arXiv subjects

Vahid Tarokh

Publications and source records attributed to Vahid Tarokh.

At least 127 records · Page 7Linked to original sources

arXiv2019

Minimax-optimal decoding of movement goals from local field potentials using complex spectral features

We consider the problem of predicting eye movement goals from local field potentials (LFP) recorded through a multielectrode array in the macaque prefrontal cortex. The monkey is tasked with performing memory-guided saccades to one of eight targets during which LFP activity is recorded and used to train a decoder. Previous reports have mainly relied on the spectral amplitude of the LFPs as a feature in the decoding step to limited success, while neglecting the phase without proper theoretical justification. This paper formulates the problem of decoding eye movement intentions in a statistically optimal framework and uses Gaussian sequence modeling and Pinsker's theorem to generate minimax-optimal estimates of the LFP signals which are later used as features in the decoding step. The approach is shown to act as a low-pass filter and each LFP in the feature space is represented via its complex Fourier coefficients after appropriate shrinking such that higher frequency components are attenuated; this way, the phase information inherently present in the LFP signal is naturally embedded into the feature space. The proposed complex spectrum-based decoder achieves prediction accuracy of up to $94\%$ at superficial electrode depths near the surface of the prefrontal cortex, which marks a significant performance improvement over conventional power spectrum-based decoders.

cs.NE↗

arXiv2019

Estimation of the Evolutionary Spectra with Application to Stationarity Test

In this work, we propose a new inference procedure for understanding non-stationary processes, under the framework of evolutionary spectra developed by Priestley. Among various frameworks of modeling non-stationary processes, the distinguishing feature of the evolutionary spectra is its focus on the physical meaning of frequency. The classical estimate of the evolutionary spectral density is based on a double-window technique consisting of a short-time Fourier transform and a smoothing. However, smoothing is known to suffer from the so-called bias leakage problem. By incorporating Thomson's multitaper method that was originally designed for stationary processes, we propose an improved estimate of the evolutionary spectral density, and analyze its bias/variance/resolution tradeoff. As an application of the new estimate, we further propose a non-parametric rank-based stationarity test, and provide various experimental studies.

stat.ME↗

arXiv2019

SGD Converges to Global Minimum in Deep Learning via Star-convex Path

Stochastic gradient descent (SGD) has been found to be surprisingly effective in training a variety of deep neural networks. However, there is still a lack of understanding on how and why SGD can train these complex networks towards a global minimum. In this study, we establish the convergence of SGD to a global minimum for nonconvex optimization problems that are commonly encountered in neural network training. Our argument exploits the following two important properties: 1) the training loss can achieve zero value (approximately), which has been widely observed in deep learning; 2) SGD follows a star-convex path, which is verified by various experiments in this paper. In such a context, our analysis shows that SGD, although has long been considered as a randomized algorithm, converges in an intrinsically deterministic manner to a global minimum.

cs.LG↗

arXiv2018

Stationary Geometric Graphical Model Selection

We consider the problem of model selection in Gaussian Markov fields in the sample deficient scenario. In many practically important cases, the underlying networks are embedded into Euclidean spaces. Using the natural geometric structure, we introduce the notion of spatially stationary distributions over geometric graphs. This directly generalizes the notion of stationary time series to the multidimensional setting lacking time axis. We show that the idea of spatial stationarity leads to a dramatic decrease in the sample complexity of the model selection compared to abstract graphs with the same level of sparsity. For geometric graphs on randomly spread vertices and edges of bounded length, we develop tight information-theoretic bounds on sample complexity and show that a finite number of independent samples is sufficient for a consistent recovery. Finally, we develop an efficient technique capable of reliably and consistently reconstructing graphs with a bounded number of measurements.

stat.ML↗

arXiv2018

Asymptotically Pseudo-Independent Matrices

We show that the family of pseudo-random matrices recently discovered by Soloveychik, Xiang, and Tarokh in their work `Symmetric Pseudo-Random Matrices' exhibits asymptotic independence. More specifically, any two sequences of matrices of matching sizes from that construction generated using sequences of different non-reciprocal primitive polynomials are asymptotically independent.

math.PR↗

arXiv2018

Model Selection Techniques -- An Overview

In the era of big data, analysts usually explore various statistical models or machine learning methods for observed data in order to facilitate scientific discoveries or gain predictive power. Whatever data and fitting procedures are employed, a crucial step is to select the most appropriate model or method from a set of candidates. Model selection is a key ingredient in data analysis for reliable and reproducible statistical inference or prediction, and thus central to scientific studies in fields such as ecology, economics, engineering, finance, political science, biology, and epidemiology. There has been a long history of model selection techniques that arise from researches in statistics, information theory, and signal processing. A considerable number of methods have been proposed, following different philosophies and exhibiting varying performances. The purpose of this article is to bring a comprehensive overview of them, in terms of their motivation, large sample performance, and applicability. We provide integrated and practically relevant discussions on theoretical properties of state-of- the-art model selection approaches. We also share our thoughts on some controversial views on the practice of model selection.

