Research map / Multi-agent systems and games / Human-AI and goal-conditioned
Deep Reinforcement Learning with Stacked Hierarchical Attention for Text-based Games
Yunqiu Xu, Meng Fang, Ling Chen, Yali Du, Joey Tianyi Zhou, Chengqi Zhang
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 sub-graph division and stacked hierarchical attention to KG-A2C | KG-A2C · Jericho sorcerer | raw score: 5.8 -> 29.4 | Improved |
| add sub-graph division and stacked hierarchical attention to KG-A2C | KG-A2C · Jericho detective | raw score: 207.9 -> 308.0 | Improved |
| add sub-graph division and stacked hierarchical attention to KG-A2C | KG-A2C · Jericho ztuu | raw score: 9.2 -> 25.2 | Improved |
| add sub-graph division and stacked hierarchical attention to KG-A2C | KG-A2C · Jericho zork1 | raw score: 34 -> 34.5 | No clear effect |
| add sub-graph division and attention to KG-A2C | KG-A2C · Jericho library | raw score: 14.3 -> 15.8 | Mixed |
| use single attention value per channel instead of grouped attention (remove GroupAttn) (ablation) | SHA-KG · Jericho learning curves | learning curve score | Improved |
| add high-level attention over full KG (ablation) | SHA-KG · Jericho learning curves | learning curve score | Improved |
| add low-level sub-graph division (ablation) | SHA-KG · Jericho learning curves | learning curve score | Improved |
| remove relational/temporal/history awareness sub-graph types (ablation) | SHA-KG · Jericho learning curves | learning curve score | Mixed |
Closest papers by meaning
- Emergent Hierarchical Reasoning in LLMs through Reinforcement Learning (ICLR 2026 · similarity 0.88)
- Prompted Policy Search: Reinforcement Learning through Linguistic and Numerical Reasoning in LLMs (NeurIPS 2025 · similarity 0.87)
- Monte Carlo Planning with Large Language Model for Text-Based Game Agents (ICLR 2025 · similarity 0.87)
- DeepSearch: Overcome the Bottleneck of Reinforcement Learning with Verifiable Rewards via Monte Carlo Tree Search (ICLR 2026 · similarity 0.86)
- $\texttt{Multi}^2$: Hierarchical Multi-Agent Decision-Making with LLM-Based Agents in Interactive Environments (ICML 2026 · similarity 0.86)
- Text-to-Decision Agent: Offline Meta-Reinforcement Learning from Natural Language Supervision (NeurIPS 2025 · similarity 0.86)
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.
