id
stringlengths
9
16
title
stringlengths
4
278
categories
stringlengths
5
104
abstract
stringlengths
6
4.09k
1301.2354
A New Approach for Solving Singular Systems in Topology Optimization Using Krylov Subspace Methods
cs.CE math.NA
In topology optimization, the design parameter with no contribution to the objective function vanishes. This causes the stiffness matrix to become singular. We show that a local optimal solution is obtained by Conjugate Residual Method and Conjugate Gradient Method even if the stiffness matrix becomes singular. We prove that CGMconverges to a local optimal solution in that case. Computer simulation shows that CGM gives the same solutions obtained by CRM in case of a cantilever beam problem.
1301.2362
Quasi-SLCA based Keyword Query Processing over Probabilistic XML Data
cs.DB
The probabilistic threshold query is one of the most common queries in uncertain databases, where a result satisfying the query must be also with probability meeting the threshold requirement. In this paper, we investigate probabilistic threshold keyword queries (PrTKQ) over XML data, which is not studied before. We first introduce the notion of quasi-SLCA and use it to represent results for a PrTKQ with the consideration of possible world semantics. Then we design a probabilistic inverted (PI) index that can be used to quickly return the qualified answers and filter out the unqualified ones based on our proposed lower/upper bounds. After that, we propose two efficient and comparable algorithms: Baseline Algorithm and PI index-based Algorithm. To accelerate the performance of algorithms, we also utilize probability density function. An empirical study using real and synthetic data sets has verified the effectiveness and the efficiency of our approaches.
1301.2369
Combinatorial and approximative analyses in a spatially random division process
cond-mat.stat-mech cs.DM cs.SI math-ph math.MP nlin.AO
For a spatial characteristic, there exist commonly fat-tail frequency distributions of fragment-size and -mass of glass, areas enclosed by city roads, and pore size/volume in random packings. In order to give a new analytical approach for the distributions, we consider a simple model which constructs a fractal-like hierarchical network based on random divisions of rectangles. The stochastic process makes a Markov chain and corresponds to directional random walks with splitting into four particles. We derive a combinatorial analytical form and its continuous approximation for the distribution of rectangle areas, and numerically show a good fitting with the actual distribution in the averaging behavior of the divisions.
1301.2375
Context-based Diversification for Keyword Queries over XML Data
cs.DB
While keyword query empowers ordinary users to search vast amount of data, the ambiguity of keyword query makes it difficult to effectively answer keyword queries, especially for short and vague keyword queries. To address this challenging problem, in this paper we propose an approach that automatically diversifies XML keyword search based on its different contexts in the XML data. Given a short and vague keyword query and XML data to be searched, we firstly derive keyword search candidates of the query by a classifical feature selection model. And then, we design an effective XML keyword search diversification model to measure the quality of each candidate. After that, three efficient algorithms are proposed to evaluate the possible generated query candidates representing the diversified search intentions, from which we can find and return top-$k$ qualified query candidates that are most relevant to the given keyword query while they can cover maximal number of distinct results.At last, a comprehensive evaluation on real and synthetic datasets demonstrates the effectiveness of our proposed diversification model and the efficiency of our algorithms.
1301.2378
Query-driven Frequent Co-occurring Term Extraction over Relational Data using MapReduce
cs.DB
In this paper we study how to efficiently compute \textit{frequent co-occurring terms} (FCT) in the results of a keyword query in parallel using the popular MapReduce framework. Taking as input a keyword query q and an integer k, an FCT query reports the k terms that are not in q, but appear most frequently in the results of the keyword query q over multiple joined relations. The returned terms of FCT search can be used to do query expansion and query refinement for traditional keyword search. Different from the method of FCT search in a single platform, our proposed approach can efficiently answer a FCT query using the MapReduce Paradigm without pre-computing the results of the original keyword query, which is run in parallel platform. In this work, we can output the final FCT search results by two MapReduce jobs: the first is to extract the statistical information of the data; and the second is to calculate the total frequency of each term based on the output of the first job. At the two MapReduce jobs, we would guarantee the load balance of mappers and the computational balance of reducers as much as possible. Analytical and experimental evaluations demonstrate the efficiency and scalability of our proposed approach using TPC-H benchmark datasets with different sizes.
1301.2405
Dating medieval English charters
stat.AP cs.CL
Deeds, or charters, dealing with property rights, provide a continuous documentation which can be used by historians to study the evolution of social, economic and political changes. This study is concerned with charters (written in Latin) dating from the tenth through early fourteenth centuries in England. Of these, at least one million were left undated, largely due to administrative changes introduced by William the Conqueror in 1066. Correctly dating such charters is of vital importance in the study of English medieval history. This paper is concerned with computer-automated statistical methods for dating such document collections, with the goal of reducing the considerable efforts required to date them manually and of improving the accuracy of assigned dates. Proposed methods are based on such data as the variation over time of word and phrase usage, and on measures of distance between documents. The extensive (and dated) Documents of Early England Data Set (DEEDS) maintained at the University of Toronto was used for this purpose.
1301.2427
Analysis of the LTE Access Reservation Protocol for Real-Time Traffic
cs.IT cs.NI math.IT
LTE is increasingly seen as a system for serving real-time Machine-to-Machine (M2M) communication needs. The asynchronous M2M user access in LTE is obtained through a two-phase access reservation protocol (contention and data phase). Existing analysis related to these protocols is based on the following assumptions: (1) there are sufficient resources in the data phase for all detected contention tokens, and (2) the base station is able to detect collisions, i.e., tokens activated by multiple users. These assumptions are not always applicable to LTE - specifically, (1) due to the variable amount of available data resources caused by variable load, and (2) detection of collisions in contention phase may not be possible. All of this affects transmission of real-time M2M traffic, where data packets have to be sent within a deadline and may have only one contention opportunity. We analyze the features of the two-phase LTE reservation protocol and derive its throughput, i.e., the number of successful transmissions in the data phase, when assumptions (1) and (2) do not hold.
1301.2444
TEI and LMF crosswalks
cs.CL
The present paper explores various arguments in favour of making the Text Encoding Initia-tive (TEI) guidelines an appropriate serialisation for ISO standard 24613:2008 (LMF, Lexi-cal Mark-up Framework) . It also identifies the issues that would have to be resolved in order to reach an appropriate implementation of these ideas, in particular in terms of infor-mational coverage. We show how the customisation facilities offered by the TEI guidelines can provide an adequate background, not only to cover missing components within the current Dictionary chapter of the TEI guidelines, but also to allow specific lexical projects to deal with local constraints. We expect this proposal to be a basis for a future ISO project in the context of the on going revision of LMF.
1301.2464
Time as a limited resource: Communication Strategy in Mobile Phone Networks
physics.soc-ph cs.SI physics.data-an
We used a large database of 9 billion calls from 20 million mobile users to examine the relationships between aggregated time spent on the phone, personal network size, tie strength and the way in which users distributed their limited time across their network (disparity). Compared to those with smaller networks, those with large networks did not devote proportionally more time to communication and had on average weaker ties (as measured by time spent communicating). Further, there were not substantially different levels of disparity between individuals, in that mobile users tend to distribute their time very unevenly across their network, with a large proportion of calls going to a small number of individuals. Together, these results suggest that there are time constraints which limit tie strength in large personal networks, and that even high levels of mobile communication do not fundamentally alter the disparity of time allocation across networks.
1301.2466
Determining token sequence mistakes in responses to questions with open text answer
cs.CL cs.CY
When learning grammar of the new language, a teacher should routinely check student's exercises for grammatical correctness. The paper describes a method of automatically detecting and reporting grammar mistakes, regarding an order of tokens in the response. It could report extra tokens, missing tokens and misplaced tokens. The method is useful when teaching language, where order of tokens is important, which includes most formal languages and some natural ones (like English). The method was implemented in a question type plug-in CorrectWriting for the widely used learning manage system Moodle.
1301.2479
Weight Distribution of a Class of Cyclic Codes with Arbitrary Number of Zeros
cs.IT math.IT
Cyclic codes have been widely used in digital communication systems and consume electronics as they have efficient encoding and decoding algorithms. The weight distribution of cyclic codes has been an important topic of study for many years. It is in general hard to determine the weight distribution of linear codes. In this paper, a class of cyclic codes with any number of zeros are described and their weight distributions are determined.
1301.2497
Update-Efficient Regenerating Codes with Minimum Per-Node Storage
cs.IT cs.DM math.IT
Regenerating codes provide an efficient way to recover data at failed nodes in distributed storage systems. It has been shown that regenerating codes can be designed to minimize the per-node storage (called MSR) or minimize the communication overhead for regeneration (called MBR). In this work, we propose a new encoding scheme for [n,d] error- correcting MSR codes that generalizes our earlier work on error-correcting regenerating codes. We show that by choosing a suitable diagonal matrix, any generator matrix of the [n,{\alpha}] Reed-Solomon (RS) code can be integrated into the encoding matrix. Hence, MSR codes with the least update complexity can be found. An efficient decoding scheme is also proposed that utilizes the [n,{\alpha}] RS code to perform data reconstruction. The proposed decoding scheme has better error correction capability and incurs the least number of node accesses when errors are present.
1301.2498
Modeling complex systems by Generalized Factor Analysis
cs.SY
We propose a new modeling paradigm for large dimensional aggregates of stochastic systems by Generalized Factor Analysis (GFA) models. These models describe the data as the sum of a flocking plus an uncorrelated idiosyncratic component. The flocking component describes a sort of collective orderly motion which admits a much simpler mathematical description than the whole ensemble while the idiosyncratic component describes weakly correlated noise. We first discuss static GFA representations and characterize in a rigorous way the properties of the two components. The extraction of the dynamic flocking component is discussed for time-stationary linear systems and for a simple classes of separable random fields.
1301.2533
A Novel Analytical Method for Evolutionary Graph Theory Problems
cs.GT cs.SI q-bio.PE
Evolutionary graph theory studies the evolutionary dynamics of populations structured on graphs. A central problem is determining the probability that a small number of mutants overtake a population. Currently, Monte Carlo simulations are used for estimating such fixation probabilities on general directed graphs, since no good analytical methods exist. In this paper, we introduce a novel deterministic framework for computing fixation probabilities for strongly connected, directed, weighted evolutionary graphs under neutral drift. We show how this framework can also be used to calculate the expected number of mutants at a given time step (even if we relax the assumption that the graph is strongly connected), how it can extend to other related models (e.g. voter model), how our framework can provide non-trivial bounds for fixation probability in the case of an advantageous mutant, and how it can be used to find a non-trivial lower bound on the mean time to fixation. We provide various experimental results determining fixation probabilities and expected number of mutants on different graphs. Among these, we show that our method consistently outperforms Monte Carlo simulations in speed by several orders of magnitude. Finally we show how our approach can provide insight into synaptic competition in neurology.
