RobustRLlib
An algorithm-centric open-source library and benchmark platform for robust reinforcement learning. It integrates multiple robust online, offline and safe RL algorithms, and provides a unified interface and diverse evaluation environments.
Timeline (last 90 days)
RobustRLlib integrates 17 representative robust online, offline and safe reinforcement learning algorithms under a unified interface.
RobustRLlib is an open-source, user-friendly platform providing a reproducible and extensible codebase.
RobustRLlib's evaluation platform covers diverse real-world shift scenarios and includes humanoid robot and vision-language-action (VLA) tasks.
RobustRLlib conducts a systematic evaluation of existing robust RL algorithms, covering 273 perturbation configurations.
RobustRLlib is positioned not only as a benchmarking tool but as the starting point of an open ecosystem.
All Facts (5)
RobustRLlib integrates 17 representative robust online, offline and safe reinforcement learning algorithms under a unified interface.
50%UnverifiedRobustRLlib is an open-source, user-friendly platform providing a reproducible and extensible codebase.
50%UnverifiedRobustRLlib's evaluation platform covers diverse real-world shift scenarios and includes humanoid robot and vision-language-action (VLA) tasks.
50%UnverifiedRobustRLlib conducts a systematic evaluation of existing robust RL algorithms, covering 273 perturbation configurations.
50%UnverifiedRobustRLlib is positioned not only as a benchmarking tool but as the starting point of an open ecosystem.
50%