Research map / Reinforcement learning / Visual and representation RL
Action-Sufficient Goal Representations
Jinu Hyeon, Woobin Park, Hongjoon Ahn, Taesup Moon
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 |
|---|---|---|---|
| use actor-based goal representation instead of value-based | HIQL OTA · OGBench cube-double play | success rate: 0.26 -> 0.42 | Improved |
| use actor-based goal representation instead of value-based | HIQL OTA · OGBench cube-triple play | success rate: 0.16 -> 0.41 | Improved |
| use actor-based goal representation instead of value-based | HIQL OTA · OGBench cube-quadruple play | success rate: 0.01 -> 0.28 | Improved |
| use actor-based goal representation instead of value-based | HIQL OTA · OGBench scene play | success rate: 0.56 -> 0.78 | Improved |
| use actor-based goal representation instead of value-based | H-Flow OTA · OGBench cube-triple noisy | success rate: 0.10 -> 0.23 | Improved |
| use actor-based goal representation instead of value-based | H-Flow OTA · OGBench scene play | success rate: 0.69 -> 0.81 | Improved |
| use actor-based goal representation instead of value-based | HIQL OTA · OGBench cube-double noisy | success rate: 0.25 -> 0.29 | Improved |
| use actor-based goal representation instead of value-based | HIQL OTA · OGBench pixel visual-cube | success rate: 0.05 -> 0.53 | Improved |
| use actor-based goal representation instead of value-based | HIQL OTA · OGBench pixel visual-scene | success rate: 0.23 -> 0.53 | Improved |
| add actor-based goal encoder (ablation) | HIQL OTA · OGBench cube-double play | success rate: 0.27 -> 0.42 | Improved |
| add actor-based goal encoder (ablation) | HIQL OTA · OGBench cube-triple noisy | success rate: 0.06 -> 0.19 | Improved |
| add value-based goal encoder (ablation) | HIQL OTA · OGBench cube-double play | success rate: 0.27 -> 0.26 | No clear effect |
Closest papers by meaning
- Dual Goal Representations (ICLR 2026 · similarity 0.90)
- Offline Goal-conditioned Reinforcement Learning with Quasimetric Representations (NeurIPS 2025 · similarity 0.88)
- Flattening Hierarchies with Policy Bootstrapping (NeurIPS 2025 · similarity 0.88)
- Unifying Goal-Conditioned RL and Unsupervised Skill Learning via Control-Maximization (NeurIPS 2026 · similarity 0.87)
- Option-aware Temporally Abstracted Value for Offline Goal-Conditioned Reinforcement Learning (NeurIPS 2025 · similarity 0.87)
- Learning Goal-Conditioned Representations for Language Reward Models (NeurIPS 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.
