 |
Center for Nonlinear Studies |
I will discuss how artificial neural networks can be used to identify phases and phase transitions in condensed matter systems via supervised learning. I will show that standard feed-forward neural networks can be trained to detect multiple types of order parameter directly from raw state configurations sampled with Monte Carlo. In addition, they can detect highly non-trivial states such as Coulomb phases, and if modified to a convolutional neural network, topological phases with no conventional order parameter. Furthermore, I will discuss the application of machine learning ideas to quantum systems. In particular, I will demonstrate that convolutional neural networks (CNN) have the potential to represent ground states of quantum many-body systems by showing that the ground state of Kitaev's toric code can be written as a CNN. Lastly, I will briefly show that machine learning devices such as the restricted Boltzmann machine can be efficiently used for quantum state tomography of highly-entangled states in arbitrary dimension.
Host: Lukasz Cincio