stat.ML↗

arXiv2018

Learning Bounds for Greedy Approximation with Explicit Feature Maps from Multiple Kernels

Nonlinear kernels can be approximated using finite-dimensional feature maps for efficient risk minimization. Due to the inherent trade-off between the dimension of the (mapped) feature space and the approximation accuracy, the key problem is to identify promising (explicit) features leading to a satisfactory out-of-sample performance. In this work, we tackle this problem by efficiently choosing such features from multiple kernels in a greedy fashion. Our method sequentially selects these explicit features from a set of candidate features using a correlation metric. We establish an out-of-sample error bound capturing the trade-off between the error in terms of explicit features (approximation error) and the error due to spectral properties of the best model in the Hilbert space associated to the combined kernel (spectral error). The result verifies that when the (best) underlying data model is sparse enough, i.e., the spectral error is negligible, one can control the test error with a small number of explicit features, that can scale poly-logarithmically with data. Our empirical results show that given a fixed number of explicit features, the method can achieve a lower test error with a smaller time cost, compared to the state-of-the-art in data-dependent random features.

cs.LG↗

arXiv2018

Bayesian model comparison with the Hyvärinen score: computation and consistency

The Bayes factor is a widely used criterion in model comparison and its logarithm is a difference of out-of-sample predictive scores under the logarithmic scoring rule. However, when some of the candidate models involve vague priors on their parameters, the log-Bayes factor features an arbitrary additive constant that hinders its interpretation. As an alternative, we consider model comparison using the Hyvärinen score. We propose a method to consistently estimate this score for parametric models, using sequential Monte Carlo methods. We show that this score can be estimated for models with tractable likelihoods as well as nonlinear non-Gaussian state-space models with intractable likelihoods. We prove the asymptotic consistency of this new model selection criterion under strong regularity assumptions in the case of non-nested models, and we provide qualitative insights for the nested case. We also use existing characterizations of proper scoring rules on discrete spaces to extend the Hyvärinen score to discrete observations. Our numerical illustrations include Lévy-driven stochastic volatility models and diffusion models for population dynamics.

stat.ME↗

arXiv2018

Sequential Detection of Regime Changes in Neural Data

The problem of detecting changes in firing patterns in neural data is studied. The problem is formulated as a quickest change detection problem. Important algorithms from the literature are reviewed. A new algorithmic technique is discussed to detect deviations from learned baseline behavior. The algorithms studied can be applied to both spike and local field potential data. The algorithms are applied to mice spike data to verify the presence of behavioral learning.

eess.SP↗

arXiv2018

Cyclostationary Statistical Models and Algorithms for Anomaly Detection Using Multi-Modal Data

A framework is proposed to detect anomalies in multi-modal data. A deep neural network-based object detector is employed to extract counts of objects and sub-events from the data. A cyclostationary model is proposed to model regular patterns of behavior in the count sequences. The anomaly detection problem is formulated as a problem of detecting deviations from learned cyclostationary behavior. Sequential algorithms are proposed to detect anomalies using the proposed model. The proposed algorithms are shown to be asymptotically efficient in a well-defined sense. The developed algorithms are applied to a multi-modal data consisting of CCTV imagery and social media posts to detect a 5K run in New York City.

eess.SP↗

arXiv2018

Large Deviations of Convex Polyominoes

Enumeration of various types of lattice polygons and in particular polyominoes is of primary importance in many machine learning, pattern recognition, and geometric analysis problems. In this work, we develop a large deviation principle for convex polyominoes under different restrictions, such as fixed area and/or perimeter.

math.PR↗

arXiv2018

Region Detection in Markov Random Fields: Gaussian Case

We consider the problem of model selection in Gaussian Markov fields in the sample deficient scenario. The benchmark information-theoretic results in the case of d-regular graphs require the number of samples to be at least proportional to the logarithm of the number of vertices to allow consistent graph recovery. When the number of samples is less than this amount, reliable detection of all edges is impossible. In many applications, it is more important to learn the distribution of the edge (coupling) parameters over the network than the specific locations of the edges. Assuming that the entire graph can be partitioned into a number of spatial regions with similar edge parameters and reasonably regular boundaries, we develop new information-theoretic sample complexity bounds and show that a bounded number of samples can be sufficient to consistently recover these regions. Finally, we introduce and analyze an efficient region growing algorithm capable of recovering the regions with high accuracy. We show that it is consistent and demonstrate its performance benefits in synthetic simulations.

