Free-energy and the brain

By Karl J. Friston, Klaas E. Stephan. Synthese, 2007

607 citations recorded by the source corpus.

If one formulates Helmholtz’s ideas about perception in terms of modern-day theories one arrives at a model of perceptual inference and learning that can explain a remarkable range of neurobiological facts. Using constructs from statistical physics it can be shown that the problems of inferring what cause our sensory inputs and learning causal regularities in the sensorium can be resolved using exactly the same principles. Furthermore, inference and learning can proceed in a biologically plausible fashion. The ensuing scheme rests on Empirical Bayes and hierarchical models of how sensory information is generated. The use of hierarchical models enables the brain to construct prior expectations in a dynamic and context-sensitive fashion. This scheme provides a principled way to understand many aspects of the brain’s organisation and responses. In this paper, we suggest that these perceptual processes are just one emergent property of systems that conform to a free-energy principle. The free-energy considered here represents a bound on the surprise inherent in any exchange with the environment, under expectations encoded by its state or configuration. A system can minimise free-energy by changing its configuration to change the way it samples the environment, or to change its expectations. These changes correspond to action and perception, respectively, and lead to an adaptive exchange with the environment that is characteristic of biological systems. This treatment implies that the system’s state and structure encode an implicit and probabilistic model of the environment. We will look at models entailed by the brain and how minimisation of free-energy can explain its dynamics and structure.

Subfields: Artificial Intelligence, Cognitive Neuroscience, Plant Science

View Paper: https://www.semanticscholar.org/paper/bf15128e48db7aaa9146328d965fc81dbdf4db3f

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Cognitive Neuroscience

In Plant Science

Citations in the corpus

Prediction, Cognition and the Brain (2009, background): “…research on prediction in perception has been greatly infl uenced by von Helmholtz who argued that sensory systems evolved in order to infer the causes of changes in sensory inputs (Friston and Stephan, 2007), thus equating perception with recognition, namely inference about the state of the world.”
The free energy principle for action and perception: A mathematical review (2017, background): “These arguments have been developed in a series of papers which have appeared over the course of the last several years [16, 17, 18, 19, 20, 21, 22, 23, 9, 24, 25, 26, 27, 28, 29].”
Towards a Biologically Plausible Backprop (2016, background): “It has been hypothesized numerous times (Hinton and Sejnowski, 1986; Friston and Stephan, 2007; Berkes et al., 2011), that, given a state of sensory information (current and past inputs), neurons are collectively performing inference, i.e., moving towards configurations that better “explain” the…”
Behavioral / Systems / Cognitive Striatal Prediction Error Modulates Cortical Coupling (2010, positive): “…to the free-energy principle, any prediction error should induce learning through synaptic plasticity, reconfiguring connection strengths such that prediction error is minimized at both sensory and motor levels, thus optimizing perceptual inference and motor actions (Friston and Stephan, 2007).”
The anticipating brain is not a scientist: the free-energy principle from an ecological-enactive perspective (2016, methodology): “One aim of this paper is to build on our earlier work in 3 For excellent introductions to predictive-coding see (Clark 2013; Rao and Ballard 1998; Friston and Stephan 2007).”
Free-Energy Minimization and the Dark-Room Problem (2012, background): “The most comprehensive such treatment is the “free-energy minimization” formulation due to Karl Friston (see e.g., Friston and Stephan, 2007; Friston, 2010a,b – see also Fiorillo, 2010; Thornton, 2010).”
Words and the World (2015, background): “Organisms that can predict their own sensory inputs at multiple spatial and temporal scales are well placed to do this, thus maintaining themselves within their species-specific window of viability (Friston & Stephan, 2007).”
Toward an Integration of Deep Learning and Neuroscience (2016, background): “The minimization of prediction error explains a number of properties of the nervous system (Friston and Stephan, 2007; Huang and Rao, 2011), and biologically plausible theories are available for how cortex could learn using prediction errors by exploiting temporal diﬀerences (O’Reilly et al.,…”
Vanilla PP for Philosophers: A Primer on Predictive Processing (2017, neutral): “This was first put forward by Karl Friston and Klaas Stephan (Friston and Stephan 2007) (two important papers extending this idea are Feldman and Friston 2010 and Hohwy 2012).”
Stochastic variational learning in recurrent spiking networks (2014, methodology): “…exploit more analytical properties of the variational approximation (explicitly computing expectations and covariances terms in the free energy) (Friston and Stephan, 2007; Jimenez Rezende et al., 2011) is that, although these analytical approximations may provide gradient estimators with much…”
An objective function for STDP (2015, background): “In addition to the value this brings in terms of fitting sequences of observed sensory inputs, we conjecture that this complies with the already well known hypothesis (Hinton and Sejnowski, 1986; Friston and Stephan, 2007; Berkes et al., 2011) that, given a state of sensory information (current and ”
Action and behavior: a free-energy formulation (2010, background): “We will see below that free-energy can be minimized by changing perceptual representations so that they approximate a posteriororconditionaldensityonthecausesofsensations.Inshort, the free-energy principle entails the Bayesian brain hypothesis (Knill and Pouget 2004; Ballard et al. 1983; Dayan et al”
Equilibrium Propagation: Bridging the Gap between Energy-Based Models and Backpropagation (2016, background): “Previous work (Hinton and Sejnowski, 1986; Friston and Stephan, 2007; Berkes et al. , 2011) has hypothesized that, given a state of sensory information, neurons are collectively performing inference: they are moving towards conﬁgurations that better ’explain’ the observed sensory data.”