The online Markov decision process (MDP) is a generalization of the classical Markov decision process that incorporates changing reward functions. In this paper, we propose practical online MDP algorithms with policy iteration and theoretically establish a sublinear regret bound. A notable advantage of the proposed algorithm is that it can be easily combined with function approximation, and thus large and possibly continuous state spaces can be efficiently handled. Through experiments, we demonstrate the usefulness of the proposed algorithm.
We theoretically and experimentally investigate tensor-based regression and classification. Our focus is regularization with various tensor norms, including the overlapped trace norm, the latent trace norm, and the scaled latent trace norm. We first give dual optimization methods using the alternating direction method of multipliers, which is computationally efficient when the number of training samples is moderate. We then theoretically derive an excess risk bound for each tensor norm and clarify their behavior. Finally, we perform extensive experiments using simulated and real data and demonstrate the superiority of tensor-based learning methods over vector- and matrix-based learning methods.
A typical goal of supervised dimension reduction is to find a low-dimensional subspace of the input space such that the projected input variables preserve maximal information about the output variables. The dependence maximization approach solves the supervised dimension reduction problem through maximizing a statistical dependence between projected input variables and output variables. A well-known statistical dependence measure is mutual information (MI) which is based on the Kullback-Leibler (KL) divergence. However, it is known that the KL divergence is sensitive to outliers. On the other hand, quadratic MI (QMI) is a variant of MI based on the $L_2$ distance which is more robust against outliers than the KL divergence, and a computationally efficient method to estimate QMI from data, called least-squares QMI (LSQMI), has been proposed recently. For these reasons, developing a supervised dimension reduction method based on LSQMI seems promising. However, not QMI itself, but the derivative of QMI is needed for subspace search in supervised dimension reduction, and the derivative of an accurate QMI estimator is not necessarily a good estimator of the derivative of QMI. In this paper, we propose to directly estimate the derivative of QMI without estimating QMI itself. We show that the direct estimation of the derivative of QMI is more accurate than the derivative of the estimated QMI. Finally, we develop a supervised dimension reduction algorithm which efficiently uses the proposed derivative estimator, and demonstrate through experiments that the proposed method is more robust against outliers than existing methods.
Log-density gradient estimation is a fundamental statistical problem and possesses various practical applications such as clustering and measuring non-Gaussianity. A naive two-step approach of first estimating the density and then taking its log-gradient is unreliable because an accurate density estimate does not necessarily lead to an accurate log-density gradient estimate. To cope with this problem, a method to directly estimate the log-density gradient without density estimation has been explored, and demonstrated to work much better than the two-step method. The objective of this paper is to further improve the performance of this direct method in multi-dimensional cases. Our idea is to regard the problem of log-density gradient estimation in each dimension as a task, and apply regularized multi-task learning to the direct log-density gradient estimator. We experimentally demonstrate the usefulness of the proposed multi-task method in log-density gradient estimation and mode-seeking clustering.
Task selection (picking an appropriate labeling task) and worker selection (assigning the labeling task to a suitable worker) are two major challenges in task assignment for crowdsourcing. Recently, worker selection has been successfully addressed by the bandit-based task assignment (BBTA) method, while task selection has not been thoroughly investigated yet. In this paper, we experimentally compare several task selection strategies borrowed from active learning literature, and show that the least confidence strategy significantly improves the performance of task assignment in crowdsourcing.
We consider a task assignment problem in crowdsourcing, which is aimed at collecting as many reliable labels as possible within a limited budget. A challenge in this scenario is how to cope with the diversity of tasks and the task-dependent reliability of workers, e.g., a worker may be good at recognizing the name of sports teams, but not be familiar with cosmetics brands. We refer to this practical setting as heterogeneous crowdsourcing. In this paper, we propose a contextual bandit formulation for task assignment in heterogeneous crowdsourcing, which is able to deal with the exploration-exploitation trade-off in worker selection. We also theoretically investigate the regret bounds for the proposed method, and demonstrate its practical usefulness experimentally.
In support vector machine (SVM) applications with unreliable data that contains a portion of outliers, non-robustness of SVMs often causes considerable performance deterioration. Although many approaches for improving the robustness of SVMs have been studied, two major challenges remain in robust SVM learning. First, robust learning algorithms are essentially formulated as non-convex optimization problems. It is thus important to develop a non-convex optimization method for robust SVM that can find a good local optimal solution. The second practical issue is how one can tune the hyperparameter that controls the balance between robustness and efficiency. Unfortunately, due to the non-convexity, robust SVM solutions with slightly different hyper-parameter values can be significantly different, which makes model selection highly unstable. In this paper, we address these two issues simultaneously by introducing a novel homotopy approach to non-convex robust SVM learning. Our basic idea is to introduce parametrized formulations of robust SVM which bridge the standard SVM and fully robust SVM via the parameter that represents the influence of outliers. We characterize the necessary and sufficient conditions of the local optimal solutions of robust SVM, and develop an algorithm that can trace a path of local optimal solutions when the influence of outliers is gradually decreased. An advantage of our homotopy approach is that it can be interpreted as simulated annealing, a common approach for finding a good local optimal solution in non-convex optimization problems. In addition, our homotopy method allows stable and efficient model selection based on the path of local optimal solutions. Empirical performances of the proposed approach are demonstrated through intensive numerical experiments both on robust classification and regression problems.
Metric learning has been shown to be highly effective to improve the performance of nearest neighbor classification. In this paper, we address the problem of metric learning for Symmetric Positive Definite (SPD) matrices such as covariance matrices, which arise in many real-world applications. Naively using standard Mahalanobis metric learning methods under the Euclidean geometry for SPD matrices is not appropriate, because the difference of SPD matrices can be a non-SPD matrix and thus the obtained solution can be uninterpretable. To cope with this problem, we propose to use a properly parameterized LogEuclidean distance and optimize the metric with respect to kernel-target alignment, which is a supervised criterion for kernel learning. Then the resulting non-trivial optimization problem is solved by utilizing the Riemannian geometry. Finally, we experimentally demonstrate the usefulness of our LogEuclidean metric learning algorithm on real-world classification tasks for EEG signals and texture patches.
Estimation of density derivatives is a versatile tool in statistical data analysis. A naive approach is to first estimate the density and then compute its derivative. However, such a two-step approach does not work well because a good density estimator does not necessarily mean a good density-derivative estimator. In this paper, we give a direct method to approximate the density derivative without estimating the density itself. Our proposed estimator allows analytic and computationally efficient approximation of multi-dimensional high-order density derivatives, with the ability that all hyper-parameters can be chosen objectively by cross-validation. We further show that the proposed density-derivative estimator is useful in improving the accuracy of non-parametric KL-divergence estimation via metric learning. The practical superiority of the proposed method is experimentally demonstrated in change detection and feature selection.
In this study, we show that a movement policy can be improved efficiently using the previous experiences of a real robot. Reinforcement Learning (RL) is becoming a popular approach to acquire a nonlinear optimal policy through trial and error. However, it is considered very difficult to apply RL to real robot control since it usually requires many learning trials. Such trials cannot be executed in real environments because unrealistic time is necessary and the real system's durability is limited. Therefore, in this study, instead of executing many learning trials, we propose to use a recently developed RL algorithm, importance-weighted PGPE, by which the robot can efficiently reuse previously sampled data to improve it's policy parameters. We apply importance-weighted PGPE to CB-i, our real humanoid robot, and show that it can learn a target reaching movement and a cart-pole swing up movement in a real environment without using any prior knowledge of the task or any carefully designed initial trajectory.