id
stringlengths
9
16
title
stringlengths
4
278
categories
stringlengths
5
104
abstract
stringlengths
6
4.09k
1206.3881
DANCo: Dimensionality from Angle and Norm Concentration
cs.LG stat.ML
In the last decades the estimation of the intrinsic dimensionality of a dataset has gained considerable importance. Despite the great deal of research work devoted to this task, most of the proposed solutions prove to be unreliable when the intrinsic dimensionality of the input dataset is high and the manifold where the points lie is nonlinearly embedded in a higher dimensional space. In this paper we propose a novel robust intrinsic dimensionality estimator that exploits the twofold complementary information conveyed both by the normalized nearest neighbor distances and by the angles computed on couples of neighboring points, providing also closed-forms for the Kullback-Leibler divergences of the respective distributions. Experiments performed on both synthetic and real datasets highlight the robustness and the effectiveness of the proposed algorithm when compared to state of the art methodologies.
1206.3897
Sampled-data design for robust control of a single qubit
quant-ph cs.SY
This paper presents a sampled-data approach for the robust control of a single qubit (quantum bit). The required robustness is defined using a sliding mode domain and the control law is designed offline and then utilized online with a single qubit having bounded uncertainties. Two classes of uncertainties are considered involving the system Hamiltonian and the coupling strength of the system-environment interaction. Four cases are analyzed in detail including without decoherence, with amplitude damping decoherence, phase damping decoherence and depolarizing decoherence. Sampling periods are specifically designed for these cases to guarantee the required robustness. Two sufficient conditions are presented for guiding the design of unitary control for the cases without decoherence and with amplitude damping decoherence. The proposed approach has potential applications in quantum error-correction and in constructing robust quantum gates.
1206.3902
On the Complexity of Existential Positive Queries
cs.LO cs.AI cs.CC
We systematically investigate the complexity of model checking the existential positive fragment of first-order logic. In particular, for a set of existential positive sentences, we consider model checking where the sentence is restricted to fall into the set; a natural question is then to classify which sentence sets are tractable and which are intractable. With respect to fixed-parameter tractability, we give a general theorem that reduces this classification question to the corresponding question for primitive positive logic, for a variety of representations of structures. This general theorem allows us to deduce that an existential positive sentence set having bounded arity is fixed-parameter tractable if and only if each sentence is equivalent to one in bounded-variable logic. We then use the lens of classical complexity to study these fixed-parameter tractable sentence sets. We show that such a set can be NP-complete, and consider the length needed by a translation from sentences in such a set to bounded-variable logic; we prove superpolynomial lower bounds on this length using the theory of compilability, obtaining an interesting type of formula size lower bound. Overall, the tools, concepts, and results of this article set the stage for the future consideration of the complexity of model checking on more expressive logics.
1206.3924
Recommendation systems in the scope of opinion formation: a model
physics.soc-ph cs.SI physics.data-an
Aggregated data in real world recommender applications often feature fat-tailed distributions of the number of times individual items have been rated or favored. We propose a model to simulate such data. The model is mainly based on social interactions and opinion formation taking place on a complex network with a given topology. A threshold mechanism is used to govern the decision making process that determines whether a user is or is not interested in an item. We demonstrate the validity of the model by fitting attendance distributions from different real data sets. The model is mathematically analyzed by investigating its master equation. Our approach provides an attempt to understand recommender system's data as a social process. The model can serve as a starting point to generate artificial data sets useful for testing and evaluating recommender systems.
1206.3933
Prediction of Emerging Technologies Based on Analysis of the U.S. Patent Citation Network
cs.SI physics.soc-ph
The network of patents connected by citations is an evolving graph, which provides a representation of the innovation process. A patent citing another implies that the cited patent reflects a piece of previously existing knowledge that the citing patent builds upon. A methodology presented here (i) identifies actual clusters of patents: i.e. technological branches, and (ii) gives predictions about the temporal changes of the structure of the clusters. A predictor, called the {citation vector}, is defined for characterizing technological development to show how a patent cited by other patents belongs to various industrial fields. The clustering technique adopted is able to detect the new emerging recombinations, and predicts emerging new technology clusters. The predictive ability of our new method is illustrated on the example of USPTO subcategory 11, Agriculture, Food, Textiles. A cluster of patents is determined based on citation data up to 1991, which shows significant overlap of the class 442 formed at the beginning of 1997. These new tools of predictive analytics could support policy decision making processes in science and technology, and help formulate recommendations for action.
1206.3953
Probabilistic Reconstruction in Compressed Sensing: Algorithms, Phase Diagrams, and Threshold Achieving Matrices
cond-mat.stat-mech cs.IT math.IT
Compressed sensing is a signal processing method that acquires data directly in a compressed form. This allows one to make less measurements than what was considered necessary to record a signal, enabling faster or more precise measurement protocols in a wide range of applications. Using an interdisciplinary approach, we have recently proposed in [arXiv:1109.4424] a strategy that allows compressed sensing to be performed at acquisition rates approaching to the theoretical optimal limits. In this paper, we give a more thorough presentation of our approach, and introduce many new results. We present the probabilistic approach to reconstruction and discuss its optimality and robustness. We detail the derivation of the message passing algorithm for reconstruction and expectation max- imization learning of signal-model parameters. We further develop the asymptotic analysis of the corresponding phase diagrams with and without measurement noise, for different distribution of signals, and discuss the best possible reconstruction performances regardless of the algorithm. We also present new efficient seeding matrices, test them on synthetic data and analyze their performance asymptotically.
1206.3959
Proceedings of the Twenty-Fifth Conference on Uncertainty in Artificial Intelligence (2009)
cs.AI
This is the Proceedings of the Twenty-Fifth Conference on Uncertainty in Artificial Intelligence, which was held in Montreal, QC, Canada, June 18 - 21 2009.
1206.3963
Small-world topology of functional connectivity in randomly connected dynamical systems
cs.SI physics.data-an physics.soc-ph q-bio.NC stat.AP
Characterization of real-world complex systems increasingly involves the study of their topological structure using graph theory. Among global network properties, small-world property, consisting in existence of relatively short paths together with high clustering of the network, is one of the most discussed and studied. When dealing with coupled dynamical systems, links among units of the system are commonly quantified by a measure of pairwise statistical dependence of observed time series (functional connectivity). We argue that the functional connectivity approach leads to upwardly biased estimates of small-world characteristics (with respect to commonly used random graph models) due to partial transitivity of the accepted functional connectivity measures such as the correlation coefficient. In particular, this may lead to observation of small-world characteristics in connectivity graphs estimated from generic randomly connected dynamical systems. The ubiquity and robustness of the phenomenon is documented by an extensive parameter study of its manifestation in a multivariate linear autoregressive process, with discussion of the potential relevance for nonlinear processes and measures.
1206.3965
On the Bivariate Nakagami-$m$ Cumulative Distribution Function: Closed-form Expression and Applications
cs.IT math.IT
In this paper, we derive exact closed-form expressions for the bivariate Nakagami-$m$ cumulative distribution function (CDF) with positive integer fading severity index $m$ in terms of a class of hypergeometric functions. Particularly, we show that the bivariate Nakagami-$m$ CDF can be expressed as a finite sum of elementary functions and bivariate confluent hypergeometric $\Phi_3$ functions. Direct applications which arise from the proposed closed-form expression are the outage probability (OP) analysis of a dual-branch selection combiner in correlated Nakagami-$m$ fading, or the calculation of the level crossing rate (LCR) and average fade duration (AFD) of a sampled Nakagami-$m$ fading envelope.
1206.3975
The Ultrasound Visualization Pipeline - A Survey
cs.GR cs.CV
Ultrasound is one of the most frequently used imaging modality in medicine. The high spatial resolution, its interactive nature and non-invasiveness makes it the first choice in many examinations. Image interpretation is one of ultrasound's main challenges. Much training is required to obtain a confident skill level in ultrasound-based diagnostics. State-of-the-art graphics techniques is needed to provide meaningful visualizations of ultrasound in real-time. In this paper we present the process-pipeline for ultrasound visualization, including an overview of the tasks performed in the specific steps. To provide an insight into the trends of ultrasound visualization research, we have selected a set of significant publications and divided them into a technique-based taxonomy covering the topics pre-processing, segmentation, registration, rendering and augmented reality. For the different technique types we discuss the difference between ultrasound-based techniques and techniques for other modalities.
1206.3988
Cooperative localization using angle of arrival measurements: sequential algorithms and non-line-of-sight suppression
cs.NI cs.MA
We investigate localization of a source based on angle of arrival (AoA) measurements made at a geographically dispersed network of cooperating receivers. The goal is to efficiently compute accurate estimates despite outliers in the AoA measurements due to multipath reflections in non-line-of-sight (NLOS) environments. Maximal likelihood (ML) location estimation in such a setting requires exhaustive testing of estimates from all possible subsets of "good" measurements, which has exponential complexity in the number of measurements. We provide a randomized algorithm that approaches ML performance with linear complexity in the number of measurements. The building block for this algorithm is a low-complexity sequential algorithm for updating the source location estimates under line-of-sight (LOS) environments. Our Bayesian framework can exploit the ability to resolve multiple paths in wideband systems to provide significant performance gains over narrowband systems in NLOS environments, and easily extends to accommodate additional information such as range measurements and prior information about location.
1206.3992
Evaluating Overlapping Communities with the Conductance of their Boundary Nodes
cs.SI physics.soc-ph
Usually the boundary of a community in a network is drawn between nodes and thus crosses its outgoing links. If we construct overlapping communities by applying the link-clustering approach nodes and links interchange their roles. Therefore, boundaries must drawn through the nodes shared by two or more communities. For the purpose of community evaluation we define a conductance of boundary nodes of overlapping communities analogously to the graph conductance of boundary-crossing links used to partition a graph into disjoint communities. We show that conductance of boundary nodes (or normalised node cut) can be deduced from ordinary graph conductance of disjoint clusters in the network's weighted line graph introduced by Evans and Lambiotte (2009) to get overlapping communities of nodes in the original network. We test whether our definition can be used to construct meaningful overlapping communities with a local greedy algorithm of link clustering. In this note we present encouraging results we obtained for Zachary's karate-club network.
1206.4020
The Theory of Bonds: A New Method for the Analysis of Linkages
math.AG cs.RO math.MG math.RA
In this paper we introduce a new technique, based on dual quaternions, for the analysis of closed linkages with revolute joints: the theory of bonds. The bond structure comprises a lot of information on closed revolute chains with a one-parametric mobility. We demonstrate the usefulness of bond theory by giving a new and transparent proof for the well-known classification of overconstrained 5R linkages.
