Research map / Reinforcement learning / Visual and representation RL
Optimistic Active Exploration of Dynamical Systems
Bhavya Sukhija, Lenart Treven, Cansu Sancaktar, Sebastian Blaes, Stelian Coros, Andreas Krause
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 |
|---|---|---|---|
| active exploration (OpAx) instead of random exploration | dynamics model learning · Pendulum-v1 GP | max epistemic uncertainty | Improved |
| active exploration (OpAx) instead of random exploration | dynamics model learning · Pendulum-v1 PE | max epistemic uncertainty | Improved |
| optimistic planner (OpAx) instead of mean planner or PETS-AE | active exploration · Pendulum-v1 GP | max epistemic uncertainty | No clear effect |
| optimistic planner (OpAx) instead of Mean-AE and PETS-AE | active exploration · Pendulum-v1 PE | max epistemic uncertainty | Improved |
| active exploration instead of random exploration | model-based RL · MountainCar downstream | task reward | Improved |
| PETS-AE trajectory sampling instead of other planners | active exploration · MountainCar downstream | task reward | Hurt |
| OpAx instead of task-specific H-UCRL on tasks H-UCRL was trained for | model-based RL · Pendulum, MountainCar, Reacher, Swimmer, Cheetah | task reward | No clear effect |
| active exploration instead of H-UCRL on unseen tasks | model-based RL · unseen downstream tasks | task reward | Improved |
| OpAx instead of CEE-US, PETS-AE, Mean-AE | active exploration · Fetch Pick & Place Construction | success rate | No clear effect |
| OpAx instead of random policy | active exploration · Fetch Pick & Place Construction | success rate | Improved |
Closest papers by meaning
- Beyond Optimism: Exploration With Partially Observable Rewards (NeurIPS 2024 · similarity 0.89)
- Efficient Model-Based Reinforcement Learning Through Optimistic Thompson Sampling (ICLR 2025 · similarity 0.88)
- NeoRL: Efficient Exploration for Nonepisodic RL (NeurIPS 2024 · similarity 0.88)
- Bayesian Exploration Networks (ICML 2024 · similarity 0.88)
- SOMBRL: Scalable and Optimistic Model-Based RL (NeurIPS 2025 · similarity 0.88)
- OVD-Explorer: Optimism Should Not Be the Sole Pursuit of Exploration in Noisy Environments (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.
