How Many Random Seeds? Statistical Power Analysis in Deep Reinforcement Learning Experiments

Abstract

Consistently checking the statistical significance of experimental results is one of the mandatory methodological steps to address the so-called "reproducibility crisis" in deep reinforcement learning. In this tutorial paper, we explain how the number of random seeds relates to the probabilities of statistical errors. For both the t-test and the bootstrap confidence interval test, we recall theoretical guidelines to determine the number of random seeds one should use to provide a statistically significant comparison of the performance of two algorithms. Finally, we discuss the influence of deviations from the assumptions usually made by statistical tests. We show that they can lead to inaccurate evaluations of statistical errors and provide guidelines to counter these negative effects. We make our code available to perform the tests.

References in corpus (5)

• Continuous control with deep reinforcement learning
• Scalable trust-region method for deep reinforcement learning using Kronecker-factored approximation
• Parameter Space Noise for Exploration
• Reproducibility of Benchmarked Deep Reinforcement Learning Tasks for Continuous Control
• Simple random search provides a competitive approach to reinforcement learning

Cited by in corpus (22)

• Towards the Systematic Reporting of the Energy and Carbon Footprints of Machine Learning
• Automated Reinforcement Learning (AutoRL): A Survey and Open Problems
• ChainerRL: A Deep Reinforcement Learning Library
• Towards Training Reproducible Deep Learning Models
• Deep Reinforcement Learning at the Edge of the Statistical Precipice
• Measuring the Reliability of Reinforcement Learning Algorithms
• Investigating Generalisation in Continuous Deep Reinforcement Learning
• Deterministic Implementations for Reproducibility in Deep Reinforcement Learning
• Optimistic Reinforcement Learning by Forward Kullback-Leibler Divergence Optimization
• Revisiting Rainbow: Promoting more Insightful and Inclusive Deep Reinforcement Learning Research
• Training Agents using Upside-Down Reinforcement Learning
• How to Make Deep RL Work in Practice
• Catalyst.RL: A Distributed Framework for Reproducible RL Research
• HiER: Highlight Experience Replay for Boosting Off-Policy Reinforcement Learning Agents
• TeachMyAgent: a Benchmark for Automatic Curriculum Learning in Deep RL
• Quantity vs. Quality: On Hyperparameter Optimization for Deep Reinforcement Learning
• A Distance-based Anomaly Detection Framework for Deep Reinforcement Learning
• On the Sensory Commutativity of Action Sequences for Embodied Agents
• DESTA: A Framework for Safe Reinforcement Learning with Markov Games of Intervention
• Control with adaptive Q-learning
• MDP Playground: An Analysis and Debug Testbed for Reinforcement Learning
• Is High Variance Unavoidable in RL? A Case Study in Continuous Control