arXiv subjects

Venugopal V. Veeravalli

Publications and source records attributed to Venugopal V. Veeravalli.

At least 55 records · Page 3Linked to original sources

arXiv2020

Quickest Detection of Growing Dynamic Anomalies in Networks

The problem of quickest growing dynamic anomaly detection in sensor networks is studied. Initially, the observations at the sensors, which are sampled sequentially by the decision maker, are generated according to a pre-change distribution. At some unknown but deterministic time instant, a dynamic anomaly emerges in the network, affecting a different set of sensors as time progresses. The observations of the affected sensors are generated from a post-change distribution. It is assumed that the number of affected sensors increases with time, and that only the initial and the final size of the anomaly are known by the decision maker. The goal is to detect the emergence of the anomaly as quickly as possible while guaranteeing a sufficiently low frequency of false alarm events. This detection problem is posed as a stochastic optimization problem by using a delay metric that is based on the worst possible path of the anomaly. A detection rule is proposed that is asymptotically optimal as the mean time to false alarm goes to infinity. Finally, numerical results are provided to validate our theoretical analysis.

math.ST↗

arXiv2019

Multi-User MABs with User Dependent Rewards for Uncoordinated Spectrum Access

Multi-user multi-armed bandits have emerged as a good model for uncoordinated spectrum access problems. In this paper we consider the scenario where users cannot communicate with each other. In addition, the environment may appear differently to different users, ${i.e.}$, the mean rewards as observed by different users for the same channel may be different. With this setup, we present a policy that achieves a regret of $O (\log{T})$. This paper has been accepted at Asilomar Conference on Signals, Systems, and Computers 2019.

cs.IT↗

arXiv2019

Information Flow Optimization in Inference Networks

The problem of maximizing the information flow through a sensor network tasked with an inference objective at the fusion center is considered. The sensor nodes take observations, compress and send them to the fusion center through a network of relays. The network imposes capacity constraints on the rate of transmission in each connection and flow conservation constraints. It is shown that this rate-constrained inference problem can be cast as a Network Utility Maximization problem by suitably defining the utility functions for each sensor, and can be solved using existing techniques. Two practical settings are analyzed: multi-terminal parameter estimation and binary hypothesis testing. It is verified via simulations that using the proposed formulation gives better inference performance than the Max-Flow solution that simply maximizes the total bit-rate to the fusion center.

math.OC↗

arXiv2019

Sequential Controlled Sensing for Composite Multihypothesis Testing

The problem of multi-hypothesis testing with controlled sensing of observations is considered. The distribution of observations collected under each control is assumed to follow a single-parameter exponential family distribution. The goal is to design a policy to find the true hypothesis with minimum expected delay while ensuring that the probability of error is below a given constraint. The decision-maker can control the delay by intelligently choosing the control for observation collection in each time slot. We derive a policy that satisfies the given constraint on the error probability. We also show that the policy is asymptotically optimal in the sense that it asymptotically achieves an information-theoretic lower bound on the expected delay.

math.ST↗

arXiv2019

Multi-User Multi-Armed Bandits for Uncoordinated Spectrum Access

A multi-user multi-armed bandit (MAB) framework is used to develop algorithms for uncoordinated spectrum access. The number of users is assumed to be unknown to each user. A stochastic setting is first considered, where the rewards on a channel are the same for each user. In contrast to prior work, it is assumed that the number of users can possibly exceed the number of channels, and that rewards can be non-zero even under collisions. The proposed algorithm consists of an estimation phase and an allocation phase. It is shown that if every user adopts the algorithm, the system wide regret is constant with time with high probability. The regret guarantees hold for any number of users and channels, in particular, even when the number of users is less than the number of channels. Next, an adversarial multi-user MAB framework is considered, where the rewards on the channels are user-dependent. It is assumed that the number of users is less than the number of channels, and that the users receive zero reward on collision. The proposed algorithm combines the Exp3.P algorithm developed in prior work for single user adversarial bandits with a collision resolution mechanism to achieve sub-linear regret. It is shown that if every user employs the proposed algorithm, the system wide regret is of the order $O(T^\frac{3}{4})$ over a horizon of time $T$. The algorithms in both stochastic and adversarial scenarios are extended to the dynamic case where the number of users in the system evolves over time and are shown to lead to sub-linear regret.

