs.CV cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Learning to localize objects with minimal supervision is an important problem in computer vision, since large fully annotated datasets are extremely costly to obtain. In this paper, we propose a new method that achieves this goal with only image-level labels of whether the objects are present or not. Our approach combines a discriminative submodular cover problem for automatically discovering a set of positive object windows with a smoothed latent SVM formulation. The latter allows us to leverage efficient quasi-Newton optimization techniques. Our experiments demonstrate that the proposed approach provides a 50% relative improvement in mean average precision over the current state-of-the-art on PASCAL VOC 2007 detection.
[ { "version": "v1", "created": "Wed, 5 Mar 2014 07:21:20 GMT" }, { "version": "v2", "created": "Fri, 14 Mar 2014 00:50:26 GMT" }, { "version": "v3", "created": "Mon, 17 Mar 2014 21:04:49 GMT" }, { "version": "v4", "created": "Thu, 15 May 2014 22:08:59 GMT" } ]
2014-05-19T00:00:00
[ [ "Song", "Hyun Oh", "" ], [ "Girshick", "Ross", "" ], [ "Jegelka", "Stefanie", "" ], [ "Mairal", "Julien", "" ], [ "Harchaoui", "Zaid", "" ], [ "Darrell", "Trevor", "" ] ]
TITLE: On learning to localize objects with minimal supervision ABSTRACT: Learning to localize objects with minimal supervision is an important problem in computer vision, since large fully annotated datasets are extremely costly to obtain. In this paper, we propose a new method that achieves this goal with only image-level labels of whether the objects are present or not. Our approach combines a discriminative submodular cover problem for automatically discovering a set of positive object windows with a smoothed latent SVM formulation. The latter allows us to leverage efficient quasi-Newton optimization techniques. Our experiments demonstrate that the proposed approach provides a 50% relative improvement in mean average precision over the current state-of-the-art on PASCAL VOC 2007 detection.
no_new_dataset
0.71145
1405.4054
Jianfeng Wang
Jianfeng Wang, Jingdong Wang, Jingkuan Song, Xin-Shun Xu, Heng Tao Shen, Shipeng Li
Optimized Cartesian $K$-Means
to appear in IEEE TKDE, accepted in Apr. 2014
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Product quantization-based approaches are effective to encode high-dimensional data points for approximate nearest neighbor search. The space is decomposed into a Cartesian product of low-dimensional subspaces, each of which generates a sub codebook. Data points are encoded as compact binary codes using these sub codebooks, and the distance between two data points can be approximated efficiently from their codes by the precomputed lookup tables. Traditionally, to encode a subvector of a data point in a subspace, only one sub codeword in the corresponding sub codebook is selected, which may impose strict restrictions on the search accuracy. In this paper, we propose a novel approach, named Optimized Cartesian $K$-Means (OCKM), to better encode the data points for more accurate approximate nearest neighbor search. In OCKM, multiple sub codewords are used to encode the subvector of a data point in a subspace. Each sub codeword stems from different sub codebooks in each subspace, which are optimally generated with regards to the minimization of the distortion errors. The high-dimensional data point is then encoded as the concatenation of the indices of multiple sub codewords from all the subspaces. This can provide more flexibility and lower distortion errors than traditional methods. Experimental results on the standard real-life datasets demonstrate the superiority over state-of-the-art approaches for approximate nearest neighbor search.
[ { "version": "v1", "created": "Fri, 16 May 2014 03:09:01 GMT" } ]
2014-05-19T00:00:00
[ [ "Wang", "Jianfeng", "" ], [ "Wang", "Jingdong", "" ], [ "Song", "Jingkuan", "" ], [ "Xu", "Xin-Shun", "" ], [ "Shen", "Heng Tao", "" ], [ "Li", "Shipeng", "" ] ]
TITLE: Optimized Cartesian $K$-Means ABSTRACT: Product quantization-based approaches are effective to encode high-dimensional data points for approximate nearest neighbor search. The space is decomposed into a Cartesian product of low-dimensional subspaces, each of which generates a sub codebook. Data points are encoded as compact binary codes using these sub codebooks, and the distance between two data points can be approximated efficiently from their codes by the precomputed lookup tables. Traditionally, to encode a subvector of a data point in a subspace, only one sub codeword in the corresponding sub codebook is selected, which may impose strict restrictions on the search accuracy. In this paper, we propose a novel approach, named Optimized Cartesian $K$-Means (OCKM), to better encode the data points for more accurate approximate nearest neighbor search. In OCKM, multiple sub codewords are used to encode the subvector of a data point in a subspace. Each sub codeword stems from different sub codebooks in each subspace, which are optimally generated with regards to the minimization of the distortion errors. The high-dimensional data point is then encoded as the concatenation of the indices of multiple sub codewords from all the subspaces. This can provide more flexibility and lower distortion errors than traditional methods. Experimental results on the standard real-life datasets demonstrate the superiority over state-of-the-art approaches for approximate nearest neighbor search.