stat.ML↗

arXiv2018

Sequential Event Detection Using Multimodal Data in Nonstationary Environments

The problem of sequential detection of anomalies in multimodal data is considered. The objective is to observe physical sensor data from CCTV cameras, and social media data from Twitter and Instagram to detect anomalous behaviors or events. Data from each modality is transformed to discrete time count data by using an artificial neural network to obtain counts of objects in CCTV images and by counting the number of tweets or Instagram posts in a geographical area. The anomaly detection problem is then formulated as a problem of quickest detection of changes in count statistics. The quickest detection problem is then solved using the framework of partially observable Markov decision processes (POMDP), and structural results on the optimal policy are obtained. The resulting optimal policy is then applied to real multimodal data collected from New York City around a 5K race to detect the race. The count data both before and after the change is found to be nonstationary in nature. The proposed mathematical approach to this problem provides a framework for event detection in such nonstationary environments and across multiple data modalities.

stat.AP↗

arXiv2018

Pseudo-Wigner Matrices

We consider the problem of generating pseudo-random matrices based on the similarity of their spectra to Wigner's semicircular law. We introduce the notion of an r-independent pseudo-Wigner matrix ensemble and prove closeness of the spectra of its matrices to the semicircular density in the Kolmogorov distance. We give an explicit construction of a family of N by N pseudo-Wigner ensembles using dual BCH codes and show that the Kolmogorov complexity of the obtained matrices is of the order of log(N) bits for a fixed designed Kolmogorov distance precision. We compare our construction to the quasi-random graphs introduced by Chung, Graham and Wilson and demonstrate that the pseudo-Wigner matrices pass stronger randomness tests than the adjacency matrices of these graphs (lifted by the mapping 0 -> 1 and 1 -> -1) do. Finally, we provide numerical simulations verifying our theoretical results.

cs.IT↗

arXiv2018

Symmetric Pseudo-Random Matrices

We consider the problem of generating symmetric pseudo-random sign (+/-1) matrices based on the similarity of their spectra to Wigner's semicircular law. Using binary m-sequences (Golomb sequences) of lengths n=2^m-1, we give a simple explicit construction of circulant n by n sign matrices and show that their spectra converge to the semicircular law when n grows. The Kolmogorov complexity of the proposed matrices equals to that of Golomb sequences and is at most 2log(n) bits.

math.PR↗

arXiv2017

On Data-Dependent Random Features for Improved Generalization in Supervised Learning

The randomized-feature approach has been successfully employed in large-scale kernel approximation and supervised learning. The distribution from which the random features are drawn impacts the number of features required to efficiently perform a learning task. Recently, it has been shown that employing data-dependent randomization improves the performance in terms of the required number of random features. In this paper, we are concerned with the randomized-feature approach in supervised learning for good generalizability. We propose the Energy-based Exploration of Random Features (EERF) algorithm based on a data-dependent score function that explores the set of possible features and exploits the promising regions. We prove that the proposed score function with high probability recovers the spectrum of the best fit within the model class. Our empirical results on several benchmark datasets further verify that our method requires smaller number of random features to achieve a certain generalization error compared to the state-of-the-art while introducing negligible pre-processing overhead. EERF can be implemented in a few lines of code and requires no additional tuning parameters.

stat.ML↗

arXiv2017

Classification of Local Field Potentials using Gaussian Sequence Model

A problem of classification of local field potentials (LFPs), recorded from the prefrontal cortex of a macaque monkey, is considered. An adult macaque monkey is trained to perform a memory-based saccade. The objective is to decode the eye movement goals from the LFP collected during a memory period. The LFP classification problem is modeled as that of classification of smooth functions embedded in Gaussian noise. It is then argued that using minimax function estimators as features would lead to consistent LFP classifiers. The theory of Gaussian sequence models allows us to represent minimax estimators as finite dimensional objects. The LFP classifier resulting from this mathematical endeavor is a spectrum based technique, where Fourier series coefficients of the LFP data, followed by appropriate shrinkage and thresholding, are used as features in a linear discriminant classifier. The classifier is then applied to the LFP data to achieve high decoding accuracy. The function classification approach taken in the paper also provides a systematic justification for using Fourier series, with shrinkage and thresholding, as features for the problem, as opposed to using the power spectrum. It also suggests that phase information is crucial to the decision making.

stat.ME↗

arXiv2017

Wavelet Shrinkage and Thresholding based Robust Classification for Brain Computer Interface

A macaque monkey is trained to perform two different kinds of tasks, memory aided and visually aided. In each task, the monkey saccades to eight possible target locations. A classifier is proposed for direction decoding and task decoding based on local field potentials (LFP) collected from the prefrontal cortex. The LFP time-series data is modeled in a nonparametric regression framework, as a function corrupted by Gaussian noise. It is shown that if the function belongs to Besov bodies, then using the proposed wavelet shrinkage and thresholding based classifier is robust and consistent. The classifier is then applied to the LFP data to achieve high decoding performance. The proposed classifier is also quite general and can be applied for the classification of other types of time-series data as well, not necessarily brain data.

stat.ME↗