Research map / Reinforcement learning / Offline RL
CORL: Research-oriented Deep Offline Reinforcement Learning Library
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 |
|---|---|---|---|
| use ReBRAC instead of IQL offline | ReBRAC · D4RL total avg | normalized score: 61.92 -> 76.04 | Helped |
| use EDAC/SAC-N (ensemble critics) instead of ReBRAC | EDAC · Gym-MuJoCo | normalized score: 89.74 -> 92.92 | Helped |
| use SAC-N/EDAC on AntMaze | SAC-N, EDAC · AntMaze | normalized score: 78.42 -> 0.00 | Hurt |
| use EDAC on Adroit instead of ReBRAC | EDAC · Adroit | normalized score: 59.39 -> 18.78 | Hurt |
| use CQL with original-paper hyperparameters | CQL · Adroit | normalized score: 1.53 | Hurt |
| online finetuning after offline pretraining for AWAC | AWAC · AntMaze/Adroit cloned | normalized score: 20.01 -> 18.58 | No clear effect |
| online finetuning after offline pretraining for CQL | CQL · AntMaze | normalized score: 48.38 -> 95.58 | Helped |
| online finetuning after offline pretraining for ReBRAC | ReBRAC · AntMaze | normalized score: 79.99 -> 78.87 | No clear effect |
| online finetuning after offline pretraining for ReBRAC | ReBRAC · Adroit cloned | normalized score: 20.33 -> 93.03 | Helped |
| online finetuning after offline pretraining for IQL | IQL · Adroit cloned | normalized score: 21.69 -> 44.99 | Helped |
| replace CQL with Cal-QL for offline-to-online | Cal-QL · AntMaze | normalized score: 95.58 -> 97.33 | No clear effect |
| transfer JAX reference hyperparameters to PyTorch implementation | IQL, CQL · Hopper / Adroit | normalized score | Hurt |
Closest papers by meaning
- A Clean Slate for Offline Reinforcement Learning (NeurIPS 2025 · similarity 0.89)
- Simple Ingredients for Offline Reinforcement Learning (ICML 2024 · similarity 0.88)
- Stealing That Free Lunch: Exposing the Limits of Dyna-Style Reinforcement Learning (ICML 2025 · similarity 0.88)
- Towards an Information Theoretic Framework of Context-Based Offline Meta-Reinforcement Learning (NeurIPS 2024 · similarity 0.88)
- Uni-O4: Unifying Online and Offline Deep Reinforcement Learning with Multi-Step On-Policy Optimization (ICLR 2024 · similarity 0.87)
- Scaling Offline RL via Efficient and Expressive Shortcut Models (NeurIPS 2025 · similarity 0.87)
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.