Research map / Reinforcement learning / RL from feedback
Master Skill Learning with Policy-Grounded Synergy of LLM-based Reward Shaping and Exploring
Yanbin Chang, Junfan Lin, Jie Jiang, Runhao Zeng, Changxin Huang, Jianqiang Li
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 |
|---|---|---|---|
| add goal-oriented reward Rg (ablation) | PoRSE · TwoCatch | MTS: 0.190 -> 0.349 | Improved |
| add LLM-designed exploration bonus Re (ablation) | PoRSE · BlockStack | MTS: 0.393 -> 0.753 | Improved |
| add dynamic policy fusion inheritance (ablation) | PoRSE · TwoCatch | MTS: 0.017 -> 0.349 | Improved |
| add dynamic policy fusion inheritance (ablation) | PoRSE · FrankaCabinet | MTS: 0.671 -> 0.957 | Improved |
| adapt reward ratio beta (vs fixed) (ablation) | PoRSE · PushBlock | MTS: 0.243 -> 0.318 | Improved |
| use task-aware LLM affordance mapping instead of SimHash (ablation) | PoRSE · 8 Bi-DexHands tasks | MTS: SwingCup 0.454 -> 0.995; DoorOpenInward 0.000 -> 0.250 | Improved |
| use adaptive refined bonus instead of fixed LLMCount bonus (ablation) | PoRSE · TwoCatch | MTS: 0.000 -> 0.349 | Improved |
| use fused Rg+Re versus single reward (ablation) | PoRSE · DoorOpenInward / TwoCatch | success rate: 0% -> 97% (DoorOpenInward), 25% (TwoCatch) | Improved |
| replace LLM-generated AFS with random AFS (ablation) | PoRSE · BlockStack | MTS: 0.753 -> 0.680 | Hurt |
| replace DeepSeek-V3 with GPT-4o-mini (ablation) | PoRSE · BlockStack | MTS: 0.753 -> 0.618 | Hurt |
| add prompt-guided task-irrelevant AFS dimensions (ablation) | PoRSE · FrankaCabinet | MTS: 0.957 -> 0.951 | No clear effect |
| increase bin count from 1000 to 3000 (ablation) | PoRSE · PushBlock | MTS: 0.378 -> 0.321 | No clear effect |
Closest papers by meaning
- Efficient Language-instructed Skill Acquisition via Reward-Policy Co-Evolution (AAAI 2025 · similarity 0.93)
- Text2Reward: Reward Shaping with Language Models for Reinforcement Learning (ICLR 2024 · similarity 0.91)
- Progress Reward Model for Reinforcement Learning via Large Language Models (NeurIPS 2025 · similarity 0.90)
- Eureka: Human-Level Reward Design via Coding Large Language Models (ICLR 2024 · similarity 0.90)
- Learning Reward for Robot Skills Using Large Language Models via Self-Alignment (ICML 2024 · similarity 0.90)
- Multi-Stage Manipulation with Demonstration-Augmented Reward, Policy, and World Model Learning (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.
