id
stringlengths
9
16
title
stringlengths
4
278
categories
stringlengths
5
104
abstract
stringlengths
6
4.09k
1311.2241
Learning Gaussian Graphical Models with Observed or Latent FVSs
cs.LG stat.ML
Gaussian Graphical Models (GGMs) or Gauss Markov random fields are widely used in many applications, and the trade-off between the modeling capacity and the efficiency of learning and inference has been an important research problem. In this paper, we study the family of GGMs with small feedback vertex sets (FVSs), where an FVS is a set of nodes whose removal breaks all the cycles. Exact inference such as computing the marginal distributions and the partition function has complexity $O(k^{2}n)$ using message-passing algorithms, where k is the size of the FVS, and n is the total number of nodes. We propose efficient structure learning algorithms for two cases: 1) All nodes are observed, which is useful in modeling social or flight networks where the FVS nodes often correspond to a small number of high-degree nodes, or hubs, while the rest of the networks is modeled by a tree. Regardless of the maximum degree, without knowing the full graph structure, we can exactly compute the maximum likelihood estimate in $O(kn^2+n^2\log n)$ if the FVS is known or in polynomial time if the FVS is unknown but has bounded size. 2) The FVS nodes are latent variables, where structure learning is equivalent to decomposing a inverse covariance matrix (exactly or approximately) into the sum of a tree-structured matrix and a low-rank matrix. By incorporating efficient inference into the learning steps, we can obtain a learning algorithm using alternating low-rank correction with complexity $O(kn^{2}+n^{2}\log n)$ per iteration. We also perform experiments using both synthetic data as well as real data of flight delays to demonstrate the modeling capacity with FVSs of various sizes.
1311.2252
Semantic Sort: A Supervised Approach to Personalized Semantic Relatedness
cs.CL cs.LG
We propose and study a novel supervised approach to learning statistical semantic relatedness models from subjectively annotated training examples. The proposed semantic model consists of parameterized co-occurrence statistics associated with textual units of a large background knowledge corpus. We present an efficient algorithm for learning such semantic models from a training sample of relatedness preferences. Our method is corpus independent and can essentially rely on any sufficiently large (unstructured) collection of coherent texts. Moreover, the approach facilitates the fitting of semantic models for specific users or groups of users. We present the results of extensive range of experiments from small to large scale, indicating that the proposed method is effective and competitive with the state-of-the-art.
1311.2271
More data speeds up training time in learning halfspaces over sparse vectors
cs.LG
The increased availability of data in recent years has led several authors to ask whether it is possible to use data as a {\em computational} resource. That is, if more data is available, beyond the sample complexity limit, is it possible to use the extra examples to speed up the computation time required to perform the learning task? We give the first positive answer to this question for a {\em natural supervised learning problem} --- we consider agnostic PAC learning of halfspaces over $3$-sparse vectors in $\{-1,1,0\}^n$. This class is inefficiently learnable using $O\left(n/\epsilon^2\right)$ examples. Our main contribution is a novel, non-cryptographic, methodology for establishing computational-statistical gaps, which allows us to show that, under a widely believed assumption that refuting random $\mathrm{3CNF}$ formulas is hard, it is impossible to efficiently learn this class using only $O\left(n/\epsilon^2\right)$ examples. We further show that under stronger hardness assumptions, even $O\left(n^{1.499}/\epsilon^2\right)$ examples do not suffice. On the other hand, we show a new algorithm that learns this class efficiently using $\tilde{\Omega}\left(n^2/\epsilon^2\right)$ examples. This formally establishes the tradeoff between sample and computational complexity for a natural supervised learning problem.
1311.2272
From average case complexity to improper learning complexity
cs.LG cs.CC
The basic problem in the PAC model of computational learning theory is to determine which hypothesis classes are efficiently learnable. There is presently a dearth of results showing hardness of learning problems. Moreover, the existing lower bounds fall short of the best known algorithms. The biggest challenge in proving complexity results is to establish hardness of {\em improper learning} (a.k.a. representation independent learning).The difficulty in proving lower bounds for improper learning is that the standard reductions from $\mathbf{NP}$-hard problems do not seem to apply in this context. There is essentially only one known approach to proving lower bounds on improper learning. It was initiated in (Kearns and Valiant 89) and relies on cryptographic assumptions. We introduce a new technique for proving hardness of improper learning, based on reductions from problems that are hard on average. We put forward a (fairly strong) generalization of Feige's assumption (Feige 02) about the complexity of refuting random constraint satisfaction problems. Combining this assumption with our new technique yields far reaching implications. In particular, 1. Learning $\mathrm{DNF}$'s is hard. 2. Agnostically learning halfspaces with a constant approximation ratio is hard. 3. Learning an intersection of $\omega(1)$ halfspaces is hard.
1311.2276
A Quantitative Evaluation Framework for Missing Value Imputation Algorithms
cs.LG
We consider the problem of quantitatively evaluating missing value imputation algorithms. Given a dataset with missing values and a choice of several imputation algorithms to fill them in, there is currently no principled way to rank the algorithms using a quantitative metric. We develop a framework based on treating imputation evaluation as a problem of comparing two distributions and show how it can be used to compute quantitative metrics. We present an efficient procedure for applying this framework to practical datasets, demonstrate several metrics derived from the existing literature on comparing distributions, and propose a new metric called Neighborhood-based Dissimilarity Score which is fast to compute and provides similar results. Results are shown on several datasets, metrics, and imputations algorithms.
1311.2296
Newton based Stochastic Optimization using q-Gaussian Smoothed Functional Algorithms
math.OC cs.IT math.IT
We present the first q-Gaussian smoothed functional (SF) estimator of the Hessian and the first Newton-based stochastic optimization algorithm that estimates both the Hessian and the gradient of the objective function using q-Gaussian perturbations. Our algorithm requires only two system simulations (regardless of the parameter dimension) and estimates both the gradient and the Hessian at each update epoch using these. We also present a proof of convergence of the proposed algorithm. In a related recent work (Ghoshdastidar et al., 2013), we presented gradient SF algorithms based on the q-Gaussian perturbations. Our work extends prior work on smoothed functional algorithms by generalizing the class of perturbation distributions as most distributions reported in the literature for which SF algorithms are known to work and turn out to be special cases of the q-Gaussian distribution. Besides studying the convergence properties of our algorithm analytically, we also show the results of several numerical simulations on a model of a queuing network, that illustrate the significance of the proposed method. In particular, we observe that our algorithm performs better in most cases, over a wide range of q-values, in comparison to Newton SF algorithms with the Gaussian (Bhatnagar, 2007) and Cauchy perturbations, as well as the gradient q-Gaussian SF algorithms (Ghoshdastidar et al., 2013).
1311.2311
Direct solutions to tropical optimization problems with nonlinear objective functions and boundary constraints
math.OC cs.SY
We examine two multidimensional optimization problems that are formulated in terms of tropical mathematics. The problems are to minimize nonlinear objective functions, which are defined through the multiplicative conjugate vector transposition on vectors of a finite-dimensional semimodule over an idempotent semifield, and subject to boundary constraints. The solution approach is implemented, which involves the derivation of the sharp bounds on the objective functions, followed by determination of vectors that yield the bound. Based on the approach, direct solutions to the problems are obtained in a compact vector form. To illustrate, we apply the results to solving constrained Chebyshev approximation and location problems, and give numerical examples.
1311.2321
On Improving the Balance between the Completion Time and Decoding Delay in Instantly Decodable Network Coded Systems
cs.IT math.IT
This paper studies the complicated interplay of the completion time (as a measure of throughput) and the decoding delay performance in instantly decodable network coded (IDNC) systems over wireless broadcast erasure channels with memory, and proposes two new algorithms that improve the balance between the completion time and decoding delay of broadcasting a block of packets. We first formulate the IDNC packet selection problem that provides joint control of the completion time and decoding delay as a statistical shortest path (SSP) problem. However, since finding the optimal packet selection policy using the SSP technique is computationally complex, we employ its geometric structure to find some guidelines and use them to propose two heuristic packet selection algorithms that can efficiently improve the balance between the completion time and decoding delay for broadcast erasure channels with a wide range of memory conditions. It is shown that each one of the two proposed algorithms is superior for a specific range of memory conditions. Furthermore, we show that the proposed algorithms achieve an improved fairness in terms of the decoding delay across all receivers.
1311.2331
Robust Adaptive Beamforming Based on Low-Complexity Shrinkage-Based Mismatch Estimation
cs.IT math.IT
In this work, we propose a low-complexity robust adaptive beamforming (RAB) technique which estimates the steering vector using a Low-Complexity Shrinkage-Based Mismatch Estimation (LOCSME) algorithm. The proposed LOCSME algorithm estimates the covariance matrix of the input data and the interference-plus-noise covariance (INC) matrix by using the Oracle Approximating Shrinkage (OAS) method. LOCSME only requires prior knowledge of the angular sector in which the actual steering vector is located and the antenna array geometry. LOCSME does not require a costly optimization algorithm and does not need to know extra information from the interferers, which avoids direction finding for all interferers. Simulations show that LOCSME outperforms previously reported RAB algorithms and has a performance very close to the optimum.
1311.2334
Embed and Conquer: Scalable Embeddings for Kernel k-Means on MapReduce
cs.LG
The kernel $k$-means is an effective method for data clustering which extends the commonly-used $k$-means algorithm to work on a similarity matrix over complex data structures. The kernel $k$-means algorithm is however computationally very complex as it requires the complete data matrix to be calculated and stored. Further, the kernelized nature of the kernel $k$-means algorithm hinders the parallelization of its computations on modern infrastructures for distributed computing. In this paper, we are defining a family of kernel-based low-dimensional embeddings that allows for scaling kernel $k$-means on MapReduce via an efficient and unified parallelization strategy. Afterwards, we propose two methods for low-dimensional embedding that adhere to our definition of the embedding family. Exploiting the proposed parallelization strategy, we present two scalable MapReduce algorithms for kernel $k$-means. We demonstrate the effectiveness and efficiency of the proposed algorithms through an empirical evaluation on benchmark data sets.
1311.2337
The Third-Order Term in the Normal Approximation for the AWGN Channel
cs.IT math.IT
This paper shows that, under the average error probability formalism, the third-order term in the normal approximation for the additive white Gaussian noise channel with a maximal or equal power constraint is at least $\frac{1}{2} \log n + O(1)$. This matches the upper bound derived by Polyanskiy-Poor-Verd\'{u} (2010).
1311.2342
Anatomy of Graph Matching based on an XQuery and RDF Implementation
cs.DB
Graphs are becoming one of the most popular data modeling paradigms since they are able to model complex relationships that cannot be easily captured using traditional data models. One of the major tasks of graph management is graph matching, which aims to find all of the subgraphs in a data graph that match a query graph. In the literature, proposals in this context are classified into two different categories: graph-at-a-time, which process the whole query graph at the same time, and vertex-at-a-time, which process a single vertex of the query graph at the same time. In this paper, we propose a new vertex-at-a-time proposal that is based on graphlets, each of which comprises a vertex of a graph, all of the immediate neighbors of that vertex, and all of the edges that relate those neighbors. Furthermore, we also use the concept of minimum hub covers, each of which comprises a subset of vertices in the query graph that account for all of the edges in that graph. We present the algorithms of our proposal and describe an implementation based on XQuery and RDF. Our evaluation results show that our proposal is appealing to perform graph matching.
