Abstract Person re‐identification (Re‐ID) is one of the most remarkable research topics that widely applied in our daily lives. For person Re‐ID in bird's eye scenes, traditional computer vision‐based methods used multiple features, for example, texture and color, of a pedestrian's head and shoulders. Those methods are difficult to cope with environments of variety and the change the appearance of different people due to the instability of feature detection. On the other hand, although recent advanced deep learning‐based methods are powerful to extract discriminative features, the requirement of a large amount of annotated training data restricts the appliable tasks. To overcome this problem, in this article, we propose a novel method fusing multiple heterogeneous features through a multi‐feature subspace representation network (MFSRN) to maximize the classification performance while keeping the disparity among features as small as possible, that is, common‐subspace constraints. We conducted comparative experiments with state‐of‐the‐art models on the bird's‐eye view person dataset, and extensive experimental results demonstrated that our proposed MFSRN could achieve better recognition performance. Furthermore, the validity and stability of the method are confirmed.

When applying a point process to a real-world problem, an appropriate intensity function model should be designed based on physical and mathematical prior knowledge. Recently, a fully trainable deep learning-based approach has been developed for temporal point processes. In this approach, a cumulative hazard function (CHF) capable of systematic computation of adaptive intensity function is modeled in a data-driven manner. However, in this approach, although many applications of point processes generate various kinds of information such as location, magnitude, and depth, the mark information of events is not considered. To overcome this limitation, we propose a fully trainable marked point process method for modeling decomposed CHFs for time and mark prediction using multistream deep neural networks. We demonstrate the effectiveness of the proposed method through experiments with synthetic and real-world event data.

As the damage caused by heavy rainfall is becoming more serious, the improvement of precipitation forecasts is highly demanded. For this purpose, arithmetic and Bayesian average-based methods have been proposed to integrate multiple 2D-grid forecasts. However, since a single weight is shared in the entire grid in these methods, local variations of the importance of forecasts could not be taken into account. Besides, although a variety of information is available in precipitation forecast, it would not be straightforwardly to incorporate the additional information in the existing methods. To overcome these problems, we propose an encoder-decoder-based image transformation method that generates a weight image that is optimized in a pixel-wise manner and additional information could be embedded as the channel of input images and feature maps. Through the experiment of precipitation forecast in the period from April 2018 to March 2019 in Japan, we will show that our proposed integration method outperforms existing methods.

When using a point process, a specific form of the model needs to be designed for intensity function, based on physical and mathematical prior knowledge about the data. Recently, a fully trainable deep learning-based approach has been developed for temporal point processes. This approach models a cumulative hazard function (CHF), which is capable of systematic computation of adaptive intensity function in a data-driven manner. However, this approach does not take the attribute information of events into account although many applications of point processes generate with a variety of marked information such as location, magnitude, and depth of seismic activity. To overcome this limitation, we propose a fully trainable marked point process method, modeling decomposed CHFs for time and mark using multi-stream deep neural networks. In addition, we also propose to encode multiple marked information into a single image and extract necessary information adaptively without detailed knowledge about the data. We show the effectiveness of our proposed method through experiments with simulated toy data and real seismic data.

Recently, anchor-based regression methods have been applied to challenging regression problems, e.g., object detection and distance estimation, and greatly improved those performances. The key idea of anchor-based regression is to solve the regression of the residuals between selected anchors and original target variable, where the variance is expected to be smaller. However, similar to an ordinary regression method, the anchor-based regression could face difficulty on a fine-grained regression and ill-posed problems where the residual variables tend to be too small and complicated to accurately predict. To overcome these problems on the anchor-based regression, we propose to introduce an adaptive residual encoding in which the too small residual is magnified, and the too-large residual is truncated using adaptively tuned sigmoidal function. Our proposed method, called ATR-Nets (Adaptive Truncated Residual-Networks) with an end-to-end architecture could control the range of the target residual to be fitted based on the regression performance, Through experiments with toy-data and the system identification for earthquake asperity models, we show the effectiveness of our proposed method.

