In the Deep learning and Brain science cluster we use advanced tools from theoretical physics to understand how neural networks learn and function, both in living organisms and in artificial systems.

The set of problems we address are at the heart of two scientific and technological frontiers: Brain Science, aiming to unravel the principles underlying brain function, and Machine Learning, aiming to design artificial systems that learn to solve highly complex tasks.

Our aim is to provide theoretical foundations for understanding how deep neural networks learn, and how the brain’s remarkable computational abilities are achieved by the collective dynamics of biological neural networks.

Our research involves a variety of analytical tools from statistical physics, nonlinear dynamics, and information theory, and is carried out in close collaboration with computer science engineers and experimental neuroscientists.