Research map / Reinforcement learning / Offline RL
Deployment-Efficient Reinforcement Learning via Model-Based Offline Optimization
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 BREMEN instead of the strongest listed baseline | Offline RL · Ant, HalfCheetah, Hopper, and Walker2d with 1M transitions | average cumulative reward | Mixed |
| Use BREMEN instead of the strongest listed baseline | Offline RL · Ant, HalfCheetah, Hopper, and Walker2d with 100K transitions | average cumulative reward | Mixed |
| Use BREMEN instead of the strongest listed baseline | Offline RL · Ant, HalfCheetah, Hopper, and Walker2d with 50K transitions | average cumulative reward | Helped |
| Use behavior-cloned policy initialization with conservative trust-region updates rather than explicit KL value penalty (ablation) | BREMEN · Offline RL, 1M dataset | average cumulative reward | Helped |
| Initialize the policy from behavior cloning rather than no behavior initialization (ablation) | BREMEN · Offline RL, 1M dataset | model error | Helped |
| Use stationary Gaussian noise rather than a learned Gaussian policy (ablation) | BREMEN · HalfCheetah deployment-efficient setting | average cumulative reward | Helped |
| Increase the number of dynamics models in the ensemble, except to K=10 (ablation) | BREMEN · Deployment-efficient and offline settings | average cumulative reward | Helped |
Closest papers by meaning
- A Clean Slate for Offline Reinforcement Learning (NeurIPS 2025 · similarity 0.90)
- Efficient Policy Evaluation with Offline Data Informed Behavior Policy Design (ICML 2024 · similarity 0.90)
- Model Selection for Off-policy Evaluation: New Algorithms and Experimental Protocol (NeurIPS 2025 · similarity 0.89)
- Iteratively Refined Behavior Regularization for Offline Reinforcement Learning (NeurIPS 2024 · similarity 0.89)
- Uni-O4: Unifying Online and Offline Deep Reinforcement Learning with Multi-Step On-Policy Optimization (ICLR 2024 · similarity 0.89)
- OCEAN-MBRL: Offline Conservative Exploration for Model-Based Offline Reinforcement Learning (AAAI 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.