cs.LG↗

arXiv2019

Information-Theoretic Understanding of Population Risk Improvement with Model Compression

We show that model compression can improve the population risk of a pre-trained model, by studying the tradeoff between the decrease in the generalization error and the increase in the empirical risk with model compression. We first prove that model compression reduces an information-theoretic bound on the generalization error; this allows for an interpretation of model compression as a regularization technique to avoid overfitting. We then characterize the increase in empirical risk with model compression using rate distortion theory. These results imply that the population risk could be improved by model compression if the decrease in generalization error exceeds the increase in empirical risk. We show through a linear regression example that such a decrease in population risk due to model compression is indeed possible. Our theoretical results further suggest that the Hessian-weighted $K$-means clustering compression approach can be improved by regularizing the distance between the clustering centers. We provide experiments with neural networks to support our theoretical assertions.

stat.ML↗

arXiv2019

Data-driven Voltage Regulation in Radial Power Distribution Systems

In this paper, we develop a data-driven voltage regulation framework for distributed energy resources (DERs) in a balanced radial power distribution system. The objective is to determine optimal DER power injections that minimize the voltage deviations from a desirable voltage range without knowing a complete power distribution system model a priori. The nonlinear relationship between the voltage magnitudes and the power injections in the power distribution system is approximated by a linear model, the parameters of which---referred to as the voltage sensitivities---can be computed directly using information on the topology and the line parameters. Assuming the knowledge of feasible topology configurations and distribution line resistance-to-reactance ratios, the true topology configuration and corresponding line parameters can be estimated effectively using a few sets of measurements on voltage magnitudes and power injections. Using the estimated voltage sensitivities, the optimal DER power injections can be readily determined by solving a convex optimization problem. The proposed framework is intrinsically adaptive to changes in system conditions such as unknown topology reconfiguration due to its data-driven nature. The effectiveness and efficiency of the proposed framework is validated via numerical simulations on the IEEE 123-bus distribution test feeder.

math.OC↗

arXiv2018

Quickest Change Detection under Transient Dynamics: Theory and Asymptotic Analysis

The problem of quickest change detection (QCD) under transient dynamics is studied, where the change from the initial distribution to the final persistent distribution does not happen instantaneously, but after a series of transient phases. The observations within the different phases are generated by different distributions. The objective is to detect the change as quickly as possible, while controlling the average run length (ARL) to false alarm, when the durations of the transient phases are completely unknown. Two algorithms are considered, the dynamic Cumulative Sum (CuSum) algorithm, proposed in earlier work, and a newly constructed weighted dynamic CuSum algorithm. Both algorithms admit recursions that facilitate their practical implementation, and they are adaptive to the unknown transient durations. Specifically, their asymptotic optimality is established with respect to both Lorden's and Pollak's criteria as the ARL to false alarm and the durations of the transient phases go to infinity at any relative rate. Numerical results are provided to demonstrate the adaptivity of the proposed algorithms, and to validate the theoretical results.

math.ST↗

arXiv2018

Degrees of Freedom in Wireless Interference Networks with Cooperative Transmission and Backhaul Load Constraints

Degrees of freedom (DoF) gains are studied in wireless networks with cooperative transmission under a backhaul load constraint that limits the average number of messages that can be delivered from a centralized controller to base station transmitters. The backhaul load is defined as the sum of all the messages available at all the transmitters per channel use, normalized by the number of users. For Wyner's linear interference network, where each transmitter is connected to the receiver having the same index as well as one succeeding receiver, the per user DoF is characterized and the optimal scheme is presented. Furthermore, it is shown that the optimal assignment of messages to transmitters is asymmetric and satisfies a local cooperation constraint, and that the optimal coding scheme relies only on one-shot cooperative zero-forcing transmit beamforming. Using insights from the analysis of Wyner's linear interference network, the results are extended to the more practical hexagonal sectored cellular network, and coding schemes based on cooperative zero-forcing are shown to deliver significant DoF gains. It is established that by allowing for cooperative transmission and a flexible message assignment that is constrained only by an average backhaul load, one can deliver the rate gains promised by information-theoretic upper bounds with practical one-shot schemes that incur little or no additional load on the backhaul. Finally, useful upper bounds on the per user DoF for schemes based on cooperative zero-forcing are presented for lower values of the average backhaul load constraint, and an optimization framework is formulated for the general converse problem.