1206.4042
The Stability of Convergence of Curve Evolutions in Vector Fields
cs.CV math.AP
Curve evolution is often used to solve computer vision problems. If the curve evolution fails to converge, we would not be able to solve the targeted problem in a lifetime. This paper studies the theoretical aspect of the convergence of a type of general curve evolutions. We establish a theory for analyzing and improving the stability of the convergence of the general curve evolutions. Based on this theory, we ascertain that the convergence of a known curve evolution is marginal stable. We propose a way of modifying the original curve evolution equation to improve the stability of the convergence according to our theory. Numerical experiments show that the modification improves the convergence of the curve evolution, which validates our theory.
1206.4074
A Linear Approximation to the chi^2 Kernel with Geometric Convergence
cs.LG cs.CV stat.ML
We propose a new analytical approximation to the $\chi^2$ kernel that converges geometrically. The analytical approximation is derived with elementary methods and adapts to the input distribution for optimal convergence rate. Experiments show the new approximation leads to improved performance in image classification and semantic segmentation tasks using a random Fourier feature approximation of the $\exp-\chi^2$ kernel. Besides, out-of-core principal component analysis (PCA) methods are introduced to reduce the dimensionality of the approximation and achieve better performance at the expense of only an additional constant factor to the time complexity. Moreover, when PCA is performed jointly on the training and unlabeled testing data, further performance improvements can be obtained. Experiments conducted on the PASCAL VOC 2010 segmentation and the ImageNet ILSVRC 2010 datasets show statistically significant improvements over alternative approximation methods.
1206.4094
Maximal-entropy random walk unifies centrality measures
physics.soc-ph cond-mat.dis-nn cond-mat.stat-mech cs.SI
In this paper analogies between different (dis)similarity matrices are derived. These matrices, which are connected to path enumeration and random walks, are used in community detection methods or in computation of centrality measures for complex networks. The focus is on a number of known centrality measures, which inherit the connections established for similarity matrices. These measures are based on the principal eigenvector of the adjacency matrix, path enumeration, as well as on the stationary state, stochastic matrix or mean first-passage times of a random walk. Particular attention is paid to the maximal-entropy random walk, which serves as a very distinct alternative to the ordinary random walk used in network analysis. The various importance measures, defined both with the use of ordinary random walk and the maximal-entropy random walk, are compared numerically on a set of benchmark graphs. It is shown that groups of centrality measures defined with the two random walks cluster into two separate families. In particular, the group of centralities for the maximal-entropy random walk, connected to the eigenvector centrality and path enumeration, is strongly distinct from all the other measures and produces largely equivalent results.
1206.4110
ConeRANK: Ranking as Learning Generalized Inequalities
cs.LG cs.IR
We propose a new data mining approach in ranking documents based on the concept of cone-based generalized inequalities between vectors. A partial ordering between two vectors is made with respect to a proper cone and thus learning the preferences is formulated as learning proper cones. A pairwise learning-to-rank algorithm (ConeRank) is proposed to learn a non-negative subspace, formulated as a polyhedral cone, over document-pair differences. The algorithm is regularized by controlling the `volume' of the cone. The experimental studies on the latest and largest ranking dataset LETOR 4.0 shows that ConeRank is competitive against other recent ranking approaches.
1206.4116
Dependence Maximizing Temporal Alignment via Squared-Loss Mutual Information
stat.ML cs.AI
The goal of temporal alignment is to establish time correspondence between two sequences, which has many applications in a variety of areas such as speech processing, bioinformatics, computer vision, and computer graphics. In this paper, we propose a novel temporal alignment method called least-squares dynamic time warping (LSDTW). LSDTW finds an alignment that maximizes statistical dependency between sequences, measured by a squared-loss variant of mutual information. The benefit of this novel information-theoretic formulation is that LSDTW can align sequences with different lengths, different dimensionality, high non-linearity, and non-Gaussianity in a computationally efficient manner. In addition, model parameters such as an initial alignment matrix can be systematically optimized by cross-validation. We demonstrate the usefulness of LSDTW through experiments on synthetic and real-world Kinect action recognition datasets.
1206.4121
The information-theoretic costs of simulating quantum measurements
quant-ph cs.IT math.IT
Winter's measurement compression theorem stands as one of the most penetrating insights of quantum information theory (QIT). In addition to making an original and profound statement about measurement in quantum theory, it also underlies several other general protocols in QIT. In this paper, we provide a full review of Winter's measurement compression theorem, detailing the information processing task, giving examples for understanding it, reviewing Winter's achievability proof, and detailing a new approach to its single-letter converse theorem. We prove an extension of the theorem to the case in which the sender is not required to receive the outcomes of the simulated measurement. The total cost of common randomness and classical communication can be lower for such a "non-feedback" simulation, and we prove a single-letter converse theorem demonstrating optimality. We then review the Devetak-Winter theorem on classical data compression with quantum side information, providing new proofs of its achievability and converse parts. From there, we outline a new protocol that we call "measurement compression with quantum side information," announced previously by two of us in our work on triple trade-offs in quantum Shannon theory. This protocol has several applications, including its part in the "classically-assisted state redistribution" protocol, which is the most general protocol on the static side of the quantum information theory tree, and its role in reducing the classical communication cost in a task known as local purity distillation. We also outline a connection between measurement compression with quantum side information and recent work on entropic uncertainty relations in the presence of quantum memory. Finally, we prove a single-letter theorem characterizing measurement compression with quantum side information when the sender is not required to obtain the measurement outcome.
1206.4123
On the Confidentiality of Information Dispersal Algorithms and Their Erasure Codes
cs.IT cs.CR cs.DC math.IT
\emph{Information Dispersal Algorithms (IDAs)} have been widely applied to reliable and secure storage and transmission of data files in distributed systems. An IDA is a method that encodes a file $F$ of size $L=|F|$ into $n$ unrecognizable pieces $F_1$, $F_2$, ..., $F_n$, each of size $L/m$ ($m<n$), so that the original file $F$ can be reconstructed from any $m$ pieces. The core of an IDA is the adopted non-systematic $m$-of-$n$ erasure code. This paper makes a systematic study on the \emph{confidentiality} of an IDA and its connection with the adopted erasure code. Two levels of confidentiality are defined: \emph{weak confidentiality} (in the case where some parts of the original file $F$ can be reconstructed explicitly from fewer than $m$ pieces) and \emph{strong confidentiality} (in the case where nothing of the original file $F$ can be reconstructed explicitly from fewer than $m$ pieces). For an IDA that adopts an arbitrary non-systematic erasure code, its confidentiality may fall into weak confidentiality. To achieve strong confidentiality, this paper explores a sufficient and feasible condition on the adopted erasure code. Then, this paper shows that Rabin's IDA has strong confidentiality. At the same time, this paper presents an effective way to construct an IDA with strong confidentiality from an arbitrary $m$-of-$(m+n)$ erasure code. Then, as an example, this paper constructs an IDA with strong confidentiality from a Reed-Solomon code, the computation complexity of which is comparable to or sometimes even lower than that of Rabin's IDA.
1206.4169
Clustered Bandits
cs.LG
We consider a multi-armed bandit setting that is inspired by real-world applications in e-commerce. In our setting, there are a few types of users, each with a specific response to the different arms. When a user enters the system, his type is unknown to the decision maker. The decision maker can either treat each user separately ignoring the previously observed users, or can attempt to take advantage of knowing that only few types exist and cluster the users according to their response to the arms. We devise algorithms that combine the usual exploration-exploitation tradeoff with clustering of users and demonstrate the value of clustering. In the process of developing algorithms for the clustered setting, we propose and analyze simple algorithms for the setup where a decision maker knows that a user belongs to one of few types, but does not know which one.
1206.4176
Energy and Spectral Efficiencies Trade-off with Filter Optimization in Multiple Access Interference-Aware
math.OC cs.IT math.IT
This work analyzes the optimized deployment of two resources scarcely available in mobile multiple access systems, i.e., spectrum and energy, as well as the impact of filter optimization in the system performance. Taking in perspective the two conflicting metrics, throughput maximization and power consumption minimization, the distributed energy efficiency (EE) cost function is formulated. Furthermore, the best energy-spectral efficiencies (EE-SE) trade-off is achieved when each node allocates exactly the power necessary to attain the best SINR response, which guarantees the maximal EE. To demonstrate the validity of our analysis, two low-complexity energy-spectral efficient algorithms, based on distributed instantaneous SINR level are developed, and the impact of single and multiuser detection filters on the EE-SE trade-off is analyzed.
1206.4185
Ant Robotics: Covering Continuous Domains by Multi-A(ge)nt Systems
cs.RO cs.AI cs.MA
In this work we present an algorithm for covering continuous connected domains by ant-like robots with very limited capabilities. The robots can mark visited places with pheromone marks and sense the level of the pheromone in their local neighborhood. In case of multiple robots these pheromone marks can be sensed by all robots and provide the only way of (indirect) communication between the robots. The robots are assumed to be memoryless, and to have no global information such as the domain map, their own position (either absolute or relative), total marked area percentage, maximal pheromone level, etc.. Despite the robots' simplicity, we show that they are able, by running a very simple rule of behavior, to ensure efficient covering of arbitrary connected domains, including non-planar and multidimensional ones. The novelty of our algorithm lies in the fact that, unlike previously proposed methods, our algorithm works on continuous domains without relying on some "induced" underlying graph, that effectively reduces the problem to a discrete case of graph covering. The algorithm guarantees complete coverage of any connected domain. We also prove that the algorithm is noise immune, i.e., it is able to cope with any initial pheromone profile (noise). In addition the algorithm provides a bounded constant time between two successive visits of the robot, and thus, is suitable for patrolling or surveillance applications.
1206.4192
Designing Incoherent Dictionaries for Compressed Sensing: Algorithm Comparison
cs.IT cs.DS math.IT
A new method presented for design of incoherent dictionaries.
