Research map / Multi-agent systems and games / Human-AI and goal-conditioned
Agents Robust to Distribution Shifts Learn Causal World Models Even Under Mediation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Proves that agents robust to distribution shifts must have learned the causal model of their environment, even when actions affect the environment through mediating variables.
Problem
Whether agents that adapt to distribution shifts must learn causal world models, especially when the agent's actions affect the environment through mediation (common in real-world settings such as robotics).
Approach
Extends prior causal-world-model results to settings with mediation, proving that robustness to distribution shifts implies the agent has learned the environment's causal model even when actions influence the environment through mediating variables.
Key results
- Proves that agents capable of adapting to distribution shifts must have learned the causal model of their environment even in the presence of mediation.
Closest papers by meaning
- Robust agents learn causal world models (ICLR 2024 · similarity 0.91)
- Curious Causality-Seeking Agents Learn Meta Causal World (NeurIPS 2025 · similarity 0.87)
- Causal Imitation for Markov Decision Processes: a Partial Identification Approach (NeurIPS 2024 · similarity 0.87)
- Systems with Switching Causal Relations: A Meta-Causal Perspective (ICLR 2025 · similarity 0.86)
- Sim-to-Real Causal Transfer: A Metric Learning Approach to Causally-Aware Interaction Representations (CVPR 2025 · similarity 0.86)
- Learning to Steer Markovian Agents under Model Uncertainty (ICLR 2025 · similarity 0.86)
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.