cs.IT↗

arXiv2018

Model change detection with application to machine learning

Model change detection is studied, in which there are two sets of samples that are independently and identically distributed (i.i.d.) according to a pre-change probabilistic model with parameter $θ$, and a post-change model with parameter $θ'$, respectively. The goal is to detect whether the change in the model is significant, i.e., whether the difference between the pre-change parameter and the post-change parameter $\|θ-θ'\|_2$ is larger than a pre-determined threshold $ρ$. The problem is considered in a Neyman-Pearson setting, where the goal is to maximize the probability of detection under a false alarm constraint. Since the generalized likelihood ratio test (GLRT) is difficult to compute in this problem, we construct an empirical difference test (EDT), which approximates the GLRT and has low computational complexity. Moreover, we provide an approximation method to set the threshold of the EDT to meet the false alarm constraint. Experiments with linear regression and logistic regression are conducted to validate the proposed algorithms.

stat.ML↗

arXiv2018

A Binning Approach to Quickest Change Detection with Unknown Post-Change Distribution

The problem of quickest detection of a change in distribution is considered under the assumption that the pre-change distribution is known, and the post-change distribution is only known to belong to a family of distributions distinguishable from a discretized version of the pre-change distribution. A sequential change detection procedure is proposed that partitions the sample space into a finite number of bins, and monitors the number of samples falling into each of these bins to detect the change. A test statistic that approximates the generalized likelihood ratio test is developed. It is shown that the proposed test statistic can be efficiently computed using a recursive update scheme, and a procedure for choosing the number of bins in the scheme is provided. Various asymptotic properties of the test statistic are derived to offer insights into its performance trade-off between average detection delay and average run length to a false alarm. Testing on synthetic and real data demonstrates that our approach is comparable or better in performance to existing non-parametric change detection methods.

stat.AP↗

arXiv2018

Quickest Detection of Dynamic Events in Networks

The problem of quickest detection of dynamic events in networks is studied. At some unknown time, an event occurs, and a number of nodes in the network are affected by the event, in that they undergo a change in the statistics of their observations. It is assumed that the event is dynamic, in that it can propagate along the edges in the network, and affect more and more nodes with time. The event propagation dynamics is assumed to be unknown. The goal is to design a sequential algorithm that can detect a "significant" event, i.e., when the event has affected no fewer than $η$ nodes, as quickly as possible, while controlling the false alarm rate. Fully connected networks are studied first, and the results are then extended to arbitrarily connected networks. The designed algorithms are shown to be adaptive to the unknown propagation dynamics, and their first-order asymptotic optimality is demonstrated as the false alarm rate goes to zero. The algorithms can be implemented with linear computational complexity in the network size at each time step, which is critical for online implementation. Numerical simulations are provided to validate the theoretical results.

eess.SP↗

arXiv2018

Active and Adaptive Sequential learning

A framework is introduced for actively and adaptively solving a sequence of machine learning problems, which are changing in bounded manner from one time step to the next. An algorithm is developed that actively queries the labels of the most informative samples from an unlabeled data pool, and that adapts to the change by utilizing the information acquired in the previous steps. Our analysis shows that the proposed active learning algorithm based on stochastic gradient descent achieves a near-optimal excess risk performance for maximum likelihood estimation. Furthermore, an estimator of the change in the learning problems using the active learning samples is constructed, which provides an adaptive sample size selection rule that guarantees the excess risk is bounded for sufficiently large number of time steps. Experiments with synthetic and real data are presented to validate our algorithm and theoretical results.

