Research map / Reinforcement learning / Visual and representation RL
Provably sample-efficient RL with side information about latent dynamics
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 the abstract simulator and inverse-dynamics state decoder to learn a robust policy | Trembling-hand decoding · Combination lock, H=5, 10, 20, 40 | largest horizon solved within 5 × 10^5 episodes | Helped |
| Add domain randomization to PPO with RND exploration bonus | PPO+RND · Combination lock, H=5, 10, 20, 40 | largest horizon solved within 5 × 10^5 episodes | Hurt |
| Use the robust policy learned from an abstract simulator | Trembling-hand decoding · Visual MiniGrid | policy value | Helped |
| Increase perturbation level η | Trembling-hand decoding · Visual MiniGrid | policy value | Hurt |
| Increase map height from 10 to 100 while holding width, horizon, and layout fixed | Trembling-hand decoding · Visual MiniGrid | policy performance | No clear effect |
| Use the robust policy rather than the abstract simulator's optimal policy | Combination lock · Target environment with η=0.1 | policy value | Helped |
Closest papers by meaning
- Rich-Observation Reinforcement Learning with Continuous Latent Dynamics (ICML 2024 · similarity 0.88)
- Reinforcement Learning Under Latent Dynamics: Toward Statistical and Algorithmic Modularity (NeurIPS 2024 · similarity 0.88)
- Overcoming the Sim-to-Real Gap: Leveraging Simulation to Learn to Explore for Real-World RL (NeurIPS 2024 · similarity 0.88)
- Learning from Reward-Free Offline Data: A Case for Planning with Latent Dynamics Models (NeurIPS 2025 · similarity 0.87)
- Foundation Policies with Hilbert Representations (ICML 2024 · similarity 0.87)
- Imagine Beyond ! Distributionally Robust Autoencoding for State Space Coverage in Online Reinforcement Learning (NeurIPS 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.