Exploration bonuses derived from the novelty of observations in an environment have become a popular approach to motivate exploration for reinforcement learning (RL) agents in the past few years. Recent methods such as curiosity-driven exploration usually estimate the novelty of new observations by the prediction errors of their system dynamics models. In this paper, we introduce the concept of optical flow estimation from the field of computer vision to the RL domain and utilize the errors from optical flow estimation to evaluate the novelty of new observations. We introduce a flow-based intrinsic curiosity module (FICM) capable of learning the motion features and understanding the observations in a more comprehensive and efficient fashion. We evaluate our method and compare it with a number of baselines on several benchmark environments, including Atari games, Super Mario Bros., and ViZDoom. Our results show that the proposed method is superior to the baselines in certain environments, especially for those featuring sophisticated moving patterns or with high-dimensional observation spaces. We further analyze the hyper-parameters used in the training phase and discuss our insights into them.

Automated theorem proving in large theories can be learned via reinforcement learning over an indefinitely growing action space. In order to select actions, one performs nearest neighbor lookups in the knowledge base to find premises to be applied. Here we address the exploration for reinforcement learning in this space. Approaches (like epsilon-greedy strategy) that sample actions uniformly do not scale to this scenario as most actions lead to dead ends and unsuccessful proofs which are not useful for training our models. In this paper, we compare approaches that select premises using randomly initialized similarity measures and mixing them with the proposals of the learned model. We evaluate these on the HOList benchmark for tactics based higher order theorem proving. We implement an automated theorem prover named DeepHOL-Zero that does not use any of the human proofs and show that our improved exploration method manages to expand the training set continuously. DeepHOL-Zero outperforms the best theorem prover trained by imitation learning alone.

Temporal-difference learning (TD), coupled with neural networks, is among the most fundamental building blocks of deep reinforcement learning. However, due to the nonlinearity in value function approximation, such a coupling leads to nonconvexity and even divergence in optimization. As a result, the global convergence of neural TD remains unclear. In this paper, we prove for the first time that neural TD converges at a sublinear rate to the global optimum of the mean-squared projected Bellman error for policy evaluation. In particular, we show how such global convergence is enabled by the overparametrization of neural networks, which also plays a vital role in the empirical success of neural TD. Beyond policy evaluation, we establish the global convergence of neural (soft) Q-learning, which is further connected to that of policy gradient algorithms.

This course is going to show you a few different ways: including the powerful A2C (Advantage Actor-Critic) algorithm, the DDPG (Deep Deterministic Policy Gradient) algorithm, and evolution strategies. Evolution strategies is a new and fresh take on reinforcement learning, that kind of throws away all the old theory in favor of a more "black box" approach, inspired by biological evolution.

Search, analyze, and visualize big data on a cluster with Elasticsearch, Logstash, Beats, Kibana, and more. Elasticsearch 7 is a powerful tool not only for powering search on big websites, but also for analyzing big data sets in a matter of milliseconds! It's an increasingly popular technology, and a valuable skill to have in today's job market. This comprehensive course covers it all, from installation to operations, with over 90 lectures including 8 hours of video.

We identify a fundamental problem in policy gradient-based methods in continuous control. As policy gradient methods require the agent's underlying probability distribution, they limit policy representation to parametric distribution classes. We show that optimizing over such sets results in local movement in the action space and thus convergence to sub-optimal solutions. We suggest a novel distributional framework, able to represent arbitrary distribution functions over the continuous action space. Using this framework, we construct a generative scheme, trained using an off-policy actor-critic paradigm, which we call the Generative Actor Critic (GAC). Compared to policy gradient methods, GAC does not require knowledge of the underlying probability distribution, thereby overcoming these limitations. Empirical evaluation shows that our approach is comparable and often surpasses current state-of-the-art baselines in continuous domains.

Reinforcement learning is an appropriate and successful method to robustly perform low-level robot control under noisy conditions. Symbolic action planning is useful to resolve causal dependencies and to break a causally complex problem down into a sequence of simpler high-level actions. A problem with the integration of both approaches is that action planning is based on discrete high-level action- and state spaces, whereas reinforcement learning is usually driven by a continuous reward function. However, recent advances in reinforcement learning, specifically, universal value function approximators and hindsight experience replay, have focused on goal-independent methods based on sparse rewards. In this article, we build on these novel methods to facilitate the integration of action planning with reinforcement learning by exploiting the reward-sparsity as a bridge between the high-level and low-level state- and control spaces. As a result, we demonstrate that the integrated neuro-symbolic method is able to solve object manipulation problems that involve tool use and non-trivial causal dependencies under noisy conditions, exploiting both data and knowledge.

A common strategy to deal with the expensive reinforcement learning (RL) of complex tasks is to decompose them into a collection of subtasks that are usually simpler to learn as well as reusable for new problems. However, when a robot learns the policies for these subtasks, common approaches treat every policy learning process separately. Therefore, all these individual (composable) policies need to be learned before tackling the learning process of the complex task through policies composition. Such composition of individual policies is usually performed sequentially, which is not suitable for tasks that require to perform the subtasks concurrently. In this paper, we propose to combine a set of composable Gaussian policies corresponding to these subtasks using a set of activation vectors, resulting in a complex Gaussian policy that is a function of the means and covariances matrices of the composable policies. Moreover, we propose an algorithm for learning both compound and composable policies within the same learning process by exploiting the off-policy data generated from the compound policy. The algorithm is built on a maximum entropy RL approach to favor exploration during the learning process. The results of the experiments show that the experience collected with the compound policy permits not only to solve the complex task but also to obtain useful composable policies that successfully perform in their respective tasks. Supplementary videos and code are available at https://sites.google.com/view/hrl-concurrent-discovery .

We develop model free PAC performance guarantees for multiple concurrent MDPs, extending recent works where a single learner interacts with multiple non-interacting agents in a noise free environment. Our framework allows noisy and resource limited communication between agents, and develops novel PAC guarantees in this extended setting. By allowing communication between the agents themselves, we suggest improved PAC-exploration algorithms that can overcome the communication noise and lead to improved sample complexity bounds. We provide a theoretically motivated algorithm that optimally combines information from the resource limited agents, thereby analyzing the interaction between noise and communication constraints that are ubiquitous in real-world systems. We present empirical results for a simple task that supports our theoretical formulations and improve upon naive information fusion methods.

Despite recent successes in Reinforcement Learning, value-based methods often suffer from high variance hindering performance. In this paper, we illustrate this in a continuous control setting where state of the art methods perform poorly whenever sensor noise is introduced. To overcome this issue, we introduce Recurrent Value Functions (RVFs) as an alternative to estimate the value function of a state. We propose to estimate the value function of the current state using the value function of past states visited along the trajectory. Due to the nature of their formulation, RVFs have a natural way of learning an emphasis function that selectively emphasizes important states. First, we establish RVF's asymptotic convergence properties in tabular settings. We then demonstrate their robustness on a partially observable domain and continuous control tasks. Finally, we provide a qualitative interpretation of the learned emphasis function.