Estimating the distance of a target object from a single image is a challenging task since a large variation in the object appearance makes the regression of the distance difficult. In this paper, to tackle such the challenge, we propose 2.5D anchors which provide good candidates of distances, based on a perspective camera model. This candidate is expected to relax the difficulty of the regression model since only the residual from the candidate distance needs to be taken into account. We show the effectiveness of our proposed anchors, by comparing with ordinary regression methods, through experiments with Pascal 3D+ TV monitor dataset and Tsukuba challenge task.

For accurate global self-localization with small memory usage, researches for the compression of the laser-scan data have been actively conducted. Main approaches to the compression are to design feature extractor based on human knowledge regarding the specific environment, e.g., office and hallway. However, in real robot navigation tasks such as a security patrol robot, the robot would be applied to a variety of environments and it is expensive if the users need to tune the design at every environment. To alleviate such problem, we propose to extend the state-of-the-art variational auto-encoder (VAE) by introducing the step-edge detector, which detects non-continuous transition emerged frequently at the laser scan data due to the limitation of distance measurement. With our proposed method, called "laserVAE", the feature extractor of the laser scan is automatically tuned given unknown environments. Through experiments with a real self-localization with 2D laser scan, we demonstrate the effectiveness of the proposed method.

Information-maximization clustering learns a probabilistic classifier in an unsupervised manner so that mutual information between feature vectors and cluster assignments is maximized. A notable advantage of this approach is that it involves only continuous optimization of model parameters, which is substantially simpler than discrete optimization of cluster assignments. However, existing methods still involve nonconvex optimization problems, and therefore finding a good local optimal solution is not straightforward in practice. In this letter, we propose an alternative information-maximization clustering method based on a squared-loss variant of mutual information. This novel approach gives a clustering solution analytically in a computationally efficient way via kernel eigenvalue decomposition. Furthermore, we provide a practical model selection procedure that allows us to objectively optimize tuning parameters included in the kernel function. Through experiments, we demonstrate the usefulness of the proposed approach.

We propose squared-loss mutual information regularization (SMIR) for multi-class probabilistic classification, following the information maximization principle. SMIR is convex under mild conditions and thus improves the nonconvexity of mutual information regularization. It offers all of the following four abilities to semi-supervised algorithms: Analytical solution, out-of-sample/multi-class classification, and probabilistic output. Furthermore, novel generalization error bounds are derived. Experiments show SMIR compares favorably with state-of-the-art methods.

Divergence estimators based on direct approximation of density ratios without going through separate approximation of numerator and denominator densities have been successfully applied to machine learning tasks that involve distribution comparison such as outlier detection, transfer learning, and two-sample homogeneity test. However, since density-ratio functions often possess high fluctuation, divergence estimation is a challenging task in practice. In this letter, we use relative divergences for distribution comparison, which involves approximation of relative density ratios. Since relative density ratios are always smoother than corresponding ordinary density ratios, our proposed method is favorable in terms of nonparametric convergence speed. Furthermore, we show that the proposed divergence estimator has asymptotic variance independent of the model complexity under a parametric setup, implying that the proposed estimator hardly overfits even with complex models. Through experiments, we demonstrate the usefulness of the proposed approach.

The policy gradient approach is a flexible and powerful reinforcement learning method particularly for problems with continuous actions such as robot control. A common challenge is how to reduce the variance of policy gradient estimates for reliable policy updates. In this letter, we combine the following three ideas and give a highly effective policy gradient method: (1) policy gradients with parameter-based exploration, a recently proposed policy search method with low variance of gradient estimates; (2) an importance sampling technique, which allows us to reuse previously gathered data in a consistent way; and (3) an optimal baseline, which minimizes the variance of gradient estimates with their unbiasedness being maintained. For the proposed method, we give a theoretical analysis of the variance of gradient estimates and show its usefulness through extensive experiments.

