Research map / Reinforcement learning / Visual and representation RL
Towards Effective Context for Meta-Reinforcement Learning: an Approach based on Contrastive 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 |
|---|---|---|---|
| Use an independent information-gain exploration policy with a contrastive context encoder | CCM · cheetah-sparse, walker-sparse, hard-point-robot | return | Helped |
| Use contrastive context encoding with dynamics prediction | PEARL · ant-mass and cheetah-mass | return | Helped |
| Use contrastive context encoding with value-function recovery | PEARL · out-of-distribution tasks | return | Helped |
| Add contrastive encoder training to value-function recovery or dynamics prediction | PEARL context encoder · continuous-control benchmarks | return | Helped |
| Remove the regularization term from intrinsic reward (ablation) | CCM · hard-point-robot | return | Hurt |
| Set intrinsic reward scale too high or too low (ablation) | CCM · cheetah-sparse | return | Hurt |
| Update context every step rather than every episode (ablation) | CCM · Meta-RL adaptation | return | Helped |
| Train the encoder with contrastive loss only, without RV or DP (ablation) | CCM · continuous-control benchmarks | return | Helped |
Closest papers by meaning
- MetaCARD: Meta-Reinforcement Learning with Task Uncertainty Feedback via Decoupled Context-Aware Reward and Dynamics Components (AAAI 2024 · similarity 0.91)
- Towards an Information Theoretic Framework of Context-Based Offline Meta-Reinforcement Learning (NeurIPS 2024 · similarity 0.91)
- Scrutinize What We Ignore: Reining In Task Representation Shift Of Context-Based Offline Meta Reinforcement Learning (ICLR 2025 · similarity 0.90)
- AMAGO-2: Breaking the Multi-Task Barrier in Meta-Reinforcement Learning with Transformers (NeurIPS 2024 · similarity 0.90)
- Decoupling Meta-Reinforcement Learning with Gaussian Task Contexts and Skills (AAAI 2024 · similarity 0.90)
- Skill-aware Mutual Information Optimisation for Zero-shot Generalisation in Reinforcement Learning (NeurIPS 2024 · similarity 0.89)
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.