1206.4221
Distributed Maximum Likelihood for Simultaneous Self-localization and Tracking in Sensor Networks
math.OC cs.DC cs.SY stat.AP
We show that the sensor self-localization problem can be cast as a static parameter estimation problem for Hidden Markov Models and we implement fully decentralized versions of the Recursive Maximum Likelihood and on-line Expectation-Maximization algorithms to localize the sensor network simultaneously with target tracking. For linear Gaussian models, our algorithms can be implemented exactly using a distributed version of the Kalman filter and a novel message passing algorithm. The latter allows each node to compute the local derivatives of the likelihood or the sufficient statistics needed for Expectation-Maximization. In the non-linear case, a solution based on local linearization in the spirit of the Extended Kalman Filter is proposed. In numerical examples we demonstrate that the developed algorithms are able to learn the localization parameters.
1206.4226
Three-User Cognitive Interference Channel: Capacity Region with Strong Interference
cs.IT math.IT
This study investigates the capacity region of a three-user cognitive radio network with two primary users and one cognitive user. A three-user Cognitive Interference Channel (C-IFC) is proposed by considering a three-user Interference Channel (IFC) where one of the transmitters has cognitive capabilities and knows the messages of the other two transmitters in a non-causal manner. First, two inner bounds on the capacity region of the three-user C-IFC are obtained based on using the schemes which allow all receivers to decode all messages with two different orders. Next, two sets of conditions are derived, under which the capacity region of the proposed model coincides with the capacity region of a three-user C-IFC in which all three messages are required at all receivers. Under these conditions, referred to as strong interference conditions, the capacity regions for the proposed three-user C-IFC are characterized. Moreover, the Gaussian three-user C-IFC is considered and the capacity results are derived for the Gaussian case. Some numerical examples are also provided.
1206.4229
Information field dynamics for simulation scheme construction
physics.comp-ph astro-ph.IM cs.IT math.IT
Information field dynamics (IFD) is introduced here as a framework to derive numerical schemes for the simulation of physical and other fields without assuming a particular sub-grid structure as many schemes do. IFD constructs an ensemble of non-parametric sub-grid field configurations from the combination of the data in computer memory, representing constraints on possible field configurations, and prior assumptions on the sub-grid field statistics. Each of these field configurations can formally be evolved to a later moment since any differential operator of the dynamics can act on fields living in continuous space. However, these virtually evolved fields need again a representation by data in computer memory. The maximum entropy principle of information theory guides the construction of updated datasets via entropic matching, optimally representing these field configurations at the later time. The field dynamics thereby become represented by a finite set of evolution equations for the data that can be solved numerically. The sub-grid dynamics is treated within an auxiliary analytic consideration and the resulting scheme acts solely on the data space. It should provide a more accurate description of the physical field dynamics than simulation schemes constructed ad-hoc, due to the more rigorous accounting of sub-grid physics and the space discretization process. Assimilation of measurement data into an IFD simulation is conceptually straightforward since measurement and simulation data can just be merged. The IFD approach is illustrated using the example of a coarsely discretized representation of a thermally excited classical Klein-Gordon field. This should pave the way towards the construction of schemes for more complex systems like turbulent hydrodynamics.
1206.4232
Enhanced active power filter control for nonlinear non-stationary reactive power compensation
math.OC cs.SY
This paper describes a method to implement Reactive Power Compensation (RPC) in power systems that possess nonlinear non-stationary current disturbances. The Empirical Mode Decomposition (EMD) introduced in the Hilbert-Huang Transform (HHT) is used to separate the disturbances from the original current waveform. These disturbances are subsequently removed. Following that, Power Factor Correction (PFC) based on the well-known p-q power theory is conducted to remove the reactive power. Both operations were implemented in a shunt Active Power Filter (APF). The EMD significantly simplifies the singulation and the removal of the current disturbances. This helps to effectively identify the fundamental current waveform. Hence, it simplifies the implementation of RPC on nonlinear non-stationary power systems.
1206.4245
On Lossless Universal Compression of Distributed Identical Sources
cs.IT math.IT
Slepian-Wolf theorem is a well-known framework that targets almost lossless compression of (two) data streams with symbol-by-symbol correlation between the outputs of (two) distributed sources. However, this paper considers a different scenario which does not fit in the Slepian-Wolf framework. We consider two identical but spatially separated sources. We wish to study the universal compression of a sequence of length $n$ from one of the sources provided that the decoder has access to (i.e., memorized) a sequence of length $m$ from the other source. Such a scenario occurs, for example, in the universal compression of data from multiple mirrors of the same server. In this setup, the correlation does not arise from symbol-by-symbol dependency of two outputs from the two sources. Instead, the sequences are correlated through the information that they contain about the unknown source parameter. We show that the finite-length nature of the compression problem at hand requires considering a notion of almost lossless source coding, where coding incurs an error probability $p_e(n)$ that vanishes with sequence length $n$. We obtain a lower bound on the average minimax redundancy of almost lossless codes as a function of the sequence length $n$ and the permissible error probability $p_e$ when the decoder has a memory of length $m$ and the encoders do not communicate. Our results demonstrate that a strict performance loss is incurred when the two encoders do not communicate even when the decoder knows the unknown parameter vector (i.e., $m \to \infty$).
1206.4275
Joint Transmit Precoding for the Relay Interference Broadcast Channel
cs.IT math.IT
Relays in cellular systems are interference limited. The highest end-to-end sum rates are achieved when the relays are jointly optimized with the transmit strategy. Unfortunately, interference couples the links together making joint optimization challenging. Further, the end-to-end multi-hop performance is sensitive to rate mismatch, when some links have a dominant first link while others have a dominant second link. This paper proposes an algorithm for designing the linear transmit precoders at the transmitters and relays of the relay interference broadcast channel, a generic model for relay-based cellular systems, to maximize the end-to-end sum-rates. First, the relays are designed to maximize the second-hop sum-rates. Next, approximate end-to-end rates that depend on the time-sharing fraction and the second-hop rates are used to formulate a sum-utility maximization problem for designing the transmitters. This problem is solved by iteratively minimizing the weighted sum of mean square errors. Finally, the norms of the transmit precoders at the transmitters are adjusted to eliminate rate mismatch. The proposed algorithm allows for distributed implementation and has fast convergence. Numerical results show that the proposed algorithm outperforms a reasonable application of single-hop interference management strategies separately on two hops.
1206.4280
Return Migration After Brain Drain: A Simulation Approach
physics.soc-ph cs.SI
The Brain Drain phenomenon is particularly heterogeneous and is characterized by peculiar specifications. It influences the economic fundamentals of both the country of origin and the host one in terms of human capital accumulation. Here, the brain drain is considered from a microeconomic perspective: more precisely we focus on the individual rational decision to return, referring it to the social capital owned by the worker. The presented model compares utility levels to justify agent migration conduct and to simulate several scenarios within a computational environment. In particular, we developed a simulation framework based on two fundamental individual features, i.e. risk aversion and initial expectation, which characterize the dynamics of different agents according to the evolution of their social contacts. Our main result is that, according to the value of risk aversion and initial expectation, the probability of return migration depends on their ratio, with a certain degree of approximation: when risk aversion is much bigger than the initial expectation, the probability of returns is maximal, while, in the opposite case, the probability for the agents to remain abroad is very high. In between, when the two values are comparable, it does exist a broad intertwined region where it is very difficult to draw any analytical forecast.
1206.4300
Quasi-Succinct Indices
cs.IR cs.DS
Compressed inverted indices in use today are based on the idea of gap compression: documents pointers are stored in increasing order, and the gaps between successive document pointers are stored using suitable codes which represent smaller gaps using less bits. Additional data such as counts and positions is stored using similar techniques. A large body of research has been built in the last 30 years around gap compression, including theoretical modeling of the gap distribution, specialized instantaneous codes suitable for gap encoding, and ad hoc document reorderings which increase the efficiency of instantaneous codes. This paper proposes to represent an index using a different architecture based on quasi-succinct representation of monotone sequences. We show that, besides being theoretically elegant and simple, the new index provides expected constant-time operations and, in practice, significant performance improvements on conjunctive, phrasal and proximity queries.
1206.4326
Joint Reconstruction of Multi-view Compressed Images
cs.MM cs.CV
The distributed representation of correlated multi-view images is an important problem that arise in vision sensor networks. This paper concentrates on the joint reconstruction problem where the distributively compressed correlated images are jointly decoded in order to improve the reconstruction quality of all the compressed images. We consider a scenario where the images captured at different viewpoints are encoded independently using common coding solutions (e.g., JPEG, H.264 intra) with a balanced rate distribution among different cameras. A central decoder first estimates the underlying correlation model from the independently compressed images which will be used for the joint signal recovery. The joint reconstruction is then cast as a constrained convex optimization problem that reconstructs total-variation (TV) smooth images that comply with the estimated correlation model. At the same time, we add constraints that force the reconstructed images to be consistent with their compressed versions. We show by experiments that the proposed joint reconstruction scheme outperforms independent reconstruction in terms of image quality, for a given target bit rate. In addition, the decoding performance of our proposed algorithm compares advantageously to state-of-the-art distributed coding schemes based on disparity learning and on the DISCOVER.
1206.4327
Social Influence in Social Advertising: Evidence from Field Experiments
cs.SI physics.soc-ph stat.AP
Social advertising uses information about consumers' peers, including peer affiliations with a brand, product, organization, etc., to target ads and contextualize their display. This approach can increase ad efficacy for two main reasons: peers' affiliations reflect unobserved consumer characteristics, which are correlated along the social network; and the inclusion of social cues (i.e., peers' association with a brand) alongside ads affect responses via social influence processes. For these reasons, responses may be increased when multiple social signals are presented with ads, and when ads are affiliated with peers who are strong, rather than weak, ties. We conduct two very large field experiments that identify the effect of social cues on consumer responses to ads, measured in terms of ad clicks and the formation of connections with the advertised entity. In the first experiment, we randomize the number of social cues present in word-of-mouth advertising, and measure how responses increase as a function of the number of cues. The second experiment examines the effect of augmenting traditional ad units with a minimal social cue (i.e., displaying a peer's affiliation below an ad in light grey text). On average, this cue causes significant increases in ad performance. Using a measurement of tie strength based on the total amount of communication between subjects and their peers, we show that these influence effects are greatest for strong ties. Our work has implications for ad optimization, user interface design, and central questions in social science research.
1206.4329
An Improved Gauss-Newtons Method based Back-propagation Algorithm for Fast Convergence
cs.AI cs.NA
The present work deals with an improved back-propagation algorithm based on Gauss-Newton numerical optimization method for fast convergence. The steepest descent method is used for the back-propagation. The algorithm is tested using various datasets and compared with the steepest descent back-propagation algorithm. In the system, optimization is carried out using multilayer neural network. The efficacy of the proposed method is observed during the training period as it converges quickly for the dataset used in test. The requirement of memory for computing the steps of algorithm is also analyzed.
