Research map / Reinforcement learning / Visual and representation RL
Robust Adversarial Reinforcement Learning
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Train with a learned adversary applying destabilizing forces | TRPO on InvertedPendulum · InvertedPendulum | cumulative reward: 1000\pm0.0 -> 1000\pm0.0 | No clear effect |
| Train with a learned adversary applying destabilizing forces | TRPO on HalfCheetah · HalfCheetah | cumulative reward: 5093\pm44 -> 5444\pm97 | Helped |
| Train with a learned adversary applying destabilizing forces | TRPO on Swimmer · Swimmer | cumulative reward: 358\pm2.4 -> 354\pm1.5 | Hurt |
| Train with a learned adversary applying destabilizing forces | TRPO on Hopper · Hopper | cumulative reward: 3614\pm2.16 -> 3590\pm7.4 | Hurt |
| Train with a learned adversary applying destabilizing forces | TRPO on Walker2D · Walker2D | cumulative reward: 5418\pm87 -> 5854\pm159 | Helped |
| Train against learned adversarial disturbances | RARL protagonist · Adversarial test disturbances on control tasks | cumulative reward | Helped |
| Train against an adversary that learns hard, destabilizing trajectories | RARL protagonist · Different training initializations on HalfCheetah, Swimmer, and Hopper | cumulative reward | Helped |
| Train with learned adversarial forces to model uncertainty | RARL protagonist · Changed test mass | cumulative reward | Helped |
| Train with learned adversarial forces to model uncertainty | RARL protagonist · Changed test friction | cumulative reward | Helped |
| Train with learned adversarial forces to model uncertainty | RARL protagonist · Jointly changed test mass and friction on Hopper and HalfCheetah | cumulative reward | Helped |
Closest papers by meaning
- Robust Adversarial Reinforcement Learning via Bounded Rationality Curricula (ICLR 2024 · similarity 0.91)
- Adversarial Diffusion for Robust Reinforcement Learning (NeurIPS 2025 · similarity 0.91)
- Off-Policy Actor-Critic for Adversarial Observation Robustness: Virtual Alternative Training via Symmetric Policy Evaluation (ICML 2025 · similarity 0.91)
- Beyond Worst-case Attacks: Robust RL with Adaptive Defense via Non-dominated Policies (ICLR 2024 · similarity 0.90)
- Enhancing Robustness in Deep Reinforcement Learning: A Lyapunov Exponent Approach (NeurIPS 2024 · similarity 0.89)
- Towards Optimal Adversarial Robust Q-learning with Bellman Infinity-error (ICML 2024 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.