The Free Energy Principle for Perception and Action: A Deep Learning Perspective

TL;DR

This paper explores how deep learning can implement the free energy principle and active inference, providing a practical guide and surveying relevant works to bridge theoretical concepts with real-world artificial agents.

Contribution

It offers a deep-learning oriented presentation of the free energy principle, surveys relevant research, and discusses implementation choices for artificial agents based on active inference.

Findings

Deep learning enables practical implementation of active inference.

Survey of relevant machine learning and active inference works.

Provides a pragmatic guide for implementing the free energy principle.

Abstract

The free energy principle, and its corollary active inference, constitute a bio-inspired theory that assumes biological agents act to remain in a restricted set of preferred states of the world, i.e., they minimize their free energy. Under this principle, biological agents learn a generative model of the world and plan actions in the future that will maintain the agent in an homeostatic state that satisfies its preferences. This framework lends itself to being realized in silico, as it comprehends important aspects that make it computationally affordable, such as variational inference and amortized planning. In this work, we investigate the tool of deep learning to design and realize artificial agents based on active inference, presenting a deep-learning oriented presentation of the free energy principle, surveying works that are relevant in both machine learning and active inference areas, and discussing the design choices that are involved in the implementation process. This manuscript probes newer perspectives for the active inference framework, grounding its theoretical aspects into more pragmatic affairs, offering a practical guide to active inference newcomers and a starting point for deep learning practitioners that would like to investigate implementations of the free energy principle.