1206.4358
Robust Detection of Dynamic Community Structure in Networks
physics.data-an cond-mat.dis-nn cs.SI physics.bio-ph physics.soc-ph q-bio.NC
We describe techniques for the robust detection of community structure in some classes of time-dependent networks. Specifically, we consider the use of statistical null models for facilitating the principled identification of structural modules in semi-decomposable systems. Null models play an important role both in the optimization of quality functions such as modularity and in the subsequent assessment of the statistical validity of identified community structure. We examine the sensitivity of such methods to model parameters and show how comparisons to null models can help identify system scales. By considering a large number of optimizations, we quantify the variance of network diagnostics over optimizations (`optimization variance') and over randomizations of network structure (`randomization variance'). Because the modularity quality function typically has a large number of nearly-degenerate local optima for networks constructed using real data, we develop a method to construct representative partitions that uses a null model to correct for statistical noise in sets of partitions. To illustrate our results, we employ ensembles of time-dependent networks extracted from both nonlinear oscillators and empirical neuroscience data.
1206.4370
Cyclic Codes from Dickson Polynomials
cs.IT math.IT
Due to their efficient encoding and decoding algorithms cyclic codes, a subclass of linear codes, have applications in consumer electronics, data storage systems, and communication systems. In this paper, Dickson polynomials of the first and second kind over finite fields are employed to construct a number of classes of cyclic codes. Lower bounds on the minimum weight of some classes of the cyclic codes are developed. The minimum weights of some other classes of the codes constructed in this paper are determined. The dimensions of the codes obtained in this paper are flexible. Most of the codes presented in this paper are optimal or almost optimal in the sense that they meet some bound on linear codes. Over ninety cyclic codes of this paper should be used to update the current database of tables of best linear codes known. Among them sixty are optimal in the sense that they meet some bound on linear codes and the rest are cyclic codes having the same parameters as the best linear code in the current database maintained at http://www.codetables.de/.
1206.4389
Improving Two-Way Selective Decode-and-forward Wireless Relaying with Energy-Efficient One-bit Soft Forwarding
cs.IT math.IT math.PR
Motivated by applications such as battery-operated wireless sensor networks (WSN), we propose an easy-to-implement energy-efficient two-way relaying scheme. In particular, we address the challenge of improving the standard two-way selective decode-and-forward protocol (TW-SDF) in terms of block-error-rate (BLER) with minor additional complexity and energy consumption. By following the principle of soft relaying, our solution is the two-way one-bit soft forwarding (TW-1bSF) protocol in which the relay forwards the one-bit quantization of a posterior information metric about the transmitted bits, associated with an appropriately designed reliability parameter. In WSN-related standards (such as IEEE802.15.6 and Bluetooth), block codes are adopted instead of convolutional and other sophisticated codes, due to their efficient decoder hardware implementation. As the second main contribution, we derive tight upper bounds on the BLER performance for both TW-SDF and TW-1bSF, when the two-way relaying network employs block codes and hard decoding. The error probability analysis confirms the superiority of TW-1bSF. Moreover, we derive the asymptotic performance gain of TW-1bSF over TW-SDF, which further suggests that the proposed protocol is a good choice, especially when long block codes are used.
1206.4391
Gray Image extraction using Fuzzy Logic
cs.CV cs.AI
Fuzzy systems concern fundamental methodology to represent and process uncertainty and imprecision in the linguistic information. The fuzzy systems that use fuzzy rules to represent the domain knowledge of the problem are known as Fuzzy Rule Base Systems (FRBS). On the other hand image segmentation and subsequent extraction from a noise-affected background, with the help of various soft computing methods, are relatively new and quite popular due to various reasons. These methods include various Artificial Neural Network (ANN) models (primarily supervised in nature), Genetic Algorithm (GA) based techniques, intensity histogram based methods etc. providing an extraction solution working in unsupervised mode happens to be even more interesting problem. Literature suggests that effort in this respect appears to be quite rudimentary. In the present article, we propose a fuzzy rule guided novel technique that is functional devoid of any external intervention during execution. Experimental results suggest that this approach is an efficient one in comparison to different other techniques extensively addressed in literature. In order to justify the supremacy of performance of our proposed technique in respect of its competitors, we take recourse to effective metrics like Mean Squared Error (MSE), Mean Absolute Error (MAE), Peak Signal to Noise Ratio (PSNR).
1206.4436
Tiling $R^{5}$ by Crosses
cs.IT math.CO math.IT
An $n$-dimensional cross comprises $2n+1$ unit cubes: the center cube and reflections in all its faces. It is well known that there is a tiling of $R^{n}$ by crosses for all $n.$ AlBdaiwi and the first author proved that if $2n+1$ is not a prime then there are $2^{\aleph_{0}}$ \ non-congruent regular (= face-to-face) tilings of $R^{n}$ by crosses, while there is a unique tiling of $R^{n}$ by crosses for $n=2,3$. They conjectured that this is always the case if $2n+1$ is a prime. To support the conjecture we prove in this paper that also for $R^{5}$ there is a unique regular, and no non-regular, tiling by crosses. So there is a unique tiling of $R^{3}$ by crosses, there are $2^{\aleph_{0}}$ tilings of $R^{4},$ but for $R^{5}$ there is again only one tiling by crosses. We guess that this result goes against our intuition that suggests "the higher the dimension of the \ space, the more freedom we get".
1206.4438
Inverse Modeling of Climate Responses of Monumental Buildings
cs.CE
The indoor climate conditions of monumental buildings are very important for the conservation of these objects. Simplified models with physical meaning are desired that are capable of simulating temperature and relative humidity. In this paper we research state-space models as methodology for the inverse modeling of climate responses of unheated monumental buildings. It is concluded that this approach is very promising for obtaining physical models and parameters of indoor climate responses. Furthermore state space models can be simulated very efficiently: the simulation duration time of a 100 year hourly based period take less than a second on an ordinary computer.
1206.4481
Parsimonious Mahalanobis Kernel for the Classification of High Dimensional Data
cs.NA cs.LG
The classification of high dimensional data with kernel methods is considered in this article. Exploit- ing the emptiness property of high dimensional spaces, a kernel based on the Mahalanobis distance is proposed. The computation of the Mahalanobis distance requires the inversion of a covariance matrix. In high dimensional spaces, the estimated covariance matrix is ill-conditioned and its inversion is unstable or impossible. Using a parsimonious statistical model, namely the High Dimensional Discriminant Analysis model, the specific signal and noise subspaces are estimated for each considered class making the inverse of the class specific covariance matrix explicit and stable, leading to the definition of a parsimonious Mahalanobis kernel. A SVM based framework is used for selecting the hyperparameters of the parsimonious Mahalanobis kernel by optimizing the so-called radius-margin bound. Experimental results on three high dimensional data sets show that the proposed kernel is suitable for classifying high dimensional data, providing better classification accuracies than the conventional Gaussian kernel.
1206.4498
On a Class of Discrete Memoryless Broadcast Interference Channels
cs.IT math.IT
We study a class of discrete memoryless broadcast interference channels (DM-BICs), where one of the broadcast receivers is subject to the interference from a point-to-point transmission. A general achievable rate region $\mathcal{R}$ based on rate splitting, superposition coding and binning at the broadcast transmitter and rate splitting at the interfering transmitter is derived. Under two partial order broadcast conditions {\em interference-oblivious less noisy} and {\em interference-cognizant less noisy}, a reduced form of $\mathcal{R}$ is shown to be equivalent to the region based on a simpler scheme that uses only superposition coding at the broadcast transmitter. Furthermore, the capacity regions of DM-BIC under the two partial order broadcast conditions are characterized respectively for the strong and very strong interference conditions.
1206.4504
Revisiting Timed Specification Theories: A Linear-Time Perspective
cs.SE cs.LO cs.SY
We consider the setting of component-based design for real-time systems with critical timing constraints. Based on our earlier work, we propose a compositional specification theory for timed automata with I/O distinction, which supports substitutive refinement. Our theory provides the operations of parallel composition for composing components at run-time, logical conjunction/disjunction for independent development, and quotient for incremental synthesis. The key novelty of our timed theory lies in a weakest congruence preserving safety as well as bounded liveness properties. We show that the congruence can be characterised by two linear-time semantics, timed-traces and timed-strategies, the latter of which is derived from a game-based interpretation of timed interaction.
1206.4509
Decentralized Estimation of Laplacian Eigenvalues in Multi-Agent Systems
cs.SY
In this paper we present a decentralized algorithm to estimate the eigenvalues of the Laplacian matrix that encodes the network topology of a multi-agent system. We consider network topologies modeled by undirected graphs. The basic idea is to provide a local interaction rule among agents so that their state trajectory is a linear combination of sinusoids oscillating only at frequencies function of the eigenvalues of the Laplacian matrix. In this way, the problem of decentralized estimation of the eigenvalues is mapped into a standard signal processing problem in which the unknowns are the finite number of frequencies at which the signal oscillates.
1206.4522
BADREX: In situ expansion and coreference of biomedical abbreviations using dynamic regular expressions
cs.CL
BADREX uses dynamically generated regular expressions to annotate term definition-term abbreviation pairs, and corefers unpaired acronyms and abbreviations back to their initial definition in the text. Against the Medstract corpus BADREX achieves precision and recall of 98% and 97%, and against a much larger corpus, 90% and 85%, respectively. BADREX yields improved performance over previous approaches, requires no training data and allows runtime customisation of its input parameters. BADREX is freely available from https://github.com/philgooch/BADREX-Biomedical-Abbreviation-Expander as a plugin for the General Architecture for Text Engineering (GATE) framework and is licensed under the GPLv3.
1206.4555
Optimal compression of hash-origin prefix trees
cs.IT cs.DB cs.DS math.CO math.IT
There is a common problem of operating on hash values of elements of some database. In this paper there will be analyzed informational content of such general task and how to practically approach such found lower boundaries. Minimal prefix tree which distinguish elements turns out to require asymptotically only about 2.77544 bits per element, while standard approaches use a few times more. While being certain of working inside the database, the cost of distinguishability can be reduced further to about 2.33275 bits per elements. Increasing minimal depth of nodes to reduce probability of false positives leads to simple relation with average depth of such random tree, which is asymptotically larger by about 1.33275 bits than lg(n) of the perfect binary tree. This asymptotic case can be also seen as a way to optimally encode n large unordered numbers - saving lg(n!) bits of information about their ordering, which can be the major part of contained information. This ability itself allows to reduce memory requirements even to about 0.693 of required in Bloom filter for the same false positive probability.