1311.2346
Some New Results on Equivalency of Collusion-Secure Properties for Reed-Solomon Codes
cs.IT math.IT math.RA
A. Silverberg (IEEE Trans. Inform. Theory 49, 2003) proposed a question on the equivalence of identifiable parent property and traceability property for Reed-Solomon code family. Earlier studies on Silverberg's problem motivate us to think of the stronger version of the question on equivalence of separation and traceability properties. Both, however, still remain open. In this article, we integrate all the previous works on this problem with an algebraic way, and present some new results. It is notable that the concept of subspace subcode of Reed-Solomon code, which was introduced in error-correcting code theory, provides an interesting prospect for our topic.
1311.2349
Providing Trustworthy Contributions via a Reputation Framework in Social Participatory Sensing Systems
cs.SI cs.IR
Social participatory sensing is a newly proposed paradigm that tries to address the limitations of participatory sensing by leveraging online social networks as an infrastructure. A critical issue in the success of this paradigm is to assure the trustworthiness of contributions provided by participants. In this paper, we propose an application-agnostic reputation framework for social participatory sensing systems. Our framework considers both the quality of contribution and the trustworthiness level of participant within the social network. These two aspects are then combined via a fuzzy inference system to arrive at a final trust rating for a contribution. A reputation score is also calculated for each participant as a resultant of the trust ratings assigned to him. We adopt the utilization of PageRank algorithm as the building block for our reputation module. Extensive simulations demonstrate the efficacy of our framework in achieving high overall trust and assigning accurate reputation scores.
1311.2378
An Empirical Evaluation of Sequence-Tagging Trainers
cs.LG
The task of assigning label sequences to a set of observed sequences is common in computational linguistics. Several models for sequence labeling have been proposed over the last few years. Here, we focus on discriminative models for sequence labeling. Many batch and online (updating model parameters after visiting each example) learning algorithms have been proposed in the literature. On large datasets, online algorithms are preferred as batch learning methods are slow. These online algorithms were designed to solve either a primal or a dual problem. However, there has been no systematic comparison of these algorithms in terms of their speed, generalization performance (accuracy/likelihood) and their ability to achieve steady state generalization performance fast. With this aim, we compare different algorithms and make recommendations, useful for a practitioner. We conclude that the selection of an algorithm for sequence labeling depends on the evaluation criterion used and its implementation simplicity.
1311.2433
Joint recovery algorithms using difference of innovations for distributed compressed sensing
cs.IT math.IT
Distributed compressed sensing is concerned with representing an ensemble of jointly sparse signals using as few linear measurements as possible. Two novel joint reconstruction algorithms for distributed compressed sensing are presented in this paper. These algorithms are based on the idea of using one of the signals as side information; this allows to exploit joint sparsity in a more effective way with respect to existing schemes. They provide gains in reconstruction quality, especially when the nodes acquire few measurements, so that the system is able to operate with fewer measurements than is required by other existing schemes. We show that the algorithms achieve better performance with respect to the state-of-the-art.
1311.2448
Recovery of Sparse Matrices via Matrix Sketching
cs.NA cs.IT math.IT
In this paper, we consider the problem of recovering an unknown sparse matrix X from the matrix sketch Y = AX B^T. The dimension of Y is less than that of X, and A and B are known matrices. This problem can be solved using standard compressive sensing (CS) theory after converting it to vector form using the Kronecker operation. In this case, the measurement matrix assumes a Kronecker product structure. However, as the matrix dimension increases the associated computational complexity makes its use prohibitive. We extend two algorithms, fast iterative shrinkage threshold algorithm (FISTA) and orthogonal matching pursuit (OMP) to solve this problem in matrix form without employing the Kronecker product. While both FISTA and OMP with matrix inputs are shown to be equivalent in performance to their vector counterparts with the Kronecker product, solving them in matrix form is shown to be computationally more efficient. We show that the computational gain achieved by FISTA with matrix inputs over its vector form is more significant compared to that achieved by OMP.
1311.2460
Vision-Guided Robot Hearing
cs.RO cs.CV
Natural human-robot interaction in complex and unpredictable environments is one of the main research lines in robotics. In typical real-world scenarios, humans are at some distance from the robot and the acquired signals are strongly impaired by noise, reverberations and other interfering sources. In this context, the detection and localisation of speakers plays a key role since it is the pillar on which several tasks (e.g.: speech recognition and speaker tracking) rely. We address the problem of how to detect and localize people that are both seen and heard by a humanoid robot. We introduce a hybrid deterministic/probabilistic model. Indeed, the deterministic component allows us to map the visual information into the auditory space. By means of the probabilistic component, the visual features guide the grouping of the auditory features in order to form AV objects. The proposed model and the associated algorithm are implemented in real-time (17 FPS) using a stereoscopic camera pair and two microphones embedded into the head of the humanoid robot NAO. We performed experiments on (i) synthetic data, (ii) a publicly available data set and (iii) data acquired using the robot. The results we obtained validate the approach and encourage us to further investigate how vision can help robot hearing.
1311.2483
Global Sensitivity Analysis with Dependence Measures
math.ST cs.LG stat.ML stat.TH
Global sensitivity analysis with variance-based measures suffers from several theoretical and practical limitations, since they focus only on the variance of the output and handle multivariate variables in a limited way. In this paper, we introduce a new class of sensitivity indices based on dependence measures which overcomes these insufficiencies. Our approach originates from the idea to compare the output distribution with its conditional counterpart when one of the input variables is fixed. We establish that this comparison yields previously proposed indices when it is performed with Csiszar f-divergences, as well as sensitivity indices which are well-known dependence measures between random variables. This leads us to investigate completely new sensitivity indices based on recent state-of-the-art dependence measures, such as distance correlation and the Hilbert-Schmidt independence criterion. We also emphasize the potential of feature selection techniques relying on such dependence measures as alternatives to screening in high dimension.
1311.2492
Notes on Elementary Spectral Graph Theory. Applications to Graph Clustering Using Normalized Cuts
cs.CV
These are notes on the method of normalized graph cuts and its applications to graph clustering. I provide a fairly thorough treatment of this deeply original method due to Shi and Malik, including complete proofs. I include the necessary background on graphs and graph Laplacians. I then explain in detail how the eigenvectors of the graph Laplacian can be used to draw a graph. This is an attractive application of graph Laplacians. The main thrust of this paper is the method of normalized cuts. I give a detailed account for K = 2 clusters, and also for K > 2 clusters, based on the work of Yu and Shi. Three points that do not appear to have been clearly articulated before are elaborated: 1. The solutions of the main optimization problem should be viewed as tuples in the K-fold cartesian product of projective space RP^{N-1}. 2. When K > 2, the solutions of the relaxed problem should be viewed as elements of the Grassmannian G(K,N). 3. Two possible Riemannian distances are available to compare the closeness of solutions: (a) The distance on (RP^{N-1})^K. (b) The distance on the Grassmannian. I also clarify what should be the necessary and sufficient conditions for a matrix to represent a partition of the vertices of a graph to be clustered.
1311.2495
The Noisy Power Method: A Meta Algorithm with Applications
cs.DS cs.LG
We provide a new robust convergence analysis of the well-known power method for computing the dominant singular vectors of a matrix that we call the noisy power method. Our result characterizes the convergence behavior of the algorithm when a significant amount noise is introduced after each matrix-vector multiplication. The noisy power method can be seen as a meta-algorithm that has recently found a number of important applications in a broad range of machine learning problems including alternating minimization for matrix completion, streaming principal component analysis (PCA), and privacy-preserving spectral analysis. Our general analysis subsumes several existing ad-hoc convergence bounds and resolves a number of open problems in multiple applications including streaming PCA and privacy-preserving singular vector computation.
1311.2503
Predictable Feature Analysis
cs.LG stat.ML
Every organism in an environment, whether biological, robotic or virtual, must be able to predict certain aspects of its environment in order to survive or perform whatever task is intended. It needs a model that is capable of estimating the consequences of possible actions, so that planning, control, and decision-making become feasible. For scientific purposes, such models are usually created in a problem specific manner using differential equations and other techniques from control- and system-theory. In contrast to that, we aim for an unsupervised approach that builds up the desired model in a self-organized fashion. Inspired by Slow Feature Analysis (SFA), our approach is to extract sub-signals from the input, that behave as predictable as possible. These "predictable features" are highly relevant for modeling, because predictability is a desired property of the needed consequence-estimating model by definition. In our approach, we measure predictability with respect to a certain prediction model. We focus here on the solution of the arising optimization problem and present a tractable algorithm based on algebraic methods which we call Predictable Feature Analysis (PFA). We prove that the algorithm finds the globally optimal signal, if this signal can be predicted with low error. To deal with cases where the optimal signal has a significant prediction error, we provide a robust, heuristically motivated variant of the algorithm and verify it empirically. Additionally, we give formal criteria a prediction-model must meet to be suitable for measuring predictability in the PFA setting and also provide a suitable default-model along with a formal proof that it meets these criteria.
1311.2504
A Semi-automated Peer-review System
cs.DL cs.HC cs.SI physics.soc-ph
A semi-supervised model of peer review is introduced that is intended to overcome the bias and incompleteness of traditional peer review. Traditional approaches are reliant on human biases, while consensus decision-making is constrained by sparse information. Here, the architecture for one potential improvement (a semi-supervised, human-assisted classifier) to the traditional approach will be introduced and evaluated. To evaluate the potential advantages of such a system, hypothetical receiver operating characteristic (ROC) curves for both approaches will be assessed. This will provide more specific indications of how automation would be beneficial in the manuscript evaluation process. In conclusion, the implications for such a system on measurements of scientific impact and improving the quality of open submission repositories will be discussed.
1311.2505
On optimal constacyclic codes
cs.IT math-ph math.IT math.MP
In this paper we investigate the class of constacyclic codes, which is a natural generalization of the class of cyclic and negacyclic codes. This class of codes is interesting in the sense that it contains codes with good or even optimal parameters. In this light, we propose constructions of families of classical block and convolutional maximum-distance-separable (MDS) constacyclic codes, as well as families of asymmetric quantum MDS codes derived from (classical-block) constacyclic codes. These results are mainly derived from the investigation of suitable properties on cyclotomic cosets of these corresponding codes.
1311.2507
Robust Beamforming for Secure Communication in Systems with Wireless Information and Power Transfer
cs.IT math.IT
This paper considers a multiuser multiple-input single-output (MISO) downlink system with simultaneous wireless information and power transfer. In particular, we focus on secure communication in the presence of passive eavesdroppers and potential eavesdroppers (idle legitimate receivers). We study the design of a resource allocation algorithm minimizing the total transmit power for the case when the legitimate receivers are able to harvest energy from radio frequency signals. Our design advocates the dual use of both artificial noise and energy signals in providing secure communication and facilitating efficient wireless energy transfer. The algorithm design is formulated as a non-convex optimization problem. The problem formulation takes into account artificial noise and energy signal generation for protecting the transmitted information against both considered types of eavesdroppers when imperfect channel state information (CSI) of the potential eavesdroppers and no CSI of the passive eavesdroppers are available at the transmitter. In light of the intractability of the problem, we reformulate the considered problem by replacing a non-convex probabilistic constraint with a convex deterministic constraint. Then, a semi-definite programming (SDP) relaxation approach is adopted to obtain the optimal solution for the reformulated problem. Furthermore, we propose a suboptimal resource allocation scheme with low computational complexity for providing communication secrecy and facilitating efficient energy transfer. Simulation results demonstrate a close-to-optimal performance achieved by the proposed schemes and significant transmit power savings by optimization of the artificial noise and energy signal generation.
