Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Task-Agnostic Online Reinforcement Learning with an Infinite Mixture of Gaussian Processes
Mengdi Xu, Wenhao Ding, Jiacheng Zhu, ZUXIN LIU, Baiming Chen, Ding Zhao
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 |
|---|---|---|---|
| remove dynamics assignment (single GP) (ablation) | Our Method · CartPole Task2 | mean accumulated reward: 169.46 -> 183.10 | Improved |
| remove dynamics assignment (single GP) (ablation) | Our Method · CartPole Task1 | mean accumulated reward: 178.04 -> 177.62 | No clear effect |
| remove dynamics assignment (single GP) (ablation) | Our Method · HalfCheetah Task1 | mean accumulated reward: 2.70 -> 36.11 | Improved |
| remove dynamics assignment (single GP) (ablation) | Our Method · HalfCheetah Task2 | mean accumulated reward: -21.26 -> 34.14 | Improved |
| remove dynamics assignment (single GP) (ablation) | Our Method · Intersection Task2 | mean accumulated reward: 39.23 -> 54.77 | Improved |
| remove dynamics assignment (single GP) (ablation) | Our Method · Intersection Task3 | mean accumulated reward: 44.25 -> 52.48 | Improved |
| DP prior with DNN (DPNN) instead of transition prior with GP | Our Method · CartPole Task4 | mean accumulated reward: 161.58 -> 182.54 | Improved |
| DP prior with DNN (DPNN) instead of transition prior with GP | Our Method · CartPole Task1 | mean accumulated reward: 187.24 -> 177.62 | Hurt |
| DP prior with DNN (DPNN) instead of transition prior with GP | Our Method · HalfCheetah Task1 | mean accumulated reward: -42.88 -> 36.11 | Improved |
| DP prior with DNN (DPNN) instead of transition prior with GP | Our Method · Intersection Task3 | mean accumulated reward: 18.64 -> 52.48 | Improved |
| pre-trained ANP instead of online GP mixture | Our Method · Intersection Task1 | mean accumulated reward: 62.68 -> 60.66 | Hurt |
| online GP mixture instead of pre-trained DNN | DNN · CartPole Task1 | mean accumulated reward: 176.57 -> 177.62 | No clear effect |
Closest papers by meaning
- Continual Reinforcement Learning by Planning with Online World Models (ICML 2025 · similarity 0.89)
- Pausing Policy Learning in Non-stationary Reinforcement Learning (ICML 2024 · similarity 0.87)
- Offline-Online Reinforcement Learning for Linear Mixture MDPs (NeurIPS 2026 · similarity 0.87)
- Decoupling Meta-Reinforcement Learning with Gaussian Task Contexts and Skills (AAAI 2024 · similarity 0.87)
- DynaMITE-RL: A Dynamic Model for Improved Temporal Meta-Reinforcement Learning (NeurIPS 2024 · similarity 0.87)
- Autonomous Option Invention for Continual Hierarchical Reinforcement Learning and Planning (AAAI 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.
