Learn more about what comes next. As game worlds grow more vast and complex, making sure they are playable and bug-free is becoming increasingly difficult for developers. And gaming companies are looking for new tools, including artificial intelligence, to help overcome the mounting challenge of testing their products. A new paper by a group of AI researchers at Electronic Arts shows that deep reinforcement learning agents can help test games and make sure they are balanced and solvable. "Adversarial Reinforcement Learning for Procedural Content Generation," the technique presented by the EA researchers, is a novel approach that addresses some of the shortcomings of previous AI methods for testing games.

Interactive reinforcement learning, where humans actively assist during an agent's learning process, has the promise to alleviate the sample complexity challenges of practical algorithms. However, the inner workings and state of the robot are typically hidden from the teacher when humans provide feedback. To create a common ground between the human and the learning robot, in this paper, we propose an Augmented Reality (AR) system that reveals the hidden state of the learning to the human users. This paper describes our system's design and implementation and concludes with a discussion on two directions for future work which we are pursuing: 1) use of our system in AI education activities at the K-12 level; and 2) development of a framework for an AR-based human-in-the-loop reinforcement learning, where the human teacher can see sensory and cognitive representations of the robot overlaid in the real world.

The goal of continuous control is to synthesize desired behaviors. In reinforcement learning (RL)-driven approaches, this is often accomplished through careful task reward engineering for efficient exploration and running an off-the-shelf RL algorithm. While reward maximization is at the core of RL, reward engineering is not the only -- sometimes nor the easiest -- way for specifying complex behaviors. In this paper, we introduce \braxlines, a toolkit for fast and interactive RL-driven behavior generation beyond simple reward maximization that includes Composer, a programmatic API for generating continuous control environments, and set of stable and well-tested baselines for two families of algorithms -- mutual information maximization (MiMax) and divergence minimization (DMin) -- supporting unsupervised skill learning and distribution sketching as other modes of behavior specification. In addition, we discuss how to standardize metrics for evaluating these algorithms, which can no longer rely on simple reward maximization. Our implementations build on a hardware-accelerated Brax simulator in Jax with minimal modifications, enabling behavior synthesis within minutes of training. We hope Braxlines can serve as an interactive toolkit for rapid creation and testing of environments and behaviors, empowering explosions of future benchmark designs and new modes of RL-driven behavior generation and their algorithmic research.

Reinforcement learning has been successful in many tasks ranging from robotic control, games, energy management etc. In complex real world environments with sparse rewards and long task horizons, sample efficiency is still a major challenge. Most complex tasks can be easily decomposed into high-level planning and low level control. Therefore, it is important to enable agents to leverage the hierarchical structure and decompose bigger tasks into multiple smaller sub-tasks. We introduce an approach where we use language to specify sub-tasks and a high-level planner issues language commands to a low level controller. The low-level controller executes the sub-tasks based on the language commands. Our experiments show that this method is able to solve complex long horizon planning tasks with limited human supervision. Using language has added benefit of interpretability and ability for expert humans to take over the high-level planning task and provide language commands if necessary.

In recent years, the growing demand for more intelligent service robots is pushing the development of mobile robot navigation algorithms to allow safe and efficient operation in a dense crowd. Reinforcement learning (RL) approaches have shown superior ability in solving sequential decision making problems, and recent work has explored its potential to learn navigation polices in a socially compliant manner. However, the expert demonstration data used in existing methods is usually expensive and difficult to obtain. In this work, we consider the task of training an RL agent without employing the demonstration data, to achieve efficient and collision-free navigation in a crowded environment. To address the sparse reward navigation problem, we propose to incorporate the hindsight experience replay (HER) and curriculum learning (CL) techniques with RL to efficiently learn the optimal navigation policy in the dense crowd. The effectiveness of our method is validated in a simulated crowd-robot coexisting environment. The results demonstrate that our method can effectively learn human-aware navigation without requiring additional demonstration data.

