Behaviour Suite for Reinforcement Learning

Abstract

This paper introduces the Behaviour Suite for Reinforcement Learning, or bsuite for short. bsuite is a collection of carefully-designed experiments that investigate core capabilities of reinforcement learning (RL) agents with two objectives. First, to collect clear, informative and scalable problems that capture key issues in the design of general and efficient learning algorithms. Second, to study agent behaviour through their performance on these shared benchmarks. To complement this effort, we open source github.com/deepmind/bsuite, which automates evaluation and analysis of any agent on bsuite. This library facilitates reproducible and accessible research on the core issues in RL, and ultimately the design of superior learning algorithms. Our code is Python, and easy to use within existing projects. We include examples with OpenAI Baselines, Dopamine as well as new reference implementations. Going forward, we hope to incorporate more excellent experiments from the research community, and commit to a periodic review of bsuite from a committee of prominent researchers.

References in corpus (8)

• DeepMind Control Suite
• Distributed Prioritized Experience Replay
• Massively Parallel Methods for Deep Reinforcement Learning
• A Collection of Definitions of Intelligence
• Benchmarking Model-Based Reinforcement Learning
• Dopamine: A Research Framework for Deep Reinforcement Learning
• Ensemble Sampling
• A Tutorial on Thompson Sampling

Cited by in corpus (19)

• Leveraging Procedural Generation to Benchmark Reinforcement Learning
• A Minimalist Approach to Offline Reinforcement Learning
• SoftGym: Benchmarking Deep Reinforcement Learning for Deformable Object Manipulation
• Discovering Reinforcement Learning Algorithms
• An empirical investigation of the challenges of real-world reinforcement learning
• CausalWorld: A Robotic Manipulation Benchmark for Causal Structure and Transfer Learning
• On Bonus-Based Exploration Methods in the Arcade Learning Environment
• Making Sense of Reinforcement Learning and Probabilistic Inference
• CORA: Benchmarks, Baselines, and Metrics as a Platform for Continual Reinforcement Learning Agents
• Evolving Neural Networks through a Reverse Encoding Tree
• Reinforcement Learning for Adaptive Mesh Refinement
• Systematic Evaluation of Causal Discovery in Visual Model Based Reinforcement Learning
• Regularized Behavior Value Estimation
• A User's Guide to Calibrating Robotics Simulators
• Visualizing MuZero Models
• ISL: A novel approach for deep exploration
• Reward is enough for convex MDPs
• Parameterized Indexed Value Function for Efficient Exploration in Reinforcement Learning
• A Study of Policy Gradient on a Class of Exactly Solvable Models