Research map / Reinforcement learning / Visual and representation RL
Unsupervised Reinforcement Learning in Multiple Environments
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 percentile-sensitive CVaR training instead of average-entropy training | Exploration pre-training on GridWorld with Slope · GWN | state-visitation entropy | Helped |
| Use percentile-sensitive CVaR training instead of average-entropy training | Exploration pre-training on GridWorld with Slope · GWS | state-visitation entropy | Mixed |
| Initialize TRPO with the percentile-sensitive exploration policy instead of the average-entropy policy | TRPO · GridWorld with Slope, GWN sparse-reward tasks | average return | Helped |
| Lower the CVaR percentile to prioritize adverse environments more (ablation) | αMEPOL · GridWorld with Slope | state-visitation entropy | Mixed |
| Use percentile-sensitive CVaR training instead of average-entropy training | Exploration pre-training on MultiGrid · MultiGrid worst-case configuration | state-visitation entropy | Helped |
| Initialize TRPO with the percentile-sensitive exploration policy instead of the average-entropy policy | TRPO · MultiGrid, GWN fine-tuning tasks | average return | Helped |
| Use percentile-sensitive CVaR training instead of average-entropy training | Exploration pre-training on Ant Stairs · Ant Stairs Up | state-visitation entropy | Helped |
| Initialize TRPO with the percentile-sensitive exploration policy instead of the average-entropy policy | TRPO · Ant Stairs Up fine-tuning tasks | average return | Helped |
| Initialize TRPO with the percentile-sensitive exploration policy instead of the average-entropy policy | TRPO · MiniGrid AdvG fine-tuning tasks | average return | Helped |
| Initialize TRPO with the percentile-sensitive exploration policy instead of the average-entropy policy | TRPO · MiniGrid EasyG fine-tuning tasks | average return | Hurt |
| Use αMEPOL exploration pre-training instead of MAML meta-training | TRPO initialization · GridWorld with Slope and MultiGrid fine-tuning tasks | average return | Helped |
Closest papers by meaning
- Towards Principled Unsupervised Multi-Agent Reinforcement Learning (NeurIPS 2025 · similarity 0.91)
- Constrained Ensemble Exploration for Unsupervised Skill Discovery (ICML 2024 · similarity 0.89)
- Foundation Policies with Hilbert Representations (ICML 2024 · similarity 0.88)
- Unsupervised Training Sequence Design: Efficient and Generalizable Agent Training (AAAI 2024 · similarity 0.87)
- Improving Regret Approximation for Unsupervised Dynamic Environment Generation (NeurIPS 2025 · similarity 0.87)
- DGPO: Discovering Multiple Strategies with Diversity-Guided Policy Optimization (AAAI 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.