1311.2524
Rich feature hierarchies for accurate object detection and semantic segmentation
cs.CV
Object detection performance, as measured on the canonical PASCAL VOC dataset, has plateaued in the last few years. The best-performing methods are complex ensemble systems that typically combine multiple low-level image features with high-level context. In this paper, we propose a simple and scalable detection algorithm that improves mean average precision (mAP) by more than 30% relative to the previous best result on VOC 2012---achieving a mAP of 53.3%. Our approach combines two key insights: (1) one can apply high-capacity convolutional neural networks (CNNs) to bottom-up region proposals in order to localize and segment objects and (2) when labeled training data is scarce, supervised pre-training for an auxiliary task, followed by domain-specific fine-tuning, yields a significant performance boost. Since we combine region proposals with CNNs, we call our method R-CNN: Regions with CNN features. We also compare R-CNN to OverFeat, a recently proposed sliding-window detector based on a similar CNN architecture. We find that R-CNN outperforms OverFeat by a large margin on the 200-class ILSVRC2013 detection dataset. Source code for the complete system is available at http://www.cs.berkeley.edu/~rbg/rcnn.
1311.2526
User recommendation in reciprocal and bipartite social networks -- a case study of online dating
cs.SI cs.IR physics.soc-ph
Many social networks in our daily life are bipartite networks built on reciprocity. How can we recommend users/friends to a user, so that the user is interested in and attractive to recommended users? In this research, we propose a new collaborative filtering model to improve user recommendations in reciprocal and bipartite social networks. The model considers a user's "taste" in picking others and "attractiveness" in being picked by others. A case study of an online dating network shows that the new model has good performance in recommending both initial and reciprocal contacts.
1311.2531
Motility at the origin of life: Its characterization and a model
cs.AI cs.NE nlin.AO nlin.PS q-bio.PE
Due to recent advances in synthetic biology and artificial life, the origin of life is currently a hot topic of research. We review the literature and argue that the two traditionally competing "replicator-first" and "metabolism-first" approaches are merging into one integrated theory of individuation and evolution. We contribute to the maturation of this more inclusive approach by highlighting some problematic assumptions that still lead to an impoverished conception of the phenomenon of life. In particular, we argue that the new consensus has so far failed to consider the relevance of intermediate timescales. We propose that an adequate theory of life must account for the fact that all living beings are situated in at least four distinct timescales, which are typically associated with metabolism, motility, development, and evolution. On this view, self-movement, adaptive behavior and morphological changes could have already been present at the origin of life. In order to illustrate this possibility we analyze a minimal model of life-like phenomena, namely of precarious, individuated, dissipative structures that can be found in simple reaction-diffusion systems. Based on our analysis we suggest that processes in intermediate timescales could have already been operative in prebiotic systems. They may have facilitated and constrained changes occurring in the faster- and slower-paced timescales of chemical self-individuation and evolution by natural selection, respectively.
1311.2540
Asymmetric numeral systems: entropy coding combining speed of Huffman coding with compression rate of arithmetic coding
cs.IT math.IT
The modern data compression is mainly based on two approaches to entropy coding: Huffman (HC) and arithmetic/range coding (AC). The former is much faster, but approximates probabilities with powers of 2, usually leading to relatively low compression rates. The latter uses nearly exact probabilities - easily approaching theoretical compression rate limit (Shannon entropy), but at cost of much larger computational cost. Asymmetric numeral systems (ANS) is a new approach to accurate entropy coding, which allows to end this trade-off between speed and rate: the recent implementation [1] provides about $50\%$ faster decoding than HC for 256 size alphabet, with compression rate similar to provided by AC. This advantage is due to being simpler than AC: using single natural number as the state, instead of two to represent a range. Beside simplifying renormalization, it allows to put the entire behavior for given probability distribution into a relatively small table: defining entropy coding automaton. The memory cost of such table for 256 size alphabet is a few kilobytes. There is a large freedom while choosing a specific table - using pseudorandom number generator initialized with cryptographic key for this purpose allows to simultaneously encrypt the data. This article also introduces and discusses many other variants of this new entropy coding approach, which can provide direct alternatives for standard AC, for large alphabet range coding, or for approximated quasi arithmetic coding.
1311.2542
Toward a unified theory of sparse dimensionality reduction in Euclidean space
cs.DS cs.CG cs.IT math.IT math.PR stat.ML
Let $\Phi\in\mathbb{R}^{m\times n}$ be a sparse Johnson-Lindenstrauss transform [KN14] with $s$ non-zeroes per column. For a subset $T$ of the unit sphere, $\varepsilon\in(0,1/2)$ given, we study settings for $m,s$ required to ensure $$ \mathop{\mathbb{E}}_\Phi \sup_{x\in T} \left|\|\Phi x\|_2^2 - 1 \right| < \varepsilon , $$ i.e. so that $\Phi$ preserves the norm of every $x\in T$ simultaneously and multiplicatively up to $1+\varepsilon$. We introduce a new complexity parameter, which depends on the geometry of $T$, and show that it suffices to choose $s$ and $m$ such that this parameter is small. Our result is a sparse analog of Gordon's theorem, which was concerned with a dense $\Phi$ having i.i.d. Gaussian entries. We qualitatively unify several results related to the Johnson-Lindenstrauss lemma, subspace embeddings, and Fourier-based restricted isometries. Our work also implies new results in using the sparse Johnson-Lindenstrauss transform in numerical linear algebra, classical and model-based compressed sensing, manifold learning, and constrained least squares problems such as the Lasso.
1311.2547
Learning Mixtures of Linear Classifiers
cs.LG stat.ML
We consider a discriminative learning (regression) problem, whereby the regression function is a convex combination of k linear classifiers. Existing approaches are based on the EM algorithm, or similar techniques, without provable guarantees. We develop a simple method based on spectral techniques and a `mirroring' trick, that discovers the subspace spanned by the classifiers' parameter vectors. Under a probabilistic assumption on the feature vector distribution, we prove that this approach has nearly optimal statistical efficiency.
1311.2549
A doubling construction for self-orthogonal codes
cs.IT math.IT
A simple construction of quaternary hermitian self-orthogonal codes with parameters $[2n+1,k+1]$ and $[2n+2,k+2]$ from a given pair of self-orthogonal $[n,k]$ codes, and its link to quantum codes is considered. As an application, an optimal quaternary linear $[28,20,6]$ dual containing code is found that yields a new optimal $[[28,12,6]]$ quantum code.
1311.2551
Social Networks and Collective Intelligence: A Return to the Agora
cs.SI cs.CY physics.soc-ph
Nowadays, acquisition of trustable information is increasingly important in both professional and private contexts. However, establishing what information is trustable and what is not, is a very challenging task. For example, how can information quality be reliably assessed? How can sources? credibility be fairly assessed? How can gatekeeping processes be found trustworthy when filtering out news and deciding ranking and priorities of traditional media? An Internet-based solution to a human-based ancient issue is being studied, and it is called Polidoxa, from Greek "poly", meaning "many" or "several" and "doxa", meaning "common belief" or "popular opinion". This old problem will be solved by means of ancient philosophies and processes with truly modern tools and technologies. This is why this work required a collaborative and interdisciplinary joint effort from researchers with very different backgrounds and institutes with significantly different agendas. Polidoxa aims at offering: 1) a trust-based search engine algorithm, which exploits stigmergic behaviours of users? network, 2) a trust-based social network, where the notion of trust derives from network activity and 3) a holonic system for bottom-up self-protection and social privacy. By presenting the Polidoxa solution, this work also describes the current state of traditional media as well as newer ones, providing an accurate analysis of major search engines such as Google and social network (e.g., Facebook). The advantages that Polidoxa offers, compared to these, are also clearly detailed and motivated. Finally, a Twitter application (Polidoxa@twitter) which enables experimentation of basic Polidoxa principles is presented.
1311.2561
Performing edge detection by difference of Gaussians using q-Gaussian kernels
cs.CV physics.comp-ph
In image processing, edge detection is a valuable tool to perform the extraction of features from an image. This detection reduces the amount of information to be processed, since the redundant information (considered less relevant) can be unconsidered. The technique of edge detection consists of determining the points of a digital image whose intensity changes sharply. This changes are due to the discontinuities of the orientation on a surface for example. A well known method of edge detection is the Difference of Gaussians (DoG). The method consists of subtracting two Gaussians, where a kernel has a standard deviation smaller than the previous one. The convolution between the subtraction of kernels and the input image results in the edge detection of this image. This paper introduces a method of extracting edges using DoG with kernels based on the q-Gaussian probability distribution, derived from the q-statistic proposed by Constantino Tsallis. To demonstrate the method's potential, we compare the introduced method with the traditional DoG using Gaussians kernels. The results showed that the proposed method can extract edges with more accurate details.
1311.2621
Determining Leishmania Infection Levels by Automatic Analysis of Microscopy Images
cs.CV
Analysis of microscopy images is one important tool in many fields of biomedical research, as it allows the quantification of a multitude of parameters at the cellular level. However, manual counting of these images is both tiring and unreliable and ultimately very time-consuming for biomedical researchers. Not only does this slow down the overall research process, it also introduces counting errors due to a lack of objectivity and consistency inherent to the researchers own human nature. This thesis addresses this issue by automatically determining infection indexes of macrophages parasite by Leishmania in microscopy images using computer vision and pattern recognition methodologies. Initially images are submitted to a pre-processing stage that consists in a normalization of illumination conditions. Three algorithms are then applied in parallel to each image. Algorithm A intends to detect macrophage nuclei and consists of segmentation via adaptive multi-threshold, and classification of resulting regions using a set of collected features. Algorithm B intends to detect parasites and is similar to Algorithm A but the adaptive multi-threshold is parameterized with a different constraints vector. Algorithm C intends to detect the macrophages and parasites cytoplasm and consists of a cut-off version of the previous two algorithms, where the classification step is skipped. Regions with multiple nuclei or parasites are processed by a voting system that employs both a Support Vector Machine and a set of region features for determining the number of objects present in each region. The previous vote is then taken into account as the number of mixtures to be used in a Gaussian Mixture Model to decluster the said region. Finally each parasite is assigned to, at most, a single macrophage using minimum Euclidean distance to a cell nucleus, thus quantifying Leishmania infection levels.
1311.2626
Second-order Shape Optimization for Geometric Inverse Problems in Vision
cs.CV
We develop a method for optimization in shape spaces, i.e., sets of surfaces modulo re-parametrization. Unlike previously proposed gradient flows, we achieve superlinear convergence rates through a subtle approximation of the shape Hessian, which is generally hard to compute and suffers from a series of degeneracies. Our analysis highlights the role of mean curvature motion in comparison with first-order schemes: instead of surface area, our approach penalizes deformation, either by its Dirichlet energy or total variation. Latter regularizer sparks the development of an alternating direction method of multipliers on triangular meshes. Therein, a conjugate-gradients solver enables us to bypass formation of the Gaussian normal equations appearing in the course of the overall optimization. We combine all of the aforementioned ideas in a versatile geometric variation-regularized Levenberg-Marquardt-type method applicable to a variety of shape functionals, depending on intrinsic properties of the surface such as normal field and curvature as well as its embedding into space. Promising experimental results are reported.