cs.LG↗

arXiv2018

Estimation of KL Divergence: Optimal Minimax Rate

The problem of estimating the Kullback-Leibler divergence $D(P\|Q)$ between two unknown distributions $P$ and $Q$ is studied, under the assumption that the alphabet size $k$ of the distributions can scale to infinity. The estimation is based on $m$ independent samples drawn from $P$ and $n$ independent samples drawn from $Q$. It is first shown that there does not exist any consistent estimator that guarantees asymptotically small worst-case quadratic risk over the set of all pairs of distributions. A restricted set that contains pairs of distributions, with density ratio bounded by a function $f(k)$ is further considered. {An augmented plug-in estimator is proposed, and its worst-case quadratic risk is shown to be within a constant factor of $(\frac{k}{m}+\frac{kf(k)}{n})^2+\frac{\log ^2 f(k)}{m}+\frac{f(k)}{n}$, if $m$ and $n$ exceed a constant factor of $k$ and $kf(k)$, respectively.} Moreover, the minimax quadratic risk is characterized to be within a constant factor of $(\frac{k}{m\log k}+\frac{kf(k)}{n\log k})^2+\frac{\log ^2 f(k)}{m}+\frac{f(k)}{n}$, if $m$ and $n$ exceed a constant factor of $k/\log(k)$ and $kf(k)/\log k$, respectively. The lower bound on the minimax quadratic risk is characterized by employing a generalized Le Cam's method. A minimax optimal estimator is then constructed by employing both the polynomial approximation and the plug-in approaches.

cs.IT↗

arXiv2017

Linear-Complexity Exponentially-Consistent Tests for Universal Outlying Sequence Detection

The problem of universal outlying sequence detection is studied, where the goal is to detect outlying sequences among $M$ sequences of samples. A sequence is considered as outlying if the observations therein are generated by a distribution different from those generating the observations in the majority of the sequences. In the universal setting, we are interested in identifying all the outlying sequences without knowing the underlying generating distributions. In this paper, a class of tests based on distribution clustering is proposed. These tests are shown to be exponentially consistent with linear time complexity in $M$. Numerical results demonstrate that our clustering-based tests achieve similar performance to existing tests, while being considerably more computationally efficient.

cs.IT↗

arXiv2016

Detecting Sparse Mixtures: Rate of Decay of Error Probability

We study the rate of decay of the probability of error for distinguishing between a sparse signal with noise, modeled as a sparse mixture, from pure noise. This problem has many applications in signal processing, evolutionary biology, bioinformatics, astrophysics and feature selection for machine learning. We let the mixture probability tend to zero as the number of observations tends to infinity and derive oracle rates at which the error probability can be driven to zero for a general class of signal and noise distributions via the likelihood ratio test. In contrast to the problem of detection of non-sparse signals, we see the log-probability of error decays sublinearly rather than linearly and is characterized through the $χ^2$-divergence rather than the Kullback-Leibler divergence for "weak" signals and can be independent of divergence for "strong" signals. Our contribution is the first characterization of the rate of decay of the error probability for this problem for both the false alarm and miss probabilities.

cs.IT↗

arXiv2015

Universal Outlying sequence detection For Continuous Observations

The following detection problem is studied, in which there are $M$ sequences of samples out of which one outlier sequence needs to be detected. Each typical sequence contains $n$ independent and identically distributed (i.i.d.) continuous observations from a known distribution $π$, and the outlier sequence contains $n$ i.i.d. observations from an outlier distribution $μ$, which is distinct from $π$, but otherwise unknown. A universal test based on KL divergence is built to approximate the maximum likelihood test, with known $π$ and unknown $μ$. A data-dependent partitions based KL divergence estimator is employed. Such a KL divergence estimator is further shown to converge to its true value exponentially fast when the density ratio satisfies $0<K_1\leq \frac{dμ}{dπ}\leq K_2$, where $K_1$ and $K_2$ are positive constants, and this further implies that the test is exponentially consistent. The performance of the test is compared with that of a recently introduced test for this problem based on the machine learning approach of maximum mean discrepancy (MMD). We identify regimes in which the KL divergence based test is better than the MMD based test.

cs.IT↗

arXiv2015

Adaptive Sequential Optimization with Applications to Machine Learning

A framework is introduced for solving a sequence of slowly changing optimization problems, including those arising in regression and classification applications, using optimization algorithms such as stochastic gradient descent (SGD). The optimization problems change slowly in the sense that the minimizers change at either a fixed or bounded rate. A method based on estimates of the change in the minimizers and properties of the optimization algorithm is introduced for adaptively selecting the number of samples needed from the distributions underlying each problem in order to ensure that the excess risk, i.e., the expected gap between the loss achieved by the approximate minimizer produced by the optimization algorithm and the exact minimizer, does not exceed a target level. Experiments with synthetic and real data are used to confirm that this approach performs well.

cs.LG↗