Research map / Reinforcement learning / Visual and representation RL
Latent exploration for Reinforcement Learning
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 |
|---|---|---|---|
| replace independent noise with Lattice | SAC · PyBullet Humanoid | reward: 2122 ± 169 -> 2742 ± 77 | Helped |
| use Lattice with episode-length noise correlation | SAC · PyBullet Humanoid | reward: 2122 ± 169 -> 2901 ± 211 | Helped |
| replace independent noise with Lattice | SAC · PyBullet Ant | reward: 3381 ± 30 -> 3544 ± 212 | No clear effect |
| replace independent noise with Lattice | SAC · PyBullet Hopper | reward: 2417 ± 106 -> 2610 ± 78 | No clear effect |
| replace independent noise with Lattice | SAC · PyBullet Walker | reward: 2741 ± 81 -> 2718 ± 92 | No clear effect |
| replace independent noise with Lattice | SAC · PyBullet Half Cheetah | reward: 2934 ± 27 -> 2900 ± 67 | No clear effect |
| Lattice instead of gSDE (resampling every episode) | SAC · PyBullet Humanoid | reward: 2460 ± 160 -> 2901 ± 211 | Helped |
| gSDE period 4 instead of SAC noise | SAC · PyBullet Ant | reward: 3381 ± 30 -> 3978 ± 14 | Helped |
| increase noise correlation period to 8 | SAC-Lattice · PyBullet Humanoid | reward: 2742 ± 77 -> 2358 ± 77 | Hurt |
| Lattice in pose tasks | PPO · MyoSuite pose tasks | solved fraction | Hurt |
| Lattice in reach and manipulation tasks | PPO · MyoSuite reach and manipulation | reward | Helped |
| Lattice exploration | PPO · MyoSuite reach and manipulation | energy: 20% to 60% | Helped |
Closest papers by meaning
- Explore to Learn: Latent Exploration Through Disentangled Synergy Patterns for Reinforcement Learning in Overactuated Control (AAAI 2026 · similarity 0.90)
- Colored Noise in PPO: Improved Exploration and Performance through Correlated Action Sampling (AAAI 2024 · similarity 0.88)
- Efficient Planning with Latent Diffusion (ICLR 2024 · similarity 0.86)
- Reinforcement Learning Under Latent Dynamics: Toward Statistical and Algorithmic Modularity (NeurIPS 2024 · similarity 0.86)
- Maximum Total Correlation Reinforcement Learning (ICML 2025 · similarity 0.85)
- Periodic Skill Discovery (NeurIPS 2025 · similarity 0.85)
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.