1301.2542
Enhancing the retrieval performance by combing the texture and edge features
cs.CV cs.IR
In this paper, anew algorithm which is based on geometrical moments and local binary patterns (LBP) for content based image retrieval (CBIR) is proposed. In geometrical moments, each vector is compared with the all other vectors for edge map generation. The same concept is utilized at LBP calculation which is generating nine LBP patterns from a given 3x3 pattern. Finally, nine LBP histograms are calculated which are used as a feature vector for image retrieval. Moments are important features used in recognition of different types of images. Two experiments have been carried out for proving the worth of our algorithm. The results after being investigated shows a significant improvement in terms of their evaluation measures as compared to LBP and other existing transform domain techniques.
1301.2556
Information field theory
astro-ph.IM cs.IT math.IT physics.data-an stat.ML
Non-linear image reconstruction and signal analysis deal with complex inverse problems. To tackle such problems in a systematic way, I present information field theory (IFT) as a means of Bayesian, data based inference on spatially distributed signal fields. IFT is a statistical field theory, which permits the construction of optimal signal recovery algorithms even for non-linear and non-Gaussian signal inference problems. IFT algorithms exploit spatial correlations of the signal fields and benefit from techniques developed to investigate quantum and statistical field theories, such as Feynman diagrams, re-normalisation calculations, and thermodynamic potentials. The theory can be used in many areas, and applications in cosmology and numerics are presented.
1301.2561
Modeling complex systems with adaptive networks
cs.SI nlin.AO physics.soc-ph
Adaptive networks are a novel class of dynamical networks whose topologies and states coevolve. Many real-world complex systems can be modeled as adaptive networks, including social networks, transportation networks, neural networks and biological networks. In this paper, we introduce fundamental concepts and unique properties of adaptive networks through a brief, non-comprehensive review of recent literature on mathematical/computational modeling and analysis of such networks. We also report our recent work on several applications of computational adaptive network modeling and analysis to real-world problems, including temporal development of search and rescue operational networks, automated rule discovery from empirical network evolution data, and cultural integration in corporate merger.
1301.2603
Robust subspace clustering
cs.LG cs.IT math.IT math.OC math.ST stat.ML stat.TH
Subspace clustering refers to the task of finding a multi-subspace representation that best fits a collection of points taken from a high-dimensional space. This paper introduces an algorithm inspired by sparse subspace clustering (SSC) [In IEEE Conference on Computer Vision and Pattern Recognition, CVPR (2009) 2790-2797] to cluster noisy data, and develops some novel theory demonstrating its correctness. In particular, the theory uses ideas from geometric functional analysis to show that the algorithm can accurately recover the underlying subspaces under minimal requirements on their orientation, and on the number of samples per subspace. Synthetic as well as real data experiments complement our theoretical study, illustrating our approach and demonstrating its effectiveness.
1301.2604
Structure and function in flow networks
physics.soc-ph cs.SI
This Letter presents a unified approach for the fundamental relationship between structure and function in flow networks by solving analytically the voltages in a resistor network, transforming the network structure to an effective all-to-all topology, and then measuring the resultant flows. Moreover, it defines a way to study the structural resilience of the graph and to detect possible communities.
1301.2609
Learning to Optimize Via Posterior Sampling
cs.LG
This paper considers the use of a simple posterior sampling algorithm to balance between exploration and exploitation when learning to optimize actions such as in multi-armed bandit problems. The algorithm, also known as Thompson Sampling, offers significant advantages over the popular upper confidence bound (UCB) approach, and can be applied to problems with finite or infinite action spaces and complicated relationships among action rewards. We make two theoretical contributions. The first establishes a connection between posterior sampling and UCB algorithms. This result lets us convert regret bounds developed for UCB algorithms into Bayesian regret bounds for posterior sampling. Our second theoretical contribution is a Bayesian regret bound for posterior sampling that applies broadly and can be specialized to many model classes. This bound depends on a new notion we refer to as the eluder dimension, which measures the degree of dependence among action rewards. Compared to UCB algorithm Bayesian regret bounds for specific model classes, our general bound matches the best available for linear models and is stronger than the best available for generalized linear models. Further, our analysis provides insight into performance advantages of posterior sampling, which are highlighted through simulation results that demonstrate performance surpassing recently proposed UCB algorithms.
1301.2613
An Analytical Framework for Heterogeneous Partial Feedback Design in Heterogeneous Multicell OFDMA Networks
cs.IT math.IT
The inherent heterogeneous structure resulting from user densities and large scale channel effects motivates heterogeneous partial feedback design in heterogeneous networks. In such emerging networks, a distributed scheduling policy which enjoys multiuser diversity as well as maintains fairness among users is favored for individual user rate enhancement and guarantees. For a system employing the cumulative distribution function based scheduling, which satisfies the two above mentioned desired features, we develop an analytical framework to investigate heterogeneous partial feedback in a general OFDMA-based heterogeneous multicell employing the best-M partial feedback strategy. Exact sum rate analysis is first carried out and closed form expressions are obtained by a novel decomposition of the probability density function of the selected user's signal-to-interference-plus-noise ratio. To draw further insight, we perform asymptotic analysis using extreme value theory to examine the effect of partial feedback on the randomness of multiuser diversity, show the asymptotic optimality of best-1 feedback, and derive an asymptotic approximation for the sum rate in order to determine the minimum required partial feedback.
1301.2628
Robust Text Detection in Natural Scene Images
cs.CV cs.IR cs.LG
Text detection in natural scene images is an important prerequisite for many content-based image analysis tasks. In this paper, we propose an accurate and robust method for detecting texts in natural scene images. A fast and effective pruning algorithm is designed to extract Maximally Stable Extremal Regions (MSERs) as character candidates using the strategy of minimizing regularized variations. Character candidates are grouped into text candidates by the ingle-link clustering algorithm, where distance weights and threshold of the clustering algorithm are learned automatically by a novel self-training distance metric learning algorithm. The posterior probabilities of text candidates corresponding to non-text are estimated with an character classifier; text candidates with high probabilities are then eliminated and finally texts are identified with a text classifier. The proposed system is evaluated on the ICDAR 2011 Robust Reading Competition dataset; the f measure is over 76% and is significantly better than the state-of-the-art performance of 71%. Experimental results on a publicly available multilingual dataset also show that our proposed method can outperform the other competitive method with the f measure increase of over 9 percent. Finally, we have setup an online demo of our proposed scene text detection system at http://kems.ustb.edu.cn/learning/yin/dtext.
1301.2629
Upper-Bounding the Capacity of Relay Communications - Part I
cs.IT math.IT
This paper focuses on the capacity of point-to-point relay communications wherein the transmitter is assisted by an intermediate relay. We detail the mathematical model of cutset and amplify and forward (AF) relaying strategy. We present the upper bound capacity of each relaying strategy from information theory viewpoint and also in networks with Gaussian channels. We exemplify various outer region capacities of the addressed strategies with two different case studies. The results exhibit that in low signal-to-noise ratio (SNR) environments the cutset performance is better than amplify and forward strategy.
1301.2634
Blind source separation methods for deconvolution of complex signals in cancer biology
q-bio.QM cs.CE q-bio.GN
Two blind source separation methods (Independent Component Analysis and Non-negative Matrix Factorization), developed initially for signal processing in engineering, found recently a number of applications in analysis of large-scale data in molecular biology. In this short review, we present the common idea behind these methods, describe ways of implementing and applying them and point out to the advantages compared to more traditional statistical approaches. We focus more specifically on the analysis of gene expression in cancer. The review is finalized by listing available software implementations for the methods described.
1301.2638
Computational Intelligence for Deepwater Reservoir Depositional Environments Interpretation
cs.NE physics.geo-ph
Predicting oil recovery efficiency of a deepwater reservoir is a challenging task. One approach to characterize a deepwater reservoir and to predict its producibility is by analyzing its depositional information. This research proposes a deposition-based stratigraphic interpretation framework for deepwater reservoir characterization. In this framework, one critical task is the identification and labeling of the stratigraphic components in the reservoir, according to their depositional environments. This interpretation process is labor intensive and can produce different results depending on the stratigrapher who performs the analysis. To relieve stratigrapher's workload and to produce more consistent results, we have developed a novel methodology to automate this process using various computational intelligence techniques. Using a well log data set, we demonstrate that the developed methodology and the designed workflow can produce finite state transducer models that interpret deepwater reservoir depositional environments adequately.
1301.2648
A New Distributed Localization Method for Sensor Networks
cs.IT cs.DC cs.NI math.IT
This paper studies the problem of determining the sensor locations in a large sensor network using relative distance (range) measurements only. Our work follows from a seminal paper by Khan et al. [1] where a distributed algorithm, known as DILOC, for sensor localization is given using the barycentric coordinate. A main limitation of the DILOC algorithm is that all sensor nodes must be inside the convex hull of the anchor nodes. In this paper, we consider a general sensor network without the convex hull assumption, which incurs challenges in determining the sign pattern of the barycentric coordinate. A criterion is developed to address this issue based on available distance measurements. Also, a new distributed algorithm is proposed to guarantee the asymptotic localization of all localizable sensor nodes.
1301.2655
Functional Regularized Least Squares Classi cation with Operator-valued Kernels
cs.LG stat.ML
Although operator-valued kernels have recently received increasing interest in various machine learning and functional data analysis problems such as multi-task learning or functional regression, little attention has been paid to the understanding of their associated feature spaces. In this paper, we explore the potential of adopting an operator-valued kernel feature space perspective for the analysis of functional data. We then extend the Regularized Least Squares Classification (RLSC) algorithm to cover situations where there are multiple functions per observation. Experiments on a sound recognition problem show that the proposed method outperforms the classical RLSC algorithm.
1301.2656
Multiple functional regression with both discrete and continuous covariates
stat.ML cs.LG
In this paper we present a nonparametric method for extending functional regression methodology to the situation where more than one functional covariate is used to predict a functional response. Borrowing the idea from Kadri et al. (2010a), the method, which support mixed discrete and continuous explanatory variables, is based on estimating a function-valued function in reproducing kernel Hilbert spaces by virtue of positive operator-valued kernels.
