Planning under Distribution Shifts with Causal POMDPs

Researchers propose a new theoretical framework for AI planning under changing conditions using causal POMDPs (Partially Observable Markov Decision Processes). The framework represents environmental changes as interventions, enabling AI systems to evaluate and adapt plans when underlying conditions shift while maintaining computational tractability.

• →New framework addresses distribution shifts in AI planning using causal knowledge representation.
• →Method enables AI systems to actively identify which environmental components have changed.
• →Framework maintains belief updates over both latent states and underlying domains.
• →Value function remains piecewise linear and convex, preserving computational tractability.
• →Approach allows evaluation of plans under hypothetical environmental changes.