1311.2634
A New Look at Dual-Hop Relaying: Performance Limits with Hardware Impairments
cs.IT math.IT
Physical transceivers have hardware impairments that create distortions which degrade the performance of communication systems. The vast majority of technical contributions in the area of relaying neglect hardware impairments and, thus, assumes ideal hardware. Such approximations make sense in low-rate systems, but can lead to very misleading results when analyzing future high-rate systems. This paper quantifies the impact of hardware impairments on dual-hop relaying, for both amplify-and-forward and decode-and-forward protocols. The outage probability (OP) in these practical scenarios is a function of the effective end-to-end signal-to-noise-and-distortion ratio (SNDR). This paper derives new closed-form expressions for the exact and asymptotic OPs, accounting for hardware impairments at the source, relay, and destination. A similar analysis for the ergodic capacity is also pursued, resulting in new upper bounds. We assume that both hops are subject to independent but non-identically distributed Nakagami-m fading. This paper validates that the performance loss is small at low rates, but otherwise can be very substantial. In particular, it is proved that for high signal-to-noise ratio (SNR), the end-to-end SNDR converges to a deterministic constant, coined the SNDR ceiling, which is inversely proportional to the level of impairments. This stands in contrast to the ideal hardware case in which the end-to-end SNDR grows without bound in the high-SNR regime. Finally, we provide fundamental design guidelines for selecting hardware that satisfies the requirements of a practical relaying system.
1311.2637
Self-Dual codes from $(-1,1)$-matrices of skew symmetric type
cs.IT math.CO math.IT
Previously, self-dual codes have been constructed from weighing matrices, and in particular from conference matrices (skew and symmetric). In this paper, codes constructed from matrices of skew symmetric type whose determinants reach the Ehlich-Wojtas' bound are presented. A necessary and sufficient condition for these codes to be self-dual is given, and examples are provided for lengths up to 52.
1311.2642
Volumetric Reconstruction Applied to Perceptual Studies of Size and Weight
cs.CV
We explore the application of volumetric reconstruction from structured-light sensors in cognitive neuroscience, specifically in the quantification of the size-weight illusion, whereby humans tend to systematically perceive smaller objects as heavier. We investigate the performance of two commercial structured-light scanning systems in comparison to one we developed specifically for this application. Our method has two main distinct features: First, it only samples a sparse series of viewpoints, unlike other systems such as the Kinect Fusion. Second, instead of building a distance field for the purpose of points-to-surface conversion directly, we pursue a first-order approach: the distance function is recovered from its gradient by a screened Poisson reconstruction, which is very resilient to noise and yet preserves high-frequency signal components. Our experiments show that the quality of metric reconstruction from structured light sensors is subject to systematic biases, and highlights the factors that influence it. Our main performance index rates estimates of volume (a proxy of size), for which we review a well-known formula applicable to incomplete meshes. Our code and data will be made publicly available upon completion of the anonymous review process.
1311.2650
Over-the-air Signaling in Cellular Networks: An Overview
cs.IT cs.NI math.IT
To improve the capacity and coverage of current cellular networks, many advanced technologies such as massive MIMO, inter-cell coordination, small cells, device-to-device communications, and so on, are under studying. Many proposed techniques have been shown to offer significant performance improvement. Thus, the enabler of those techniques is of great importance. That is the necessary signaling which guarantee the operation of those techniques. The design and transmission of those signaling, especially the over-the-air (OTA) signaling, is challenging. In this article, we provide an overview of the OTA signaling in cellular networks to provide insights on the design of OTA signaling. Specifically, we first give a brief introduction of the OTA signaling in long term evolution (LTE), and then we discuss the challenges and requirements in designing the OTA signaling in cellular networks in detail. To better understand the OTA signaling, we give two important classifications of OTA signaling and address their properties and applications. Finally, we propose a signature-based signaling named (single-tone signaling, STS) which can be used for inter-cell OTA signaling and is especially useful and robust in multi-signal scenario. Simulation results are given to compare the detection performance of different OTA signaling.
1311.2651
MIMO Broadcast Channel with an Unknown Eavesdropper: Secrecy Degrees of Freedom
cs.IT math.IT
We study a multi-antenna broadcast channel with two legitimate receivers and an external eavesdropper. We assume that the channel matrix of the eavesdropper is unknown to the legitimate terminals but satisfies a maximum rank constraint. As our main result we characterize the associated secrecy degrees of freedom for the broadcast channel with common and private messages. We show that a direct extension of the single-user wiretap codebook does not achieve the secrecy degrees of freedom. Our proposed optimal scheme involves decomposing the signal space into a common subspace, which can be observed by both receivers, and private subspaces which can be observed by only one of the receivers, and carefully transmitting a subset of messages in each subspace. We also consider the case when each user's private message must additionally remain confidential from the other legitimate receiver and characterize the s.d.o.f.\ region in this case.
1311.2663
DinTucker: Scaling up Gaussian process models on multidimensional arrays with billions of elements
cs.LG cs.DC stat.ML
Infinite Tucker Decomposition (InfTucker) and random function prior models, as nonparametric Bayesian models on infinite exchangeable arrays, are more powerful models than widely-used multilinear factorization methods including Tucker and PARAFAC decomposition, (partly) due to their capability of modeling nonlinear relationships between array elements. Despite their great predictive performance and sound theoretical foundations, they cannot handle massive data due to a prohibitively high training time. To overcome this limitation, we present Distributed Infinite Tucker (DINTUCKER), a large-scale nonlinear tensor decomposition algorithm on MAPREDUCE. While maintaining the predictive accuracy of InfTucker, it is scalable on massive data. DINTUCKER is based on a new hierarchical Bayesian model that enables local training of InfTucker on subarrays and information integration from all local training results. We use distributed stochastic gradient descent, coupled with variational inference, to train this model. We apply DINTUCKER to multidimensional arrays with billions of elements from applications in the "Read the Web" project (Carlson et al., 2010) and in information security and compare it with the state-of-the-art large-scale tensor decomposition method, GigaTensor. On both datasets, DINTUCKER achieves significantly higher prediction accuracy with less computational time.
1311.2669
Distance-based and continuum Fano inequalities with applications to statistical estimation
cs.IT math.IT math.ST stat.TH
In this technical note, we give two extensions of the classical Fano inequality in information theory. The first extends Fano's inequality to the setting of estimation, providing lower bounds on the probability that an estimator of a discrete quantity is within some distance $t$ of the quantity. The second inequality extends our bound to a continuum setting and provides a volume-based bound. We illustrate how these inequalities lead to direct and simple proofs of several statistical minimax lower bounds.
1311.2670
Social Network Integration: Towards Constructing the Social Graph
cs.SI physics.soc-ph
In this work, we formulate the problem of social network integration. It takes multiple observed social networks as input and returns an integrated global social graph where each node corresponds to a real person. The key challenge for social network integration is to discover the correspondences or interlinks across different social networks. We engaged an in-depth analysis across three online social networks, AMiner, Linkedin, and Videolectures in order to address what reveals users' social identity, whether the social factors consistent across different social networks and how we can leverage these information to perform integration. We proposed a unified framework for the social network integration task. It crawls data from multiple social networks and further discovers accounts correspond to the same real person from the obtained networks. We use a probabilistic model to determine such correspondence, it incorporates features like the consistency of social status and social ties across different, as well as one-to-one mapping constraint and logical transitivity to jointly make the prediction. Empirical experiments verify the effectiveness of our method.
1311.2677
Sampling Based Approaches to Handle Imbalances in Network Traffic Dataset for Machine Learning Techniques
cs.NI cs.CR cs.LG
Network traffic data is huge, varying and imbalanced because various classes are not equally distributed. Machine learning (ML) algorithms for traffic analysis uses the samples from this data to recommend the actions to be taken by the network administrators as well as training. Due to imbalances in dataset, it is difficult to train machine learning algorithms for traffic analysis and these may give biased or false results leading to serious degradation in performance of these algorithms. Various techniques can be applied during sampling to minimize the effect of imbalanced instances. In this paper various sampling techniques have been analysed in order to compare the decrease in variation in imbalances of network traffic datasets sampled for these algorithms. Various parameters like missing classes in samples, probability of sampling of the different instances have been considered for comparison.
1311.2694
Hypothesis Testing for Automated Community Detection in Networks
stat.ML cs.LG cs.SI math.ST physics.soc-ph stat.TH
Community detection in networks is a key exploratory tool with applications in a diverse set of areas, ranging from finding communities in social and biological networks to identifying link farms in the World Wide Web. The problem of finding communities or clusters in a network has received much attention from statistics, physics and computer science. However, most clustering algorithms assume knowledge of the number of clusters k. In this paper we propose to automatically determine k in a graph generated from a Stochastic Blockmodel. Our main contribution is twofold; first, we theoretically establish the limiting distribution of the principal eigenvalue of the suitably centered and scaled adjacency matrix, and use that distribution for our hypothesis test. Secondly, we use this test to design a recursive bipartitioning algorithm. Using quantifiable classification tasks on real world networks with ground truth, we show that our algorithm outperforms existing probabilistic models for learning overlapping clusters, and on unlabeled networks, we show that we uncover nested community structure.
1311.2698
Packet Travel Times in Wireless Relay Chains under Spatially and Temporally Dependent Interference
cs.NI cs.IT math.IT
We investigate the statistics of the number of time slots $T$ that it takes a packet to travel through a chain of wireless relays. Derivations are performed assuming an interference model for which interference possesses spatiotemporal dependency properties. When using this model, results are harder to arrive at analytically, but they are more realistic than the ones obtained in many related works that are based on independent interference models. First, we present a method for calculating the distribution of $T$. As the required computations are extensive, we also obtain simple expressions for the expected value $\mathrm{E} [T]$ and variance $\mathrm{var} [T]$. Finally, we calculate the asymptotic limit of the average speed of the packet. Our numerical results show that spatiotemporal dependence has a significant impact on the statistics of the travel time $T$. In particular, we show that, with respect to the independent interference case, $\mathrm{E} [T]$ and $\mathrm{var} [T]$ increase, whereas the packet speed decreases.
1311.2702
Verifiable Source Code Documentation in Controlled Natural Language
cs.SE cs.AI cs.CL cs.HC cs.LO
Writing documentation about software internals is rarely considered a rewarding activity. It is highly time-consuming and the resulting documentation is fragile when the software is continuously evolving in a multi-developer setting. Unfortunately, traditional programming environments poorly support the writing and maintenance of documentation. Consequences are severe as the lack of documentation on software structure negatively impacts the overall quality of the software product. We show that using a controlled natural language with a reasoner and a query engine is a viable technique for verifying the consistency and accuracy of documentation and source code. Using ACE, a state-of-the-art controlled natural language, we present positive results on the comprehensibility and the general feasibility of creating and verifying documentation. As a case study, we used automatic documentation verification to identify and fix severe flaws in the architecture of a non-trivial piece of software. Moreover, a user experiment shows that our language is faster and easier to learn and understand than other formal languages for software documentation.
