Learning a phase diagram from dynamics

Time series data contains useful information on the phase of a system. Here we propose the use of recurrent neural networks (LSTM) to learn and extract such information in order to classify phases and locate phase boundaries. We demonstrate this on a many-body localized model, and attempt to interpret the learned behavior by looking at individual LSTM cells. We also discuss the validity of the learned model and investigate its limits.

Event Type: 
Scientific Area(s): 
Event Date: 
Monday, April 23, 2018 - 14:00 to 15:30
Time Room
Room #: