Research map / Reinforcement learning / Visual and representation RL
Adjacency Constraint for Efficient Hierarchical Reinforcement Learning
Tianren Zhang, Shangqi Guo, Tian Tan, Xiaolin Hu, Feng Chen
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 k-step adjacency constraint to goal-conditioned HRL (vs Vanilla) | HRAC · Discrete tasks and Ant tasks | episode reward / success rate | Improved |
| use learned adjacency instead of oracle adjacency (ablation) | HRAC · Ablation tasks | learning curve performance | No clear effect |
| use adjacency matrix-based training instead of NoAdj trajectory-pair training (ablation) | HRAC · Ablation tasks | learning curve performance | Improved |
| use differentiable adjacency loss instead of negative-reward penalty (ablation) | HRAC · Ablation tasks | learning curve performance | Improved |
| increase balancing coefficient eta (ablation) | HRAC · Ablation tasks | learning curve performance | Improved |
| add adjacency constraint vs HIRO under additive state noise sigma=0.1 | HRAC · Stochastic Ant Gather/Maze | learning curve performance | Improved |
| use HIRO reward shaping with sparse rewards | HIRO · Key-Chest and Ant Maze Sparse | episode reward | Hurt |
| add hindsight experience replay to low level | HRL-HER · Locomotion tasks | success rate | Hurt |
| use hierarchy (HIRO) vs flat DDPG | HIRO · FetchPush and FetchPickAndPlace | success rate | Improved |
Closest papers by meaning
- Generating Adjacency-Constrained Subgoals in Hierarchical Reinforcement Learning (NeurIPS 2020 · similarity 0.98)
- Sample Complexity of Goal-Conditioned Hierarchical Reinforcement Learning (NeurIPS 2023 · similarity 0.90)
- Hierarchical Reinforcement Learning with Targeted Causal Interventions (ICML 2025 · similarity 0.88)
- Reconciling Spatial and Temporal Abstractions for Goal Representation (ICLR 2024 · similarity 0.88)
- Probabilistic Subgoal Representations for Hierarchical Reinforcement Learning (ICML 2024 · similarity 0.87)
- Constrained Reinforcement Learning Under Model Mismatch (ICML 2024 · 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.
