Research map / Reinforcement learning / Offline RL
Recovering from Out-of-sample States via Inverse Dynamics in Offline 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 |
|---|---|---|---|
| OSR-v vs OSR (value-based actor loss) | OSR · D4RL MuJoCo-v2 average | normalized return: 76.7 -> 78.9 | Helped |
| OSR-v vs OSR | OSR · D4RL AntMaze-v0 average | normalized return: 64.3 -> 63.4 | No clear effect |
| OSR-v vs SDC | SDC · D4RL MuJoCo-v2 average | normalized return: 70.2 -> 78.9 | Helped |
| OSR vs CQL (inverse dynamics state recovery) | CQL · D4RL MuJoCo-v2 average | normalized return: 66.5 -> 76.7 | Helped |
| OSR vs CQL | CQL · D4RL AntMaze-v0 average | normalized return: 50.6 -> 64.3 | Helped |
| OSR vs SDC on AntMaze | SDC · D4RL AntMaze-v0 average | normalized return: 61.2 -> 64.3 | Helped |
| OSR vs SDC on Halfcheetah medium-expert | SDC · Halfcheetah-medium-expert | normalized return: 101.3 -> 94.7 | Hurt |
| OSR vs SDC on Walker2d medium-expert | SDC · Walker2d-medium-expert | normalized return: 105.3 -> 114.3 | Helped |
| OSR vs CQL under large external force | CQL · Hopper-OOS-large | performance decrease (%): 51.0 -> 6.2 | Helped |
| OSR vs SDC under large external force | SDC · Hopper-OOS-large | performance decrease (%): 20.9 -> 6.2 | Helped |
| OSR vs SDC under large external force | SDC · Walker2d-OOS-large | performance decrease (%): 9.2 -> 3.7 | Helped |
| OSR vs SDC under large external force | SDC · Halfcheetah-OOS-large | performance decrease (%): 17.8 -> 3.2 | Helped |
Closest papers by meaning
- State Deviation Correction for Offline Reinforcement Learning (AAAI 2022 · similarity 0.91)
- Beyond OOD State Actions: Supported Cross-Domain Offline Reinforcement Learning (AAAI 2024 · similarity 0.89)
- Dynamic Uncertainty Estimation for Offline Reinforcement Learning (AAAI 2025 · similarity 0.89)
- Hybrid Reinforcement Learning from Offline Observation Alone (ICML 2024 · similarity 0.89)
- Iteratively Refined Behavior Regularization for Offline Reinforcement Learning (NeurIPS 2024 · similarity 0.89)
- ODRL: A Benchmark for Off-Dynamics Reinforcement Learning (NeurIPS 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.