1301.2659
A Triclustering Approach for Time Evolving Graphs
cs.LG cs.SI stat.ML
This paper introduces a novel technique to track structures in time evolving graphs. The method is based on a parameter free approach for three-dimensional co-clustering of the source vertices, the target vertices and the time. All these features are simultaneously segmented in order to build time segments and clusters of vertices whose edge distributions are similar and evolve in the same way over the time segments. The main novelty of this approach lies in that the time segments are directly inferred from the evolution of the edge distribution between the vertices, thus not requiring the user to make an a priori discretization. Experiments conducted on a synthetic dataset illustrate the good behaviour of the technique, and a study of a real-life dataset shows the potential of the proposed approach for exploratory data analysis.
1301.2678
Verification of Agent-Based Artifact Systems
cs.MA cs.AI cs.LO
Artifact systems are a novel paradigm for specifying and implementing business processes described in terms of interacting modules called artifacts. Artifacts consist of data and lifecycles, accounting respectively for the relational structure of the artifacts' states and their possible evolutions over time. In this paper we put forward artifact-centric multi-agent systems, a novel formalisation of artifact systems in the context of multi-agent systems operating on them. Differently from the usual process-based models of services, the semantics we give explicitly accounts for the data structures on which artifact systems are defined. We study the model checking problem for artifact-centric multi-agent systems against specifications written in a quantified version of temporal-epistemic logic expressing the knowledge of the agents in the exchange. We begin by noting that the problem is undecidable in general. We then identify two noteworthy restrictions, one syntactical and one semantical, that enable us to find bisimilar finite abstractions and therefore reduce the model checking problem to the instance on finite models. Under these assumptions we show that the model checking problem for these systems is EXPSPACE-complete. We then introduce artifact-centric programs, compact and declarative representations of the programs governing both the artifact system and the agents. We show that, while these in principle generate infinite-state systems, under natural conditions their verification problem can be solved on finite abstractions that can be effectively computed from the programs. Finally we exemplify the theoretical results of the paper through a mainstream procurement scenario from the artifact systems literature.
1301.2683
BliStr: The Blind Strategymaker
cs.AI cs.LG cs.LO
BliStr is a system that automatically develops strategies for E prover on a large set of problems. The main idea is to interleave (i) iterated low-timelimit local search for new strategies on small sets of similar easy problems with (ii) higher-timelimit evaluation of the new strategies on all problems. The accumulated results of the global higher-timelimit runs are used to define and evolve the notion of "similar easy problems", and to control the selection of the next strategy to be improved. The technique was used to significantly strengthen the set of E strategies used by the MaLARea, PS-E, E-MaLeS, and E systems in the CASC@Turing 2012 competition, particularly in the Mizar division. Similar improvement was obtained on the problems created from the Flyspeck corpus.
1301.2696
Reduced-Rank Space-Time Interference Suppression with Joint Iterative Least Squares Algorithms for Spread Spectrum Systems
cs.IT math.IT
This paper presents novel adaptive space-time reduced-rank interference suppression least squares algorithms based on joint iterative optimization of parameter vectors. The proposed space-time reduced-rank scheme consists of a joint iterative optimization of a projection matrix that performs dimensionality reduction and an adaptive reduced-rank parameter vector that yields the symbol estimates. The proposed techniques do not require singular value decomposition (SVD) and automatically find the best set of basis for reduced-rank processing. We present least squares (LS) expressions for the design of the projection matrix and the reduced-rank parameter vector and we conduct an analysis of the convergence properties of the LS algorithms. We then develop recursive least squares (RLS) adaptive algorithms for their computationally efficient estimation and an algorithm for automatically adjusting the rank of the proposed scheme. A convexity analysis of the LS algorithms is carried out along with the development of a proof of convergence for the proposed algorithms. Simulations for a space-time interference suppression application with a DS-CDMA system show that the proposed scheme outperforms in convergence and tracking the state-of-the-art reduced-rank schemes at a comparable complexity.
1301.2697
Adaptive Reduced-Rank Equalization Algorithms Based on Alternating Optimization Design Techniques for Multi-Antenna Systems
cs.IT math.IT
This paper presents a novel adaptive reduced-rank {multi-input multi-output} (MIMO) equalization scheme and algorithms based on alternating optimization design techniques for MIMO spatial multiplexing systems. The proposed reduced-rank equalization structure consists of a joint iterative optimization of two equalization stages, namely, a transformation matrix that performs dimensionality reduction and a reduced-rank estimator that retrieves the desired transmitted symbol. The proposed reduced-rank architecture is incorporated into an equalization structure that allows both decision feedback and linear schemes for mitigating the inter-antenna and inter-symbol interference. We develop alternating least squares (LS) expressions for the design of the transformation matrix and the reduced-rank estimator along with computationally efficient alternating recursive least squares (RLS) adaptive estimation algorithms. We then present an algorithm for automatically adjusting the model order of the proposed scheme. An analysis of the LS algorithms is carried out along with sufficient conditions for convergence and a proof of convergence of the proposed algorithms to the reduced-rank Wiener filter. Simulations show that the proposed equalization algorithms outperform the existing reduced-rank and full-rank algorithms, while requiring a comparable computational cost.
1301.2698
Evaluating community structure in large network with random walks
cs.SI physics.soc-ph
Community structure is one of the most important properties of networks. Most community algorithms are not suitable for large networks because of their time consuming. In fact there are lots of networks with millons even billons of nodes. In such case, most algorithms running in time O(n2logn) or even larger are not practical. What we need are linear or approximately linear time algorithm. Rising in response to such needs, we propose a quick methods to evaluate community structure in networks and then put forward a local community algorithm with nearly linear time based on random walks. Using our community evaluating measure, we could find some difference results from measures used before, i.e., the Newman Modularity. Our algorithm are effective in small benchmark networks with small less accuracy than more complex algorithms but a great of advantage in time consuming for large networks, especially super large networks.
1301.2710
A High-Order Sliding Mode Observer: Torpedo Guidance Application
cs.SY
The guidance of a torpedo represents a hard task because of the smooth nonlinear aspect of this system and because of the extreme external disturbances. The torpedo guidance reposes on the speed and the position control. In fact, the control approach which is very solicited for the electromechanical systems is the sliding mode control (SMC) which proved its effectiveness through the different studies. The SMC is robust versus disturbances and model uncertainties; however, a sharp discontinuous control is needed which induces the chattering phenomenon. The angular velocity measurement is a hard task because of the high level of disturbances. In this way, the sliding mode observer could be a solution for the velocity estimation instead of a sensor. This article deals with torpedo guidance by SMC to reach the desired path in a short time and with high precision quality. Simulation results show that this control strategy and observer can attain excellent control performances with no chattering problem.
1301.2711
A Sliding Mode Multimodel Control for a Sensorless Photovoltaic System
cs.SY hep-ex
In this work we will talk about a new control test using the sliding mode control with a nonlinear sliding mode observer, which are very solicited in tracking problems, for a sensorless photovoltaic panel. In this case, the panel system will has as a set point the sun position at every second during the day for a period of five years; then the tracker, using sliding mode multimodel controller and a sliding mode observer, will track these positions to make the sunrays orthogonal to the photovoltaic cell that produces more energy. After sunset, the tracker goes back to the initial position (which of sunrise). Experimental measurements show that this autonomic dual axis Sun Tracker increases the power production by over 40%.
1301.2714
A Sliding Mode-Multimodel Control with Sliding Mode Observer for a Sensorless Pumping System
cs.SY
This work deals with the design of a sliding mode observer with a multi-surfaces sliding mode multimodel control (SM-MMC) for a mechanical sensorless pumping system. The observer is designed to estimate the speed and the mechanical position of the DC motor operating in the process. Robustness tests validated by simulation show the effectiveness of the sliding mode observer associated with this control approach (SM-MMC).
1301.2715
Binocular disparity as an explanation for the moon illusion
cs.CV physics.pop-ph
We present another explanation for the moon illusion, the phenomenon in which the moon looks larger near the horizon than near the zenith. In our model of the moon illusion, the sky is considered a spatially-contiguous and geometrically-smooth surface. When an object such as the moon breaks the contiguity of the surface, instead of perceiving the object as appearing through a hole in the surface, humans perceive an occlusion of the surface. Binocular vision dictates that the moon is distant, but this perception model contradicts our binocular vision, dictating that the moon is closer than the sky. To resolve the contradiction, the brain distorts the projections of the moon to increase the binocular disparity, which results in an increase in the perceived size of the moon. The degree of distortion depends upon the apparent distance to the sky, which is influenced by the surrounding objects and the condition of the sky. As the apparent distance to the sky decreases, the illusion becomes stronger. At the horizon, apparent distance to the sky is minimal, whereas at the zenith, few distance cues are present, causing difficulty with distance estimation and weakening the illusion.
1301.2722
Distributed Consensus Formation Through Unconstrained Gossiping
cs.SY math.OC
Gossip algorithms are widely used to solve the distributed consensus problem, but issues can arise when nodes receive multiple signals either at the same time or before they are able to finish processing their current work load. Specifically, a node may assume a new state that represents a linear combination of all received signals; even if such a state makes no sense in the problem domain. As a solution to this problem, we introduce the notion of conflict resolution for gossip algorithms and prove that their application leads to a valid consensus state when the underlying communication network possesses certain properties. We also introduce a methodology based on absorbing Markov chains for analyzing gossip algorithms that make use of these conflict resolution algorithms. This technique allows us to calculate both the probabilities of converging to a specific consensus state and the time that such convergence is expected to take. Finally, we make use of simulation to validate our methodology and explore the temporal behavior of gossip algorithms as the size of the network, the number of states per node, and the network density increase.
1301.2725
Robust High Dimensional Sparse Regression and Matching Pursuit
stat.ML cs.IT cs.LG math.IT math.ST stat.TH
We consider high dimensional sparse regression, and develop strategies able to deal with arbitrary -- possibly, severe or coordinated -- errors in the covariance matrix $X$. These may come from corrupted data, persistent experimental errors, or malicious respondents in surveys/recommender systems, etc. Such non-stochastic error-in-variables problems are notoriously difficult to treat, and as we demonstrate, the problem is particularly pronounced in high-dimensional settings where the primary goal is {\em support recovery} of the sparse regressor. We develop algorithms for support recovery in sparse regression, when some number $n_1$ out of $n+n_1$ total covariate/response pairs are {\it arbitrarily (possibly maliciously) corrupted}. We are interested in understanding how many outliers, $n_1$, we can tolerate, while identifying the correct support. To the best of our knowledge, neither standard outlier rejection techniques, nor recently developed robust regression algorithms (that focus only on corrupted response variables), nor recent algorithms for dealing with stochastic noise or erasures, can provide guarantees on support recovery. Perhaps surprisingly, we also show that the natural brute force algorithm that searches over all subsets of $n$ covariate/response pairs, and all subsets of possible support coordinates in order to minimize regression error, is remarkably poor, unable to correctly identify the support with even $n_1 = O(n/k)$ corrupted points, where $k$ is the sparsity. This is true even in the basic setting we consider, where all authentic measurements and noise are independent and sub-Gaussian. In this setting, we provide a simple algorithm -- no more computationally taxing than OMP -- that gives stronger performance guarantees, recovering the support with up to $n_1 = O(n/(\sqrt{k} \log p))$ corrupted points, where $p$ is the dimension of the signal to be recovered.
