Research map / Multi-agent systems and games / Human-AI and goal-conditioned
Interactive Semantic Parsing for If-Then Recipes via Hierarchical Reinforcement Learning
Ziyu Yao, Xiujun Li, Jianfeng Gao, Brian Sadler, Huan Sun
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 |
|---|---|---|---|
| Use interactive hierarchical policies instead of one-shot parsing | If-Then recipe parsing · Simulation Eval, All | C+F Acc: 0.374 -> 0.894 | Improved |
| Use interactive hierarchical policies instead of one-shot parsing | If-Then recipe parsing · Human Eval, VI-3/4 | C+F Acc: 0.206 -> 0.634$^*$ | Improved |
| Replace confidence-threshold querying with learned policies | LAM-rule · Simulation Eval, VI-3/4 | C+F Acc: 0.721 -> 0.878$^*$ | Improved |
| Replace confidence-threshold querying with learned policies | LAM-rule · Simulation Eval, VI-3/4 | #Asks: 5.568 -> 2.615$^*$ | Improved |
| Train an interactive policy with supervised examples rather than RL | LAM-sup · Simulation Eval, All | C+F Acc: 0.809 -> 0.894$^* | Hurt |
| Learn the subtask order instead of using fixed order (ablation) | HRL · Simulation Eval, All | C+F Acc: 0.881 -> 0.894$^*$ | Improved |
| Learn the subtask order instead of using fixed order (ablation) | HRL · Simulation Eval, All | #Asks: 2.272 -> 2.069$^* | Improved |
| Learn the subtask order instead of using fixed order (ablation) | HRL · Human Eval, VI-3/4 | C+F Acc: 0.581 -> 0.634$^* | Improved |
| Learn the subtask order instead of using fixed order (ablation) | HRL · Human Eval, VI-3/4 | #Asks: 2.306$^* -> 2.221$^* | Improved |
Closest papers by meaning
- ToolSearcher: Optimizing Tool Selection at Scale via Reinforcement Learning (NeurIPS 2026 · similarity 0.86)
- MA-RLHF: Reinforcement Learning from Human Feedback with Macro Actions (ICLR 2025 · similarity 0.85)
- MARS-SQL: A Multi-Agent Reinforcement Learning Framework For Text-To-SQL (ICML 2026 · similarity 0.85)
- SPELL: Self-Play Reinforcement Learning for Evolving Long-Context Language Models (ICLR 2026 · similarity 0.85)
- Search-R2: Enhancing Search-Integrated Reasoning via Actor-Refiner Collaboration (ICML 2026 · similarity 0.85)
- Joint Optimization of Tool Creation and Use for Large Language Model Agents (NeurIPS 2026 · similarity 0.85)
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.
