Our group seeks to understand how computation emerges from the complex dynamics and connectivity of artificial and biological systems. We employ a variety of approaches rooted in statistical physics and applied math, focusing on the following three main directions.
- Statistical Mechanics of Machine Learning: we use tools from spin glasses and random matrix theory to predict the typical learning and generalization performance of neural networks trained on synthetic and real tasks.
- Dynamics and learning in recurrent neural networks: we use concepts and methods from statistical physics and control theory to analyze learning in recurrent neural networks. We are recently focused on understanding quantitatively the role of heterogeneity in RNNs, using both dynamical mean field theory and numerical approaches. In parallel, we build data-driven models for neural population dynamics, using large-scale recordings in behaving animals performing tasks.
- Efficiency of neural computation: we work towards the development of a mathematical formalism that characterizes fundamental tradeoffs between computation and energy consumption in both rate and spiking networks performing function. To do so, we build on concepts from stochastic thermodynamics and information theory.