[Paper Review] A free energy principle for a particular physics

Karl Friston|arXiv (Cornell University)|Jun 24, 2019

Advanced Thermodynamics and Statistical Mechanics197 references234 citations

TL;DR

The paper outlines a theory where Markov blankets enable a Bayesian mechanics framework that interprets internal states as inferences about external states, connecting quantum, statistical, and classical mechanics under a free energy principle.

ABSTRACT

This monograph attempts a theory of every 'thing' that can be distinguished from other things in a statistical sense. The ensuing statistical independencies, mediated by Markov blankets, speak to a recursive composition of ensembles (of things) at increasingly higher spatiotemporal scales. This decomposition provides a description of small things; e.g., quantum mechanics - via the Schrodinger equation, ensembles of small things - via statistical mechanics and related fluctuation theorems, through to big things - via classical mechanics. These descriptions are complemented with a Bayesian mechanics for autonomous or active things. Although this work provides a formulation of every thing, its main contribution is to examine the implications of Markov blankets for self-organisation to nonequilibrium steady-state. In brief, we recover an information geometry and accompanying free energy principle that allows one to interpret the internal states of something as representing or making inferences about its external states. The ensuing Bayesian mechanics is compatible with quantum, statistical and classical mechanics and may offer a formal description of lifelike particles.

Motivation & Objective

• Motivate a theory of everything distinguished in a statistical sense via Markov blankets.
• Describe how recursive composition of ensembles across scales yields small- and large-scale physical descriptions.
• Introduce Bayesian mechanics for autonomous or active entities and relate it to self-organization at nonequilibrium steady state.

Proposed method

• Develop an information-geometric view leading to a free energy principle.
• Show how internal states can represent or infer external states within Markov blankets.
• Derive Bayesian mechanics compatible with quantum, statistical, and classical mechanics.

Experimental results

Research questions

• RQ1 Can Markov blankets support a universal statistical decomposition of physical systems across scales?
• RQ2 How does the free energy principle via Bayesian mechanics describe self-organization at nonequilibrium steady states?
• RQ3 In what sense can internal states be interpreted as inferences about external states across quantum, statistical, and classical regimes?

Key findings

• Proposes a unifying information-geometric framework grounded in Markov blankets.
• Derives a free energy principle that enables internal states to encode inferences about external states.
• Shows compatibility of Bayesian mechanics with quantum, statistical, and classical descriptions.