Research map / Reinforcement learning / Policy gradient and actor-critic
Taming Aleatoric Impulse in Off-Policy Reinforcement Learning
Zhouyang Yu, Guojian Zhan, Yang Guan, Jingliang Duan, Letian Tao, Shengbo Eben Li
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| add aleatoric pessimism (beta_init > 0) (ablation) | DSAC-AID · Humanoid-v3 | TAR: removing it degrades performance by 15.2% | Improved |
| add aleatoric pessimism (beta_init > 0) (ablation) | DSAC-AID · DMC Humanoid-run | TAR: removing it degrades performance by 9.7% | Improved |
| add pessimism annealing (l_beta > 0) (ablation) | DSAC-AID · Humanoid-v3 | TAR: drastic performance collapse without it | Improved |
| add pessimism annealing (l_beta > 0) (ablation) | DSAC-AID · DMC Humanoid-run | TAR: removing it drops 10.8% | Improved |
| add optimistic exploration (lambda_UB > 0) (ablation) | DSAC-AID · Humanoid-v3 | TAR: removing it gives a deficit of 11.4% | Improved |
| add optimistic exploration (lambda_UB > 0) (ablation) | DSAC-AID · DMC Humanoid-run | TAR: removing it gives a deficit of 14.8% | Improved |
| vary beta_init and l_beta over several settings (ablation) | DSAC-AID · DMC Humanoid-run | TAR | No clear effect |
| replace baseline with DSAC-AID | BRO (strongest baseline) · AVG. Gym-MuJoCo | normalized TAR: 0.793 -> 0.918 | Improved |
| replace baseline with DSAC-AID | DoubleGum · AVG. DMC Suite | normalized TAR: 0.858 -> 0.926 | Improved |
Closest papers by meaning
- Tactical Optimism and Pessimism for Deep Reinforcement Learning (NeurIPS 2021 · similarity 0.89)
- Efficient Exploration via Epistemic-Risk-Seeking Policy Optimization (ICML 2023 · similarity 0.89)
- Decoupled Policy Actor-Critic: Bridging Pessimism and Risk Awareness in Reinforcement Learning (AAAI 2025 · similarity 0.88)
- Seizing Serendipity: Exploiting the Value of Past Success in Off-Policy Actor-Critic (ICML 2024 · similarity 0.87)
- Compositional Conservatism: A Transductive Approach in Offline Reinforcement Learning (ICLR 2024 · similarity 0.87)
- Selective Uncertainty Propagation in Offline RL (AAAI 2025 · similarity 0.86)
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.
