In many real life problems, objects are described by large number of binary features. For instance, documents are characterized by presence or absence of certain keywords; cancer patients are characterized by presence or absence of certain mutations etc. In such cases, grouping together similar objects/profiles based on such high dimensional binary features is desirable, but challenging. Here, I present a Bayesian non parametric algorithm for clustering high dimensional binary data. It uses a Dirichlet Process (DP) mixture model and simulated annealing to not only cluster binary data, but also find optimal number of clusters in the data. The performance of the algorithm was evaluated and compared with other algorithms using simulated datasets. It outperformed all other clustering methods that were tested in the simulation studies. It was also used to cluster real datasets arising from document analysis, handwritten image analysis and cancer research. It successfully divided a set of documents based on their topics, hand written images based on different styles of writing digits and identified tissue and mutation specificity of chemotherapy treatments.

Nonparametric extension of tensor regression is proposed. Nonlinearity in a high-dimensional tensor space is broken into simple local functions by incorporating low-rank tensor decomposition. Compared to naive nonparametric approaches, our formulation considerably improves the convergence rate of estimation while maintaining consistency with the same function class under specific conditions. To estimate local functions, we develop a Bayesian estimator with the Gaussian process prior. Experimental results show its theoretical properties and high performance in terms of predicting a summary statistic of a real complex network.

We study the problem of distributed multi-task learning with shared representation, where each machine aims to learn a separate, but related, task in an unknown shared low-dimensional subspaces, i.e. when the predictor matrix has low rank. We consider a setting where each task is handled by a different machine, with samples for the task available locally on the machine, and study communication-efficient methods for exploiting the shared structure.

We present the first differentially private algorithms for reinforcement learning, which apply to the task of evaluating a fixed policy. We establish two approaches for achieving differential privacy, provide a theoretical analysis of the privacy and utility of the two algorithms, and show promising results on simple empirical examples.

Multi-instance data, in which each object (bag) contains a collection of instances, are widespread in machine learning, computer vision, bioinformatics, signal processing, and social sciences. We present a maximum entropy (ME) framework for learning from multi-instance data. In this approach each bag is represented as a distribution using the principle of ME. We introduce the concept of confidence-constrained ME (CME) to simultaneously learn the structure of distribution space and infer each distribution. The shared structure underlying each density is used to learn from instances inside each bag. The proposed CME is free of tuning parameters. We devise a fast optimization algorithm capable of handling large scale multi-instance data. In the experimental section, we evaluate the performance of the proposed approach in terms of exact rank recovery in the space of distributions and compare it with the regularized ME approach. Moreover, we compare the performance of CME with Multi-Instance Learning (MIL) state-of-the-art algorithms and show a comparable performance in terms of accuracy with reduced computational complexity.

Standard Gaussian Process (GP) regression, a powerful machine learning tool, is computationally expensive when it is applied to large datasets, and potentially inaccurate when data points are sparsely distributed in a high-dimensional feature space. To address these challenges, a new multiscale, sparsified GP algorithm is formulated, with the goal of application to large scientific computing datasets. In this approach, the data is partitioned into clusters and the cluster centers are used to define a reduced training set, resulting in an improvement over standard GPs in terms of training and evaluation costs. Further, a hierarchical technique is used to adaptively map the local covariance representation to the underlying sparsity of the feature space, leading to improved prediction accuracy when the data distribution is highly non-uniform. A theoretical investigation of the computational complexity of the algorithm is presented. The efficacy of this method is then demonstrated on smooth and discontinuous analytical functions and on data from a direct numerical simulation of turbulent combustion.

Feature selection (FS) has become an indispensable task in dealing with today's highly complex pattern recognition problems with massive number of features. In this study, we propose a new wrapper approach for FS based on binary simultaneous perturbation stochastic approximation (BSPSA). This pseudo-gradient descent stochastic algorithm starts with an initial feature vector and moves toward the optimal feature vector via successive iterations. In each iteration, the current feature vector's individual components are perturbed simultaneously by random offsets from a qualified probability distribution. We present computational experiments on datasets with numbers of features ranging from a few dozens to thousands using three widely-used classifiers as wrappers: nearest neighbor, decision tree, and linear support vector machine. We compare our methodology against the full set of features as well as a binary genetic algorithm and sequential FS methods using cross-validated classification error rate and AUC as the performance criteria. Our results indicate that features selected by BSPSA compare favorably to alternative methods in general and BSPSA can yield superior feature sets for datasets with tens of thousands of features by examining an extremely small fraction of the solution space. We are not aware of any other wrapper FS methods that are computationally feasible with good convergence properties for such large datasets.

We show that gradient descent converges to a local minimizer, almost surely with random initialization. This is proved by applying the Stable Manifold Theorem from dynamical systems theory.

Previously in 2014, we proposed the Nearest Descent (ND) method, capable of generating an efficient Graph, called the in-tree (IT). Due to some beautiful and effective features, this IT structure proves well suited for data clustering. Although there exist some redundant edges in IT, they usually have salient features and thus it is not hard to remove them. Subsequently, in order to prevent the seemingly redundant edges from occurring, we proposed the Nearest Neighbor Descent (NND) by adding the "Neighborhood" constraint on ND. Consequently, clusters automatically emerged, without the additional requirement of removing the redundant edges. However, NND proved still not perfect, since it brought in a new yet worse problem, the "over-partitioning" problem. Now, in this paper, we propose a method, called the Hierarchical Nearest Neighbor Descent (H-NND), which overcomes the over-partitioning problem of NND via using the hierarchical strategy. Specifically, H-NND uses ND to effectively merge the over-segmented sub-graphs or clusters that NND produces. Like ND, H-NND also generates the IT structure, in which the redundant edges once again appear. This seemingly comes back to the situation that ND faces. However, compared with ND, the redundant edges in the IT structure generated by H-NND generally become more salient, thus being much easier and more reliable to be identified even by the simplest edge-removing method which takes the edge length as the only measure. In other words, the IT structure constructed by H-NND becomes more fitted for data clustering. We prove this on several clustering datasets of varying shapes, dimensions and attributes. Besides, compared with ND, H-NND generally takes less computation time to construct the IT data structure for the input data.

We present a methodology for probabilistic load forecasting that is based on lasso (least absolute shrinkage and selection operator) estimation. The model considered can be regarded as a bivariate time-varying threshold autoregressive(AR) process for the hourly electric load and temperature. The joint modeling approach incorporates the temperature effects directly, and reflects daily, weekly, and annual seasonal patterns and public holiday effects. We provide two empirical studies, one based on the probabilistic load forecasting track of the Global Energy Forecasting Competition 2014 (GEFCom2014-L), and the other based on another recent probabilistic load forecasting competition that follows a setup similar to that of GEFCom2014-L. In both empirical case studies, the proposed methodology outperforms two multiple linear regression based benchmarks from among the top eight entries to GEFCom2014-L.