no_new_dataset
0.7114
1307.0044
Maria Gorlatova
Maria Gorlatova and John Sarik and Guy Grebla and Mina Cong and Ioannis Kymissis and Gil Zussman
Movers and Shakers: Kinetic Energy Harvesting for the Internet of Things
15 pages, 11 figures
null
null
null
cs.ET cs.NI cs.PF
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Numerous energy harvesting wireless devices that will serve as building blocks for the Internet of Things (IoT) are currently under development. However, there is still only limited understanding of the properties of various energy sources and their impact on energy harvesting adaptive algorithms. Hence, we focus on characterizing the kinetic (motion) energy that can be harvested by a wireless node with an IoT form factor and on developing energy allocation algorithms for such nodes. In this paper, we describe methods for estimating harvested energy from acceleration traces. To characterize the energy availability associated with specific human activities (e.g., relaxing, walking, cycling), we analyze a motion dataset with over 40 participants. Based on acceleration measurements that we collected for over 200 hours, we study energy generation processes associated with day-long human routines. We also briefly summarize our experiments with moving objects. We develop energy allocation algorithms that take into account practical IoT node design considerations, and evaluate the algorithms using the collected measurements. Our observations provide insights into the design of motion energy harvesters, IoT nodes, and energy harvesting adaptive algorithms.
[ { "version": "v1", "created": "Fri, 28 Jun 2013 22:40:11 GMT" }, { "version": "v2", "created": "Mon, 23 Sep 2013 20:49:06 GMT" }, { "version": "v3", "created": "Wed, 14 May 2014 22:34:35 GMT" } ]
2014-05-16T00:00:00
[ [ "Gorlatova", "Maria", "" ], [ "Sarik", "John", "" ], [ "Grebla", "Guy", "" ], [ "Cong", "Mina", "" ], [ "Kymissis", "Ioannis", "" ], [ "Zussman", "Gil", "" ] ]
TITLE: Movers and Shakers: Kinetic Energy Harvesting for the Internet of Things ABSTRACT: Numerous energy harvesting wireless devices that will serve as building blocks for the Internet of Things (IoT) are currently under development. However, there is still only limited understanding of the properties of various energy sources and their impact on energy harvesting adaptive algorithms. Hence, we focus on characterizing the kinetic (motion) energy that can be harvested by a wireless node with an IoT form factor and on developing energy allocation algorithms for such nodes. In this paper, we describe methods for estimating harvested energy from acceleration traces. To characterize the energy availability associated with specific human activities (e.g., relaxing, walking, cycling), we analyze a motion dataset with over 40 participants. Based on acceleration measurements that we collected for over 200 hours, we study energy generation processes associated with day-long human routines. We also briefly summarize our experiments with moving objects. We develop energy allocation algorithms that take into account practical IoT node design considerations, and evaluate the algorithms using the collected measurements. Our observations provide insights into the design of motion energy harvesters, IoT nodes, and energy harvesting adaptive algorithms.
no_new_dataset
0.689306
1402.1500
Eran Shaham Mr.
Eran Shaham, David Sarne, Boaz Ben-Moshe
Co-clustering of Fuzzy Lagged Data
Under consideration for publication in Knowledge and Information Systems. The final publication is available at Springer via http://dx.doi.org/10.1007/s10115-014-0758-7
null
10.1007/s10115-014-0758-7
null
cs.AI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The paper focuses on mining patterns that are characterized by a fuzzy lagged relationship between the data objects forming them. Such a regulatory mechanism is quite common in real life settings. It appears in a variety of fields: finance, gene expression, neuroscience, crowds and collective movements are but a limited list of examples. Mining such patterns not only helps in understanding the relationship between objects in the domain, but assists in forecasting their future behavior. For most interesting variants of this problem, finding an optimal fuzzy lagged co-cluster is an NP-complete problem. We thus present a polynomial-time Monte-Carlo approximation algorithm for mining fuzzy lagged co-clusters. We prove that for any data matrix, the algorithm mines a fuzzy lagged co-cluster with fixed probability, which encompasses the optimal fuzzy lagged co-cluster by a maximum 2 ratio columns overhead and completely no rows overhead. Moreover, the algorithm handles noise, anti-correlations, missing values and overlapping patterns. The algorithm was extensively evaluated using both artificial and real datasets. The results not only corroborate the ability of the algorithm to efficiently mine relevant and accurate fuzzy lagged co-clusters, but also illustrate the importance of including the fuzziness in the lagged-pattern model.
[ { "version": "v1", "created": "Thu, 6 Feb 2014 21:02:16 GMT" }, { "version": "v2", "created": "Thu, 15 May 2014 12:01:08 GMT" } ]
2014-05-16T00:00:00
[ [ "Shaham", "Eran", "" ], [ "Sarne", "David", "" ], [ "Ben-Moshe", "Boaz", "" ] ]
TITLE: Co-clustering of Fuzzy Lagged Data ABSTRACT: The paper focuses on mining patterns that are characterized by a fuzzy lagged relationship between the data objects forming them. Such a regulatory mechanism is quite common in real life settings. It appears in a variety of fields: finance, gene expression, neuroscience, crowds and collective movements are but a limited list of examples. Mining such patterns not only helps in understanding the relationship between objects in the domain, but assists in forecasting their future behavior. For most interesting variants of this problem, finding an optimal fuzzy lagged co-cluster is an NP-complete problem. We thus present a polynomial-time Monte-Carlo approximation algorithm for mining fuzzy lagged co-clusters. We prove that for any data matrix, the algorithm mines a fuzzy lagged co-cluster with fixed probability, which encompasses the optimal fuzzy lagged co-cluster by a maximum 2 ratio columns overhead and completely no rows overhead. Moreover, the algorithm handles noise, anti-correlations, missing values and overlapping patterns. The algorithm was extensively evaluated using both artificial and real datasets. The results not only corroborate the ability of the algorithm to efficiently mine relevant and accurate fuzzy lagged co-clusters, but also illustrate the importance of including the fuzziness in the lagged-pattern model.
