In the machine learning and data science (MLDS) unit, we focus on developing fundamental machine learning algorithms and solving important scientific problems using machine learning. We are currently interested in statistical modeling for high-dimensional data including kernel and deep learning models and geometric machine learning algorithms, including graph neural networks (GNN) and optimal transport problems. In addition to developing ML models, we focus on developing new machine learning methods to automatically find a new scientific discoveries from data.

The Model-Based Evolutionary Genomics Unit works at the crossroads of computational and evolutionary biology. Our long-term goal is to achieve an integrative understanding of the evolution of Life on Earth and the origins and emergence of complexity across different biological scales, from individual proteins to ecosystems. To move towards this goal, we develop and apply model-driven evolutionary genomics methods to reconstruct the Tree of Life and the major evolutionary transitions that have occurred along its branches.

The Neural Computation Unit develops algorithms that elucidate the brain’s mechanisms for robust and flexible learning. The unit focuses on how the brain processes reinforcement learning, in...

The unit studies natural time series data, from NLP to genome sequences to cephalopod camouflage, with dynamical systems and machine learning methods, letting the data lead the way.

We study in theory and experiments the engineering of quantum devices built from different subsystems that can collectively perform beyond the individual capabilities of their parts.

We investigate how behavioural contexts tune olfactory information processing in the mouse brain. Methods used include neurophysiology, imaging, circuit analysis, behaviour.

The Theory of Quantum Matter (TQM) Unit carries out research into a wide range of problems in condensed matter and statistical physics.