1311.2705
Quantum Stabilizer Codes from Maximal Curves
cs.IT math.AG math.IT
A curve attaining the Hasse-Weil bound is called a maximal curve. Usually classical error-correcting codes obtained from a maximal curve have good parameters. However, the quantum stabilizer codes obtained from such classical error-correcting codes via Euclidean or Hermitian self-orthogonality do not always possess good parameters. In this paper, the Hermitian self-orthogonality of algebraic geometry codes obtained from two maximal curves is investigated. It turns out that the stabilizer quantum codes produced from such Hermitian self-orthogonal classical codes have good parameters.
1311.2745
Sparse Phase Retrieval: Uniqueness Guarantees and Recovery Algorithms
cs.IT math.IT math.OC
The problem of signal recovery from its Fourier transform magnitude is of paramount importance in various fields of engineering and has been around for over 100 years. Due to the absence of phase information, some form of additional information is required in order to be able to uniquely identify the signal of interest. In this work, we focus our attention on discrete-time sparse signals (of length $n$). We first show that, if the DFT dimension is greater than or equal to $2n$, almost all signals with {\em aperiodic} support can be uniquely identified by their Fourier transform magnitude (up to time-shift, conjugate-flip and global phase). Then, we develop an efficient Two-stage Sparse Phase Retrieval algorithm (TSPR), which involves: (i) identifying the support, i.e., the locations of the non-zero components, of the signal using a combinatorial algorithm (ii) identifying the signal values in the support using a convex algorithm. We show that TSPR can {\em provably} recover most $O(n^{1/2-\eps})$-sparse signals (up to a time-shift, conjugate-flip and global phase). We also show that, for most $O(n^{1/4-\eps})$-sparse signals, the recovery is {\em robust} in the presence of measurement noise. Numerical experiments complement our theoretical analysis and verify the effectiveness of TSPR.
1311.2746
Deep neural networks for single channel source separation
cs.NE cs.LG
In this paper, a novel approach for single channel source separation (SCSS) using a deep neural network (DNN) architecture is introduced. Unlike previous studies in which DNN and other classifiers were used for classifying time-frequency bins to obtain hard masks for each source, we use the DNN to classify estimated source spectra to check for their validity during separation. In the training stage, the training data for the source signals are used to train a DNN. In the separation stage, the trained DNN is utilized to aid in estimation of each source in the mixed signal. Single channel source separation problem is formulated as an energy minimization problem where each source spectra estimate is encouraged to fit the trained DNN model and the mixed signal spectrum is encouraged to be written as a weighted sum of the estimated source spectra. The proposed approach works regardless of the energy scale differences between the source signals in the training and separation stages. Nonnegative matrix factorization (NMF) is used to initialize the DNN estimate for each source. The experimental results show that using DNN initialized by NMF for source separation improves the quality of the separated signal compared with using NMF for source separation.
1311.2795
Complete solution of a constrained tropical optimization problem with application to location analysis
math.OC cs.SY
We present a multidimensional optimization problem that is formulated and solved in the tropical mathematics setting. The problem consists of minimizing a nonlinear objective function defined on vectors over an idempotent semifield by means of a conjugate transposition operator, subject to constraints in the form of linear vector inequalities. A complete direct solution to the problem under fairly general assumptions is given in a compact vector form suitable for both further analysis and practical implementation. We apply the result to solve a multidimensional minimax single facility location problem with Chebyshev distance and with inequality constraints imposed on the feasible location area.
1311.2796
Mixed Human-Robot Team Surveillance
math.OC cs.RO cs.SY
We study the mixed human-robot team design in a system theoretic setting using the context of a surveillance mission. The three key coupled components of a mixed team design are (i) policies for the human operator, (ii) policies to account for erroneous human decisions, and (iii) policies to control the automaton. In this paper, we survey elements of human decision-making, including evidence aggregation, situational awareness, fatigue, and memory effects. We bring together the models for these elements in human decision-making to develop a single coherent model for human decision-making in a two-alternative choice task. We utilize the developed model to design efficient attention allocation policies for the human operator. We propose an anomaly detection algorithm that utilizes potentially erroneous decision by the operator to ascertain an anomalous region among the set of regions surveilled. Finally, we propose a stochastic vehicle routing policy that surveils an anomalous region with high probability. Our mixed team design relies on the certainty-equivalent receding-horizon control framework.
1311.2799
Aggregation of Affine Estimators
math.ST cs.LG stat.TH
We consider the problem of aggregating a general collection of affine estimators for fixed design regression. Relevant examples include some commonly used statistical estimators such as least squares, ridge and robust least squares estimators. Dalalyan and Salmon (2012) have established that, for this problem, exponentially weighted (EW) model selection aggregation leads to sharp oracle inequalities in expectation, but similar bounds in deviation were not previously known. While results indicate that the same aggregation scheme may not satisfy sharp oracle inequalities with high probability, we prove that a weaker notion of oracle inequality for EW that holds with high probability. Moreover, using a generalization of the newly introduced $Q$-aggregation scheme we also prove sharp oracle inequalities that hold with high probability. Finally, we apply our results to universal aggregation and show that our proposed estimator leads simultaneously to all the best known bounds for aggregation, including $\ell_q$-aggregation, $q \in (0,1)$, with high probability.
1311.2838
A PAC-Bayesian bound for Lifelong Learning
stat.ML cs.LG
Transfer learning has received a lot of attention in the machine learning community over the last years, and several effective algorithms have been developed. However, relatively little is known about their theoretical properties, especially in the setting of lifelong learning, where the goal is to transfer information to tasks for which no data have been observed so far. In this work we study lifelong learning from a theoretical perspective. Our main result is a PAC-Bayesian generalization bound that offers a unified view on existing paradigms for transfer learning, such as the transfer of parameters or the transfer of low-dimensional representations. We also use the bound to derive two principled lifelong learning algorithms, and we show that these yield results comparable with existing methods.
1311.2850
Virtual Modules in Discrete-Event Systems: Achieving Modular Diagnosability
cs.SY
This paper deals with the problem of enforcing modular diagnosability for discrete-event systems that don't satisfy this property by their natural modularity. We introduce an approach to achieve this property combining existing modules into new virtual modules. An underlining mathematical problem is to find a partition of a set, such that the partition satisfies the required property. The time complexity of such problem is very high. To overcome it, the paper introduces a structural analysis of the system's modules. In the analysis we focus on the case when the modules participate in diagnosis with their observations, rather then the case when indistinguishable observations are blocked due to concurrency.
1311.2852
Quantifying unique information
cs.IT math.IT
We propose new measures of shared information, unique information and synergistic information that can be used to decompose the multi-information of a pair of random variables $(Y,Z)$ with a third random variable $X$. Our measures are motivated by an operational idea of unique information which suggests that shared information and unique information should depend only on the pair marginal distributions of $(X,Y)$ and $(X,Z)$. Although this invariance property has not been studied before, it is satisfied by other proposed measures of shared information. The invariance property does not uniquely determine our new measures, but it implies that the functions that we define are bounds to any other measures satisfying the same invariance property. We study properties of our measures and compare them to other candidate measures.
1311.2854
Spectral Clustering via the Power Method -- Provably
cs.LG cs.NA
Spectral clustering is one of the most important algorithms in data mining and machine intelligence; however, its computational complexity limits its application to truly large scale data analysis. The computational bottleneck in spectral clustering is computing a few of the top eigenvectors of the (normalized) Laplacian matrix corresponding to the graph representing the data to be clustered. One way to speed up the computation of these eigenvectors is to use the "power method" from the numerical linear algebra literature. Although the power method has been empirically used to speed up spectral clustering, the theory behind this approach, to the best of our knowledge, remains unexplored. This paper provides the \emph{first} such rigorous theoretical justification, arguing that a small number of power iterations suffices to obtain near-optimal partitionings using the approximate eigenvectors. Specifically, we prove that solving the $k$-means clustering problem on the approximate eigenvectors obtained via the power method gives an additive-error approximation to solving the $k$-means problem on the optimal eigenvectors.
1311.2878
Selection Effects in Online Sharing: Consequences for Peer Adoption
cs.SI physics.soc-ph
Most models of social contagion take peer exposure to be a corollary of adoption, yet in many settings, the visibility of one's adoption behavior happens through a separate decision process. In online systems, product designers can define how peer exposure mechanisms work: adoption behaviors can be shared in a passive, automatic fashion, or occur through explicit, active sharing. The consequences of these mechanisms are of substantial practical and theoretical interest: passive sharing may increase total peer exposure but active sharing may expose higher quality products to peers who are more likely to adopt. We examine selection effects in online sharing through a large-scale field experiment on Facebook that randomizes whether or not adopters share Offers (coupons) in a passive manner. We derive and estimate a joint discrete choice model of adopters' sharing decisions and their peers' adoption decisions. Our results show that active sharing enables a selection effect that exposes peers who are more likely to adopt than the population exposed under passive sharing. We decompose the selection effect into two distinct mechanisms: active sharers expose peers to higher quality products, and the peers they share with are more likely to adopt independently of product quality. Simulation results show that the user-level mechanism comprises the bulk of the selection effect. The study's findings are among the first to address downstream peer effects induced by online sharing mechanisms, and can inform design in settings where a surplus of sharing could be viewed as costly.
1311.2886
A Fuzzy AHP Approach for Supplier Selection Problem: A Case Study in a Gear Motor Company
cs.AI
Suuplier selection is one of the most important functions of a purchasing department. Since by deciding the best supplier, companies can save material costs and increase competitive advantage.However this decision becomes compilcated in case of multiple suppliers, multiple conflicting criteria, and imprecise parameters. In addition the uncertainty and vagueness of the experts' opinion is the prominent characteristic of the problem. therefore an extensively used multi criteria decision making tool Fuzzy AHP can be utilized as an approach for supplier selection problem. This paper reveals the application of Fuzzy AHP in a gear motor company determining the best supplier with respect to selected criteria. the contribution of this study is not only the application of the Fuzzy AHP methodology for supplier selection problem, but also releasing a comprehensive literature review of multi criteria decision making problems. In addition by stating the steps of Fuzzy AHP clearly and numerically, this study can be a guide of the methodology to be implemented to other multiple criteria decision making problems.
1311.2887
Are all Social Networks Structurally Similar? A Comparative Study using Network Statistics and Metrics
cs.SI
The modern age has seen an exponential growth of social network data available on the web. Analysis of these networks reveal important structural information about these networks in particular and about our societies in general. More often than not, analysis of these networks is concerned in identifying similarities among social networks and how they are different from other networks such as protein interaction networks, computer networks and food web. In this paper, our objective is to perform a critical analysis of different social networks using structural metrics in an effort to highlight their similarities and differences. We use five different social network datasets which are contextually and semantically different from each other. We then analyze these networks using a number of different network statistics and metrics. Our results show that although these social networks have been constructed from different contexts, they are structurally similar. We also review the snowball sampling method and show its vulnerability against different network metrics.
1311.2889
Reinforcement Learning for Matrix Computations: PageRank as an Example
cs.LG cs.SI stat.ML
Reinforcement learning has gained wide popularity as a technique for simulation-driven approximate dynamic programming. A less known aspect is that the very reasons that make it effective in dynamic programming can also be leveraged for using it for distributed schemes for certain matrix computations involving non-negative matrices. In this spirit, we propose a reinforcement learning algorithm for PageRank computation that is fashioned after analogous schemes for approximate dynamic programming. The algorithm has the advantage of ease of distributed implementation and more importantly, of being model-free, i.e., not dependent on any specific assumptions about the transition probabilities in the random web-surfer model. We analyze its convergence and finite time behavior and present some supporting numerical experiments.
