A machine-learning tool called diffusion maps has been used to identify topological phase transitions in experimental data. The research was done by teams led by Mordechai Segev and Ronen Talmon at ...