no_new_dataset
0.711425
1405.2798
Jun Wang
Jun Wang, Ke Sun, Fei Sha, Stephane Marchand-Maillet, Alexandros Kalousis
Two-Stage Metric Learning
Accepted for publication in ICML 2014
null
null
null
cs.LG cs.AI stat.ML
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
In this paper, we present a novel two-stage metric learning algorithm. We first map each learning instance to a probability distribution by computing its similarities to a set of fixed anchor points. Then, we define the distance in the input data space as the Fisher information distance on the associated statistical manifold. This induces in the input data space a new family of distance metric with unique properties. Unlike kernelized metric learning, we do not require the similarity measure to be positive semi-definite. Moreover, it can also be interpreted as a local metric learning algorithm with well defined distance approximation. We evaluate its performance on a number of datasets. It outperforms significantly other metric learning methods and SVM.
[ { "version": "v1", "created": "Mon, 12 May 2014 15:18:15 GMT" } ]
2014-05-16T00:00:00
[ [ "Wang", "Jun", "" ], [ "Sun", "Ke", "" ], [ "Sha", "Fei", "" ], [ "Marchand-Maillet", "Stephane", "" ], [ "Kalousis", "Alexandros", "" ] ]
TITLE: Two-Stage Metric Learning ABSTRACT: In this paper, we present a novel two-stage metric learning algorithm. We first map each learning instance to a probability distribution by computing its similarities to a set of fixed anchor points. Then, we define the distance in the input data space as the Fisher information distance on the associated statistical manifold. This induces in the input data space a new family of distance metric with unique properties. Unlike kernelized metric learning, we do not require the similarity measure to be positive semi-definite. Moreover, it can also be interpreted as a local metric learning algorithm with well defined distance approximation. We evaluate its performance on a number of datasets. It outperforms significantly other metric learning methods and SVM.
no_new_dataset
0.711224
1405.2998
Alexander Romanovsky
Alexander Romanovsky, Marc-Olivier Killijian
The Tenth European Dependable Computer Conference
EDCC-2014
null
null
null
cs.SE
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The 21st century society relies on computing systems more than ever. Computers are no longer simply machines that are used by organizations or at home. They are embedded everywhere, from cell phones to cars or industrial control devices, and large-scale cloud computing providers are sharing them among many organizations in an unprecedented scale. As computers have become indispensable, their failures may significantly perturb our daily lives. The increased hardware and software complexity, as well as the scaling of manufacturing technologies towards nanometer size devices, pose new challenges to the developers. As a consequence the development, testing, and benchmarking of dependable systems has become a vital topic of research, both for academia and industry. EDCC is the leading European conference for presenting and discussing the latest research in dependable computing. As in previous years, its tenth edition aims at providing a European-hosted venue for researchers and practitioners from all over the world to present and discuss their latest research results on dependability, security, fault-tolerance, and testing. Original papers are solicited on theory, techniques, systems, and tools for the design, validation, operation and evaluation of dependable and secure computing systems, covering any fault model, from traditional hardware and software faults to accidental and malicious human interactions.
[ { "version": "v1", "created": "Mon, 12 May 2014 23:42:30 GMT" }, { "version": "v2", "created": "Thu, 15 May 2014 00:31:59 GMT" } ]
2014-05-16T00:00:00
[ [ "Romanovsky", "Alexander", "" ], [ "Killijian", "Marc-Olivier", "" ] ]
TITLE: The Tenth European Dependable Computer Conference ABSTRACT: The 21st century society relies on computing systems more than ever. Computers are no longer simply machines that are used by organizations or at home. They are embedded everywhere, from cell phones to cars or industrial control devices, and large-scale cloud computing providers are sharing them among many organizations in an unprecedented scale. As computers have become indispensable, their failures may significantly perturb our daily lives. The increased hardware and software complexity, as well as the scaling of manufacturing technologies towards nanometer size devices, pose new challenges to the developers. As a consequence the development, testing, and benchmarking of dependable systems has become a vital topic of research, both for academia and industry. EDCC is the leading European conference for presenting and discussing the latest research in dependable computing. As in previous years, its tenth edition aims at providing a European-hosted venue for researchers and practitioners from all over the world to present and discuss their latest research results on dependability, security, fault-tolerance, and testing. Original papers are solicited on theory, techniques, systems, and tools for the design, validation, operation and evaluation of dependable and secure computing systems, covering any fault model, from traditional hardware and software faults to accidental and malicious human interactions.
no_new_dataset
0.700332
1405.3727
Sweta Rai
Sweta Rai
Student Dropout Risk Assessment in Undergraduate Course at Residential University
arXiv admin note: text overlap with arXiv:1202.4815, arXiv:1203.3832, arXiv:1002.1144 by other authors
null
null
null
cs.CY cs.AI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Student dropout prediction is an indispensable for numerous intelligent systems to measure the education system and success rate of any university as well as throughout the university in the world. Therefore, it becomes essential to develop efficient methods for prediction of the students at risk of dropping out, enabling the adoption of proactive process to minimize the situation. Thus, this research work propose a prototype machine learning tool which can automatically recognize whether the student will continue their study or drop their study using classification technique based on decision tree and extract hidden information from large data about what factors are responsible for dropout student. Further the contribution of factors responsible for dropout risk was studied using discriminant analysis and to extract interesting correlations, frequent patterns, associations or casual structures among significant datasets, Association rule mining was applied. In this study, the descriptive statistics analysis was carried out to measure the quality of data using SPSS 20.0 statistical software and application of decision tree and association rule were carried out by using WEKA data mining tool.
[ { "version": "v1", "created": "Thu, 15 May 2014 02:35:41 GMT" } ]
2014-05-16T00:00:00
[ [ "Rai", "Sweta", "" ] ]
TITLE: Student Dropout Risk Assessment in Undergraduate Course at Residential University ABSTRACT: Student dropout prediction is an indispensable for numerous intelligent systems to measure the education system and success rate of any university as well as throughout the university in the world. Therefore, it becomes essential to develop efficient methods for prediction of the students at risk of dropping out, enabling the adoption of proactive process to minimize the situation. Thus, this research work propose a prototype machine learning tool which can automatically recognize whether the student will continue their study or drop their study using classification technique based on decision tree and extract hidden information from large data about what factors are responsible for dropout student. Further the contribution of factors responsible for dropout risk was studied using discriminant analysis and to extract interesting correlations, frequent patterns, associations or casual structures among significant datasets, Association rule mining was applied. In this study, the descriptive statistics analysis was carried out to measure the quality of data using SPSS 20.0 statistical software and application of decision tree and association rule were carried out by using WEKA data mining tool.
no_new_dataset
0.708213
1405.3866
Max Jaderberg
Max Jaderberg, Andrea Vedaldi, Andrew Zisserman
Speeding up Convolutional Neural Networks with Low Rank Expansions
null
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The focus of this paper is speeding up the evaluation of convolutional neural networks. While delivering impressive results across a range of computer vision and machine learning tasks, these networks are computationally demanding, limiting their deployability. Convolutional layers generally consume the bulk of the processing time, and so in this work we present two simple schemes for drastically speeding up these layers. This is achieved by exploiting cross-channel or filter redundancy to construct a low rank basis of filters that are rank-1 in the spatial domain. Our methods are architecture agnostic, and can be easily applied to existing CPU and GPU convolutional frameworks for tuneable speedup performance. We demonstrate this with a real world network designed for scene text character recognition, showing a possible 2.5x speedup with no loss in accuracy, and 4.5x speedup with less than 1% drop in accuracy, still achieving state-of-the-art on standard benchmarks.
[ { "version": "v1", "created": "Thu, 15 May 2014 14:41:06 GMT" } ]
2014-05-16T00:00:00
[ [ "Jaderberg", "Max", "" ], [ "Vedaldi", "Andrea", "" ], [ "Zisserman", "Andrew", "" ] ]
TITLE: Speeding up Convolutional Neural Networks with Low Rank Expansions ABSTRACT: The focus of this paper is speeding up the evaluation of convolutional neural networks. While delivering impressive results across a range of computer vision and machine learning tasks, these networks are computationally demanding, limiting their deployability. Convolutional layers generally consume the bulk of the processing time, and so in this work we present two simple schemes for drastically speeding up these layers. This is achieved by exploiting cross-channel or filter redundancy to construct a low rank basis of filters that are rank-1 in the spatial domain. Our methods are architecture agnostic, and can be easily applied to existing CPU and GPU convolutional frameworks for tuneable speedup performance. We demonstrate this with a real world network designed for scene text character recognition, showing a possible 2.5x speedup with no loss in accuracy, and 4.5x speedup with less than 1% drop in accuracy, still achieving state-of-the-art on standard benchmarks.
no_new_dataset
0.711005
1405.3410
Tieming Chen
Tieming Chen, Xu Zhang, Shichao Jin, Okhee Kim
Efficient classification using parallel and scalable compressed model and Its application on intrusion detection
null
null
10.1016/j.eswa.2014.04.009
null
cs.LG cs.CR
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
In order to achieve high efficiency of classification in intrusion detection, a compressed model is proposed in this paper which combines horizontal compression with vertical compression. OneR is utilized as horizontal com-pression for attribute reduction, and affinity propagation is employed as vertical compression to select small representative exemplars from large training data. As to be able to computationally compress the larger volume of training data with scalability, MapReduce based parallelization approach is then implemented and evaluated for each step of the model compression process abovementioned, on which common but efficient classification methods can be directly used. Experimental application study on two publicly available datasets of intrusion detection, KDD99 and CMDC2012, demonstrates that the classification using the compressed model proposed can effectively speed up the detection procedure at up to 184 times, most importantly at the cost of a minimal accuracy difference with less than 1% on average.
[ { "version": "v1", "created": "Wed, 14 May 2014 08:47:31 GMT" } ]
2014-05-15T00:00:00
[ [ "Chen", "Tieming", "" ], [ "Zhang", "Xu", "" ], [ "Jin", "Shichao", "" ], [ "Kim", "Okhee", "" ] ]
TITLE: Efficient classification using parallel and scalable compressed model and Its application on intrusion detection ABSTRACT: In order to achieve high efficiency of classification in intrusion detection, a compressed model is proposed in this paper which combines horizontal compression with vertical compression. OneR is utilized as horizontal com-pression for attribute reduction, and affinity propagation is employed as vertical compression to select small representative exemplars from large training data. As to be able to computationally compress the larger volume of training data with scalability, MapReduce based parallelization approach is then implemented and evaluated for each step of the model compression process abovementioned, on which common but efficient classification methods can be directly used. Experimental application study on two publicly available datasets of intrusion detection, KDD99 and CMDC2012, demonstrates that the classification using the compressed model proposed can effectively speed up the detection procedure at up to 184 times, most importantly at the cost of a minimal accuracy difference with less than 1% on average.
no_new_dataset
0.708629
1405.2941
Jiang Wang Mr.
Jiang wang, Xiaohan Nie, Yin Xia, Ying Wu, Song-Chun Zhu
Cross-view Action Modeling, Learning and Recognition
CVPR 2014
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Existing methods on video-based action recognition are generally view-dependent, i.e., performing recognition from the same views seen in the training data. We present a novel multiview spatio-temporal AND-OR graph (MST-AOG) representation for cross-view action recognition, i.e., the recognition is performed on the video from an unknown and unseen view. As a compositional model, MST-AOG compactly represents the hierarchical combinatorial structures of cross-view actions by explicitly modeling the geometry, appearance and motion variations. This paper proposes effective methods to learn the structure and parameters of MST-AOG. The inference based on MST-AOG enables action recognition from novel views. The training of MST-AOG takes advantage of the 3D human skeleton data obtained from Kinect cameras to avoid annotating enormous multi-view video frames, which is error-prone and time-consuming, but the recognition does not need 3D information and is based on 2D video input. A new Multiview Action3D dataset has been created and will be released. Extensive experiments have demonstrated that this new action representation significantly improves the accuracy and robustness for cross-view action recognition on 2D videos.
[ { "version": "v1", "created": "Mon, 12 May 2014 20:21:53 GMT" } ]
2014-05-14T00:00:00
[ [ "wang", "Jiang", "" ], [ "Nie", "Xiaohan", "" ], [ "Xia", "Yin", "" ], [ "Wu", "Ying", "" ], [ "Zhu", "Song-Chun", "" ] ]
TITLE: Cross-view Action Modeling, Learning and Recognition ABSTRACT: Existing methods on video-based action recognition are generally view-dependent, i.e., performing recognition from the same views seen in the training data. We present a novel multiview spatio-temporal AND-OR graph (MST-AOG) representation for cross-view action recognition, i.e., the recognition is performed on the video from an unknown and unseen view. As a compositional model, MST-AOG compactly represents the hierarchical combinatorial structures of cross-view actions by explicitly modeling the geometry, appearance and motion variations. This paper proposes effective methods to learn the structure and parameters of MST-AOG. The inference based on MST-AOG enables action recognition from novel views. The training of MST-AOG takes advantage of the 3D human skeleton data obtained from Kinect cameras to avoid annotating enormous multi-view video frames, which is error-prone and time-consuming, but the recognition does not need 3D information and is based on 2D video input. A new Multiview Action3D dataset has been created and will be released. Extensive experiments have demonstrated that this new action representation significantly improves the accuracy and robustness for cross-view action recognition on 2D videos.
new_dataset
0.71561
1405.3080
Tong Zhang
Peilin Zhao, Tong Zhang
Accelerating Minibatch Stochastic Gradient Descent using Stratified Sampling
null
null
null
null
stat.ML cs.LG math.OC
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Stochastic Gradient Descent (SGD) is a popular optimization method which has been applied to many important machine learning tasks such as Support Vector Machines and Deep Neural Networks. In order to parallelize SGD, minibatch training is often employed. The standard approach is to uniformly sample a minibatch at each step, which often leads to high variance. In this paper we propose a stratified sampling strategy, which divides the whole dataset into clusters with low within-cluster variance; we then take examples from these clusters using a stratified sampling technique. It is shown that the convergence rate can be significantly improved by the algorithm. Encouraging experimental results confirm the effectiveness of the proposed method.
[ { "version": "v1", "created": "Tue, 13 May 2014 09:45:49 GMT" } ]
2014-05-14T00:00:00
[ [ "Zhao", "Peilin", "" ], [ "Zhang", "Tong", "" ] ]
TITLE: Accelerating Minibatch Stochastic Gradient Descent using Stratified Sampling ABSTRACT: Stochastic Gradient Descent (SGD) is a popular optimization method which has been applied to many important machine learning tasks such as Support Vector Machines and Deep Neural Networks. In order to parallelize SGD, minibatch training is often employed. The standard approach is to uniformly sample a minibatch at each step, which often leads to high variance. In this paper we propose a stratified sampling strategy, which divides the whole dataset into clusters with low within-cluster variance; we then take examples from these clusters using a stratified sampling technique. It is shown that the convergence rate can be significantly improved by the algorithm. Encouraging experimental results confirm the effectiveness of the proposed method.
no_new_dataset
0.709416
1405.3094
Olfa Boubaker
Olfa Boubaker
The inverted Pendulum: A fundamental Benchmark in Control Theory and Robotics
IEEE International Conference on Education and e-Learning Innovations (ICEELI), 1-3 July 2012, Sousse, Tunisia
null
10.1109/ICEELI.2012.6360606
null
cs.RO cs.SY
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
For at least fifty years, the inverted pendulum has been the most popular benchmark, among others, for teaching and researches in control theory and robotics. This paper presents the key motivations for the use of that system and explains, in details, the main reflections on how the inverted pendulum benchmark gives an effective and efficient application. Several real experiences, virtual models and web-based remote control laboratories will be presented with emphasis on the practical design implementation of this system. A bibliographical survey of different design control approaches and trendy robotic problems will be presented through applications to the inverted pendulum system. In total, 150 references in the open literature, dating back to 1960, are compiled to provide an overall picture of historical, current and challenging developments.
[ { "version": "v1", "created": "Tue, 13 May 2014 10:38:02 GMT" } ]
2014-05-14T00:00:00
[ [ "Boubaker", "Olfa", "" ] ]
TITLE: The inverted Pendulum: A fundamental Benchmark in Control Theory and Robotics ABSTRACT: For at least fifty years, the inverted pendulum has been the most popular benchmark, among others, for teaching and researches in control theory and robotics. This paper presents the key motivations for the use of that system and explains, in details, the main reflections on how the inverted pendulum benchmark gives an effective and efficient application. Several real experiences, virtual models and web-based remote control laboratories will be presented with emphasis on the practical design implementation of this system. A bibliographical survey of different design control approaches and trendy robotic problems will be presented through applications to the inverted pendulum system. In total, 150 references in the open literature, dating back to 1960, are compiled to provide an overall picture of historical, current and challenging developments.
no_new_dataset
0.709177
1405.3210
Jeremy Kun
Jeremy Kun, Rajmonda Caceres, Kevin Carter
Locally Boosted Graph Aggregation for Community Detection
arXiv admin note: substantial text overlap with arXiv:1401.3258
null
null
null
cs.LG cs.SI physics.soc-ph
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Learning the right graph representation from noisy, multi-source data has garnered significant interest in recent years. A central tenet of this problem is relational learning. Here the objective is to incorporate the partial information each data source gives us in a way that captures the true underlying relationships. To address this challenge, we present a general, boosting-inspired framework for combining weak evidence of entity associations into a robust similarity metric. Building on previous work, we explore the extent to which different local quality measurements yield graph representations that are suitable for community detection. We present empirical results on a variety of datasets demonstrating the utility of this framework, especially with respect to real datasets where noise and scale present serious challenges. Finally, we prove a convergence theorem in an ideal setting and outline future research into other application domains.
[ { "version": "v1", "created": "Tue, 13 May 2014 16:08:55 GMT" } ]
2014-05-14T00:00:00
[ [ "Kun", "Jeremy", "" ], [ "Caceres", "Rajmonda", "" ], [ "Carter", "Kevin", "" ] ]
TITLE: Locally Boosted Graph Aggregation for Community Detection ABSTRACT: Learning the right graph representation from noisy, multi-source data has garnered significant interest in recent years. A central tenet of this problem is relational learning. Here the objective is to incorporate the partial information each data source gives us in a way that captures the true underlying relationships. To address this challenge, we present a general, boosting-inspired framework for combining weak evidence of entity associations into a robust similarity metric. Building on previous work, we explore the extent to which different local quality measurements yield graph representations that are suitable for community detection. We present empirical results on a variety of datasets demonstrating the utility of this framework, especially with respect to real datasets where noise and scale present serious challenges. Finally, we prove a convergence theorem in an ideal setting and outline future research into other application domains.
no_new_dataset
0.709177
1210.4460
Yaman Aksu Ph.D.
Yaman Aksu
Fast SVM-based Feature Elimination Utilizing Data Radius, Hard-Margin, Soft-Margin
Incomplete but good, again. To Apr 28 version, made few misc text and notation improvements including typo corrections, probably mostly in Appendix, but probably best to read in whole again. New results for one of the datasets (Leukemia gene dataset)
null
null
null
stat.ML cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Margin maximization in the hard-margin sense, proposed as feature elimination criterion by the MFE-LO method, is combined here with data radius utilization to further aim to lower generalization error, as several published bounds and bound-related formulations pertaining to lowering misclassification risk (or error) pertain to radius e.g. product of squared radius and weight vector squared norm. Additionally, we propose additional novel feature elimination criteria that, while instead being in the soft-margin sense, too can utilize data radius, utilizing previously published bound-related formulations for approaching radius for the soft-margin sense, whereby e.g. a focus was on the principle stated therein as "finding a bound whose minima are in a region with small leave-one-out values may be more important than its tightness". These additional criteria we propose combine radius utilization with a novel and computationally low-cost soft-margin light classifier retraining approach we devise named QP1; QP1 is the soft-margin alternative to the hard-margin LO. We correct an error in the MFE-LO description, find MFE-LO achieves the highest generalization accuracy among the previously published margin-based feature elimination (MFE) methods, discuss some limitations of MFE-LO, and find our novel methods herein outperform MFE-LO, attain lower test set classification error rate. On several datasets that each both have a large number of features and fall into the `large features few samples' dataset category, and on datasets with lower (low-to-intermediate) number of features, our novel methods give promising results. Especially, among our methods the tunable ones, that do not employ (the non-tunable) LO approach, can be tuned more aggressively in the future than herein, to aim to demonstrate for them even higher performance than herein.
[ { "version": "v1", "created": "Tue, 16 Oct 2012 15:54:36 GMT" }, { "version": "v2", "created": "Wed, 9 Jan 2013 16:28:17 GMT" }, { "version": "v3", "created": "Mon, 28 Apr 2014 21:15:46 GMT" }, { "version": "v4", "created": "Sun, 11 May 2014 11:47:07 GMT" } ]
2014-05-13T00:00:00
[ [ "Aksu", "Yaman", "" ] ]
TITLE: Fast SVM-based Feature Elimination Utilizing Data Radius, Hard-Margin, Soft-Margin ABSTRACT: Margin maximization in the hard-margin sense, proposed as feature elimination criterion by the MFE-LO method, is combined here with data radius utilization to further aim to lower generalization error, as several published bounds and bound-related formulations pertaining to lowering misclassification risk (or error) pertain to radius e.g. product of squared radius and weight vector squared norm. Additionally, we propose additional novel feature elimination criteria that, while instead being in the soft-margin sense, too can utilize data radius, utilizing previously published bound-related formulations for approaching radius for the soft-margin sense, whereby e.g. a focus was on the principle stated therein as "finding a bound whose minima are in a region with small leave-one-out values may be more important than its tightness". These additional criteria we propose combine radius utilization with a novel and computationally low-cost soft-margin light classifier retraining approach we devise named QP1; QP1 is the soft-margin alternative to the hard-margin LO. We correct an error in the MFE-LO description, find MFE-LO achieves the highest generalization accuracy among the previously published margin-based feature elimination (MFE) methods, discuss some limitations of MFE-LO, and find our novel methods herein outperform MFE-LO, attain lower test set classification error rate. On several datasets that each both have a large number of features and fall into the `large features few samples' dataset category, and on datasets with lower (low-to-intermediate) number of features, our novel methods give promising results. Especially, among our methods the tunable ones, that do not employ (the non-tunable) LO approach, can be tuned more aggressively in the future than herein, to aim to demonstrate for them even higher performance than herein.
no_new_dataset
0.711663
1210.4567
Jacob Eisenstein
David Bamman, Jacob Eisenstein, and Tyler Schnoebelen
Gender identity and lexical variation in social media
submission version
Journal of Sociolinguistics 18 (2014) 135-160
10.1111/josl.12080
null
cs.CL
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We present a study of the relationship between gender, linguistic style, and social networks, using a novel corpus of 14,000 Twitter users. Prior quantitative work on gender often treats this social variable as a female/male binary; we argue for a more nuanced approach. By clustering Twitter users, we find a natural decomposition of the dataset into various styles and topical interests. Many clusters have strong gender orientations, but their use of linguistic resources sometimes directly conflicts with the population-level language statistics. We view these clusters as a more accurate reflection of the multifaceted nature of gendered language styles. Previous corpus-based work has also had little to say about individuals whose linguistic styles defy population-level gender patterns. To identify such individuals, we train a statistical classifier, and measure the classifier confidence for each individual in the dataset. Examining individuals whose language does not match the classifier's model for their gender, we find that they have social networks that include significantly fewer same-gender social connections and that, in general, social network homophily is correlated with the use of same-gender language markers. Pairing computational methods and social theory thus offers a new perspective on how gender emerges as individuals position themselves relative to audiences, topics, and mainstream gender norms.
[ { "version": "v1", "created": "Tue, 16 Oct 2012 20:22:56 GMT" }, { "version": "v2", "created": "Mon, 12 May 2014 15:04:32 GMT" } ]
2014-05-13T00:00:00
[ [ "Bamman", "David", "" ], [ "Eisenstein", "Jacob", "" ], [ "Schnoebelen", "Tyler", "" ] ]
TITLE: Gender identity and lexical variation in social media ABSTRACT: We present a study of the relationship between gender, linguistic style, and social networks, using a novel corpus of 14,000 Twitter users. Prior quantitative work on gender often treats this social variable as a female/male binary; we argue for a more nuanced approach. By clustering Twitter users, we find a natural decomposition of the dataset into various styles and topical interests. Many clusters have strong gender orientations, but their use of linguistic resources sometimes directly conflicts with the population-level language statistics. We view these clusters as a more accurate reflection of the multifaceted nature of gendered language styles. Previous corpus-based work has also had little to say about individuals whose linguistic styles defy population-level gender patterns. To identify such individuals, we train a statistical classifier, and measure the classifier confidence for each individual in the dataset. Examining individuals whose language does not match the classifier's model for their gender, we find that they have social networks that include significantly fewer same-gender social connections and that, in general, social network homophily is correlated with the use of same-gender language markers. Pairing computational methods and social theory thus offers a new perspective on how gender emerges as individuals position themselves relative to audiences, topics, and mainstream gender norms.