1206.4557
A model of competition among more than two languages
physics.soc-ph cond-mat.stat-mech cs.SI
We extend the Abrams-Strogatz model for competition between two languages [Nature 424, 900 (2003)] to the case of n(>=2) competing states (i.e., languages). Although the Abrams-Strogatz model for n=2 can be interpreted as modeling either majority preference or minority aversion, the two mechanisms are distinct when n>=3. We find that the condition for the coexistence of different states is independent of n under the pure majority preference, whereas it depends on n under the pure minority aversion. We also show that the stable coexistence equilibrium and stable monopoly equilibria can be multistable under the minority aversion and not under the majority preference. Furthermore, we obtain the phase diagram of the model when the effects of the majority preference and minority aversion are mixed, under the condition that different states have the same attractiveness. We show that the multistability is a generic property of the model facilitated by large n.
1206.4560
Residual Component Analysis: Generalising PCA for more flexible inference in linear-Gaussian models
cs.LG stat.ML
Probabilistic principal component analysis (PPCA) seeks a low dimensional representation of a data set in the presence of independent spherical Gaussian noise. The maximum likelihood solution for the model is an eigenvalue problem on the sample covariance matrix. In this paper we consider the situation where the data variance is already partially explained by other actors, for example sparse conditional dependencies between the covariates, or temporal correlations leaving some residual variance. We decompose the residual variance into its components through a generalised eigenvalue problem, which we call residual component analysis (RCA). We explore a range of new algorithms that arise from the framework, including one that factorises the covariance of a Gaussian density into a low-rank and a sparse-inverse component. We illustrate the ideas on the recovery of a protein-signaling network, a gene expression time-series data set and the recovery of the human skeleton from motion capture 3-D cloud data.
1206.4572
Autocorrelations of Binary Sequences and Run Structure
cs.IT math.CO math.IT
We analyze the connection between the autocorrelation of a binary sequence and its run structure given by the run length encoding. We show that both the periodic and the aperiodic autocorrelation of a binary sequence can be formulated in terms of the run structure. The run structure is given by the consecutive runs of the sequence. Let C=(C(0), C(1),...,C(n)) denote the autocorrelation vector of a binary sequence. We prove that the kth component of the second order difference operator of C can be directly calculated by using the consecutive runs of total length k. In particular this shows that the kth autocorrelation is already determined by all consecutive runs of total length L<k. In the aperiodic case we show how the run vector R can be efficiently calculated and give a characterization of skew-symmetric sequences in terms of their run length encoding.
1206.4588
An Evolutionary Approach to Drug-Design Using Quantam Binary Particle Swarm Optimization Algorithm
cs.NE cs.CE
The present work provides a new approach to evolve ligand structures which represent possible drug to be docked to the active site of the target protein. The structure is represented as a tree where each non-empty node represents a functional group. It is assumed that the active site configuration of the target protein is known with position of the essential residues. In this paper the interaction energy of the ligands with the protein target is minimized. Moreover, the size of the tree is difficult to obtain and it will be different for different active sites. To overcome the difficulty, a variable tree size configuration is used for designing ligands. The optimization is done using a quantum discrete PSO. The result using fixed length and variable length configuration are compared.
1206.4599
A Unified Robust Classification Model
cs.LG stat.ML
A wide variety of machine learning algorithms such as support vector machine (SVM), minimax probability machine (MPM), and Fisher discriminant analysis (FDA), exist for binary classification. The purpose of this paper is to provide a unified classification model that includes the above models through a robust optimization approach. This unified model has several benefits. One is that the extensions and improvements intended for SVM become applicable to MPM and FDA, and vice versa. Another benefit is to provide theoretical results to above learning methods at once by dealing with the unified model. We give a statistical interpretation of the unified classification model and propose a non-convex optimization algorithm that can be applied to non-convex variants of existing learning methods.
1206.4600
Bayesian Nonexhaustive Learning for Online Discovery and Modeling of Emerging Classes
cs.LG stat.ML
We present a framework for online inference in the presence of a nonexhaustively defined set of classes that incorporates supervised classification with class discovery and modeling. A Dirichlet process prior (DPP) model defined over class distributions ensures that both known and unknown class distributions originate according to a common base distribution. In an attempt to automatically discover potentially interesting class formations, the prior model is coupled with a suitably chosen data model, and sequential Monte Carlo sampling is used to perform online inference. Our research is driven by a biodetection application, where a new class of pathogen may suddenly appear, and the rapid increase in the number of samples originating from this class indicates the onset of an outbreak.
1206.4601
Convex Multitask Learning with Flexible Task Clusters
cs.LG stat.ML
Traditionally, multitask learning (MTL) assumes that all the tasks are related. This can lead to negative transfer when tasks are indeed incoherent. Recently, a number of approaches have been proposed that alleviate this problem by discovering the underlying task clusters or relationships. However, they are limited to modeling these relationships at the task level, which may be restrictive in some applications. In this paper, we propose a novel MTL formulation that captures task relationships at the feature-level. Depending on the interactions among tasks and features, the proposed method construct different task clusters for different features, without even the need of pre-specifying the number of clusters. Computationally, the proposed formulation is strongly convex, and can be efficiently solved by accelerated proximal methods. Experiments are performed on a number of synthetic and real-world data sets. Under various degrees of task relationships, the accuracy of the proposed method is consistently among the best. Moreover, the feature-specific task clusters obtained agree with the known/plausible task structures of the data.
1206.4602
Quasi-Newton Methods: A New Direction
cs.NA cs.LG stat.ML
Four decades after their invention, quasi-Newton methods are still state of the art in unconstrained numerical optimization. Although not usually interpreted thus, these are learning algorithms that fit a local quadratic approximation to the objective function. We show that many, including the most popular, quasi-Newton methods can be interpreted as approximations of Bayesian linear regression under varying prior assumptions. This new notion elucidates some shortcomings of classical algorithms, and lights the way to a novel nonparametric quasi-Newton method, which is able to make more efficient use of available information at computational cost similar to its predecessors.
1206.4603
Latent Collaborative Retrieval
cs.IR cs.AI
Retrieval tasks typically require a ranking of items given a query. Collaborative filtering tasks, on the other hand, learn to model user's preferences over items. In this paper we study the joint problem of recommending items to a user with respect to a given query, which is a surprisingly common task. This setup differs from the standard collaborative filtering one in that we are given a query x user x item tensor for training instead of the more traditional user x item matrix. Compared to document retrieval we do have a query, but we may or may not have content features (we will consider both cases) and we can also take account of the user's profile. We introduce a factorized model for this new task that optimizes the top-ranked items returned for the given query and user. We report empirical results where it outperforms several baselines.
1206.4604
Learning the Experts for Online Sequence Prediction
cs.LG cs.AI
Online sequence prediction is the problem of predicting the next element of a sequence given previous elements. This problem has been extensively studied in the context of individual sequence prediction, where no prior assumptions are made on the origin of the sequence. Individual sequence prediction algorithms work quite well for long sequences, where the algorithm has enough time to learn the temporal structure of the sequence. However, they might give poor predictions for short sequences. A possible remedy is to rely on the general model of prediction with expert advice, where the learner has access to a set of $r$ experts, each of which makes its own predictions on the sequence. It is well known that it is possible to predict almost as well as the best expert if the sequence length is order of $\log(r)$. But, without firm prior knowledge on the problem, it is not clear how to choose a small set of {\em good} experts. In this paper we describe and analyze a new algorithm that learns a good set of experts using a training set of previously observed sequences. We demonstrate the merits of our approach by applying it on the task of click prediction on the web.
1206.4606
TrueLabel + Confusions: A Spectrum of Probabilistic Models in Analyzing Multiple Ratings
cs.LG cs.AI stat.ML
This paper revisits the problem of analyzing multiple ratings given by different judges. Different from previous work that focuses on distilling the true labels from noisy crowdsourcing ratings, we emphasize gaining diagnostic insights into our in-house well-trained judges. We generalize the well-known DawidSkene model (Dawid & Skene, 1979) to a spectrum of probabilistic models under the same "TrueLabel + Confusion" paradigm, and show that our proposed hierarchical Bayesian model, called HybridConfusion, consistently outperforms DawidSkene on both synthetic and real-world data sets.
1206.4607
Distributed Tree Kernels
cs.LG stat.ML
In this paper, we propose the distributed tree kernels (DTK) as a novel method to reduce time and space complexity of tree kernels. Using a linear complexity algorithm to compute vectors for trees, we embed feature spaces of tree fragments in low-dimensional spaces where the kernel computation is directly done with dot product. We show that DTKs are faster, correlate with tree kernels, and obtain a statistically similar performance in two natural language processing tasks.
1206.4608
A Hybrid Algorithm for Convex Semidefinite Optimization
cs.LG cs.DS cs.NA stat.ML
We present a hybrid algorithm for optimizing a convex, smooth function over the cone of positive semidefinite matrices. Our algorithm converges to the global optimal solution and can be used to solve general large-scale semidefinite programs and hence can be readily applied to a variety of machine learning problems. We show experimental results on three machine learning problems (matrix completion, metric learning, and sparse PCA) . Our approach outperforms state-of-the-art algorithms.
1206.4609
On multi-view feature learning
cs.CV cs.LG stat.ML
Sparse coding is a common approach to learning local features for object recognition. Recently, there has been an increasing interest in learning features from spatio-temporal, binocular, or other multi-observation data, where the goal is to encode the relationship between images rather than the content of a single image. We provide an analysis of multi-view feature learning, which shows that hidden variables encode transformations by detecting rotation angles in the eigenspaces shared among multiple image warps. Our analysis helps explain recent experimental results showing that transformation-specific features emerge when training complex cell models on videos. Our analysis also shows that transformation-invariant features can emerge as a by-product of learning representations of transformations.
