Research map / Reinforcement learning / Visual and representation RL
Hierarchical Reinforcement Learning by Discovering Intrinsic Options
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 |
|---|---|---|---|
| learn task-agnostic options with the HIDIO intrinsic worker objective and jointly train the scheduler | HIDIO · Pusher, Reacher, GoalTask, KickBall | success rate | Helped |
| use Action, StateAction, or StateDiff features instead of other discriminator features (ablation) | HIDIO worker · Pusher, Reacher, GoalTask, KickBall | success rate | Helped |
| use Soft rather than Hard worker discounting (ablation) | HIDIO worker · Pusher, Reacher, GoalTask, KickBall | success rate | No clear effect |
| use Soft rather than Hard worker discounting (ablation) | HIDIO worker with State or StateConcat features · Reacher | training stability | Hurt |
| increase option length from K=1 to K=3 (ablation) | HIDIO · Pusher, Reacher, GoalTask, KickBall | success rate | Helped |
| increase option length from K=3 to K=5 (ablation) | HIDIO · GoalTask | success rate | Helped |
| increase option length from K=3 to K=5 (ablation) | HIDIO · Pusher, Reacher, KickBall | success rate | No clear effect |
| jointly train the worker and scheduler instead of pre-training and freezing the worker (ablation) | HIDIO · KickBall | success rate | Helped |
| pre-train the worker and freeze it rather than jointly training (ablation) | HIDIO · Pusher | sample efficiency | Mixed |
Closest papers by meaning
- Direct Preference Optimization for Primitive-Enabled Hierarchical RL: A Bilevel Approach (ICLR 2026 · similarity 0.89)
- Data-Efficient Hierarchical Reinforcement Learning (NeurIPS 2018 · similarity 0.89)
- Sub-policy Adaptation for Hierarchical Reinforcement Learning (ICLR 2020 · similarity 0.89)
- Hierarchical Reinforcement Learning via Advantage-Weighted Information Maximization (ICLR 2019 · similarity 0.89)
- PIPER: Primitive-Informed Preference-based Hierarchical Reinforcement Learning via Hindsight Relabeling (ICML 2024 · similarity 0.87)
- Hierarchical Reinforcement Learning with Targeted Causal Interventions (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.