Improved Machine Learning Algorithm for Predicting Ground State Properties

Explore an advanced machine learning algorithm for predicting ground state properties of quantum many-body systems in this 1 hour 15 minute lecture by Laura Lewis from Caltech. Delivered as part of the Quantum Summer Cluster Workshop at the Simons Institute, discover how this classical ML approach incorporates an inductive bias encoding geometric locality to efficiently predict properties of gapped local Hamiltonians. Learn about the algorithm's ability to make predictions after training on only O(log(n)) data from Hamiltonians in the same quantum phase of matter, a significant improvement over previous methods requiring O(n^c) data. Examine the algorithm's O(n log n) scaling for training and prediction time, and review numerical experiments on physical systems with up to 45 qubits that demonstrate its effectiveness with small training datasets.