1206.4610
Manifold Relevance Determination
cs.LG cs.CV stat.ML
In this paper we present a fully Bayesian latent variable model which exploits conditional nonlinear(in)-dependence structures to learn an efficient latent representation. The latent space is factorized to represent shared and private information from multiple views of the data. In contrast to previous approaches, we introduce a relaxation to the discrete segmentation and allow for a "softly" shared latent space. Further, Bayesian techniques allow us to automatically estimate the dimensionality of the latent spaces. The model is capable of capturing structure underlying extremely high dimensional spaces. This is illustrated by modelling unprocessed images with tenths of thousands of pixels. This also allows us to directly generate novel images from the trained model by sampling from the discovered latent spaces. We also demonstrate the model by prediction of human pose in an ambiguous setting. Our Bayesian framework allows us to perform disambiguation in a principled manner by including latent space priors which incorporate the dynamic nature of the data.
1206.4611
A Convex Feature Learning Formulation for Latent Task Structure Discovery
cs.LG stat.ML
This paper considers the multi-task learning problem and in the setting where some relevant features could be shared across few related tasks. Most of the existing methods assume the extent to which the given tasks are related or share a common feature space to be known apriori. In real-world applications however, it is desirable to automatically discover the groups of related tasks that share a feature space. In this paper we aim at searching the exponentially large space of all possible groups of tasks that may share a feature space. The main contribution is a convex formulation that employs a graph-based regularizer and simultaneously discovers few groups of related tasks, having close-by task parameters, as well as the feature space shared within each group. The regularizer encodes an important structure among the groups of tasks leading to an efficient algorithm for solving it: if there is no feature space under which a group of tasks has close-by task parameters, then there does not exist such a feature space for any of its supersets. An efficient active set algorithm that exploits this simplification and performs a clever search in the exponentially large space is presented. The algorithm is guaranteed to solve the proposed formulation (within some precision) in a time polynomial in the number of groups of related tasks discovered. Empirical results on benchmark datasets show that the proposed formulation achieves good generalization and outperforms state-of-the-art multi-task learning algorithms in some cases.
1206.4612
Exact Soft Confidence-Weighted Learning
cs.LG
In this paper, we propose a new Soft Confidence-Weighted (SCW) online learning scheme, which enables the conventional confidence-weighted learning method to handle non-separable cases. Unlike the previous confidence-weighted learning algorithms, the proposed soft confidence-weighted learning method enjoys all the four salient properties: (i) large margin training, (ii) confidence weighting, (iii) capability to handle non-separable data, and (iv) adaptive margin. Our experimental results show that the proposed SCW algorithms significantly outperform the original CW algorithm. When comparing with a variety of state-of-the-art algorithms (including AROW, NAROW and NHERD), we found that SCW generally achieves better or at least comparable predictive accuracy, but enjoys significant advantage of computational efficiency (i.e., smaller number of updates and lower time cost).
1206.4613
Near-Optimal BRL using Optimistic Local Transitions
cs.AI cs.LG stat.ML
Model-based Bayesian Reinforcement Learning (BRL) allows a found formalization of the problem of acting optimally while facing an unknown environment, i.e., avoiding the exploration-exploitation dilemma. However, algorithms explicitly addressing BRL suffer from such a combinatorial explosion that a large body of work relies on heuristic algorithms. This paper introduces BOLT, a simple and (almost) deterministic heuristic algorithm for BRL which is optimistic about the transition function. We analyze BOLT's sample complexity, and show that under certain parameters, the algorithm is near-optimal in the Bayesian sense with high probability. Then, experimental results highlight the key differences of this method compared to previous work.
1206.4614
Information-theoretic Semi-supervised Metric Learning via Entropy Regularization
cs.LG stat.ML
We propose a general information-theoretic approach called Seraph (SEmi-supervised metRic leArning Paradigm with Hyper-sparsity) for metric learning that does not rely upon the manifold assumption. Given the probability parameterized by a Mahalanobis distance, we maximize the entropy of that probability on labeled data and minimize it on unlabeled data following entropy regularization, which allows the supervised and unsupervised parts to be integrated in a natural and meaningful way. Furthermore, Seraph is regularized by encouraging a low-rank projection induced from the metric. The optimization of Seraph is solved efficiently and stably by an EM-like scheme with the analytical E-Step and convex M-Step. Experiments demonstrate that Seraph compares favorably with many well-known global and local metric learning methods.
1206.4615
Levy Measure Decompositions for the Beta and Gamma Processes
stat.ME cs.LG math.ST stat.TH
We develop new representations for the Levy measures of the beta and gamma processes. These representations are manifested in terms of an infinite sum of well-behaved (proper) beta and gamma distributions. Further, we demonstrate how these infinite sums may be truncated in practice, and explicitly characterize truncation errors. We also perform an analysis of the characteristics of posterior distributions, based on the proposed decompositions. The decompositions provide new insights into the beta and gamma processes (and their generalizations), and we demonstrate how the proposed representation unifies some properties of the two. This paper is meant to provide a rigorous foundation for and new perspectives on Levy processes, as these are of increasing importance in machine learning.
1206.4616
A Hierarchical Dirichlet Process Model with Multiple Levels of Clustering for Human EEG Seizure Modeling
stat.AP cs.LG stat.ML
Driven by the multi-level structure of human intracranial electroencephalogram (iEEG) recordings of epileptic seizures, we introduce a new variant of a hierarchical Dirichlet Process---the multi-level clustering hierarchical Dirichlet Process (MLC-HDP)---that simultaneously clusters datasets on multiple levels. Our seizure dataset contains brain activity recorded in typically more than a hundred individual channels for each seizure of each patient. The MLC-HDP model clusters over channels-types, seizure-types, and patient-types simultaneously. We describe this model and its implementation in detail. We also present the results of a simulation study comparing the MLC-HDP to a similar model, the Nested Dirichlet Process and finally demonstrate the MLC-HDP's use in modeling seizures across multiple patients. We find the MLC-HDP's clustering to be comparable to independent human physician clusterings. To our knowledge, the MLC-HDP model is the first in the epilepsy literature capable of clustering seizures within and between patients.
1206.4617
Continuous Inverse Optimal Control with Locally Optimal Examples
cs.LG cs.AI stat.ML
Inverse optimal control, also known as inverse reinforcement learning, is the problem of recovering an unknown reward function in a Markov decision process from expert demonstrations of the optimal policy. We introduce a probabilistic inverse optimal control algorithm that scales gracefully with task dimensionality, and is suitable for large, continuous domains where even computing a full policy is impractical. By using a local approximation of the reward function, our method can also drop the assumption that the demonstrations are globally optimal, requiring only local optimality. This allows it to learn from examples that are unsuitable for prior methods.
1206.4618
Compact Hyperplane Hashing with Bilinear Functions
cs.LG stat.ML
Hyperplane hashing aims at rapidly searching nearest points to a hyperplane, and has shown practical impact in scaling up active learning with SVMs. Unfortunately, the existing randomized methods need long hash codes to achieve reasonable search accuracy and thus suffer from reduced search speed and large memory overhead. To this end, this paper proposes a novel hyperplane hashing technique which yields compact hash codes. The key idea is the bilinear form of the proposed hash functions, which leads to higher collision probability than the existing hyperplane hash functions when using random projections. To further increase the performance, we propose a learning based framework in which the bilinear functions are directly learned from the data. This results in short yet discriminative codes, and also boosts the search performance over the random projection based solutions. Large-scale active learning experiments carried out on two datasets with up to one million samples demonstrate the overall superiority of the proposed approach.
1206.4619
Inductive Kernel Low-rank Decomposition with Priors: A Generalized Nystrom Method
cs.LG
Low-rank matrix decomposition has gained great popularity recently in scaling up kernel methods to large amounts of data. However, some limitations could prevent them from working effectively in certain domains. For example, many existing approaches are intrinsically unsupervised, which does not incorporate side information (e.g., class labels) to produce task specific decompositions; also, they typically work "transductively", i.e., the factorization does not generalize to new samples, so the complete factorization needs to be recomputed when new samples become available. To solve these problems, in this paper we propose an"inductive"-flavored method for low-rank kernel decomposition with priors. We achieve this by generalizing the Nystr\"om method in a novel way. On the one hand, our approach employs a highly flexible, nonparametric structure that allows us to generalize the low-rank factors to arbitrarily new samples; on the other hand, it has linear time and space complexities, which can be orders of magnitudes faster than existing approaches and renders great efficiency in learning a low-rank kernel decomposition. Empirical results demonstrate the efficacy and efficiency of the proposed method.
1206.4620
Improved Information Gain Estimates for Decision Tree Induction
cs.LG stat.ML
Ensembles of classification and regression trees remain popular machine learning methods because they define flexible non-parametric models that predict well and are computationally efficient both during training and testing. During induction of decision trees one aims to find predicates that are maximally informative about the prediction target. To select good predicates most approaches estimate an information-theoretic scoring function, the information gain, both for classification and regression problems. We point out that the common estimation procedures are biased and show that by replacing them with improved estimators of the discrete and the differential entropy we can obtain better decision trees. In effect our modifications yield improved predictive performance and are simple to implement in any decision tree code.
1206.4621
Path Integral Policy Improvement with Covariance Matrix Adaptation
cs.LG
There has been a recent focus in reinforcement learning on addressing continuous state and action problems by optimizing parameterized policies. PI2 is a recent example of this approach. It combines a derivation from first principles of stochastic optimal control with tools from statistical estimation theory. In this paper, we consider PI2 as a member of the wider family of methods which share the concept of probability-weighted averaging to iteratively update parameters to optimize a cost function. We compare PI2 to other members of the same family - Cross-Entropy Methods and CMAES - at the conceptual level and in terms of performance. The comparison suggests the derivation of a novel algorithm which we call PI2-CMA for "Path Integral Policy Improvement with Covariance Matrix Adaptation". PI2-CMA's main advantage is that it determines the magnitude of the exploration noise automatically.
1206.4622
A Graphical Model Formulation of Collaborative Filtering Neighbourhood Methods with Fast Maximum Entropy Training
cs.LG cs.IR stat.ML
Item neighbourhood methods for collaborative filtering learn a weighted graph over the set of items, where each item is connected to those it is most similar to. The prediction of a user's rating on an item is then given by that rating of neighbouring items, weighted by their similarity. This paper presents a new neighbourhood approach which we call item fields, whereby an undirected graphical model is formed over the item graph. The resulting prediction rule is a simple generalization of the classical approaches, which takes into account non-local information in the graph, allowing its best results to be obtained when using drastically fewer edges than other neighbourhood approaches. A fast approximate maximum entropy training method based on the Bethe approximation is presented, which uses a simple gradient ascent procedure. When using precomputed sufficient statistics on the Movielens datasets, our method is faster than maximum likelihood approaches by two orders of magnitude.
