Research map / Large language models / Agents, code and math
A General Highly Accurate Online Planning Method Integrating Large Language Models into Nested Rollout Policy Adaptation for Dialogue Tasks
Hui Wang, Fafa Zhang, Xiaoyu Zhang, Chaoxu Mu
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 nesting level from 1 to 2 (ablation) | NRPA-GD (ChatGPT) · P4G static | win rate vs ChatGPT (%): 68.61% -> 77.49% | Improved |
| increase nesting level from 1 to 2 (ablation) | NRPA-GD (ChatGPT) · ESConv static | win rate vs ChatGPT (%): 63.40% -> 74.50% | Improved |
| increase nesting level from 1 to 2 (run time cost) (ablation) | NRPA-GD · P4G static | run time: 239 s -> 1039 s | Hurt |
| replace GDP-Zero MCTS with NRPA-GD Level 1 | GDP-Zero · P4G static | win rate vs ChatGPT (%): 58.66% -> 68.61% | Improved |
| replace GDP-Zero MCTS with NRPA-GD Level 1 (run time) | GDP-Zero · P4G static | run time: 636 s -> 239 s | Improved |
| replace DPDP (System 1) with NRPA-GD Level 1 | DPDP · CIMA | AT: 2.24 -> 1.08 | Improved |
| replace DPDP (System 1) with NRPA-GD Level 1 | DPDP · CraigslistBargain | SL: 0.4108 -> 0.6371 | Improved |
| increase nesting level from 1 to 2 (ablation) | NRPA-GD (ChatGPT) · CraigslistBargain | SL: 0.6371 -> 0.5161 | Mixed |
| increase nesting level from 1 to 2 (ablation) | NRPA-GD (GPT-4o-mini) · ESConv | SR: 0.9461 -> 1.0000 | Improved |
| increase nesting level from 1 to 2 (ablation) | NRPA-GD (Llama3.1-8b) · CraigslistBargain | SL: 0.5166 -> 0.4938 | Mixed |
| increase nesting level from 1 to 2 (ablation) | NRPA-GD (Qwen3-0.6b) · CIMA | AT: 1.22 -> 1.08 | Improved |
| use 0.6B Qwen3 backbone with NRPA-GD Level 2 | NRPA-GD · CraigslistBargain | SL: 0.6307 | Improved |
Closest papers by meaning
- Plug-and-Play Policy Planner for Large Language Model Powered Dialogue Agents (ICLR 2024 · similarity 0.91)
- Simulation-Free Hierarchical Latent Policy Planning for Proactive Dialogues (AAAI 2025 · similarity 0.90)
- Planning without Search: Refining Frontier LLMs with Offline Goal-Conditioned RL (NeurIPS 2025 · similarity 0.88)
- Multimodal Policy Internalization for Conversational Agents (ICLR 2026 · similarity 0.87)
- Exploring Conversational Adaptability: Assessing the Proficiency of Large Language Models in Dynamic Alignment with Updated User Intent (AAAI 2025 · similarity 0.87)
- Regressing the Relative Future: Efficient Policy Optimization for Multi-turn RLHF (ICLR 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.
