The approximate posterior over policies is a softmax function of the negative expected free energy.

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• The marginal free energy currently stands as the most biologically plausible approximation.claim0.788
  Evaluation of various free energy approximations, Section 4.2.
• By minimising variational free energy, agents perform approximate Bayesian inference.claim0.778
  Describes the epistemic function of variational free energy.
• The most likely (i.e., best) policies minimise expected free energy.claim0.777
  Decision-making rule in active inference.
• Winner take-all architectures of decision-making are already commonplace in computational neuroscience, and the softmax function provides a smooth approximation.claim0.777
  Neural plausibility argument for softmax policy selection.
• Any system minimizing free energy will appear to engage in implicit Bayesian inference of hidden external causes.hypothesis0.776
  Predicts that internal states encode posterior beliefs about external world through gradient descent on free energy.
• Softmax Policy Priorconcept0.763
  Policies assigned probability via softmax of expected free energy; enables self-evidencing behavior.
• The softmax function may be interpreted as performing lateral inhibition, leading to sharper inferences.claim0.760
  Interpretation of neural implementation in Section 5.1.
• Curiosity, insight, decision-making, and diverse phenomena can all be accommodated by a single imperative: minimization of expected free energy (resolution of uncertainty).claim0.759
  Central thesis of the paper unifying cognitive phenomena under one objective function