1206.4623
On the Size of the Online Kernel Sparsification Dictionary
cs.LG stat.ML
We analyze the size of the dictionary constructed from online kernel sparsification, using a novel formula that expresses the expected determinant of the kernel Gram matrix in terms of the eigenvalues of the covariance operator. Using this formula, we are able to connect the cardinality of the dictionary with the eigen-decay of the covariance operator. In particular, we show that under certain technical conditions, the size of the dictionary will always grow sub-linearly in the number of data points, and, as a consequence, the kernel linear regressor constructed from the resulting dictionary is consistent.
1206.4624
Robust Multiple Manifolds Structure Learning
cs.LG stat.ML
We present a robust multiple manifolds structure learning (RMMSL) scheme to robustly estimate data structures under the multiple low intrinsic dimensional manifolds assumption. In the local learning stage, RMMSL efficiently estimates local tangent space by weighted low-rank matrix factorization. In the global learning stage, we propose a robust manifold clustering method based on local structure learning results. The proposed clustering method is designed to get the flattest manifolds clusters by introducing a novel curved-level similarity function. Our approach is evaluated and compared to state-of-the-art methods on synthetic data, handwritten digit images, human motion capture data and motorbike videos. We demonstrate the effectiveness of the proposed approach, which yields higher clustering accuracy, and produces promising results for challenging tasks of human motion segmentation and motion flow learning from videos.
1206.4625
Optimizing F-measure: A Tale of Two Approaches
cs.LG
F-measures are popular performance metrics, particularly for tasks with imbalanced data sets. Algorithms for learning to maximize F-measures follow two approaches: the empirical utility maximization (EUM) approach learns a classifier having optimal performance on training data, while the decision-theoretic approach learns a probabilistic model and then predicts labels with maximum expected F-measure. In this paper, we investigate the theoretical justifications and connections for these two approaches, and we study the conditions under which one approach is preferable to the other using synthetic and real datasets. Given accurate models, our results suggest that the two approaches are asymptotically equivalent given large training and test sets. Nevertheless, empirically, the EUM approach appears to be more robust against model misspecification, and given a good model, the decision-theoretic approach appears to be better for handling rare classes and a common domain adaptation scenario.
1206.4626
On-Line Portfolio Selection with Moving Average Reversion
cs.CE cs.LG q-fin.PM
On-line portfolio selection has attracted increasing interests in machine learning and AI communities recently. Empirical evidences show that stock's high and low prices are temporary and stock price relatives are likely to follow the mean reversion phenomenon. While the existing mean reversion strategies are shown to achieve good empirical performance on many real datasets, they often make the single-period mean reversion assumption, which is not always satisfied in some real datasets, leading to poor performance when the assumption does not hold. To overcome the limitation, this article proposes a multiple-period mean reversion, or so-called Moving Average Reversion (MAR), and a new on-line portfolio selection strategy named "On-Line Moving Average Reversion" (OLMAR), which exploits MAR by applying powerful online learning techniques. From our empirical results, we found that OLMAR can overcome the drawback of existing mean reversion algorithms and achieve significantly better results, especially on the datasets where the existing mean reversion algorithms failed. In addition to superior trading performance, OLMAR also runs extremely fast, further supporting its practical applicability to a wide range of applications.
1206.4627
Convergence Rates of Biased Stochastic Optimization for Learning Sparse Ising Models
cs.LG stat.ML
We study the convergence rate of stochastic optimization of exact (NP-hard) objectives, for which only biased estimates of the gradient are available. We motivate this problem in the context of learning the structure and parameters of Ising models. We first provide a convergence-rate analysis of deterministic errors for forward-backward splitting (FBS). We then extend our analysis to biased stochastic errors, by first characterizing a family of samplers and providing a high probability bound that allows understanding not only FBS, but also proximal gradient (PG) methods. We derive some interesting conclusions: FBS requires only a logarithmically increasing number of random samples in order to converge (although at a very low rate); the required number of random samples is the same for the deterministic and the biased stochastic setting for FBS and basic PG; accelerated PG is not guaranteed to converge in the biased stochastic setting.
1206.4628
Robust PCA in High-dimension: A Deterministic Approach
cs.LG stat.ML
We consider principal component analysis for contaminated data-set in the high dimensional regime, where the dimensionality of each observation is comparable or even more than the number of observations. We propose a deterministic high-dimensional robust PCA algorithm which inherits all theoretical properties of its randomized counterpart, i.e., it is tractable, robust to contaminated points, easily kernelizable, asymptotic consistent and achieves maximal robustness -- a breakdown point of 50%. More importantly, the proposed method exhibits significantly better computational efficiency, which makes it suitable for large-scale real applications.
1206.4629
Multiple Kernel Learning from Noisy Labels by Stochastic Programming
cs.LG
We study the problem of multiple kernel learning from noisy labels. This is in contrast to most of the previous studies on multiple kernel learning that mainly focus on developing efficient algorithms and assume perfectly labeled training examples. Directly applying the existing multiple kernel learning algorithms to noisily labeled examples often leads to suboptimal performance due to the incorrect class assignments. We address this challenge by casting multiple kernel learning from noisy labels into a stochastic programming problem, and presenting a minimax formulation. We develop an efficient algorithm for solving the related convex-concave optimization problem with a fast convergence rate of $O(1/T)$ where $T$ is the number of iterations. Empirical studies on UCI data sets verify both the effectiveness of the proposed framework and the efficiency of the proposed optimization algorithm.
1206.4630
Efficient Decomposed Learning for Structured Prediction
cs.LG
Structured prediction is the cornerstone of several machine learning applications. Unfortunately, in structured prediction settings with expressive inter-variable interactions, exact inference-based learning algorithms, e.g. Structural SVM, are often intractable. We present a new way, Decomposed Learning (DecL), which performs efficient learning by restricting the inference step to a limited part of the structured spaces. We provide characterizations based on the structure, target parameters, and gold labels, under which DecL is equivalent to exact learning. We then show that in real world settings, where our theoretical assumptions may not completely hold, DecL-based algorithms are significantly more efficient and as accurate as exact learning.
1206.4631
A Poisson convolution model for characterizing topical content with word frequency and exclusivity
cs.LG cs.CL cs.IR stat.ME stat.ML
An ongoing challenge in the analysis of document collections is how to summarize content in terms of a set of inferred themes that can be interpreted substantively in terms of topics. The current practice of parametrizing the themes in terms of most frequent words limits interpretability by ignoring the differential use of words across topics. We argue that words that are both common and exclusive to a theme are more effective at characterizing topical content. We consider a setting where professional editors have annotated documents to a collection of topic categories, organized into a tree, in which leaf-nodes correspond to the most specific topics. Each document is annotated to multiple categories, at different levels of the tree. We introduce a hierarchical Poisson convolution model to analyze annotated documents in this setting. The model leverages the structure among categories defined by professional editors to infer a clear semantic description for each topic in terms of words that are both frequent and exclusive. We carry out a large randomized experiment on Amazon Turk to demonstrate that topic summaries based on the FREX score are more interpretable than currently established frequency based summaries, and that the proposed model produces more efficient estimates of exclusivity than with currently models. We also develop a parallelized Hamiltonian Monte Carlo sampler that allows the inference to scale to millions of documents.
1206.4632
A Complete Analysis of the l_1,p Group-Lasso
cs.LG math.OC stat.ML
The Group-Lasso is a well-known tool for joint regularization in machine learning methods. While the l_{1,2} and the l_{1,\infty} version have been studied in detail and efficient algorithms exist, there are still open questions regarding other l_{1,p} variants. We characterize conditions for solutions of the l_{1,p} Group-Lasso for all p-norms with 1 <= p <= \infty, and we present a unified active set algorithm. For all p-norms, a highly efficient projected gradient algorithm is presented. This new algorithm enables us to compare the prediction performance of many variants of the Group-Lasso in a multi-task learning setting, where the aim is to solve many learning problems in parallel which are coupled via the Group-Lasso constraint. We conduct large-scale experiments on synthetic data and on two real-world data sets. In accordance with theoretical characterizations of the different norms we observe that the weak-coupling norms with p between 1.5 and 2 consistently outperform the strong-coupling norms with p >> 2.
1206.4633
Fast Bounded Online Gradient Descent Algorithms for Scalable Kernel-Based Online Learning
cs.LG stat.ML
Kernel-based online learning has often shown state-of-the-art performance for many online learning tasks. It, however, suffers from a major shortcoming, that is, the unbounded number of support vectors, making it non-scalable and unsuitable for applications with large-scale datasets. In this work, we study the problem of bounded kernel-based online learning that aims to constrain the number of support vectors by a predefined budget. Although several algorithms have been proposed in literature, they are neither computationally efficient due to their intensive budget maintenance strategy nor effective due to the use of simple Perceptron algorithm. To overcome these limitations, we propose a framework for bounded kernel-based online learning based on an online gradient descent approach. We propose two efficient algorithms of bounded online gradient descent (BOGD) for scalable kernel-based online learning: (i) BOGD by maintaining support vectors using uniform sampling, and (ii) BOGD++ by maintaining support vectors using non-uniform sampling. We present theoretical analysis of regret bound for both algorithms, and found promising empirical performance in terms of both efficacy and efficiency by comparing them to several well-known algorithms for bounded kernel-based online learning on large-scale datasets.
1206.4634
Artist Agent: A Reinforcement Learning Approach to Automatic Stroke Generation in Oriental Ink Painting
cs.LG cs.GR stat.ML
Oriental ink painting, called Sumi-e, is one of the most appealing painting styles that has attracted artists around the world. Major challenges in computer-based Sumi-e simulation are to abstract complex scene information and draw smooth and natural brush strokes. To automatically find such strokes, we propose to model the brush as a reinforcement learning agent, and learn desired brush-trajectories by maximizing the sum of rewards in the policy search framework. We also provide elaborate design of actions, states, and rewards tailored for a Sumi-e agent. The effectiveness of our proposed approach is demonstrated through simulated Sumi-e experiments.
