Research map / Reinforcement learning / RL from feedback
SHADOW: Dynamic-Aware Credit Assignment Against Long-Horizon Tasks
Yuze Liu, Chaochao Lu, Chao Yang
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 dynamic-inconsistent separation (DIS) of observation-identical states (ablation) | SHADOW · ALFWorld (Qwen2.5-1.5B) | success rate (%): 89.8 -> 96.1 | Improved |
| add GAE local dynamic advantage estimator (ablation) | SHADOW · ALFWorld (Qwen2.5-1.5B) | success rate (%): 93.5 -> 96.1 | Improved |
| add dynamic-consistent aggregation (DCA) (ablation) | SHADOW · ALFWorld (Qwen2.5-1.5B) | success rate (%): 95.1 -> 96.1 | No clear effect |
| replace GiGPO with SHADOW | GiGPO · ALFWorld (Qwen2.5-1.5B) | success rate (%): 86.7 -> 96.1 | Improved |
| replace GiGPO with SHADOW | GiGPO · ALFWorld (Qwen2.5-7B) | success rate (%): 90.1 -> 97.7 | Improved |
| replace GiGPO with SHADOW | GiGPO · WebShop (Qwen2.5-1.5B) | success rate (%): 71.4 -> 76.6 | Improved |
| replace GiGPO with SHADOW | GiGPO · WebShop (Qwen2.5-7B) | success rate (%): 76.3 -> 81.5 | Improved |
| replace GiGPO with SHADOW on Cool subtask | GiGPO · ALFWorld Cool (Qwen2.5-1.5B) | success rate (%): 90.1 -> 88.5 | No clear effect |
Closest papers by meaning
- Latent Reward: LLM-Empowered Credit Assignment in Episodic Reinforcement Learning (AAAI 2025 · similarity 0.88)
- Segment Policy Optimization: Effective Segment-Level Credit Assignment in RL for Large Language Models (NeurIPS 2025 · similarity 0.88)
- Sequence Compression Speeds Up Credit Assignment in Reinforcement Learning (ICML 2024 · similarity 0.88)
- Group-Graph Policy Optimization for Long-Horizon Agentic Reinforcement Learning (NeurIPS 2026 · similarity 0.88)
- Knowing When to Ask: Segment-Level Credit Assignment for LLM Tool Use (NeurIPS 2026 · similarity 0.88)
- Not Only Where, But When: Temporal Scheduling for RLVR (NeurIPS 2026 · 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.
