## Planning as Inference ## Overview Planning as inference casts action selection as probabilistic inference under a generative model. Desired outcomes are encoded as priors (preferences) and plans (policies) are inferred by minimizing expected free energy. ## Core Formulation ```math P(\pi) \propto \exp\big(-\gamma\,G(\pi)\big),\quad G(\pi)=\sum_{\tau} G(\pi,\tau) ``` - [[../mathematics/expected_free_energy]]: balances epistemic and pragmatic value - [[../mathematics/policy_selection]]: softmax over negative EFE - [[../mathematics/variational_free_energy]]: perceptual inference objective ## Connections - [[active_inference]]: unified scheme for perception and action - [[predictive_coding]]: message passing implementation - [[../mathematics/active_inference_pomdp]]: discrete POMDP instantiation ## See Also - [[../mathematics/epistemic_value]] — Epistemic value computation - [[../mathematics/pragmatic_value]] — Pragmatic value computation