1206.4635
Deep Mixtures of Factor Analysers
cs.LG stat.ML
An efficient way to learn deep density models that have many layers of latent variables is to learn one layer at a time using a model that has only one layer of latent variables. After learning each layer, samples from the posterior distributions for that layer are used as training data for learning the next layer. This approach is commonly used with Restricted Boltzmann Machines, which are undirected graphical models with a single hidden layer, but it can also be used with Mixtures of Factor Analysers (MFAs) which are directed graphical models. In this paper, we present a greedy layer-wise learning algorithm for Deep Mixtures of Factor Analysers (DMFAs). Even though a DMFA can be converted to an equivalent shallow MFA by multiplying together the factor loading matrices at different levels, learning and inference are much more efficient in a DMFA and the sharing of each lower-level factor loading matrix by many different higher level MFAs prevents overfitting. We demonstrate empirically that DMFAs learn better density models than both MFAs and two types of Restricted Boltzmann Machine on a wide variety of datasets.
1206.4636
Modeling Latent Variable Uncertainty for Loss-based Learning
cs.LG cs.AI cs.CV
We consider the problem of parameter estimation using weakly supervised datasets, where a training sample consists of the input and a partially specified annotation, which we refer to as the output. The missing information in the annotation is modeled using latent variables. Previous methods overburden a single distribution with two separate tasks: (i) modeling the uncertainty in the latent variables during training; and (ii) making accurate predictions for the output and the latent variables during testing. We propose a novel framework that separates the demands of the two tasks using two distributions: (i) a conditional distribution to model the uncertainty of the latent variables for a given input-output pair; and (ii) a delta distribution to predict the output and the latent variables for a given input. During learning, we encourage agreement between the two distributions by minimizing a loss-based dissimilarity coefficient. Our approach generalizes latent SVM in two important ways: (i) it models the uncertainty over latent variables instead of relying on a pointwise estimate; and (ii) it allows the use of loss functions that depend on latent variables, which greatly increases its applicability. We demonstrate the efficacy of our approach on two challenging problems---object detection and action detection---using publicly available datasets.
1206.4637
Learning to Identify Regular Expressions that Describe Email Campaigns
cs.LG cs.CL stat.ML
This paper addresses the problem of inferring a regular expression from a given set of strings that resembles, as closely as possible, the regular expression that a human expert would have written to identify the language. This is motivated by our goal of automating the task of postmasters of an email service who use regular expressions to describe and blacklist email spam campaigns. Training data contains batches of messages and corresponding regular expressions that an expert postmaster feels confident to blacklist. We model this task as a learning problem with structured output spaces and an appropriate loss function, derive a decoder and the resulting optimization problem, and a report on a case study conducted with an email service.
1206.4638
Efficient Euclidean Projections onto the Intersection of Norm Balls
cs.LG stat.ML
Using sparse-inducing norms to learn robust models has received increasing attention from many fields for its attractive properties. Projection-based methods have been widely applied to learning tasks constrained by such norms. As a key building block of these methods, an efficient operator for Euclidean projection onto the intersection of $\ell_1$ and $\ell_{1,q}$ norm balls $(q=2\text{or}\infty)$ is proposed in this paper. We prove that the projection can be reduced to finding the root of an auxiliary function which is piecewise smooth and monotonic. Hence, a bisection algorithm is sufficient to solve the problem. We show that the time complexity of our solution is $O(n+g\log g)$ for $q=2$ and $O(n\log n)$ for $q=\infty$, where $n$ is the dimensionality of the vector to be projected and $g$ is the number of disjoint groups; we confirm this complexity by experimentation. Empirical study reveals that our method achieves significantly better performance than classical methods in terms of running time and memory usage. We further show that embedded with our efficient projection operator, projection-based algorithms can solve regression problems with composite norm constraints more efficiently than other methods and give superior accuracy.
1206.4639
Adaptive Regularization for Weight Matrices
cs.LG cs.AI
Algorithms for learning distributions over weight-vectors, such as AROW were recently shown empirically to achieve state-of-the-art performance at various problems, with strong theoretical guaranties. Extending these algorithms to matrix models pose challenges since the number of free parameters in the covariance of the distribution scales as $n^4$ with the dimension $n$ of the matrix, and $n$ tends to be large in real applications. We describe, analyze and experiment with two new algorithms for learning distribution of matrix models. Our first algorithm maintains a diagonal covariance over the parameters and can handle large covariance matrices. The second algorithm factors the covariance to capture inter-features correlation while keeping the number of parameters linear in the size of the original matrix. We analyze both algorithms in the mistake bound model and show a superior precision performance of our approach over other algorithms in two tasks: retrieving similar images, and ranking similar documents. The factored algorithm is shown to attain faster convergence rate.
1206.4640
Stability of matrix factorization for collaborative filtering
cs.NA cs.LG stat.ML
We study the stability vis a vis adversarial noise of matrix factorization algorithm for matrix completion. In particular, our results include: (I) we bound the gap between the solution matrix of the factorization method and the ground truth in terms of root mean square error; (II) we treat the matrix factorization as a subspace fitting problem and analyze the difference between the solution subspace and the ground truth; (III) we analyze the prediction error of individual users based on the subspace stability. We apply these results to the problem of collaborative filtering under manipulator attack, which leads to useful insights and guidelines for collaborative filtering system design.
1206.4641
Total Variation and Euler's Elastica for Supervised Learning
cs.LG cs.CV stat.ML
In recent years, total variation (TV) and Euler's elastica (EE) have been successfully applied to image processing tasks such as denoising and inpainting. This paper investigates how to extend TV and EE to the supervised learning settings on high dimensional data. The supervised learning problem can be formulated as an energy functional minimization under Tikhonov regularization scheme, where the energy is composed of a squared loss and a total variation smoothing (or Euler's elastica smoothing). Its solution via variational principles leads to an Euler-Lagrange PDE. However, the PDE is always high-dimensional and cannot be directly solved by common methods. Instead, radial basis functions are utilized to approximate the target function, reducing the problem to finding the linear coefficients of basis functions. We apply the proposed methods to supervised learning tasks (including binary classification, multi-class classification, and regression) on benchmark data sets. Extensive experiments have demonstrated promising results of the proposed methods.
1206.4642
Fast Computation of Subpath Kernel for Trees
cs.DS cs.LG stat.ML
The kernel method is a potential approach to analyzing structured data such as sequences, trees, and graphs; however, unordered trees have not been investigated extensively. Kimura et al. (2011) proposed a kernel function for unordered trees on the basis of their subpaths, which are vertical substructures of trees responsible for hierarchical information in them. Their kernel exhibits practically good performance in terms of accuracy and speed; however, linear-time computation is not guaranteed theoretically, unlike the case of the other unordered tree kernel proposed by Vishwanathan and Smola (2003). In this paper, we propose a theoretically guaranteed linear-time kernel computation algorithm that is practically fast, and we present an efficient prediction algorithm whose running time depends only on the size of the input tree. Experimental results show that the proposed algorithms are quite efficient in practice.
1206.4643
Lightning Does Not Strike Twice: Robust MDPs with Coupled Uncertainty
cs.LG cs.GT cs.SY
We consider Markov decision processes under parameter uncertainty. Previous studies all restrict to the case that uncertainties among different states are uncoupled, which leads to conservative solutions. In contrast, we introduce an intuitive concept, termed "Lightning Does not Strike Twice," to model coupled uncertain parameters. Specifically, we require that the system can deviate from its nominal parameters only a bounded number of times. We give probabilistic guarantees indicating that this model represents real life situations and devise tractable algorithms for computing optimal control policies using this concept.
1206.4644
Groupwise Constrained Reconstruction for Subspace Clustering
cs.LG stat.ML
Reconstruction based subspace clustering methods compute a self reconstruction matrix over the samples and use it for spectral clustering to obtain the final clustering result. Their success largely relies on the assumption that the underlying subspaces are independent, which, however, does not always hold in the applications with increasing number of subspaces. In this paper, we propose a novel reconstruction based subspace clustering model without making the subspace independence assumption. In our model, certain properties of the reconstruction matrix are explicitly characterized using the latent cluster indicators, and the affinity matrix used for spectral clustering can be directly built from the posterior of the latent cluster indicators instead of the reconstruction matrix. Experimental results on both synthetic and real-world datasets show that the proposed model can outperform the state-of-the-art methods.
1206.4645
Ensemble Methods for Convex Regression with Applications to Geometric Programming Based Circuit Design
cs.LG cs.NA stat.ME stat.ML
Convex regression is a promising area for bridging statistical estimation and deterministic convex optimization. New piecewise linear convex regression methods are fast and scalable, but can have instability when used to approximate constraints or objective functions for optimization. Ensemble methods, like bagging, smearing and random partitioning, can alleviate this problem and maintain the theoretical properties of the underlying estimator. We empirically examine the performance of ensemble methods for prediction and optimization, and then apply them to device modeling and constraint approximation for geometric programming based circuit design.
1206.4646
Partial-Hessian Strategies for Fast Learning of Nonlinear Embeddings
cs.LG stat.ML
Stochastic neighbor embedding (SNE) and related nonlinear manifold learning algorithms achieve high-quality low-dimensional representations of similarity data, but are notoriously slow to train. We propose a generic formulation of embedding algorithms that includes SNE and other existing algorithms, and study their relation with spectral methods and graph Laplacians. This allows us to define several partial-Hessian optimization strategies, characterize their global and local convergence, and evaluate them empirically. We achieve up to two orders of magnitude speedup over existing training methods with a strategy (which we call the spectral direction) that adds nearly no overhead to the gradient and yet is simple, scalable and applicable to several existing and future embedding algorithms.
1206.4647
Active Learning for Matching Problems
cs.LG cs.AI cs.IR
Effective learning of user preferences is critical to easing user burden in various types of matching problems. Equally important is active query selection to further reduce the amount of preference information users must provide. We address the problem of active learning of user preferences for matching problems, introducing a novel method for determining probabilistic matchings, and developing several new active learning strategies that are sensitive to the specific matching objective. Experiments with real-world data sets spanning diverse domains demonstrate that matching-sensitive active learning
1206.4648
Two-Manifold Problems with Applications to Nonlinear System Identification
cs.LG
Recently, there has been much interest in spectral approaches to learning manifolds---so-called kernel eigenmap methods. These methods have had some successes, but their applicability is limited because they are not robust to noise. To address this limitation, we look at two-manifold problems, in which we simultaneously reconstruct two related manifolds, each representing a different view of the same data. By solving these interconnected learning problems together, two-manifold algorithms are able to succeed where a non-integrated approach would fail: each view allows us to suppress noise in the other, reducing bias. We propose a class of algorithms for two-manifold problems, based on spectral decomposition of cross-covariance operators in Hilbert space, and discuss when two-manifold problems are useful. Finally, we demonstrate that solving a two-manifold problem can aid in learning a nonlinear dynamical system from limited data.