Research map / Reinforcement learning / Offline RL
Fine-tuning Behavioral Cloning Policies with Preference‑Based Reinforcement Learning
Mael Macuglia, Paul Friedrich, Giorgia Ramponi
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 |
|---|---|---|---|
| increase offline expert trajectories from 10 to 1000 (ablation) | BRIDGE · Gridworld (clean demos) | confidence set size (%): 99.9 -> 8.0 | Improved |
| reduce demonstration noise from 20% to 0% at 1000 trajectories (ablation) | BRIDGE · Gridworld | confidence set size (%): 95.4 -> 8.0 | Improved |
| add offline-constrained confidence set vs pure online PbRL | BRIDGE · StarMDP, Gridworld, Reacher, Ant | cumulative regret | Improved |
| shrink confidence radius too far (ablation) | BRIDGE · Reacher | regret | Mixed |
| increase offline trajectories (ablation) | BRIDGE · Reacher | regret | Improved |
| use embedding closer to true reward (reacher_reward) (ablation) | BRIDGE · Reacher | regret | Improved |
Closest papers by meaning
- Should I Run Offline Reinforcement Learning or Behavioral Cloning? (ICLR 2022 · similarity 0.90)
- Preference Elicitation for Offline Reinforcement Learning (ICLR 2025 · similarity 0.88)
- When should we prefer Decision Transformers for Offline Reinforcement Learning? (ICLR 2024 · similarity 0.88)
- Efficient Preference-Based Reinforcement Learning: Randomized Exploration meets Experimental Design (NeurIPS 2025 · similarity 0.88)
- Adversarial Policy Optimization for Offline Preference-based Reinforcement Learning (ICLR 2025 · similarity 0.88)
- Policy-labeled Preference Learning: Is Preference Enough for RLHF? (ICML 2025 · 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.
