The science underneath
The free energy principle
In shortKarl Friston's idea that every living thing keeps itself going by predicting its world and its own body.
The idea
Karl Friston proposed a general idea about everything that is alive: to keep existing, an organism has to stay within a small set of states (temperature, blood sugar, connection with others). To do that it needs a model of the world and of its own body, and it uses that model in two ways: it changes the model when the world does not match what it expected, or it changes the world, by acting, so that it matches. Perceiving, learning, feeling and acting become parts of the same process.
What to try
- Faced with a behaviour, ask: what did this person expect, and what did the behaviour do for them?
- Notice there are two ways out of a gap between what was expected and what happens: change the expectation, or act on the situation (including avoiding it).
- Changes of routine cost more when the prediction was strong and long-standing.
The common mistake
Using the principle as a ready-made explanation (‘high free energy’). It is a general frame; explaining a particular person is done with what they expect, feel and do, in particular situations.
Want to know more?
Start here
Short, plain pieces by the researchers and good writers.
Books
How the predictive brain shapes what we see, feel and do, with many everyday examples.
See on Amazon ›Amazon links are affiliate links: if you buy, Bounded Uncertainty receives a small commission at no cost to you.
The science behind itHow strong the evidence is, the model, videos and references.
How strong the evidence is
A way of thinking A lens for interpreting what a child does, not a rule from trials.
Formally, a system that persists minimises an upper bound (variational free energy) on the surprisal of its observations, given a generative model. Free energy decomposes into complexity minus accuracy, a trade-off between explaining the data and keeping the model simple. For practice, three consequences matter: behaviour is inference (selecting policies that reduce expected free energy, which includes epistemic and pragmatic value); the precision assigned to predictions and errors decides what changes; and context (body, people, environment) is part of the model. There is debate about whether the principle is falsifiable and about the status of its claims; we treat it as a lens that organises hypotheses, not as proof.
More technical books
The technical textbook on active inference, for those who want the formal model.
See on Amazon ›Amazon links are affiliate links: if you buy, Bounded Uncertainty receives a small commission at no cost to you.
Videos
External videos, chosen because they agree with these ideas. Bounded Uncertainty has no link to the channels.
References
- Friston, K. (2010). The free-energy principle: a unified brain theory? Nature Reviews Neuroscience, 11, 127–138.
- Friston, K. (2013). Life as we know it. Journal of the Royal Society Interface, 10, 20130475.
- Parr, T., Pezzulo, G., & Friston, K. J. (2022). Active Inference: The Free Energy Principle in Mind, Brain, and Behavior. MIT Press.
Found this useful? Buy Ricardo a coffee