In this letter, we address the problem of controlling energy storage systems (ESSs) for arbitrage in real-time electricity markets under price uncertainty. We first formulate this problem as a Markov decision process, and then develop a deep reinforcement learning based algorithm to learn a stochastic control policy that maps a set of available information processed by a recurrent neural network to ESSs' charging/discharging actions. Finally, we verify the effectiveness of our algorithm using real-time electricity prices from PJM.

Navigating urban environments represents a complex task for automated vehicles. They must reach their goal safely and efficiently while considering a multitude of traffic participants. We propose a modular decision making algorithm to autonomously navigate intersections, addressing challenges of existing rule-based and reinforcement learning (RL) approaches. We first present a safe RL algorithm relying on a model-checker to ensure safety guarantees. To make the decision strategy robust to perception errors and occlusions, we introduce a belief update technique using a learning based approach. Finally, we use a scene decomposition approach to scale our algorithm to environments with multiple traffic participants. We empirically demonstrate that our algorithm outperforms rule-based methods and reinforcement learning techniques on a complex intersection scenario.

Rather than proposing a new method, this paper investigates an issue present in existing learning algorithms. We study the learning dynamics of reinforcement learning (RL), specifically a characteristic coupling between learning and data generation that arises because RL agents control their future data distribution. In the presence of function approximation, this coupling can lead to a problematic type of 'ray interference', characterized by learning dynamics that sequentially traverse a number of performance plateaus, effectively constraining the agent to learn one thing at a time even when learning in parallel is better. We establish the conditions under which ray interference occurs, show its relation to saddle points and obtain the exact learning dynamics in a restricted setting. We characterize a number of its properties and discuss possible remedies.

Assembling the compound targets is an open problem for achieving good learning performance in Reinforcement Learning (RL). TD(λ), which uses a single parameter controlled geometric sequence as the weights of the n-step returns, stands out from the sea of compound update methods for its efficient incremental updates and its interesting mathematical properties. Empirical studies show that different λ's yield different performance. Furthermore, it is expected that adapting λ appropriately during the learning boosts performance in terms of convergence speed and accuracy. The goal of this paper is to find a method that optimizes the overall target error for all the states. We first derive a new state meta-objective for optimizing the bias-variance tradeoff and show that the meta-objective proposed in an existing work [1] is actually a special case of the newly proposed objective. Then, we propose a trust-region style method to tackle the difficulties of optimizing the meta-objective and prove its equivalence to optimizing the overall target error, given appropriate assumptions. In experiments, we observe that the proposed method MTA has generally significantly improved empirical performance over the existing method and baselines.

Learning to take actions based on observations is a core requirement for artificial agents to be able to be successful and robust at their task. Reinforcement Learning (RL) is a well-known technique for learning such policies. However, current RL algorithms often have to deal with reward shaping, have difficulties generalizing to other environments and are most often sample inefficient. In this paper, we explore active inference and the free energy principle, a normative theory from neuroscience that explains how self-organizing biological systems operate by maintaining a model of the world and casting action selection as an inference problem. We apply this concept to a typical problem known to the RL community, the mountain car problem, and show how active inference encompasses both RL and learning from demonstrations.

Machine learning has been widely applied to various applications, some of which involve training with privacy-sensitive data. A modest number of data breaches have been studied, including credit card information in natural language data and identities from face dataset. However, most of these studies focus on supervised learning models. As deep reinforcement learning (DRL) has been deployed in a number of real-world systems, such as indoor robot navigation, whether trained DRL policies can leak private information requires in-depth study. To explore such privacy breaches in general, we mainly propose two methods: environment dynamics search via genetic algorithm and candidate inference based on shadow policies. We conduct extensive experiments to demonstrate such privacy vulnerabilities in DRL under various settings. We leverage the proposed algorithms to infer floor plans from some trained Grid World navigation DRL agents with LiDAR perception. The proposed algorithm can correctly infer most of the floor plans and reaches an average recovery rate of 95.83% using policy gradient trained agents. In addition, we are able to recover the robot configuration in continuous control environments and an autonomous driving simulator with high accuracy. To the best of our knowledge, this is the first work to investigate privacy leakage in DRL settings and we show that DRL-based agents do potentially leak privacy-sensitive information from the trained policies.