1311.2891
The More, the Merrier: the Blessing of Dimensionality for Learning Large Gaussian Mixtures
cs.LG cs.DS stat.ML
In this paper we show that very large mixtures of Gaussians are efficiently learnable in high dimension. More precisely, we prove that a mixture with known identical covariance matrices whose number of components is a polynomial of any fixed degree in the dimension n is polynomially learnable as long as a certain non-degeneracy condition on the means is satisfied. It turns out that this condition is generic in the sense of smoothed complexity, as soon as the dimensionality of the space is high enough. Moreover, we prove that no such condition can possibly exist in low dimension and the problem of learning the parameters is generically hard. In contrast, much of the existing work on Gaussian Mixtures relies on low-dimensional projections and thus hits an artificial barrier. Our main result on mixture recovery relies on a new "Poissonization"-based technique, which transforms a mixture of Gaussians to a linear map of a product distribution. The problem of learning this map can be efficiently solved using some recent results on tensor decompositions and Independent Component Analysis (ICA), thus giving an algorithm for recovering the mixture. In addition, we combine our low-dimensional hardness results for Gaussian mixtures with Poissonization to show how to embed difficult instances of low-dimensional Gaussian mixtures into the ICA setting, thus establishing exponential information-theoretic lower bounds for underdetermined ICA in low dimension. To the best of our knowledge, this is the first such result in the literature. In addition to contributing to the problem of Gaussian mixture learning, we believe that this work is among the first steps toward better understanding the rare phenomenon of the "blessing of dimensionality" in the computational aspects of statistical inference.
1311.2897
Exponential Stability of Homogeneous Positive Systems of Degree One With Time-Varying Delays
cs.SY
While the asymptotic stability of positive linear systems in the presence of bounded time delays has been thoroughly investigated, the theory for nonlinear positive systems is considerably less well-developed. This paper presents a set of conditions for establishing delay-independent stability and bounding the decay rate of a significant class of nonlinear positive systems which includes positive linear systems as a special case. Specifically, when the time delays have a known upper bound, we derive necessary and sufficient conditions for exponential stability of (a) continuous-time positive systems whose vector fields are homogeneous and cooperative, and (b) discrete-time positive systems whose vector fields are homogeneous and order preserving. We then present explicit expressions that allow us to quantify the impact of delays on the decay rate and show that the best decay rate of positive linear systems that our bounds provide can be found via convex optimization. Finally, we extend the results to general linear systems with time-varying delays.
1311.2901
Visualizing and Understanding Convolutional Networks
cs.CV
Large Convolutional Network models have recently demonstrated impressive classification performance on the ImageNet benchmark. However there is no clear understanding of why they perform so well, or how they might be improved. In this paper we address both issues. We introduce a novel visualization technique that gives insight into the function of intermediate feature layers and the operation of the classifier. We also perform an ablation study to discover the performance contribution from different model layers. This enables us to find model architectures that outperform Krizhevsky \etal on the ImageNet classification benchmark. We show our ImageNet model generalizes well to other datasets: when the softmax classifier is retrained, it convincingly beats the current state-of-the-art results on Caltech-101 and Caltech-256 datasets.
1311.2903
Protocol Design and Stability Analysis of Cooperative Cognitive Radio Users
cs.NI cs.IT math.IT
A single cognitive radio transmitter--receiver pair shares the spectrum with two primary users communicating with their respective receivers. Each primary user has a local traffic queue, whereas the cognitive user has three queues; one storing its own traffic while the other two are relaying queues used to store primary relayed packets admitted from the two primary users. A new cooperative cognitive medium access control protocol for the described network is proposed, where the cognitive user exploits the idle periods of the primary spectrum bands. Traffic arrival to each relaying queue is controlled using a tuneable admittance factor, while relaying queues service scheduling is controlled via channel access probabilities assigned to each queue based on the band of operation. The stability region of the proposed protocol is characterized shedding light on its maximum expected throughput. Numerical results demonstrate the performance gains of the proposed cooperative cognitive protocol.
1311.2906
Centrality in Interconnected Multilayer Networks
physics.soc-ph cond-mat.dis-nn cs.SI
Real-world complex systems exhibit multiple levels of relationships. In many cases, they require to be modeled by interconnected multilayer networks, characterizing interactions on several levels simultaneously. It is of crucial importance in many fields, from economics to biology, from urban planning to social sciences, to identify the most (or the less) influent nodes in a network. However, defining the centrality of actors in an interconnected structure is not trivial. In this paper, we capitalize on the tensorial formalism, recently proposed to characterize and investigate this kind of complex topologies, to show how several centrality measures -- well-known in the case of standard ("monoplex") networks -- can be extended naturally to the realm of interconnected multiplexes. We consider diagnostics widely used in different fields, e.g., computer science, biology, communication and social sciences, to cite only some of them. We show, both theoretically and numerically, that using the weighted monoplex obtained by aggregating the multilayer network leads, in general, to relevant differences in ranking the nodes by their importance.
1311.2911
Exploring universal patterns in human home-work commuting from mobile phone data
cs.SI cs.CY physics.soc-ph
Home-work commuting has always attracted significant research attention because of its impact on human mobility. One of the key assumptions in this domain of study is the universal uniformity of commute times. However, a true comparison of commute patterns has often been hindered by the intrinsic differences in data collection methods, which make observation from different countries potentially biased and unreliable. In the present work, we approach this problem through the use of mobile phone call detail records (CDRs), which offers a consistent method for investigating mobility patterns in wholly different parts of the world. We apply our analysis to a broad range of datasets, at both the country and city scale. Additionally, we compare these results with those obtained from vehicle GPS traces in Milan. While different regions have some unique commute time characteristics, we show that the home-work time distributions and average values within a single region are indeed largely independent of commute distance or country (Portugal, Ivory Coast, and Boston)--despite substantial spatial and infrastructural differences. Furthermore, a comparative analysis demonstrates that such distance-independence holds true only if we consider multimodal commute behaviors--as consistent with previous studies. In car-only (Milan GPS traces) and car-heavy (Saudi Arabia) commute datasets, we see that commute time is indeed influenced by commute distance.
1311.2912
A Misanthropic Reinterpretation of the Chinese Room Problem
cs.AI
The chinese room problem asks if computers can think; I ask here if most humans can.
1311.2914
A novel local search based on variable-focusing for random K-SAT
cs.AI cond-mat.dis-nn
We introduce a new local search algorithm for satisfiability problems. Usual approaches focus uniformly on unsatisfied clauses. The new method works by picking uniformly random variables in unsatisfied clauses. A Variable-based Focused Metropolis Search (V-FMS) is then applied to random 3-SAT. We show that it is quite comparable in performance to the clause-based FMS. Consequences for algorithmic design are discussed.
1311.2971
Approximate Inference in Continuous Determinantal Point Processes
stat.ML cs.LG stat.ME
Determinantal point processes (DPPs) are random point processes well-suited for modeling repulsion. In machine learning, the focus of DPP-based models has been on diverse subset selection from a discrete and finite base set. This discrete setting admits an efficient sampling algorithm based on the eigendecomposition of the defining kernel matrix. Recently, there has been growing interest in using DPPs defined on continuous spaces. While the discrete-DPP sampler extends formally to the continuous case, computationally, the steps required are not tractable in general. In this paper, we present two efficient DPP sampling schemes that apply to a wide range of kernel functions: one based on low rank approximations via Nystrom and random Fourier feature techniques and another based on Gibbs sampling. We demonstrate the utility of continuous DPPs in repulsive mixture modeling and synthesizing human poses spanning activity spaces.
1311.2972
Learning Mixtures of Discrete Product Distributions using Spectral Decompositions
stat.ML cs.CC cs.IT cs.LG math.IT
We study the problem of learning a distribution from samples, when the underlying distribution is a mixture of product distributions over discrete domains. This problem is motivated by several practical applications such as crowd-sourcing, recommendation systems, and learning Boolean functions. The existing solutions either heavily rely on the fact that the number of components in the mixtures is finite or have sample/time complexity that is exponential in the number of components. In this paper, we introduce a polynomial time/sample complexity method for learning a mixture of $r$ discrete product distributions over $\{1, 2, \dots, \ell\}^n$, for general $\ell$ and $r$. We show that our approach is statistically consistent and further provide finite sample guarantees. We use techniques from the recent work on tensor decompositions for higher-order moment matching. A crucial step in these moment matching methods is to construct a certain matrix and a certain tensor with low-rank spectral decompositions. These tensors are typically estimated directly from the samples. The main challenge in learning mixtures of discrete product distributions is that these low-rank tensors cannot be obtained directly from the sample moments. Instead, we reduce the tensor estimation problem to: $a$) estimating a low-rank matrix using only off-diagonal block elements; and $b$) estimating a tensor using a small number of linear measurements. Leveraging on recent developments in matrix completion, we give an alternating minimization based method to estimate the low-rank matrix, and formulate the tensor completion problem as a least-squares problem.
1311.2978
Authorship Attribution Using Word Network Features
cs.CL
In this paper, we explore a set of novel features for authorship attribution of documents. These features are derived from a word network representation of natural language text. As has been noted in previous studies, natural language tends to show complex network structure at word level, with low degrees of separation and scale-free (power law) degree distribution. There has also been work on authorship attribution that incorporates ideas from complex networks. The goal of our paper is to explore properties of these complex networks that are suitable as features for machine-learning-based authorship attribution of documents. We performed experiments on three different datasets, and obtained promising results.
1311.2987
Learning Input and Recurrent Weight Matrices in Echo State Networks
cs.LG
Echo State Networks (ESNs) are a special type of the temporally deep network model, the Recurrent Neural Network (RNN), where the recurrent matrix is carefully designed and both the recurrent and input matrices are fixed. An ESN uses the linearity of the activation function of the output units to simplify the learning of the output matrix. In this paper, we devise a special technique that take advantage of this linearity in the output units of an ESN, to learn the input and recurrent matrices. This has not been done in earlier ESNs due to their well known difficulty in learning those matrices. Compared to the technique of BackPropagation Through Time (BPTT) in learning general RNNs, our proposed method exploits linearity of activation function in the output units to formulate the relationships amongst the various matrices in an RNN. These relationships results in the gradient of the cost function having an analytical form and being more accurate. This would enable us to compute the gradients instead of obtaining them by recursion as in BPTT. Experimental results on phone state classification show that learning one or both the input and recurrent matrices in an ESN yields superior results compared to traditional ESNs that do not learn these matrices, especially when longer time steps are used.
1311.3001
Informed Source Separation: A Bayesian Tutorial
stat.ML cs.LG
Source separation problems are ubiquitous in the physical sciences; any situation where signals are superimposed calls for source separation to estimate the original signals. In this tutorial I will discuss the Bayesian approach to the source separation problem. This approach has a specific advantage in that it requires the designer to explicitly describe the signal model in addition to any other information or assumptions that go into the problem description. This leads naturally to the idea of informed source separation, where the algorithm design incorporates relevant information about the specific problem. This approach promises to enable researchers to design their own high-quality algorithms that are specifically tailored to the problem at hand.
