Research map / Reinforcement learning / Offline RL
Constrained Policy Optimization with Explicit Behavior Density For Offline Reinforcement Learning
Jing Zhang, Chi Zhang, Wenjia Wang, Bingyi Jing
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 |
|---|---|---|---|
| Use CPED instead of SPOT | offline RL · Gym-MuJoCo tasks (Total) | total average normalized score: 773.0 -> 791.7{ $\pm$37.8} | Improved |
| Use CPED instead of IQL | offline RL · AntMaze tasks (Total) | total average normalized score: 392.4 -> 376.8{ $\pm$23.8} | Hurt |
| Use CPED instead of SPOT | offline RL · halfcheetah-medium | average normalized score: 58.4 -> 61.8{ $\pm$1.6} | Improved |
| Use CPED instead of SPOT | offline RL · hopper-medium | average normalized score: 86.0 -> 100.1{ $\pm$2.8} | Improved |
| Use CPED instead of SPOT | offline RL · walker2d-medium | average normalized score: 86.4 -> 90.2{ $\pm$1.7} | Improved |
| Use CPED instead of SPOT | offline RL · halfcheetah-medium-replay | average normalized score: 52.2 -> 55.8{ $\pm$2.9} | Improved |
| Use CPED instead of SPOT | offline RL · hopper-medium-replay | average normalized score: 100.2 -> 98.1{ $\pm$2.1} | Hurt |
| Use CPED instead of SPOT | offline RL · walker2d-medium-replay | average normalized score: 91.6 -> 91.9{ $\pm$0.9} | Improved |
| Use CPED instead of IQL | offline RL · AntMaze umaze-diverse | average normalized score: 65.6 -> 55.6{ $\pm$2.2} | Hurt |
| Use CPED instead of IQL | offline RL · AntMaze medium-play | average normalized score: 76.4 -> 85.1{ $\pm$3.4} | Improved |
| Use CPED instead of IQL | offline RL · AntMaze medium-diverse | average normalized score: 72.8 -> 72.1{ $\pm$2.9} | No clear effect |
| Use CPED instead of IQL | offline RL · AntMaze large-play | average normalized score: 42.0 -> 34.9{ $\pm$5.3} | Hurt |
Closest papers by meaning
- Generative OOD-regularized Model-based Policy Optimization (NeurIPS 2026 · similarity 0.89)
- Constrained Latent Action Policies for Model-Based Offline Reinforcement Learning (NeurIPS 2024 · similarity 0.89)
- Offline Behavior Distillation (NeurIPS 2024 · similarity 0.89)
- Learning from Sparse Offline Datasets via Conservative Density Estimation (ICLR 2024 · similarity 0.89)
- Harnessing Density Ratios for Online Reinforcement Learning (ICLR 2024 · similarity 0.89)
- Iteratively Refined Behavior Regularization for Offline Reinforcement Learning (NeurIPS 2024 · similarity 0.88)
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.
