Karl J. Friston
University College London
- Research field
- Theoretical neuroscience, brain imaging methods, the free-energy principle
Karl Friston proposed that perception, learning, and action can all be understood as one process, the reduction of a quantity called free energy that bounds how surprising the brain's sensory input is. In the 2009 paper with Stefan Kiebel the two authors build a simulated brain from the equations of this principle and let it listen to artificial birdsong: it recognizes the songs, sorts them into categories, and produces a burst of prediction error when a song stops early, as real brains do when an expected sound is left out. The unit presents the principle as a unifying proposal and this paper as its clearest worked demonstration, next to the objections its critics raise.
In the masterclass
Section 7, after the statement of the free-energy principle and the criticism of it, shows how he and Stefan Kiebel turned the principle into a simulated brain that recognizes birdsong.
The research is presented in Unit 15: The Predictive Brain, in a research spotlight box that explains the result at the level of the unit and links it to what you have just learned.