Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Saute RL: Almost Surely Safe Reinforcement Learning Using State Augmentation

About

Satisfying safety constraints almost surely (or with probability one) can be critical for the deployment of Reinforcement Learning (RL) in real-life applications. For example, plane landing and take-off should ideally occur with probability one. We address the problem by introducing Safety Augmented (Saute) Markov Decision Processes (MDPs), where the safety constraints are eliminated by augmenting them into the state-space and reshaping the objective. We show that Saute MDP satisfies the Bellman equation and moves us closer to solving Safe RL with constraints satisfied almost surely. We argue that Saute MDP allows viewing the Safe RL problem from a different perspective enabling new features. For instance, our approach has a plug-and-play nature, i.e., any RL algorithm can be "Sauteed". Additionally, state augmentation allows for policy generalization across safety constraints. We finally show that Saute RL algorithms can outperform their state-of-the-art counterparts when constraint satisfaction is of high importance.

Aivar Sootla, Alexander I. Cowen-Rivers, Taher Jafferjee, Ziyan Wang, David Mguni, Jun Wang, Haitham Bou-Ammar• 2022

Related benchmarks

TaskDatasetResultRank
Safety-constrained Reinforcement LearningSafety-Gym SafetyPointCircle1 (evaluation)
Average Reward39.6
15
Velocity ControlSafety Gymnasium Hopper
Return1.66e+3
8
Velocity ControlSafety Gymnasium Swimmer
Return111.6
8
NavigationSafety Gymnasium PointCircle1
Return42.4
8
NavigationSafety Gymnasium PointPush1
Return0.3
8
Velocity ControlSafety Gymnasium Ant
Return2.71e+3
8
Velocity ControlSafety Gymnasium HalfCheetah
Return2.79e+3
8
Safe Reinforcement LearningBullet Safety Gym AntRun (eval environment)
Reward652
8
NavigationSafety Gymnasium PointPush2
Return0.1
8
NavigationSafety Gymnasium CarPush2
Return0.3
8
Showing 10 of 34 rows

Other info

Follow for update