This work studies the question of Representation Learning in RL: how can we learn a compact low-dimensional representation such that on top of the representation we can perform RL procedures such as exploration and exploitation, in a sample efficient manner. We focus on the low-rank Markov Decision Processes (MDPs) where the transition dynamics correspond to a low-rank transition matrix. Unlike prior works that assume the representation is known (e.g., linear MDPs), here we need to learn the representation for the low-rank MDP. We study both the online RL and offline RL settings. For the online setting, operating with the same computational oracles used in FLAMBE (Agarwal et.al), the state-of-art algorithm for learning representations in low-rank MDPs, we propose an algorithm REP-UCB Upper Confidence Bound driven Representation learning for RL), which significantly improves the sample complexity from $\widetilde{O}( A^9 d^7 / (\epsilon^{10} (1-\gamma)^{22}))$ for FLAMBE to $\widetilde{O}( A^4 d^4 / (\epsilon^2 (1-\gamma)^{3}) )$ with $d$ being the rank of the transition matrix (or dimension of the ground truth representation), $A$ being the number of actions, and $\gamma$ being the discounted factor. Notably, REP-UCB is simpler than FLAMBE, as it directly balances the interplay between representation learning, exploration, and exploitation, while FLAMBE is an explore-then-commit style approach and has to perform reward-free exploration step-by-step forward in time. For the offline RL setting, we develop an algorithm that leverages pessimism to learn under a partial coverage condition: our algorithm is able to compete against any policy as long as it is covered by the offline distribution.

Humans decompose novel complex tasks into simpler ones to exploit previously learned skills. Analogously, hierarchical reinforcement learning seeks to leverage lower-level policies for simple tasks to solve complex ones. However, because each lower-level policy induces a different distribution of states, transitioning from one lower-level policy to another may fail due to an unexpected starting state. We introduce transition policies that smoothly connect lower-level policies by producing a distribution of states and actions that matches what is expected by the next policy. Training transition policies is challenging because the natural reward signal -- whether the next policy can execute its subtask successfully -- is sparse. By training transition policies via adversarial inverse reinforcement learning to match the distribution of expected states and actions, we avoid relying on task-based reward. To further improve performance, we use deep Q-learning with a binary action space to determine when to switch from a transition policy to the next pre-trained policy, using the success or failure of the next subtask as the reward. Although the reward is still sparse, the problem is less severe due to the simple binary action space. We demonstrate our method on continuous bipedal locomotion and arm manipulation tasks that require diverse skills. We show that it smoothly connects the lower-level policies, achieving higher success rates than previous methods that search for successful trajectories based on a reward function, but do not match the state distribution.

Recently, deep reinforcement learning (RL) has achieved remarkable empirical success by integrating deep neural networks into RL frameworks. However, these algorithms often require a large number of training samples and admit little theoretical understanding. To mitigate these issues, we propose a theoretically principled nearest neighbor (NN) function approximator that can improve the value networks in deep RL methods. Inspired by human similarity judgments, the NN approximator estimates the action values using rollouts on past observations and can provably obtain a small regret bound that depends only on the intrinsic complexity of the environment. We present (1) Nearest Neighbor Actor-Critic (NNAC), an online policy gradient algorithm that demonstrates the practicality of combining function approximation with deep RL, and (2) a plug-and-play NN update module that aids the training of existing deep RL methods. Experiments on classical control and MuJoCo locomotion tasks show that the NN-accelerated agents achieve higher sample efficiency and stability than the baseline agents. Based on its theoretical benefits, we believe that the NN approximator can be further applied to other complex domains to speed-up learning.

Machine learning has successfully framed many sequential decision making problems as either supervised prediction, or optimal decision-making policy identification via reinforcement learning. In data-constrained offline settings, both approaches may fail as they assume fully optimal behavior or rely on exploring alternatives that may not exist. We introduce an inherently different approach that identifies possible ``dead-ends'' of a state space. We focus on the condition of patients in the intensive care unit, where a ``medical dead-end'' indicates that a patient will expire, regardless of all potential future treatment sequences. We postulate ``treatment security'' as avoiding treatments with probability proportional to their chance of leading to dead-ends, present a formal proof, and frame discovery as an RL problem. We then train three independent deep neural models for automated state construction, dead-end discovery and confirmation. Our empirical results discover that dead-ends exist in real clinical data among septic patients, and further reveal gaps between secure treatments and those that were administered.

Over the recent years, vast progress has been made in Offline Reinforcement Learning (Offline-RL) for various decision-making domains: from finance to robotics. However, comparing and reporting new Offline-RL algorithms has been noted as underdeveloped: (1) use of unlimited online evaluation budget for hyperparameter search (2) sidestepping offline policy selection (3) ad-hoc performance statistics reporting. In this work, we propose an evaluation technique addressing these issues, Expected Online Performance, that provides a performance estimate for a best-found policy given a fixed online evaluation budget. Using our approach, we can estimate the number of online evaluations required to surpass a given behavioral policy performance. Applying it to several Offline-RL baselines, we find that with a limited online evaluation budget, (1) Behavioral Cloning constitutes a strong baseline over various expert levels and data regimes, and (2) offline uniform policy selection is competitive with value-based approaches. We hope the proposed technique will make it into the toolsets of Offline-RL practitioners to help them arrive at informed conclusions when deploying RL in real-world systems.