Research map / Reinforcement learning / Visual and representation RL
Discovering Mixture Skills for Unsupervised Reinforcement Learning
Nelson Ma, Junyu Xuan, Guangquan Zhang, Jie Lu
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 DiMS instead of CIC on Jaco manipulation tasks | DiMS · URLB Jaco | normalised IQM return: p = 0.03 | Improved |
| use DiMS instead of CIC on Walker locomotion tasks | DiMS · URLB Walker | normalised IQM return | Hurt |
| add online joint GMVAE training instead of phased explore-discover-learn stages (ablation) | DiMS · URLB (Jaco tasks especially) | normalised IQM return | Improved |
| use GMVAE mixture latent instead of discrete VQ-VAE (ablation) | DiMS · URLB | normalised IQM return | Improved |
| use skill-agnostic exploration bonus (RND or APT) instead of a second exploration latent (CIC or CSD) (ablation) | DiMS · URLB (all three domains) | normalised IQM return | Improved |
| use RND instead of APT as exploration bonus (ablation) | DiMS · URLB Jaco | normalised IQM return | Improved |
Closest papers by meaning
- Constrained Ensemble Exploration for Unsupervised Skill Discovery (ICML 2024 · similarity 0.90)
- Skill Disentanglement in Reproducing Kernel Hilbert Space (AAAI 2025 · similarity 0.90)
- Efficient Skill Discovery via Regret-Aware Optimization (ICML 2025 · similarity 0.89)
- Disentangled Unsupervised Skill Discovery for Efficient Hierarchical Reinforcement Learning (NeurIPS 2024 · similarity 0.89)
- Leveraging Skills from Unlabeled Prior Data for Efficient Online Exploration (ICML 2025 · similarity 0.88)
- Task Adaptation from Skills: Information Geometry, Disentanglement, and New Objectives for Unsupervised Reinforcement Learning (ICLR 2024 · 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.