1311.3009
A Construction of New Quantum MDS Codes
cs.IT math.IT
It has been a great challenge to construct new quantum MDS codes. In particular, it is very hard to construct quantum MDS codes with relatively large minimum distance. So far, except for some sparse lengths, all known $q$-ary quantum MDS codes have minimum distance less than or equal to $q/2+1$. In the present paper, we provide a construction of quantum MDS codes with minimum distance bigger than $q/2+1$. In particular, we show existence of $q$-ary quantum MDS codes with length $n=q^2+1$ and minimum distance $d$ for any $d\le q-1$ and $d= q+1$(this result extends those given in \cite{Gu11,Jin1,KZ12}); and with length $(q^2+2)/3$ and minimum distance $d$ for any $d\le (2q+2)/3$ if $3|(q+1)$. Our method is through Hermitian self-orthogonal codes. The main idea of constructing Hermitian self-orthogonal codes is based on the solvability in $\F_q$ of a system of homogenous equations over $\F_{q^2}$.
1311.3011
Cornell SPF: Cornell Semantic Parsing Framework
cs.CL
The Cornell Semantic Parsing Framework (SPF) is a learning and inference framework for mapping natural language to formal representation of its meaning.
1311.3023
Asynchronous Distributed Downlink Beamforming and Power Control in Multi-cell Networks
cs.IT math.IT
In this paper, we consider a multi-cell network where every base station (BS) serves multiple users with an antenna array. Each user is associated with only one BS and has a single antenna. Assume that only long-term channel state information (CSI) is available in the system. The objective is to minimize the network downlink transmission power needed to meet the users' signal-to-interference-plus-noise ratio (SINR) requirements. For this objective, we propose an asynchronous distributed beamforming and power control algorithm which provides the same optimal solution as given by centralized algorithms. To design the algorithm, the power minimization problem is formulated mathematically as a non-convex problem. For distributed implementation, the non-convex problem is cast into the dual decomposition framework. Resorting to the theory about matrix pencil, a novel asynchronous iterative method is proposed for solving the dual of the non-convex problem. The methods for beamforming and power control are obtained by investigating the primal problem. At last, simulation results are provided to demonstrate the convergence and performance of the algorithm.
1311.3037
Practical Characterization of Large Networks Using Neighborhood Information
cs.SI cs.CY physics.soc-ph
Characterizing large online social networks (OSNs) through node querying is a challenging task. OSNs often impose severe constraints on the query rate, hence limiting the sample size to a small fraction of the total network. Various ad-hoc subgraph sampling methods have been proposed, but many of them give biased estimates and no theoretical basis on the accuracy. In this work, we focus on developing sampling methods for OSNs where querying a node also reveals partial structural information about its neighbors. Our methods are optimized for NoSQL graph databases (if the database can be accessed directly), or utilize Web API available on most major OSNs for graph sampling. We show that our sampling method has provable convergence guarantees on being an unbiased estimator, and it is more accurate than current state-of-the-art methods. We characterize metrics such as node label density estimation and edge label density estimation, two of the most fundamental network characteristics from which other network characteristics can be derived. We evaluate our methods on-the-fly over several live networks using their native APIs. Our simulation studies over a variety of offline datasets show that by including neighborhood information, our method drastically (4-fold) reduces the number of samples required to achieve the same estimation accuracy of state-of-the-art methods.
1311.3045
Joint Power and Admission Control: Non-Convex $L_q$ Approximation and An Effective Polynomial Time Deflation Approach
cs.IT cs.NI math.IT math.OC
In an interference limited network, joint power and admission control (JPAC) aims at supporting a maximum number of links at their specified signal to interference plus noise ratio (SINR) targets while using a minimum total transmission power. Various convex approximation deflation approaches have been developed for the JPAC problem. In this paper, we propose an effective polynomial time non-convex approximation deflation approach for solving the problem. The approach is based on the non-convex $\ell_q$-minimization approximation of an equivalent sparse $\ell_0$-minimization reformulation of the JPAC problem where $q\in(0,1).$ We show that, for any instance of the JPAC problem, there exists a $\bar q\in(0,1)$ such that it can be exactly solved by solving its $\ell_q$-minimization approximation problem with any $q\in(0, \bar q]$. We also show that finding the global solution of the $\ell_q$ approximation problem is NP-hard. Then, we propose a potential reduction interior-point algorithm, which can return an $\epsilon$-KKT solution of the NP-hard $\ell_q$-minimization approximation problem in polynomial time. The returned solution can be used to check the simultaneous supportability of all links in the network and to guide an iterative link removal procedure, resulting in the polynomial time non-convex approximation deflation approach for the JPAC problem. Numerical simulations show that the proposed approach outperforms the existing convex approximation approaches in terms of the number of supported links and the total transmission power, particularly exhibiting a quite good performance in selecting which subset of links to support.
1311.3062
Ants: Mobile Finite State Machines
cs.DC cs.MA
Consider the Ants Nearby Treasure Search (ANTS) problem introduced by Feinerman, Korman, Lotker, and Sereni (PODC 2012), where $n$ mobile agents, initially placed at the origin of an infinite grid, collaboratively search for an adversarially hidden treasure. In this paper, the model of Feinerman et al. is adapted such that the agents are controlled by a (randomized) finite state machine: they possess a constant-size memory and are able to communicate with each other through constant-size messages. Despite the restriction to constant-size memory, we show that their collaborative performance remains the same by presenting a distributed algorithm that matches a lower bound established by Feinerman et al. on the run-time of any ANTS algorithm.
1311.3064
Ranking users, papers and authors in online scientific communities
cs.SI cs.DL cs.IR physics.soc-ph
The ever-increasing quantity and complexity of scientific production have made it difficult for researchers to keep track of advances in their own fields. This, together with growing popularity of online scientific communities, calls for the development of effective information filtering tools. We propose here a method to simultaneously compute reputation of users and quality of scientific artifacts in an online scientific community. Evaluation on artificially-generated data and real data from the Econophysics Forum is used to determine the method's best-performing variants. We show that when the method is extended by considering author credit, its performance improves on multiple levels. In particular, top papers have higher citation count and top authors have higher $h$-index than top papers and top authors chosen by other algorithms.
1311.3076
An Efficient Method for Recognizing the Low Quality Fingerprint Verification by Means of Cross Correlation
cs.CV
In this paper, we propose an efficient method to provide personal identification using fingerprint to get better accuracy even in noisy condition. The fingerprint matching based on the number of corresponding minutia pairings, has been in use for a long time, which is not very efficient for recognizing the low quality fingerprints. To overcome this problem, correlation technique is used. The correlation-based fingerprint verification system is capable of dealing with low quality images from which no minutiae can be extracted reliably and with fingerprints that suffer from non-uniform shape distortions, also in case of damaged and partial images. Orientation Field Methodology (OFM) has been used as a preprocessing module, and it converts the images into a field pattern based on the direction of the ridges, loops and bifurcations in the image of a fingerprint. The input image is then Cross Correlated (CC) with all the images in the cluster and the highest correlated image is taken as the output. The result gives a good recognition rate, as the proposed scheme uses Cross Correlation of Field Orientation (CCFO = OFM + CC) for fingerprint identification.
1311.3085
Community detection thresholds and the weak Ramanujan property
cs.SI
Decelle et al.\cite{Decelle11} conjectured the existence of a sharp threshold for community detection in sparse random graphs drawn from the stochastic block model. Mossel et al.\cite{Mossel12} established the negative part of the conjecture, proving impossibility of meaningful detection below the threshold. However the positive part of the conjecture remained elusive so far. Here we solve the positive part of the conjecture. We introduce a modified adjacency matrix $B$ that counts self-avoiding paths of a given length $\ell$ between pairs of nodes and prove that for logarithmic $\ell$, the leading eigenvectors of this modified matrix provide non-trivial detection, thereby settling the conjecture. A key step in the proof consists in establishing a {\em weak Ramanujan property} of matrix $B$. Namely, the spectrum of $B$ consists in two leading eigenvalues $\rho(B)$, $\lambda_2$ and $n-2$ eigenvalues of a lower order $O(n^{\epsilon}\sqrt{\rho(B)})$ for all $\epsilon>0$, $\rho(B)$ denoting $B$'s spectral radius. $d$-regular graphs are Ramanujan when their second eigenvalue verifies $|\lambda|\le 2 \sqrt{d-1}$. Random $d$-regular graphs have a second largest eigenvalue $\lambda$ of $2\sqrt{d-1}+o(1)$ (see Friedman\cite{friedman08}), thus being {\em almost} Ramanujan. Erd\H{o}s-R\'enyi graphs with average degree $d$ at least logarithmic ($d=\Omega(\log n)$) have a second eigenvalue of $O(\sqrt{d})$ (see Feige and Ofek\cite{Feige05}), a slightly weaker version of the Ramanujan property. However this spectrum separation property fails for sparse ($d=O(1)$) Erd\H{o}s-R\'enyi graphs. Our result thus shows that by constructing matrix $B$ through neighborhood expansion, we regularize the original adjacency matrix to eventually recover a weak form of the Ramanujan property.
1311.3087
A Fractal and Scale-free Model of Complex Networks with Hub Attraction Behaviors
physics.soc-ph cs.SI
It is widely believed that fractality of complex networks origins from hub repulsion behaviors (anticorrelation or disassortativity), which means large degree nodes tend to connect with small degree nodes. This hypothesis was demonstrated by a dynamical growth model, which evolves as the inverse renormalization procedure proposed by Song et al. Now we find that the dynamical growth model is based on the assumption that all the cross-boxes links has the same probability e to link to the most connected nodes inside each box. Therefore, we modify the growth model by adopting the flexible probability e, which makes hubs have higher probability to connect with hubs than non-hubs. With this model, we find some fractal and scale-free networks have hub attraction behaviors (correlation or assortativity). The results are the counter-examples of former beliefs.
1311.3123
Polar Codes for Arbitrary DMCs and Arbitrary MACs
cs.IT math.IT
Polar codes are constructed for arbitrary channels by imposing an arbitrary quasigroup structure on the input alphabet. Just as with "usual" polar codes, the block error probability under successive cancellation decoding is $o(2^{-N^{1/2-\epsilon}})$, where $N$ is the block length. Encoding and decoding for these codes can be implemented with a complexity of $O(N\log N)$. It is shown that the same technique can be used to construct polar codes for arbitrary multiple access channels (MAC) by using an appropriate Abelian group structure. Although the symmetric sum capacity is achieved by this coding scheme, some points in the symmetric capacity region may not be achieved. In the case where the channel is a combination of linear channels, we provide a necessary and sufficient condition characterizing the channels whose symmetric capacity region is preserved by the polarization process. We also provide a sufficient condition for having a maximal loss in the dominant face.