1301.2750
Load-aware Channel Selection for 802.11 WLANs with Limited Measurement
cs.NI cs.IT cs.PF math.IT
It has been known that load unaware channel selection in 802.11 networks results in high level interference, and can significantly reduce the network throughput. In current implementation, the only way to determine the traffic load on a channel is to measure that channel for a certain duration of time. Therefore, in order to find the best channel with the minimum load all channels have to be measured, which is costly and can cause unacceptable communication interruptions between the AP and the stations. In this paper, we propose a learning based approach which aims to find the channel with the minimum load by measuring only limited number of channels. Our method uses Gaussian Process Regressing to accurately track the traffic load on each channel based on the previous measured load. We confirm the performance of our algorithm by using experimental data, and show that the time consumed for the load measurement can be reduced up to 46% compared to the case where all channels are monitored.
1301.2774
Crowd Labeling: a survey
cs.AI
Recently, there has been a burst in the number of research projects on human computation via crowdsourcing. Multiple choice (or labeling) questions could be referred to as a common type of problem which is solved by this approach. As an application, crowd labeling is applied to find true labels for large machine learning datasets. Since crowds are not necessarily experts, the labels they provide are rather noisy and erroneous. This challenge is usually resolved by collecting multiple labels for each sample, and then aggregating them to estimate the true label. Although the mechanism leads to high-quality labels, it is not actually cost-effective. As a result, efforts are currently made to maximize the accuracy in estimating true labels, while fixing the number of acquired labels. This paper surveys methods to aggregate redundant crowd labels in order to estimate unknown true labels. It presents a unified statistical latent model where the differences among popular methods in the field correspond to different choices for the parameters of the model. Afterwards, algorithms to make inference on these models will be surveyed. Moreover, adaptive methods which iteratively collect labels based on the previously collected labels and estimated models will be discussed. In addition, this paper compares the distinguished methods, and provides guidelines for future work required to address the current open issues.
1301.2785
A comparison of SVM and RVM for Document Classification
cs.IR cs.LG
Document classification is a task of assigning a new unclassified document to one of the predefined set of classes. The content based document classification uses the content of the document with some weighting criteria to assign it to one of the predefined classes. It is a major task in library science, electronic document management systems and information sciences. This paper investigates document classification by using two different classification techniques (1) Support Vector Machine (SVM) and (2) Relevance Vector Machine (RVM). SVM is a supervised machine learning technique that can be used for classification task. In its basic form, SVM represents the instances of the data into space and tries to separate the distinct classes by a maximum possible wide gap (hyper plane) that separates the classes. On the other hand RVM uses probabilistic measure to define this separation space. RVM uses Bayesian inference to obtain succinct solution, thus RVM uses significantly fewer basis functions. Experimental studies on three standard text classification datasets reveal that although RVM takes more training time, its classification is much better as compared to SVM.
1301.2811
Cutting Recursive Autoencoder Trees
cs.CL cs.AI
Deep Learning models enjoy considerable success in Natural Language Processing. While deep architectures produce useful representations that lead to improvements in various tasks, they are often difficult to interpret. This makes the analysis of learned structures particularly difficult. In this paper, we rely on empirical tests to see whether a particular structure makes sense. We present an analysis of the Semi-Supervised Recursive Autoencoder, a well-known model that produces structural representations of text. We show that for certain tasks, the structure of the autoencoder can be significantly reduced without loss of classification accuracy and we evaluate the produced structures using human judgment.
1301.2820
Clustering Learning for Robotic Vision
cs.CV
We present the clustering learning technique applied to multi-layer feedforward deep neural networks. We show that this unsupervised learning technique can compute network filters with only a few minutes and a much reduced set of parameters. The goal of this paper is to promote the technique for general-purpose robotic vision systems. We report its use in static image datasets and object tracking datasets. We show that networks trained with clustering learning can outperform large networks trained for many hours on complex datasets.
1301.2831
Beyond consistent reconstructions: optimality and sharp bounds for generalized sampling, and application to the uniform resampling problem
math.NA cs.IT math.IT
Generalized sampling is a recently developed linear framework for sampling and reconstruction in separable Hilbert spaces. It allows one to recover any element in any finite-dimensional subspace given finitely many of its samples with respect to an arbitrary frame. Unlike more common approaches for this problem, such as the consistent reconstruction technique of Eldar et al, it leads to completely stable numerical methods possessing both guaranteed stability and accuracy. The purpose of this paper is twofold. First, we give a complete and formal analysis of generalized sampling, the main result of which being the derivation of new, sharp bounds for the accuracy and stability of this approach. Such bounds improve those given previously, and result in a necessary and sufficient condition, the stable sampling rate, which guarantees a priori a good reconstruction. Second, we address the topic of optimality. Under some assumptions, we show that generalized sampling is an optimal, stable reconstruction. Correspondingly, whenever these assumptions hold, the stable sampling rate is a universal quantity. In the final part of the paper we illustrate our results by applying generalized sampling to the so-called uniform resampling problem.
1301.2840
Unsupervised Feature Learning for low-level Local Image Descriptors
cs.CV cs.LG stat.ML
Unsupervised feature learning has shown impressive results for a wide range of input modalities, in particular for object classification tasks in computer vision. Using a large amount of unlabeled data, unsupervised feature learning methods are utilized to construct high-level representations that are discriminative enough for subsequently trained supervised classification algorithms. However, it has never been \emph{quantitatively} investigated yet how well unsupervised learning methods can find \emph{low-level representations} for image patches without any additional supervision. In this paper we examine the performance of pure unsupervised methods on a low-level correspondence task, a problem that is central to many Computer Vision applications. We find that a special type of Restricted Boltzmann Machines (RBMs) performs comparably to hand-crafted descriptors. Additionally, a simple binarization scheme produces compact representations that perform better than several state-of-the-art descriptors.
1301.2851
Efficient algorithm to study interconnected networks
physics.comp-ph cond-mat.stat-mech cs.SI physics.soc-ph
Interconnected networks have been shown to be much more vulnerable to random and targeted failures than isolated ones, raising several interesting questions regarding the identification and mitigation of their risk. The paradigm to address these questions is the percolation model, where the resilience of the system is quantified by the dependence of the size of the largest cluster on the number of failures. Numerically, the major challenge is the identification of this cluster and the calculation of its size. Here, we propose an efficient algorithm to tackle this problem. We show that the algorithm scales as O(N log N), where N is the number of nodes in the network, a significant improvement compared to O(N^2) for a greedy algorithm, what permits studying much larger networks. Our new strategy can be applied to any network topology and distribution of interdependencies, as well as any sequence of failures.
1301.2857
SpeedRead: A Fast Named Entity Recognition Pipeline
cs.CL
Online content analysis employs algorithmic methods to identify entities in unstructured text. Both machine learning and knowledge-base approaches lie at the foundation of contemporary named entities extraction systems. However, the progress in deploying these approaches on web-scale has been been hampered by the computational cost of NLP over massive text corpora. We present SpeedRead (SR), a named entity recognition pipeline that runs at least 10 times faster than Stanford NLP pipeline. This pipeline consists of a high performance Penn Treebank- compliant tokenizer, close to state-of-art part-of-speech (POS) tagger and knowledge-based named entity recognizer.
1301.2860
Rateless Resilient Network Coding Against Byzantine Adversaries
cs.IT math.IT
This paper considers rateless network error correction codes for reliable multicast in the presence of adversarial errors. Most existing network error correction codes are designed for a given network capacity and maximum number of errors known a priori to the encoder and decoder. However, in certain practical settings it may be necessary to operate without such a priori knowledge. We present rateless coding schemes for two adversarial models, where the source sends more redundancy over time, until decoding succeeds. The first model assumes there is a secret channel between the source and the destination that the adversaries cannot overhear. The rate of the channel is negligible compared to the main network. In the second model, instead of a secret channel, the source and destination share random secrets independent of the input information. The amount of secret information required is negligible compared to the amount of information sent. Both schemes are optimal in that decoding succeeds with high probability when the total amount of information received by the sink satisfies the cut set bound with respect to the amount of message and error information. The schemes are distributed, polynomial-time and end-to-end in that other than the source and destination nodes, other intermediate nodes carry out classical random linear network coding.
1301.2866
Generalized Multiscale Finite Element Methods (GMsFEM)
math.NA cs.CE cs.NA math.AP
In this paper, we propose a general approach called Generalized Multiscale Finite Element Method (GMsFEM) for performing multiscale simulations for problems without scale separation over a complex input space. As in multiscale finite element methods (MsFEMs), the main idea of the proposed approach is to construct a small dimensional local solution space that can be used to generate an efficient and accurate approximation to the multiscale solution with a potentially high dimensional input parameter space. In the proposed approach, we present a general procedure to construct the offline space that is used for a systematic enrichment of the coarse solution space in the online stage. The enrichment in the online stage is performed based on a spectral decomposition of the offline space. In the online stage, for any input parameter, a multiscale space is constructed to solve the global problem on a coarse grid. The online space is constructed via a spectral decomposition of the offline space and by choosing the eigenvectors corresponding to the largest eigenvalues. The computational saving is due to the fact that the construction of the online multiscale space for any input parameter is fast and this space can be re-used for solving the forward problem with any forcing and boundary condition. Compared with the other approaches where global snapshots are used, the local approach that we present in this paper allows us to eliminate unnecessary degrees of freedom on a coarse-grid level. We present various examples in the paper and some numerical results to demonstrate the effectiveness of our method.
1301.2884
Wavelet-based Scale Saliency
cs.CV
Both pixel-based scale saliency (PSS) and basis project methods focus on multiscale analysis of data content and structure. Their theoretical relations and practical combination are previously discussed. However, no models have ever been proposed for calculating scale saliency on basis-projected descriptors since then. This paper extend those ideas into mathematical models and implement them in the wavelet-based scale saliency (WSS). While PSS uses pixel-value descriptors, WSS treats wavelet sub-bands as basis descriptors. The paper discusses different wavelet descriptors: discrete wavelet transform (DWT), wavelet packet transform (DWPT), quaternion wavelet transform (QWT) and best basis quaternion wavelet packet transform (QWPTBB). WSS saliency maps of different descriptors are generated and compared against other saliency methods by both quantitative and quanlitative methods. Quantitative results, ROC curves, AUC values and NSS values are collected from simulations on Bruce and Kootstra image databases with human eye-tracking data as ground-truth. Furthermore, qualitative visual results of saliency maps are analyzed and compared against each other as well as eye-tracking data inclusive in the databases.
