Research map / Reinforcement learning / Visual and representation RL
Causality-driven Hierarchical Structure Discovery for Reinforcement Learning
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 causality-driven hierarchy discovery instead of randomness-driven discovery | CDHRL · 2D-Minecraft and Eden | task performance and learning speed | Helped |
| Use causality-driven exploration rather than random-action exploration | causality discovery in 2D-Minecraft · 2D-Minecraft | causal graph accuracy (SHD and SID) | Helped |
| Add subgoal-policy-assisted intervention sampling | causality discovery in 2D-Minecraft · 2D-Minecraft | SID | Helped |
| Order subgoals using discovered causal dependencies | CDHRL · Eden | subgoal learning speed and stability | Helped |
| Use causal dependencies to reduce the subgoal policy action space | CDHRL · Eden | subgoal training efficiency | Helped |
| Use a partially discovered causal graph before convergence (ablation) | CDHRL · 2D-Minecraft | exploration capability | Helped |
| Remove 10% of effective environment variables (ablation) | CDHRL · 2D-Minecraft and Eden | task performance | No clear effect |
| Allow noisy environment variables in the input (ablation) | CDHRL · 2D-Minecraft and Eden | exploration efficiency | No clear effect |
Closest papers by meaning
- Hierarchical Reinforcement Learning with Targeted Causal Interventions (ICML 2025 · similarity 0.93)
- On Efficiency in Hierarchical Reinforcement Learning (NeurIPS 2020 · similarity 0.88)
- Causality-Aware Efficient Exploration for Cooperative Multi-Agent Reinforcement Learning (AAAI 2026 · similarity 0.88)
- Enhancing Exploration and Exploitation in Hierarchical Reinforcement Learning with Subgoal Graph Learning (AAAI 2026 · similarity 0.88)
- Structural Information-based Hierarchical Diffusion for Offline Reinforcement Learning (NeurIPS 2025 · similarity 0.87)
- Hierarchical Reinforcement Learning with Uncertainty-Guided Diffusional Subgoals (ICML 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.