From recognising pictures of cats on the internet to defeating world champions in the game of Go, the power of Machine Learning is increasingly visible in our everyday lives. Many of the core ideas in modern machine learning can be traced to insights from studying learning in animals. However, the benefits from in both directions, with brain scientists using theories from optimal control, Bayesian decision-making, and predictive coding, to make sense of the changes that take place in the brain during learning. BCBT has featured many great talks about learning. Here we feature two talks, from Matthew Botvinick and Marc Toussaint, that applies principles from the study of predictive inference to understand learning in the brain, we also feature one of several talks that examine how the cerebellum contributes to learning skilled movements.

Featured Lectures

Marc ToussaintMarc Toussaint (2013)

Decision making and planning as probabilistic inference

Mathew BotvinickMathew Botvinick (2010)

Goal-directed decision making as probabilistic inference

Narender RamnaniNarender Ramnani (2015)

Cerebellar contributions to instrumental learning

Other Lectures

Will we ever reverse-engineer animal cognition?

Tugging at tool use

The proactive brain: predictions in visual cognition

State estimation and the cerebellum

Perceptual learning of parametric face categories leads to the integration of high-level class-based information but not to high-level pop-out
Searching for the cerebellar algorithm: the adaptive-filter hypothesis

Optical approaches to monitoring cerebellar processing of unexpected events