Relevance of Karl J. Friston's Free Energy Principle

Click For Summary

Discussion Overview

The discussion centers on the relevance of Karl J. Friston's Free Energy Principle in neuroscience and biology. Participants explore its implications, understanding, and connections to other concepts in physics and inference.

Discussion Character

  • Exploratory
  • Technical explanation
  • Conceptual clarification
  • Debate/contested

Main Points Raised

  • Patrick questions the relevance and fruitfulness of the Free Energy Principle in neuroscience and biology.
  • Some participants express difficulty in understanding the model, suggesting that it may derive from thermodynamic concepts.
  • References to articles and papers are shared to provide further context and explanation of the Free Energy Principle and related concepts.
  • One participant mentions the principle of least action in classical physics and poses an inverse question regarding dynamical equations.
  • Another participant discusses the use of variational language in approximate inference, linking it to the Free Energy Principle.

Areas of Agreement / Disagreement

Participants do not reach a consensus on the validity or relevance of the Free Energy Principle, with some expressing confusion and others providing references for further exploration. Multiple viewpoints and interpretations remain present in the discussion.

Contextual Notes

Some participants indicate limitations in their understanding of the Free Energy Principle, and there are references to complex mathematical and theoretical frameworks that may not be fully resolved in the discussion.

Biology news on Phys.org
Ryan_m_b said:
I can't comment on how valid the model is because I don't understand it but here is an article which may explain it in better detail:
indeed, it is not easy to understand this approach, which seem to derive from the concept of thermodynamics :
https://en.wikipedia.org/wiki/Thermodynamic_free_energy

Thank for the article

Patrick
 
In classical physics, dynamical equations can be derived from a principle of least action.

We can ask the inverse question, which dynamical equations mininimize an action? This question has been addressed by Tonti (and others).
http://www.dic.univ.trieste.it/perspage/tonti/DEPOSITO/Nonlinear.pdf
http://www.dic.univ.trieste.it/perspage/tonti/DEPOSITO/Rassias.pdf
http://www.dic.univ.trieste.it/perspage/tonti/DEPOSITO/Tonti-russi.pdf

Here is some related work on energy functions for "self-organizing maps".
http://www.ncbi.nlm.nih.gov/pubmed/1606243
 
Last edited by a moderator:
Perhaps closer to the technicalities of the article, many forms of approximate inference can be stated using variational language.

http://www.merl.com/publications/docs/TR2001-22.pdf
Understanding Belief Propagation and its Generalizations
Jonathan S. Yedidia, William T. Freeman, and Yair Weiss

http://www.cs.princeton.edu/courses/archive/spr06/cos598C/papers/YedidaFreemanWeiss2004.pdf
Constructing Free Energy Approximations and Generalized Belief Propagation Algorithms
Jonathan S. Yedidia, William T. Freeman, and Yair Weiss

https://www.eecs.berkeley.edu/~wainwrig/Papers/WaiJor08_FTML.pdf
Graphical Models, Exponential Families, and Variational Inference
Martin J. Wainwright and Michael I. Jordan

http://www.cs.berkeley.edu/~jordan/papers/variational-intro.pdf
An Introduction to Variational Methods for Graphical Models
Michael Jordan, Zoubin Ghahramani, Tomi Jaakkola, Lawrence Saul
 

Similar threads

  • · Replies 18 ·
Replies
18
Views
3K
  • · Replies 48 ·
2
Replies
48
Views
5K
  • · Replies 35 ·
2
Replies
35
Views
5K
Replies
0
Views
3K
  • · Replies 14 ·
Replies
14
Views
3K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 1 ·
Replies
1
Views
3K
  • · Replies 16 ·
Replies
16
Views
3K
  • · Replies 2 ·
Replies
2
Views
5K
  • · Replies 2 ·
Replies
2
Views
2K