The use of target networks has been a popular and key component of recent deep Q-learning algorithms for reinforcement learning, yet little is known from the theory side. In this work, we introduce a new family of target-based temporal difference (TD) learning algorithms and provide theoretical analysis on their convergences. In contrast to the standard TD-learning, target-based TD algorithms maintain two separate learning parameters-the target variable and online variable. Particularly, we introduce three members in the family, called the averaging TD, double TD, and periodic TD, where the target variable is updated through an averaging, symmetric, or periodic fashion, mirroring those techniques used in deep Q-learning practice. We establish asymptotic convergence analyses for both averaging TD and double TD and a finite sample analysis for periodic TD. In addition, we also provide some simulation results showing potentially superior convergence of these target-based TD algorithms compared to the standard TD-learning. While this work focuses on linear function approximation and policy evaluation setting, we consider this as a meaningful step towards the theoretical understanding of deep Q-learning variants with target networks.

Machine learning can help us in solving problems in the context big data analysis and classification, as well as in playing complex games such as Go. But can it also be used to find novel protocols and algorithms for applications such as large-scale quantum communication? Here we show that machine learning can be used to identify central quantum protocols, including teleportation, entanglement purification and the quantum repeater. These schemes are of importance in long-distance quantum communication, and their discovery has shaped the field of quantum information processing. However, the usefulness of learning agents goes beyond the mere re-production of known protocols; the same approach allows one to find improved solutions to long-distance communication problems, in particular when dealing with asymmetric situations where channel noise and segment distance are non-uniform. Our findings are based on the use of projective simulation, a model of a learning agent that combines reinforcement learning and decision making in a physically motivated framework. The learning agent is provided with a universal gate set, and the desired task is specified via a reward scheme. From a technical perspective, the learning agent has to deal with stochastic environments and reactions. We utilize an idea reminiscent of hierarchical skill acquisition, where solutions to sub-problems are learned and re-used in the overall scheme. This is of particular importance in the development of long-distance communication schemes, and opens the way for using machine learning in the design and implementation of quantum networks.

Opioids are the preferred medications for the treatment of pain in the intensive care unit. While undertreatment leads to unrelieved pain and poor clinical outcomes, excessive use of opioids puts patients at risk of experiencing multiple adverse effects. In this work, we present a sequential decision making framework for opioid dosing based on deep reinforcement learning. It provides real-time clinically interpretable dosing recommendations, personalized according to each patient's evolving pain and physiological condition. We focus on morphine, one of the most commonly prescribed opioids. To train and evaluate the model, we used retrospective data from the publicly available MIMIC-3 database. Our results demonstrate that reinforcement learning may be used to aid decision making in the intensive care setting by providing personalized pain management interventions.

In an episodic Markov Decision Process (MDP) problem, an online algorithm chooses from a set of actions in a sequence of $H$ trials, where $H$ is the episode length, in order to maximize the total payoff of the chosen actions. Q-learning, as the most popular model-free reinforcement learning (RL) algorithm, directly parameterizes and updates value functions without explicitly modeling the environment. Recently, [Jin et al. 2018] studies the sample complexity of Q-learning with finite states and actions. Their algorithm achieves nearly optimal regret, which shows that Q-learning can be made sample efficient. However, MDPs with large discrete states and actions [Silver et al. 2016] or continuous spaces [Mnih et al. 2013] cannot learn efficiently in this way. Hence, it is critical to develop new algorithms to solve this dilemma with provable guarantee on the sample complexity. With this motivation, we propose a novel algorithm that works for MDPs with a more general setting, which has infinitely many states and actions and assumes that the payoff function and transition kernel are Lipschitz continuous. We also provide corresponding theory justification for our algorithm. It achieves the regret $\tilde{\mathcal{O}}(K^{\frac{d+1}{d+2}}\sqrt{H^3}),$ where $K$ denotes the number of episodes and $d$ denotes the dimension of the joint space. To the best of our knowledge, this is the first analysis in the model-free setting whose established regret matches the lower bound up to a logarithmic factor.