Multi-label classification allows a sample to belong to multiple classes simultaneously, which is often the case in real-world applications such as audio tagging, image annotation, video search, and text mining. In such a multi-label scenario, taking into account correlation between multiple labels can boost the classification accuracy. However, this in turn makes classifier training more challenging because handling multiple labels tends to induce a high-dimensional optimization problem. In this paper, we propose a highly scalable multi-label classifier based on a computationally efficient classification algorithm called the least-squares probabilistic classifier. Through experiments, we show the usefulness of our proposed method.

Divergence estimators based on direct approximation of density-ratios without going through separate approximation of numerator and denominator densities have been successfully applied to machine learning tasks that involve distribution comparison such as outlier detection, transfer learning, and two-sample homogeneity test. However, since density-ratio functions often possess high fluctuation, divergence estimation is still a challenging task in practice. In this paper, we propose to use relative divergences for distribution comparison, which involves approximation of relative density-ratios. Since relative density-ratios are always smoother than corresponding ordinary density-ratios, our proposed method is favorable in terms of the non-parametric convergence speed. Furthermore, we show that the proposed divergence estimator has asymptotic variance independent of the model complexity under a parametric setup, implying that the proposed estimator hardly overfits even with complex models. Through experiments, we demonstrate the usefulness of the proposed approach.

Policy gradient is a useful model-free reinforcement learning approach, but it tends to suffer from instability of gradient estimates. In this paper, we analyze and improve the stability of policy gradient methods. We first prove that the variance of gradient estimates in the PGPE(policy gradients with parameter-based exploration) method is smaller than that of the classical REINFORCE method under a mild assumption. We then derive the optimal baseline for PGPE, which contributes to further reducing the variance. We also theoretically show that PGPE with the optimal baseline is more preferable than REINFORCE with the optimal baseline in terms of the variance of gradient estimates. Finally, we demonstrate the usefulness of the improved PGPE method through experiments.

Direct policy search is a promising reinforcement learning framework, in particular for controlling continuous, high-dimensional systems. Policy search often requires a large number of samples for obtaining a stable policy update estimator, and this is prohibitive when the sampling cost is expensive. In this letter, we extend an expectation-maximization-based policy search method so that previously collected samples can be efficiently reused. The usefulness of the proposed method, reward-weighted regression with sample reuse (R3), is demonstrated through robot learning experiments. (This letter is an extended version of our earlier conference paper: Hachiya, Peters, & Sugiyama, 2009 .).

Off-policy reinforcement learning is aimed at efficiently using data samples gathered from a policy that is different from the currently optimized policy. A common approach is to use importance sampling techniques for compensating for the bias of value function estimators caused by the difference between the data-sampling policy and the target policy. However, existing off-policy methods often do not take the variance of the value function estimators explicitly into account and therefore their performance tends to be unstable. To cope with this problem, we propose using an adaptive importance sampling technique which allows us to actively control the trade-off between bias and variance. We further provide a method for optimally determining the trade-off parameter based on a variant of cross-validation. The usefulness of the proposed approach is demonstrated through simulated swing-up inverted-pendulum problem.

Least-squares policy iteration is a useful reinforcement learning method in robotics due to its computational efficiency. However, it tends to be sensitive to outliers in observed rewards. In this paper, we propose an alternative method that employs the absolute loss for enhancing robustness and reliability. The proposed method is formulated as a linear programming problem which can be solved efficiently by standard optimization software, so the computational advantage is not sacrificed for gaining robustness and reliability. We demonstrate the usefulness of the proposed approach through simulated robot-control tasks.

The least squares approach works efficiently in value function approximation, given appropriate basis functions. Because of its smoothness, the Gaussian kernel is a popular and useful choice as a basis function. However, it does not allow for discontinuity which typically arises in real-world reinforcement learning tasks. In this paper, we propose a new basis function based on geodesic Gaussian kernels, which exploits the non-linear manifold structure induced by the Markov decision processes. The usefulness of the proposed method is successfully demonstrated in a simulated robot arm control and Khepera robot navigation.