Research map / Reinforcement learning / Visual and representation RL
Unsupervised Skill Discovery via Recurrent Skill Training
Zheyuan Jiang, Jingyue Gao, Jianyu 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 |
|---|---|---|---|
| train skills one after another recurrently with RND-based state coverage reward instead of in parallel | ReST vs DIAYN/DADS · 2D navigation mazes (DoorMaze, CenterMaze, 4RoomMaze, 9RoomMaze) | state coverage | Improved |
| recurrent skill training instead of parallel training | ReST vs DIAYN/DADS · 2D navigation mazes | mutual information I(S;Z) | No clear effect |
| use independent neural networks per skill instead of latent-conditioned network in DIAYN and DADS (ablation) | DIAYN, DADS · 2D navigation mazes | state coverage and mutual information | No clear effect |
| recurrent skill training in place of parallel training | ReST vs DIAYN/DADS · HalfCheetah, Hopper, Walker2d | x position over time (semi-quantitative) | Improved |
| train only one skill per epoch instead of all skills in parallel | ReST · skill discovery training | sample complexity | Hurt |
Closest papers by meaning
- Constrained Ensemble Exploration for Unsupervised Skill Discovery (ICML 2024 · similarity 0.90)
- Periodic Skill Discovery (NeurIPS 2025 · similarity 0.90)
- Exploration by Learning Diverse Skills through Successor State Representations (NeurIPS 2024 · similarity 0.89)
- Skill Disentanglement in Reproducing Kernel Hilbert Space (AAAI 2025 · similarity 0.89)
- Efficient Skill Discovery via Regret-Aware Optimization (ICML 2025 · similarity 0.89)
- SkiLD: Unsupervised Skill Discovery Guided by Factor Interactions (NeurIPS 2024 · similarity 0.89)
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.