1301.2907
Conditions on the generator for forging ElGamal signature
cs.CR cs.IT math.IT
This paper describes new conditions on parameters selection that lead to an efficient algorithm for forging ElGamal digital signature. Our work is inspired by Bleichenbacher's ideas.
1301.2935
Novel Subcarrier-pair based Opportunistic DF Protocol for Cooperative Downlink OFDMA
cs.IT cs.NI math.IT
A novel subcarrier-pair based opportunistic DF protocol is proposed for cooperative downlink OFDMA transmission aided by a decode-and-forward (DF) relay. Specifically, user message bits are transmitted in two consecutive equal-duration time slots. A subcarrier in the first slot can be paired with a subcarrier in the second slot for the DF relay-aided transmission to a user. In particular, the source and the relay can transmit simultaneously to implement beamforming at the subcarrier in the second slot for the relay-aided transmission. Each unpaired subcarrier in either the first or second slot is used by the source for direct transmission to a user without the relay's assistance. The sum rate maximized resource allocation (RA) problem is addressed for this protocol under a total power constraint. It is shown that the novel protocol leads to a maximum sum rate greater than or equal to that for a benchmark one, which does not allow the source to implement beamforming at the subcarrier in the second slot for the relay-aided transmission. Then, a polynomial-complexity RA algorithm is developed to find an (at least approximately) optimum resource allocation (i.e., source/relay power, subcarrier pairing and assignment to users) for either the proposed or benchmark protocol. Numerical experiments illustrate that the novel protocol can lead to a much greater sum rate than the benchmark one.
1301.2941
Power minimization for OFDM Transmission with Subcarrier-pair based Opportunistic DF Relaying
cs.IT cs.NI math.IT
This paper develops a sum-power minimized resource allocation (RA) algorithm subject to a sum-rate constraint for cooperative orthogonal frequency division modulation (OFDM) transmission with subcarrier-pair based opportunistic decode-and-forward (DF) relaying. The improved DF protocol first proposed in [1] is used with optimized subcarrier pairing. Instrumental to the RA algorithm design is appropriate definition of variables to represent source/relay power allocation, subcarrier pairing and transmission-mode selection elegantly, so that after continuous relaxation, the dual method and the Hungarian algorithm can be used to find an (at least approximately) optimum RA with polynomial complexity. Moreover, the bisection method is used to speed up the search of the optimum Lagrange multiplier for the dual method. Numerical results are shown to illustrate the power-reduction benefit of the improved DF protocol with optimized subcarrier pairing.
1301.2944
Competing of Sznajd and voter dynamics in the Watts-Strogatz network
physics.soc-ph cs.SI physics.comp-ph
We investigate the Watts-Strogatz network with the clustering coefficient C dependent on the rewiring probability. The network is an area of two opposite contact processes, where nodes can be in two states, S or D. One of the processes is governed by the Sznajd dynamics: if there are two connected nodes in D-state, all their neighbors become D with probability p. For the opposite process it is sufficient to have only one neighbor in state S; this transition occurs with probability 1. The concentration of S-nodes changes abruptly at given value of the probability p. The result is that for small p, in clusterized networks the activation of S nodes prevails. This result is explained by a comparison of two limit cases: the Watts-Strogatz network without rewiring, where C=0.5, and the Bethe lattice where C=0.
1301.2952
The Anatomy of a Scientific Rumor
physics.soc-ph cond-mat.stat-mech cs.SI
The announcement of the discovery of a Higgs boson-like particle at CERN will be remembered as one of the milestones of the scientific endeavor of the 21st century. In this paper we present a study of information spreading processes on Twitter before, during and after the announcement of the discovery of a new particle with the features of the elusive Higgs boson on 4th July 2012. We report evidence for non-trivial spatio-temporal patterns in user activities at individual and global level, such as tweeting, re-tweeting and replying to existing tweets. We provide a possible explanation for the observed time-varying dynamics of user activities during the spreading of this scientific "rumor". We model the information spreading in the corresponding network of individuals who posted a tweet related to the Higgs boson discovery. Finally, we show that we are able to reproduce the global behavior of about 500,000 individuals with remarkable accuracy.
1301.2957
Characters and patterns of communities in networks
cs.SI physics.soc-ph
In this paper, we propose some new notions to characterize and analyze the communities. The new notions are general characters of the communities or local structures of networks. At first, we introduce the notions of internal dominating set and external dominating set of a community. We show that most communities in real networks have a small internal dominating set and a small external dominating set, and that the internal dominating set of a community keeps much of the information of the community. Secondly, based on the notions of the internal dominating set and the external dominating set, we define an internal slope (ISlope, for short) and an external slope (ESlope, for short) to measure the internal heterogeneity and external heterogeneity of a community respectively. We show that the internal slope (ISlope) of a community largely determines the structure of the community, that most communities in real networks are heterogeneous, meaning that most of the communities have a core/periphery structure, and that both ISlopes and ESlopes (reflecting the structure of communities) of all the communities of a network approximately follow a normal distribution. Therefore typical values of both ISolpes and ESoples of all the communities of a given network are in a narrow interval, and there is only a small number of communities having ISlopes or ESlopes out of the range of typical values of the ISlopes and ESlopes of the network. Finally, we show that all the communities of the real networks we studied, have a three degree separation phenomenon, that is, the average distance of communities is approximately 3, implying a general property of true communities for many real networks, and that good community finding algorithms find communities that amplify clustering coefficients of the networks, for many real networks.
1301.2959
New elements for a network (including brain) general theory during learning period
nlin.AO cs.NE nlin.CD
This study deals with the evolution of the so called 'intelligent' networks (insect society without leader, cells of an organism, brain,...) during their learning period. First we summarize briefly the Version 2 (published in French), whose the main characteristics are: 1) A network connected to its environment is considered as immersed into an information field created by this environment which so dictates to it the learning constraints. 2) The used formalism draws one's inspiration from the one of the Quantum field theory (Principle of stationary action, gauge fields, invariance by symmetry transformations,...). 3) We obtain Lagrange equations whose solutions describe the network evolution during the whole learning period. 4) Then, while proceeding with the same formalism inspiration, we suggest other study ways capable of evolving the knowledge in the considered scope. In a second part, after a reminder of the points to be improved, we exhibit the Version 5 which brings, we think, relevant improvements. Indeed: 5) We consider the weighted averages of the variables; this introduces probabilities. 6) We define two observables (L average of information flux and A activity of the network) which could be measured and so be compared with experimental results. 7) We find that L , weighted average of information flows, is an invariant. 8) Finally, we propose two expressions for the conactance, from which we deduce the corresponding Lagrange equations which have to be solved to know the evolution of the considered weighted averages. But, at the present stage, we think that we can progress only by carrying out experiments (see projects like Human brain project) and discovering invariants, symmetries which would allow us, like in Physics, to classify networks and above all to understand better the connections between them. Indeed, and that is what we propose among the future research ways, the underlying problem is to understand how, after their learning period, several networks can connect together to produce, in the brain case for instance, what we call mental states.
1301.2995
Measuring Cultural Dynamics Through the Eurovision Song Contest
physics.soc-ph cs.SI physics.data-an
Measuring culture and its dynamics through surveys has important limitations, but the emerging field of computational social science allows us to overcome them by analyzing large-scale datasets. In this article, we study cultural dynamics through the votes in the Eurovision song contest, which are decided by a crowd-based scheme in which viewers vote through mobile phone messages. Taking into account asymmetries and imperfect perception of culture, we measure cultural relations among European countries in terms of cultural affinity. We propose the Friend-or-Foe coefficient, a metric to measure voting biases among participants of a Eurovision contest. We validate how this metric represents cultural affinity through its relation with known cultural distances, and through numerical analysis of biased Eurovision contests. We apply this metric to the historical set of Eurovision contests from 1975 to 2012, finding new patterns of stronger modularity than using votes alone. Furthermore, we define a measure of polarization that, when applied to empirical data, shows a sharp increase within EU countries during 2010 and 2011. We empirically validate the relation between this polarization and economic indicators in the EU, showing how political decisions influence both the economy and the way citizens relate to the culture of other EU members.
1301.3003
On the Vector Linear Solvability of Networks and Discrete Polymatroids
cs.IT math.IT
We consider the vector linear solvability of networks over a field $\mathbb{F}_q.$ It is well known that a scalar linear solution over $\mathbb{F}_q$ exists for a network if and only if the network is \textit{matroidal} with respect to a \textit{matroid} representable over $\mathbb{F}_q.$ A \textit{discrete polymatroid} is the multi-set analogue of a matroid. In this paper, a \textit{discrete polymatroidal} network is defined and it is shown that a vector linear solution over a field $\mathbb{F}_q$ exists for a network if and only if the network is discrete polymatroidal with respect to a discrete polymatroid representable over $\mathbb{F}_q.$ An algorithm to construct networks starting from a discrete polymatroid is provided. Every representation over $\mathbb{F}_q$ for the discrete polymatroid, results in a vector linear solution over $\mathbb{F}_q$ for the constructed network. Examples which illustrate the construction algorithm are provided, in which the resulting networks admit vector linear solution but no scalar linear solution over $\mathbb{F}_q.$
1301.3021
Accurate detection of moving targets via random sensor arrays and Kerdock codes
math.NA cs.IT math.IT
The detection and parameter estimation of moving targets is one of the most important tasks in radar. Arrays of randomly distributed antennas have been popular for this purpose for about half a century. Yet, surprisingly little rigorous mathematical theory exists for random arrays that addresses fundamental question such as how many targets can be recovered, at what resolution, at which noise level, and with which algorithm. In a different line of research in radar, mathematicians and engineers have invested significant effort into the design of radar transmission waveforms which satisfy various desirable properties. In this paper we bring these two seemingly unrelated areas together. Using tools from compressive sensing we derive a theoretical framework for the recovery of targets in the azimuth-range-Doppler domain via random antennas arrays. In one manifestation of our theory we use Kerdock codes as transmission waveforms and exploit some of their peculiar properties in our analysis. Our paper provides two main contributions: (i) We derive the first rigorous mathematical theory for the detection of moving targets using random sensor arrays. (ii) The transmitted waveforms satisfy a variety of properties that are very desirable and important from a practical viewpoint. Thus our approach does not just lead to useful theoretical insights, but is also of practical importance. Various extensions of our results are derived and numerical simulations confirming our theory are presented.
