Research map / Reinforcement learning / Visual and representation RL
Invariant Causal Prediction for Block MDPs
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 |
|---|---|---|---|
| Select the causal state variables {x1, x2} and exclude x3 | Linear MISA · Linear model learning with novel hard interventions on x3 | generalization error | Helped |
| Include x3 in the full-state linear predictor | Full-state linear predictor · Linear model learning with novel hard interventions on x3 | generalization error | Hurt |
| Learn an invariant representation from two training environments rather than one | MISA · Cheetah Run model learning on evaluation environments | model error | Helped |
| Train on a single environment | MISA · Cheetah Run model learning on evaluation environments | model error | Hurt |
| Combine data from two environments and train one model over all data instead of using invariant representation learning | Cheetah Run model learning · Cheetah Run evaluation environments | model error | Hurt |
| Use IRM's gradient penalty for invariance | Cheetah Run model learning · Cheetah Run evaluation environments | model error | Hurt |
| Learn an invariant representation for imitation across camera angles | MISA · Cheetah Run imitation learning on evaluation camera angles | test error | Helped |
| Train SAC with an invariant state representation | SAC · Cartpole Swingup with correlated spurious dimensions | generalization gap | Helped |
| Train SAC without noise on the ground-truth states | SAC baseline · Cartpole Swingup | performance | Hurt |
Closest papers by meaning
- On Generalization Across Environments In Multi-Objective Reinforcement Learning (ICLR 2025 · similarity 0.88)
- Towards Generalizable Reinforcement Learning via Causality-Guided Self-Adaptive Representations (ICLR 2025 · similarity 0.88)
- Building Minimal and Reusable Causal State Abstractions for Reinforcement Learning (AAAI 2024 · similarity 0.87)
- Horizon Generalization in Reinforcement Learning (ICLR 2025 · similarity 0.87)
- What Effects the Generalization in Visual Reinforcement Learning: Policy Consistency with Truncated Return Prediction (AAAI 2024 · similarity 0.87)
- Tackling Non-Stationarity in Reinforcement Learning via Causal-Origin Representation (ICML 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.