Research map / Reinforcement learning / Visual and representation RL
Hierarchical Reinforcement Learning with Advantage-Based Auxiliary Rewards
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 high-level advantage as auxiliary reward to adapt low-level skills | HAAR · Ant Maze | success rate | Helped |
| Use high-level advantage as auxiliary reward to adapt low-level skills | HAAR · Swimmer Maze | success rate | Helped |
| Use high-level advantage as auxiliary reward to adapt low-level skills | HAAR · Ant Gather | average return | Helped |
| Use skill-length annealing (ablation) | HAAR · Ant Maze, Swimmer Maze, and Ant Gather | learning speed | Helped |
| Use skill-length annealing instead of keeping skill length fixed (ablation) | HAAR · Ant Maze, Swimmer Maze, and Ant Gather | final outcome | No clear effect |
| Optimize high- and low-level policies concurrently on one batch rather than alternately (ablation) | HAAR · Ant Maze | training speed | Helped |
| Optimize high- and low-level policies concurrently on one batch rather than alternately (ablation) | HAAR · Ant Maze | final success rate | No clear effect |
| Initialize low-level skills with random policies instead of pre-trained skills (ablation) | HAAR · Ant Maze | performance | Hurt |
| Transfer both high- and low-level policies from the source task (ablation) | HAAR · Mirrored maze and spiral maze | learning speed | Helped |
| Transfer only low-level skills from the source task (ablation) | HAAR · Mirrored maze and spiral maze | learning speed | Helped |
Closest papers by meaning
- Hierarchical Reinforcement Learning via Advantage-Weighted Information Maximization (ICLR 2019 · similarity 0.92)
- Sub-policy Adaptation for Hierarchical Reinforcement Learning (ICLR 2020 · similarity 0.91)
- On Efficiency in Hierarchical Reinforcement Learning (NeurIPS 2020 · similarity 0.90)
- Data-Efficient Hierarchical Reinforcement Learning (NeurIPS 2018 · similarity 0.89)
- Deep Reinforcement Learning from Hierarchical Preference Design (ICML 2025 · similarity 0.88)
- Hierarchical Reinforcement Learning with Targeted Causal Interventions (ICML 2025 · similarity 0.88)
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.