1301.3106
Topological Interference Management through Index Coding
cs.IT math.IT
This work studies linear interference networks, both wired and wireless, with no channel state information at the transmitters (CSIT) except a coarse knowledge of the end-to-end one-hop topology of the network that only allows a distinction between weak (zero) and significant (non-zero) channels and no further knowledge of the channel coefficients' realizations. The network capacity (wired) and DoF (wireless) are found to be bounded above by the capacity of an index coding problem for which the antidote graph is the complement of the given interference graph. The problems are shown to be equivalent under linear solutions. An interference alignment perspective is then used to translate the existing index coding solutions into the wired network capacity and wireless network DoF solutions, as well as to find new and unified solutions to different classes of all three problems.
1301.3118
A parallel implementation of a derivative pricing model incorporating SABR calibration and probability lookup tables
cs.DC cs.CE q-fin.CP
We describe a high performance parallel implementation of a derivative pricing model, within which we introduce a new parallel method for the calibration of the industry standard SABR (stochastic-\alpha \beta \rho) stochastic volatility model using three strike inputs. SABR calibration involves a non-linear three dimensional minimisation and parallelisation is achieved by incorporating several assumptions unique to the SABR class of models. Our calibration method is based on principles of surface intersection, guarantees convergence to a unique solution and operates by iteratively refining a two dimensional grid with local mesh refinement. As part of our pricing model we additionally present a fast parallel iterative algorithm for the creation of dynamically sized cumulative probability lookup tables that are able to cap maximum estimated linear interpolation error. We optimise performance for probability distributions that exhibit clustering of linear interpolation error. We also make an empirical assessment of error propagation through our pricing model as a result of changes in accuracy parameters within the pricing model's multiple algorithmic steps. Algorithms are implemented on a GPU (graphics processing unit) using Nvidia's Fermi architecture. The pricing model targets the evaluation of spread options using copula methods, however the presented algorithms can be applied to a wider class of financial instruments.
1301.3120
An edge density definition of overlapping and weighted graph communities
physics.soc-ph cs.SI
Community detection in networks refers to the process of seeking strongly internally connected groups of nodes which are weakly externally connected. In this work, we introduce and study a community definition based on internal edge density. Beginning with the simple concept that edge density equals number of edges divided by maximal number of edges, we apply this definition to a variety of node and community arrangements to show that our definition yields sensible results. Our community definition is equivalent to that of the Absolute Potts Model community detection method (Phys. Rev. E 81, 046114 (2010)), and the performance of that method validates the usefulness of our definition across a wide variety of network types. We discuss how this definition can be extended to weighted, and multigraphs, and how the definition is capable of handling overlapping communities and local algorithms. We further validate our definition against the recently proposed Affiliation Graph Model (arXiv:1205.6228 [cs.SI]) and show that we can precisely solve these benchmarks. More than proposing an end-all community definition, we explain how studying the detailed properties of community definitions is important in order to validate that definitions do not have negative analytic properties. We urge that community definitions be separated from community detection algorithms and propose that community definitions be further evaluated by criteria such as these.
1301.3154
Coherent Quantum Filtering for Physically Realizable Linear Quantum Plants
quant-ph cs.SY math.OC math.PR
The paper is concerned with a problem of coherent (measurement-free) filtering for physically realizable (PR) linear quantum plants. The state variables of such systems satisfy canonical commutation relations and are governed by linear quantum stochastic differential equations, dynamically equivalent to those of an open quantum harmonic oscillator. The problem is to design another PR quantum system, connected unilaterally to the output of the plant and playing the role of a quantum filter, so as to minimize a mean square discrepancy between the dynamic variables of the plant and the output of the filter. This coherent quantum filtering (CQF) formulation is a simplified feedback-free version of the coherent quantum LQG control problem which remains open despite recent studies. The CQF problem is transformed into a constrained covariance control problem which is treated by using the Frechet differentiation of an appropriate Lagrange function with respect to the matrices of the filter.
1301.3174
Loss Visibility Optimized Real-time Video Transmission over MIMO Systems
cs.IT cs.MM math.IT
The structured nature of video data motivates introducing video-aware decisions that make use of this structure for improved video transmission over wireless networks. In this paper, we introduce an architecture for real-time video transmission over multiple-input multiple-output (MIMO) wireless communication systems using loss visibility side information. We quantify the perceptual importance of a packet through the packet loss visibility and use the loss visibility distribution to provide a notion of relative packet importance. To jointly achieve video quality and low latency, we define the optimization objective function as the throughput weighted by the loss visibility of each packet, a proxy for the total perceptual value of successful packets per unit time. We solve the problem of mapping video packets to MIMO subchannels and adapting per-stream rates to maximize the proposed objective. We show that the solution enables jointly reaping gains in terms of improved video quality and lower latency. Optimized packet-stream mapping enables transmission of more relevant packets over more reliable streams while unequal modulation opportunistically increases the transmission rate on the stronger streams to enable low latency delivery of high priority packets. We extend the solution to capture codebook-based limited feedback and MIMO mode adaptation. Results show that the composite quality and throughput gains are significant under full channel state information as well as limited feedback. Tested on H.264-encoded video sequences, for a 4x4 MIMO with 3 spatial streams, the proposed architecture achieves 8 dB power reduction for the same video quality and supports 2.4x higher throughput due to unequal modulation. Furthermore, the gains are achieved at the expense of few bits of cross-layer overhead rather than a complex cross-layer design.
1301.3192
Matrix Approximation under Local Low-Rank Assumption
cs.LG stat.ML
Matrix approximation is a common tool in machine learning for building accurate prediction models for recommendation systems, text mining, and computer vision. A prevalent assumption in constructing matrix approximations is that the partially observed matrix is of low-rank. We propose a new matrix approximation model where we assume instead that the matrix is only locally of low-rank, leading to a representation of the observed matrix as a weighted sum of low-rank matrices. We analyze the accuracy of the proposed local low-rank modeling. Our experiments show improvements in prediction accuracy in recommendation tasks.
1301.3193
Learning Graphical Model Parameters with Approximate Marginal Inference
cs.LG cs.CV
Likelihood based-learning of graphical models faces challenges of computational-complexity and robustness to model mis-specification. This paper studies methods that fit parameters directly to maximize a measure of the accuracy of predicted marginals, taking into account both model and inference approximations at training time. Experiments on imaging problems suggest marginalization-based learning performs better than likelihood-based approximations on difficult problems where the model being fit is approximate in nature.
1301.3195
Audio Classical Composer Identification by Deep Neural Network
cs.NE cs.IR
Audio Classical Composer Identification (ACC) is an important problem in Music Information Retrieval (MIR) which aims at identifying the composer for audio classical music clips. The famous annual competition, Music Information Retrieval Evaluation eXchange (MIREX), also takes it as one of the four training&testing tasks. We built a hybrid model based on Deep Belief Network (DBN) and Stacked Denoising Autoencoder (SDA) to identify the composer from audio signal. As a matter of copyright, sponsors of MIREX cannot publish their data set. We built a comparable data set to test our model. We got an accuracy of 76.26% in our data set which is better than some pure models and shallow models. We think our method is promising even though we test it in a different data set, since our data set is comparable to that in MIREX by size. We also found that samples from different classes become farther away from each other when transformed by more layers in our model.
1301.3214
The Manifold of Human Emotions
cs.CL
Sentiment analysis predicts the presence of positive or negative emotions in a text document. In this paper, we consider higher dimensional extensions of the sentiment concept, which represent a richer set of human emotions. Our approach goes beyond previous work in that our model contains a continuous manifold rather than a finite set of human emotions. We investigate the resulting model, compare it to psychological observations, and explore its predictive capabilities.
1301.3220
A Low-Complexity Encoding of Quasi-Cyclic Codes Based on Galois Fourier Transform
cs.IT math.IT
The encoding complexity of a general (en,ek) quasi-cyclic code is O[(e^2)(n-k)k]. This paper presents a novel low-complexity encoding algorithm for quasi-cyclic (QC) codes based on matrix transformation. First, a message vector is encoded into a transformed codeword in the transform domain. Then, the transmitted codeword is obtained from the transformed codeword by the inverse Galois Fourier transform. For binary QC codes, a simple and fast mapping is required to post-process the transformed codeword such that the transmitted codeword is binary as well. The complexity of our proposed encoding algorithm is O[e(n-k)k] symbol operations for non-binary codes and O[ek(n-k)(log_2 e)] bit operations for binary codes. These complexities are much lower than their traditional counterpart O[(e^2)(n-k)k]. For example, our complexity of encoding a 64-ary (4095,2160) QC code is only 1.59% of that of traditional encoding, and our complexities of encoding the binary (4095, 2160) and (8176, 7154) QC codes are respectively 9.52% and 1.77% of those of traditional encoding. We also study the application of our low-complexity encoding algorithm to one of the most important subclasses of QC codes, namely QC-LDPC codes, especially when their parity-check matrices are rank deficient.
1301.3224
Efficient Learning of Domain-invariant Image Representations
cs.LG
We present an algorithm that learns representations which explicitly compensate for domain mismatch and which can be efficiently realized as linear classifiers. Specifically, we form a linear transformation that maps features from the target (test) domain to the source (training) domain as part of training the classifier. We optimize both the transformation and classifier parameters jointly, and introduce an efficient cost function based on misclassification loss. Our method combines several features previously unavailable in a single algorithm: multi-class adaptation through representation learning, ability to map across heterogeneous feature spaces, and scalability to large datasets. We present experiments on several image datasets that demonstrate improved accuracy and computational advantages compared to previous approaches.
1301.3226
The Expressive Power of Word Embeddings
cs.LG cs.CL stat.ML
We seek to better understand the difference in quality of the several publicly released embeddings. We propose several tasks that help to distinguish the characteristics of different embeddings. Our evaluation of sentiment polarity and synonym/antonym relations shows that embeddings are able to capture surprisingly nuanced semantics even in the absence of sentence structure. Moreover, benchmarking the embeddings shows great variance in quality and characteristics of the semantics captured by the tested embeddings. Finally, we show the impact of varying the number of dimensions and the resolution of each dimension on the effective useful features captured by the embedding space. Our contributions highlight the importance of embeddings for NLP tasks and the effect of their quality on the final results.
