Quantum many-body systems remain among the most challenging frontiers of modern physics, where the complexity of microscopic interactions and the correlations they produce make predictive calculations notoriously difficult. Recent advances in machine learning are beginning to change how we represent and solve the Schrödinger equation from first principles.

In this talk, I will introduce neural quantum states: variational wave functions parameterized by artificial neural networks. I will show how these representations can capture correlations that are difficult for traditional methods, and discuss examples where they achieve competitive accuracy for strongly interacting fermionic systems, including nuclei and ultracold Fermi gases.

I will also give some intuition for what these networks are learning, and how this relates to familiar concepts in many-body theory. This perspective suggests a route to predicting properties of quantum matter directly from underlying interactions.

This lecture was made possible by the William C. Ferguson Fund.