no_new_dataset
0.550124
1307.5102
Stefano Gonella
Stefano Gonella, Jarvis D. Haupt
Automated Defect Localization via Low Rank Plus Outlier Modeling of Propagating Wavefield Data
16 pages, 9 figures, Submitted to the IEEE Transactions on Ultrasonics, Ferroelectrics and Frequency Control on August 30th 2012
IEEE Transactions on Ultrasonics, Ferroelectrics and Frequency Control, v. 60, n.12, pp. 2553 - 2565
10.1109/TUFFC.2013.2854
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
This work proposes an agnostic inference strategy for material diagnostics, conceived within the context of laser-based non-destructive evaluation methods, which extract information about structural anomalies from the analysis of acoustic wavefields measured on the structure's surface by means of a scanning laser interferometer. The proposed approach couples spatiotemporal windowing with low rank plus outlier modeling, to identify a priori unknown deviations in the propagating wavefields caused by material inhomogeneities or defects, using virtually no knowledge of the structural and material properties of the medium. This characteristic makes the approach particularly suitable for diagnostics scenarios where the mechanical and material models are complex, unknown, or unreliable. We demonstrate our approach in a simulated environment using benchmark point and line defect localization problems based on propagating flexural waves in a thin plate.
[ { "version": "v1", "created": "Fri, 19 Jul 2013 00:06:59 GMT" } ]
2014-05-13T00:00:00
[ [ "Gonella", "Stefano", "" ], [ "Haupt", "Jarvis D.", "" ] ]
TITLE: Automated Defect Localization via Low Rank Plus Outlier Modeling of Propagating Wavefield Data ABSTRACT: This work proposes an agnostic inference strategy for material diagnostics, conceived within the context of laser-based non-destructive evaluation methods, which extract information about structural anomalies from the analysis of acoustic wavefields measured on the structure's surface by means of a scanning laser interferometer. The proposed approach couples spatiotemporal windowing with low rank plus outlier modeling, to identify a priori unknown deviations in the propagating wavefields caused by material inhomogeneities or defects, using virtually no knowledge of the structural and material properties of the medium. This characteristic makes the approach particularly suitable for diagnostics scenarios where the mechanical and material models are complex, unknown, or unreliable. We demonstrate our approach in a simulated environment using benchmark point and line defect localization problems based on propagating flexural waves in a thin plate.
no_new_dataset
0.712407
1401.1124
Yu Chen
Yu Chen, Weicheng Xie, Xiufen Zou
A binary differential evolution algorithm learning from explored solutions
null
null
null
null
cs.NE
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Although real-coded differential evolution (DE) algorithms can perform well on continuous optimization problems (CoOPs), it is still a challenging task to design an efficient binary-coded DE algorithm. Inspired by the learning mechanism of particle swarm optimization (PSO) algorithms, we propose a binary learning differential evolution (BLDE) algorithm that can efficiently locate the global optimal solutions by learning from the last population. Then, we theoretically prove the global convergence of BLDE, and compare it with some existing binary-coded evolutionary algorithms (EAs) via numerical experiments. Numerical results show that BLDE is competitive to the compared EAs, and meanwhile, further study is performed via the change curves of a renewal metric and a refinement metric to investigate why BLDE cannot outperform some compared EAs for several selected benchmark problems. Finally, we employ BLDE solving the unit commitment problem (UCP) in power systems to show its applicability in practical problems.
[ { "version": "v1", "created": "Mon, 6 Jan 2014 15:54:16 GMT" }, { "version": "v2", "created": "Sat, 10 May 2014 13:01:45 GMT" } ]
2014-05-13T00:00:00
[ [ "Chen", "Yu", "" ], [ "Xie", "Weicheng", "" ], [ "Zou", "Xiufen", "" ] ]
TITLE: A binary differential evolution algorithm learning from explored solutions ABSTRACT: Although real-coded differential evolution (DE) algorithms can perform well on continuous optimization problems (CoOPs), it is still a challenging task to design an efficient binary-coded DE algorithm. Inspired by the learning mechanism of particle swarm optimization (PSO) algorithms, we propose a binary learning differential evolution (BLDE) algorithm that can efficiently locate the global optimal solutions by learning from the last population. Then, we theoretically prove the global convergence of BLDE, and compare it with some existing binary-coded evolutionary algorithms (EAs) via numerical experiments. Numerical results show that BLDE is competitive to the compared EAs, and meanwhile, further study is performed via the change curves of a renewal metric and a refinement metric to investigate why BLDE cannot outperform some compared EAs for several selected benchmark problems. Finally, we employ BLDE solving the unit commitment problem (UCP) in power systems to show its applicability in practical problems.
no_new_dataset
0.707632