1301.3235
Robust control of quantum gates via sequential convex programming
quant-ph cs.SY
Resource tradeoffs can often be established by solving an appropriate robust optimization problem for a variety of scenarios involving constraints on optimization variables and uncertainties. Using an approach based on sequential convex programming, we demonstrate that quantum gate transformations can be made substantially robust against uncertainties while simultaneously using limited resources of control amplitude and bandwidth. Achieving such a high degree of robustness requires a quantitative model that specifies the range and character of the uncertainties. Using a model of a controlled one-qubit system for illustrative simulations, we identify robust control fields for a universal gate set and explore the tradeoff between the worst-case gate fidelity and the field fluence. Our results demonstrate that, even for this simple model, there exist a rich variety of control design possibilities. In addition, we study the effect of noise represented by a stochastic uncertainty model.
1301.3248
Sparse Recovery with Coherent Tight Frames via Analysis Dantzig Selector and Analysis LASSO
cs.IT math.IT math.NA
This article considers recovery of signals that are sparse or approximately sparse in terms of a (possibly) highly overcomplete and coherent tight frame from undersampled data corrupted with additive noise. We show that the properly constrained $l_1$-analysis, called analysis Dantzig selector, stably recovers a signal which is nearly sparse in terms of a tight frame provided that the measurement matrix satisfies a restricted isometry property adapted to the tight frame. As a special case, we consider the Gaussian noise. Further, under a sparsity scenario, with high probability, the recovery error from noisy data is within a log-like factor of the minimax risk over the class of vectors which are at most $s$ sparse in terms of the tight frame. Similar results for the analysis LASSO are showed. The above two algorithms provide guarantees only for noise that is bounded or bounded with high probability (for example, Gaussian noise). However, when the underlying measurements are corrupted by sparse noise, these algorithms perform suboptimally. We demonstrate robust methods for reconstructing signals that are nearly sparse in terms of a tight frame in the presence of bounded noise combined with sparse noise. The analysis in this paper is based on the restricted isometry property adapted to a tight frame, which is a natural extension to the standard restricted isometry property.
1301.3258
New variant of ElGamal signature scheme
cs.CR cs.IT math.IT
In this paper, a new variant of ElGamal signature scheme is presented and its security analyzed. We also give, for its theoretical interest, a general form of the signature equation.
1301.3323
Auto-pooling: Learning to Improve Invariance of Image Features from Image Sequences
cs.CV cs.LG
Learning invariant representations from images is one of the hardest challenges facing computer vision. Spatial pooling is widely used to create invariance to spatial shifting, but it is restricted to convolutional models. In this paper, we propose a novel pooling method that can learn soft clustering of features from image sequences. It is trained to improve the temporal coherence of features, while keeping the information loss at minimum. Our method does not use spatial information, so it can be used with non-convolutional models too. Experiments on images extracted from natural videos showed that our method can cluster similar features together. When trained by convolutional features, auto-pooling outperformed traditional spatial pooling on an image classification task, even though it does not use the spatial topology of features.
1301.3342
Barnes-Hut-SNE
cs.LG cs.CV stat.ML
The paper presents an O(N log N)-implementation of t-SNE -- an embedding technique that is commonly used for the visualization of high-dimensional data in scatter plots and that normally runs in O(N^2). The new implementation uses vantage-point trees to compute sparse pairwise similarities between the input data objects, and it uses a variant of the Barnes-Hut algorithm - an algorithm used by astronomers to perform N-body simulations - to approximate the forces between the corresponding points in the embedding. Our experiments show that the new algorithm, called Barnes-Hut-SNE, leads to substantial computational advantages over standard t-SNE, and that it makes it possible to learn embeddings of data sets with millions of objects.
1301.3347
Multi-agent learning using Fictitious Play and Extended Kalman Filter
cs.MA cs.LG math.OC stat.ML
Decentralised optimisation tasks are important components of multi-agent systems. These tasks can be interpreted as n-player potential games: therefore game-theoretic learning algorithms can be used to solve decentralised optimisation tasks. Fictitious play is the canonical example of these algorithms. Nevertheless fictitious play implicitly assumes that players have stationary strategies. We present a novel variant of fictitious play where players predict their opponents' strategies using Extended Kalman filters and use their predictions to update their strategies. We show that in 2 by 2 games with at least one pure Nash equilibrium and in potential games where players have two available actions, the proposed algorithm converges to the pure Nash equilibrium. The performance of the proposed algorithm was empirically tested, in two strategic form games and an ad-hoc sensor network surveillance problem. The proposed algorithm performs better than the classic fictitious play algorithm in these games and therefore improves the performance of game-theoretical learning in decentralised optimisation.
1301.3369
Self-synchronizing pulse position modulation with error tolerance
cs.IT math.CO math.IT
Pulse position modulation (PPM) is a popular signal modulation technique which creates M-ary data by means of the position of a pulse within a time interval. While PPM and its variations have great advantages in many contexts, this type of modulation is vulnerable to loss of synchronization, potentially causing a severe error floor or throughput penalty even when little or no noise is assumed. Another disadvantage is that this type of modulation typically offers no error correction mechanism on its own, making them sensitive to intersymbol interference and environmental noise. In this paper we propose a coding theoretic variation of PPM that allows for significantly more efficient symbol and frame synchronization as well as strong error correction. The proposed scheme can be divided into a synchronization layer and a modulation layer. This makes our technique compatible with major existing techniques such as standard PPM, multipluse PPM, and expurgated PPM as well in that the scheme can be realized by adding a simple synchronization layer to one of these standard techniques. We also develop a generalization of expurgated PPM suited for the modulation layer of the proposed self-synchronizing modulation scheme. This generalized PPM can also be used as stand-alone error-correcting PPM with a larger number of available symbols.
1301.3375
On the Identifiability of Overcomplete Dictionaries via the Minimisation Principle Underlying K-SVD
cs.IT math.IT
This article gives theoretical insights into the performance of K-SVD, a dictionary learning algorithm that has gained significant popularity in practical applications. The particular question studied here is when a dictionary $\Phi\in \mathbb{R}^{d \times K}$ can be recovered as local minimum of the minimisation criterion underlying K-SVD from a set of $N$ training signals $y_n =\Phi x_n$. A theoretical analysis of the problem leads to two types of identifiability results assuming the training signals are generated from a tight frame with coefficients drawn from a random symmetric distribution. First, asymptotic results showing, that in expectation the generating dictionary can be recovered exactly as a local minimum of the K-SVD criterion if the coefficient distribution exhibits sufficient decay. Second, based on the asymptotic results it is demonstrated that given a finite number of training samples $N$, such that $N/\log N = O(K^3d)$, except with probability $O(N^{-Kd})$ there is a local minimum of the K-SVD criterion within distance $O(KN^{-1/4})$ to the generating dictionary.
1301.3385
Recurrent Online Clustering as a Spatio-Temporal Feature Extractor in DeSTIN
cs.CV
This paper presents a basic enhancement to the DeSTIN deep learning architecture by replacing the explicitly calculated transition tables that are used to capture temporal features with a simpler, more scalable mechanism. This mechanism uses feedback of state information to cluster over a space comprised of both the spatial input and the current state. The resulting architecture achieves state-of-the-art results on the MNIST classification benchmark.
1301.3388
Confluently Persistent Sets and Maps
cs.DS cs.DB
Ordered sets and maps play important roles as index structures in relational data models. When a shared index in a multi-user system is modified concurrently, the current state of the index will diverge into multiple versions containing the local modifications performed in each work flow. The confluent persistence problem arises when versions should be melded in commit and refresh operations so that modifications performed by different users become merged. Confluently Persistent Sets and Maps are functional binary search trees that support efficient set operations both when operands are disjoint and when they are overlapping. Treap properties with hash values as priorities are maintained and with hash-consing of nodes a unique representation is provided. Non-destructive set merge algorithms that skip inspection of equal subtrees and a conflict detecting meld algorithm based on set merges are presented. The meld algorithm is used in commit and refresh operations. With m modifications in one flow and n items in total, the expected cost of the operations is O(m log(n/m)).
1301.3389
The Diagonalized Newton Algorithm for Nonnegative Matrix Factorization
cs.NA cs.LG
Non-negative matrix factorization (NMF) has become a popular machine learning approach to many problems in text mining, speech and image processing, bio-informatics and seismic data analysis to name a few. In NMF, a matrix of non-negative data is approximated by the low-rank product of two matrices with non-negative entries. In this paper, the approximation quality is measured by the Kullback-Leibler divergence between the data and its low-rank reconstruction. The existence of the simple multiplicative update (MU) algorithm for computing the matrix factors has contributed to the success of NMF. Despite the availability of algorithms showing faster convergence, MU remains popular due to its simplicity. In this paper, a diagonalized Newton algorithm (DNA) is proposed showing faster convergence while the implementation remains simple and suitable for high-rank problems. The DNA algorithm is applied to various publicly available data sets, showing a substantial speed-up on modern hardware.
1301.3391
Feature grouping from spatially constrained multiplicative interaction
cs.LG
We present a feature learning model that learns to encode relationships between images. The model is defined as a Gated Boltzmann Machine, which is constrained such that hidden units that are nearby in space can gate each other's connections. We show how frequency/orientation "columns" as well as topographic filter maps follow naturally from training the model on image pairs. The model also helps explain why square-pooling models yield feature groups with similar grouping properties. Experimental results on synthetic image transformations show that spatially constrained gating is an effective way to reduce the number of parameters and thereby to regularize a transformation-learning model.
1301.3452
Exponential communication gap between weak and strong classical simulations of quantum communication
quant-ph cs.IT math.IT
The most trivial way to simulate classically the communication of a quantum state is to transmit the classical description of the quantum state itself. However, this requires an infinite amount of classical communication if the simulation is exact. A more intriguing and potentially less demanding strategy would encode the full information about the quantum state into the probability distribution of the communicated variables, so that this information is never sent in each single shot. This kind of simulation is called weak, as opposed to strong simulations, where the quantum state is communicated in individual shots. In this paper, we introduce a bounded-error weak protocol for simulating the communication of an arbitrary number of qubits and a subsequent two-outcome measurement consisting of an arbitrary pure state projector and its complement. This protocol requires an amount of classical communication independent of the number of qubits and proportional to Delta^{-1}, where Delta is the error and a free parameter of the protocol. Conversely, a bounded-error strong protocol requires an amount of classical communication growing exponentially with the number of qubits for a fixed error. Our result improves a previous protocol, based on the Johnson-Lindenstrauss lemma, with communication cost scaling as Delta^{-2} log Delta^{-1}.
1301.3457
A Geometric Descriptor for Cell-Division Detection
cs.CV
We describe a method for cell-division detection based on a geometric-driven descriptor that can be represented as a 5-layers processing network, based mainly on wavelet filtering and a test for mirror symmetry between pairs of pixels. After the centroids of the descriptors are computed for a sequence of frames, the two-steps piecewise constant function that best fits the sequence of centroids determines the frame where the division occurs.
1301.3461
Factorized Topic Models
cs.LG cs.CV cs.IR
In this paper we present a modification to a latent topic model, which makes the model exploit supervision to produce a factorized representation of the observed data. The structured parameterization separately encodes variance that is shared between classes from variance that is private to each class by the introduction of a new prior over the topic space. The approach allows for a more eff{}icient inference and provides an intuitive interpretation of the data in terms of an informative signal together with structured noise. The factorized representation is shown to enhance inference performance for image, text, and video classification.
1301.3468
Boltzmann Machines and Denoising Autoencoders for Image Denoising
stat.ML cs.CV cs.LG
Image denoising based on a probabilistic model of local image patches has been employed by various researchers, and recently a deep (denoising) autoencoder has been proposed by Burger et al. [2012] and Xie et al. [2012] as a good model for this. In this paper, we propose that another popular family of models in the field of deep learning, called Boltzmann machines, can perform image denoising as well as, or in certain cases of high level of noise, better than denoising autoencoders. We empirically evaluate the two models on three different sets of images with different types and levels of noise. Throughout the experiments we also examine the effect of the depth of the models. The experiments confirmed our claim and revealed that the performance can be improved by adding more hidden layers, especially when the level of noise is high.
1301.3476
Pushing Stochastic Gradient towards Second-Order Methods -- Backpropagation Learning with Transformations in Nonlinearities
cs.LG cs.CV stat.ML
Recently, we proposed to transform the outputs of each hidden neuron in a multi-layer perceptron network to have zero output and zero slope on average, and use separate shortcut connections to model the linear dependencies instead. We continue the work by firstly introducing a third transformation to normalize the scale of the outputs of each hidden neuron, and secondly by analyzing the connections to second order optimization methods. We show that the transformations make a simple stochastic gradient behave closer to second-order optimization methods and thus speed up learning. This is shown both in theory and with experiments. The experiments on the third transformation show that while it further increases the speed of learning, it can also hurt performance by converging to a worse local optimum, where both the inputs and outputs of many hidden neurons are close to zero.
1301.3485
A Semantic Matching Energy Function for Learning with Multi-relational Data
cs.LG
Large-scale relational learning becomes crucial for handling the huge amounts of structured data generated daily in many application domains ranging from computational biology or information retrieval, to natural language processing. In this paper, we present a new neural network architecture designed to embed multi-relational graphs into a flexible continuous vector space in which the original data is kept and enhanced. The network is trained to encode the semantics of these graphs in order to assign high probabilities to plausible components. We empirically show that it reaches competitive performance in link prediction on standard datasets from the literature.
1301.3488
Various improvements to text fingerprinting
cs.DS cs.DM cs.IR
Let s = s_1 .. s_n be a text (or sequence) on a finite alphabet \Sigma of size \sigma. A fingerprint in s is the set of distinct characters appearing in one of its substrings. The problem considered here is to compute the set {\cal F} of all fingerprints of all substrings of s in order to answer efficiently certain questions on this set. A substring s_i .. s_j is a maximal location for a fingerprint f in F (denoted by <i,j>) if the alphabet of s_i .. s_j is f and s_{i-1}, s_{j+1}, if defined, are not in f. The set of maximal locations ins is {\cal L} (it is easy to see that |{\cal L}| \leq n \sigma). Two maximal locations <i,j> and <k,l> such that s_i .. s_j = s_k .. s_l are named {\em copies}, and the quotient set of {\cal L} according to the copy relation is denoted by {\cal L}_C. We present new exact and approximate efficient algorithms and data structures for the following three problems: (1) to compute {\cal F}; (2) given f as a set of distinct characters in \Sigma, to answer if f represents a fingerprint in {\cal F}; (3) given f, to find all maximal locations of f in s.
1301.3509
Kidney Exchange in Dynamic Sparse Heterogenous Pools
cs.DS cs.SI
Current kidney exchange pools are of moderate size and thin, as they consist of many highly sensitized patients. Creating a thicker pool can be done by waiting for many pairs to arrive. We analyze a simple class of matching algorithms that search periodically for allocations. We find that if only 2-way cycles are conducted, in order to gain a significant amount of matches over the online scenario (matching each time a new incompatible pair joins the pool) the waiting period should be "very long". If 3-way cycles are also allowed we find regimes in which waiting for a short period also increases the number of matches considerably. Finally, a significant increase of matches can be obtained by using even one non-simultaneous chain while still matching in an online fashion. Our theoretical findings and data-driven computational experiments lead to policy recommendations.
1301.3516
Learnable Pooling Regions for Image Classification
cs.CV cs.LG
Biologically inspired, from the early HMAX model to Spatial Pyramid Matching, pooling has played an important role in visual recognition pipelines. Spatial pooling, by grouping of local codes, equips these methods with a certain degree of robustness to translation and deformation yet preserving important spatial information. Despite the predominance of this approach in current recognition systems, we have seen little progress to fully adapt the pooling strategy to the task at hand. This paper proposes a model for learning task dependent pooling scheme -- including previously proposed hand-crafted pooling schemes as a particular instantiation. In our work, we investigate the role of different regularization terms showing that the smooth regularization term is crucial to achieve strong performance using the presented architecture. Finally, we propose an efficient and parallel method to train the model. Our experiments show improved performance over hand-crafted pooling schemes on the CIFAR-10 and CIFAR-100 datasets -- in particular improving the state-of-the-art to 56.29% on the latter.
1301.3524
How good is the Electricity benchmark for evaluating concept drift adaptation
cs.LG
In this correspondence, we will point out a problem with testing adaptive classifiers on autocorrelated data. In such a case random change alarms may boost the accuracy figures. Hence, we cannot be sure if the adaptation is working well.
1301.3527
Block Coordinate Descent for Sparse NMF
cs.LG cs.NA
Nonnegative matrix factorization (NMF) has become a ubiquitous tool for data analysis. An important variant is the sparse NMF problem which arises when we explicitly require the learnt features to be sparse. A natural measure of sparsity is the L$_0$ norm, however its optimization is NP-hard. Mixed norms, such as L$_1$/L$_2$ measure, have been shown to model sparsity robustly, based on intuitive attributes that such measures need to satisfy. This is in contrast to computationally cheaper alternatives such as the plain L$_1$ norm. However, present algorithms designed for optimizing the mixed norm L$_1$/L$_2$ are slow and other formulations for sparse NMF have been proposed such as those based on L$_1$ and L$_0$ norms. Our proposed algorithm allows us to solve the mixed norm sparsity constraints while not sacrificing computation time. We present experimental evidence on real-world datasets that shows our new algorithm performs an order of magnitude faster compared to the current state-of-the-art solvers optimizing the mixed norm and is suitable for large-scale datasets.
1301.3528
An Efficient Sufficient Dimension Reduction Method for Identifying Genetic Variants of Clinical Significance
q-bio.GN cs.LG stat.ML
Fast and cheaper next generation sequencing technologies will generate unprecedentedly massive and highly-dimensional genomic and epigenomic variation data. In the near future, a routine part of medical record will include the sequenced genomes. A fundamental question is how to efficiently extract genomic and epigenomic variants of clinical utility which will provide information for optimal wellness and interference strategies. Traditional paradigm for identifying variants of clinical validity is to test association of the variants. However, significantly associated genetic variants may or may not be usefulness for diagnosis and prognosis of diseases. Alternative to association studies for finding genetic variants of predictive utility is to systematically search variants that contain sufficient information for phenotype prediction. To achieve this, we introduce concepts of sufficient dimension reduction and coordinate hypothesis which project the original high dimensional data to very low dimensional space while preserving all information on response phenotypes. We then formulate clinically significant genetic variant discovery problem into sparse SDR problem and develop algorithms that can select significant genetic variants from up to or even ten millions of predictors with the aid of dividing SDR for whole genome into a number of subSDR problems defined for genomic regions. The sparse SDR is in turn formulated as sparse optimal scoring problem, but with penalty which can remove row vectors from the basis matrix. To speed up computation, we develop the modified alternating direction method for multipliers to solve the sparse optimal scoring problem which can easily be implemented in parallel. To illustrate its application, the proposed method is applied to simulation data and the NHLBI's Exome Sequencing Project dataset
1301.3530
The Neural Representation Benchmark and its Evaluation on Brain and Machine
cs.NE cs.CV cs.LG q-bio.NC
A key requirement for the development of effective learning representations is their evaluation and comparison to representations we know to be effective. In natural sensory domains, the community has viewed the brain as a source of inspiration and as an implicit benchmark for success. However, it has not been possible to directly test representational learning algorithms directly against the representations contained in neural systems. Here, we propose a new benchmark for visual representations on which we have directly tested the neural representation in multiple visual cortical areas in macaque (utilizing data from [Majaj et al., 2012]), and on which any computer vision algorithm that produces a feature space can be tested. The benchmark measures the effectiveness of the neural or machine representation by computing the classification loss on the ordered eigendecomposition of a kernel matrix [Montavon et al., 2011]. In our analysis we find that the neural representation in visual area IT is superior to visual area V4. In our analysis of representational learning algorithms, we find that three-layer models approach the representational performance of V4 and the algorithm in [Le et al., 2012] surpasses the performance of V4. Impressively, we find that a recent supervised algorithm [Krizhevsky et al., 2012] achieves performance comparable to that of IT for an intermediate level of image variation difficulty, and surpasses IT at a higher difficulty level. We believe this result represents a major milestone: it is the first learning algorithm we have found that exceeds our current estimate of IT representation performance. We hope that this benchmark will assist the community in matching the representational performance of visual cortex and will serve as an initial rallying point for further correspondence between representations derived in brains and machines.
1301.3533
Sparse Penalty in Deep Belief Networks: Using the Mixed Norm Constraint
cs.NE cs.LG stat.ML
Deep Belief Networks (DBN) have been successfully applied on popular machine learning tasks. Specifically, when applied on hand-written digit recognition, DBNs have achieved approximate accuracy rates of 98.8%. In an effort to optimize the data representation achieved by the DBN and maximize their descriptive power, recent advances have focused on inducing sparse constraints at each layer of the DBN. In this paper we present a theoretical approach for sparse constraints in the DBN using the mixed norm for both non-overlapping and overlapping groups. We explore how these constraints affect the classification accuracy for digit recognition in three different datasets (MNIST, USPS, RIMES) and provide initial estimations of their usefulness by altering different parameters such as the group size and overlap percentage.