1311.3141
Cloud Compute-and-Forward with Relay Cooperation
cs.IT cs.GT cs.NI math.IT
We study a cloud network with M distributed receiving antennas and L users, which transmit their messages towards a centralized decoder (CD), where M>=L. We consider that the cloud network applies the Compute-and-Forward (C&F) protocol, where L antennas/relays are selected to decode integer equations of the transmitted messages. In this work, we focus on the best relay selection and the optimization of the Physical-Layer Network Coding (PNC) at the relays, aiming at the throughput maximization of the network. Existing literature optimizes PNC with respect to the maximization of the minimum rate among users. The proposed strategy maximizes the sum rate of the users allowing nonsymmetric rates, while the optimal solution is explored with the aid of the Pareto frontier. The problem of relay selection is matched to a coalition formation game, where the relays and the CD cooperate in order to maximize their profit. Efficient coalition formation algorithms are proposed, which perform joint relay selection and PNC optimization. Simulation results show that a considerable improvement is achieved compared to existing results, both in terms of the network sum rate and the players' profits.
1311.3157
Multiple Closed-Form Local Metric Learning for K-Nearest Neighbor Classifier
cs.LG
Many researches have been devoted to learn a Mahalanobis distance metric, which can effectively improve the performance of kNN classification. Most approaches are iterative and computational expensive and linear rigidity still critically limits metric learning algorithm to perform better. We proposed a computational economical framework to learn multiple metrics in closed-form.
1311.3175
Architecture of an Ontology-Based Domain-Specific Natural Language Question Answering System
cs.CL cs.IR
Question answering (QA) system aims at retrieving precise information from a large collection of documents against a query. This paper describes the architecture of a Natural Language Question Answering (NLQA) system for a specific domain based on the ontological information, a step towards semantic web question answering. The proposed architecture defines four basic modules suitable for enhancing current QA capabilities with the ability of processing complex questions. The first module was the question processing, which analyses and classifies the question and also reformulates the user query. The second module allows the process of retrieving the relevant documents. The next module processes the retrieved documents, and the last module performs the extraction and generation of a response. Natural language processing techniques are used for processing the question and documents and also for answer extraction. Ontology and domain knowledge are used for reformulating queries and identifying the relations. The aim of the system is to generate short and specific answer to the question that is asked in the natural language in a specific domain. We have achieved 94 % accuracy of natural language question answering in our implementation.
1311.3192
Localizing Grasp Affordances in 3-D Points Clouds Using Taubin Quadric Fitting
cs.RO
Perception-for-grasping is a challenging problem in robotics. Inexpensive range sensors such as the Microsoft Kinect provide sensing capabilities that have given new life to the effort of developing robust and accurate perception methods for robot grasping. This paper proposes a new approach to localizing enveloping grasp affordances in 3-D point clouds efficiently. The approach is based on modeling enveloping grasp affordances as a cylindrical shells that corresponds to the geometry of the robot hand. A fast and accurate fitting method for quadratic surfaces is the core of our approach. An evaluation on a set of cluttered environments shows high precision and recall statistics. Our results also show that the approach compares favorably with some alternatives, and that it is efficient enough to be employed for robot grasping in real-time.
1311.3198
Sound, Complete and Minimal UCQ-Rewriting for Existential Rules
cs.AI cs.LO
We address the issue of Ontology-Based Data Access, with ontologies represented in the framework of existential rules, also known as Datalog+/-. A well-known approach involves rewriting the query using ontological knowledge. We focus here on the basic rewriting technique which consists of rewriting the initial query into a union of conjunctive queries. First, we study a generic breadth-first rewriting algorithm, which takes as input any rewriting operator, and define properties of rewriting operators that ensure the correctness of the algorithm. Then, we focus on piece-unifiers, which provide a rewriting operator with the desired properties. Finally, we propose an implementation of this framework and report some experiments.
1311.3211
Stochastic inference with deterministic spiking neurons
q-bio.NC cond-mat.dis-nn cs.NE physics.bio-ph stat.ML
The seemingly stochastic transient dynamics of neocortical circuits observed in vivo have been hypothesized to represent a signature of ongoing stochastic inference. In vitro neurons, on the other hand, exhibit a highly deterministic response to various types of stimulation. We show that an ensemble of deterministic leaky integrate-and-fire neurons embedded in a spiking noisy environment can attain the correct firing statistics in order to sample from a well-defined target distribution. We provide an analytical derivation of the activation function on the single cell level; for recurrent networks, we examine convergence towards stationarity in computer simulations and demonstrate sample-based Bayesian inference in a mixed graphical model. This establishes a rigorous link between deterministic neuron models and functional stochastic dynamics on the network level.
1311.3220
Chaotic Arithmetic Coding for Secure Video Multicast
cs.MM cs.CR cs.IT math.IT
Arithmetic Coding (AC) is widely used for the entropy coding of text and video data. It involves recursive partitioning of the range [0,1) in accordance with the relative probabilities of occurrence of the input symbols. A data (image or video) encryption scheme based on arithmetic coding called as Chaotic Arithmetic Coding (CAC) has been presented in previous works. In CAC, a large number of chaotic maps can be used to perform coding, each achieving Shannon optimal compression performance. The exact choice of map is governed by a key. CAC has the effect of scrambling the intervals without making any changes to the width of interval in which the codeword must lie, thereby allowing encryption without sacrificing any coding efficiency. In this paper, we use a redundancy in CAC procedure for secure multicast of videos where multiple users are distributed with different keys to decode same encrypted file. By encrypting once, we can generate multiple keys, either of which can be used to decrypt the encoded file. This is very suitable for video distribution over Internet where a single video can be distributed to multiple clients in a privacy preserving manner.
1311.3269
On a non-local spectrogram for denoising one-dimensional signals
cs.CV
In previous works, we investigated the use of local filters based on partial differential equations (PDE) to denoise one-dimensional signals through the image processing of time-frequency representations, such as the spectrogram. In this image denoising algorithms, the particularity of the image was hardly taken into account. We turn, in this paper, to study the performance of non-local filters, like Neighborhood or Yaroslavsky filters, in the same problem. We show that, for certain iterative schemes involving the Neighborhood filter, the computational time is drastically reduced with respect to Yaroslavsky or nonlinear PDE based filters, while the outputs of the filtering processes are similar. This is heuristically justified by the connection between the (fast) Neighborhood filter applied to a spectrogram and the corresponding Nonlocal Means filter (accurate) applied to the Wigner-Ville distribution of the signal. This correspondence holds only for time-frequency representations of one-dimensional signals, not to usual images, and in this sense the particularity of the image is exploited. We compare though a series of experiments on synthetic and biomedical signals the performance of local and non-local filters.
1311.3284
A family of optimal locally recoverable codes
cs.IT math.IT
A code over a finite alphabet is called locally recoverable (LRC) if every symbol in the encoding is a function of a small number (at most $r$) other symbols. We present a family of LRC codes that attain the maximum possible value of the distance for a given locality parameter and code cardinality. The codewords are obtained as evaluations of specially constructed polynomials over a finite field, and reduce to a Reed-Solomon code if the locality parameter $r$ is set to be equal to the code dimension. The size of the code alphabet for most parameters is only slightly greater than the code length. The recovery procedure is performed by polynomial interpolation over $r$ points. We also construct codes with several disjoint recovering sets for every symbol. This construction enables the system to conduct several independent and simultaneous recovery processes of a specific symbol by accessing different parts of the codeword. This property enables high availability of frequently accessed data ("hot data").
1311.3287
Nonparametric Estimation of Multi-View Latent Variable Models
cs.LG stat.ML
Spectral methods have greatly advanced the estimation of latent variable models, generating a sequence of novel and efficient algorithms with strong theoretical guarantees. However, current spectral algorithms are largely restricted to mixtures of discrete or Gaussian distributions. In this paper, we propose a kernel method for learning multi-view latent variable models, allowing each mixture component to be nonparametric. The key idea of the method is to embed the joint distribution of a multi-view latent variable into a reproducing kernel Hilbert space, and then the latent parameters are recovered using a robust tensor power method. We establish that the sample complexity for the proposed method is quadratic in the number of latent components and is a low order polynomial in the other relevant parameters. Thus, our non-parametric tensor approach to learning latent variable models enjoys good sample and computational efficiencies. Moreover, the non-parametric tensor power method compares favorably to EM algorithm and other existing spectral algorithms in our experiments.
1311.3312
Synthetic Data Generation using Benerator Tool
cs.DB
Datasets of different characteristics are needed by the research community for experimental purposes. However, real data may be difficult to obtain due to privacy concerns. Moreover, real data may not meet specific characteristics which are needed to verify new approaches under certain conditions. Given these limitations, the use of synthetic data is a viable alternative to complement the real data. In this report, we describe the process followed to generate synthetic data using Benerator, a publicly available tool. The results show that the synthetic data preserves a high level of accuracy compared to the original data. The generated datasets correspond to microdata containing records with social, economic and demographic data which mimics the distribution of aggregated statistics from the 2011 Irish Census data.
1311.3315
Sparse Matrix Factorization
cs.LG stat.ML
We investigate the problem of factorizing a matrix into several sparse matrices and propose an algorithm for this under randomness and sparsity assumptions. This problem can be viewed as a simplification of the deep learning problem where finding a factorization corresponds to finding edges in different layers and values of hidden units. We prove that under certain assumptions for a sparse linear deep network with $n$ nodes in each layer, our algorithm is able to recover the structure of the network and values of top layer hidden units for depths up to $\tilde O(n^{1/6})$. We further discuss the relation among sparse matrix factorization, deep learning, sparse recovery and dictionary learning.
1311.3318
A Study of Actor and Action Semantic Retention in Video Supervoxel Segmentation
cs.CV
Existing methods in the semantic computer vision community seem unable to deal with the explosion and richness of modern, open-source and social video content. Although sophisticated methods such as object detection or bag-of-words models have been well studied, they typically operate on low level features and ultimately suffer from either scalability issues or a lack of semantic meaning. On the other hand, video supervoxel segmentation has recently been established and applied to large scale data processing, which potentially serves as an intermediate representation to high level video semantic extraction. The supervoxels are rich decompositions of the video content: they capture object shape and motion well. However, it is not yet known if the supervoxel segmentation retains the semantics of the underlying video content. In this paper, we conduct a systematic study of how well the actor and action semantics are retained in video supervoxel segmentation. Our study has human observers watching supervoxel segmentation videos and trying to discriminate both actor (human or animal) and action (one of eight everyday actions). We gather and analyze a large set of 640 human perceptions over 96 videos in 3 different supervoxel scales. Furthermore, we conduct machine recognition experiments on a feature defined on supervoxel segmentation, called supervoxel shape context, which is inspired by the higher order processes in human perception. Our ultimate findings suggest that a significant amount of semantics have been well retained in the video supervoxel segmentation and can be used for further video analysis.
1311.3353
SUNNY: a Lazy Portfolio Approach for Constraint Solving
cs.AI
*** To appear in Theory and Practice of Logic Programming (TPLP) *** Within the context of constraint solving, a portfolio approach allows one to exploit the synergy between different solvers in order to create a globally better solver. In this paper we present SUNNY: a simple and flexible algorithm that takes advantage of a portfolio of constraint solvers in order to compute --- without learning an explicit model --- a schedule of them for solving a given Constraint Satisfaction Problem (CSP). Motivated by the performance reached by SUNNY vs. different simulations of other state of the art approaches, we developed sunny-csp, an effective portfolio solver that exploits the underlying SUNNY algorithm in order to solve a given CSP. Empirical tests conducted on exhaustive benchmarks of MiniZinc models show that the actual performance of SUNNY conforms to the predictions. This is encouraging both for improving the power of CSP portfolio solvers and for trying to export them to fields such as Answer Set Programming and Constraint Logic Programming.