PIRSA:19040090

Can we trust phase diagrams produced by artificial neural networks?

Wetzel, S. (2019). Can we trust phase diagrams produced by artificial neural networks?. Perimeter Institute. https://pirsa.org/19040090

Wetzel, Sebastian. Can we trust phase diagrams produced by artificial neural networks?. Perimeter Institute, Apr. 11, 2019, https://pirsa.org/19040090 

          @misc{ pirsa_PIRSA:19040090,
            doi = {10.48660/19040090},
            url = {https://pirsa.org/19040090},
            author = {Wetzel, Sebastian},
            keywords = {Other},
            language = {en},
            title = {Can we trust phase diagrams produced by artificial neural networks?},
            publisher = {Perimeter Institute},
            year = {2019},
            month = {apr},
            note = {PIRSA:19040090 see, \url{https://pirsa.org}}
          }

Sebastian Wetzel

Mitacs

Talk number
PIRSA:19040090
Collection
Talk Type
Subject
Abstract

So far artificial neural networks have been applied to discover phase diagrams in many different physical models. However, none of these studies have revealed any fundamentally new physics. A major problem is that these neural networks are mainly considered as black box algorithms. On the journey to detect new physics it is important to interpret what artificial neural networks learn. On the one hand this allows us to judge whether to trust the results, and on the other hand this can give us insight to possible new physics. In this talk I will 
discuss applications to different models where we successfully interpreted what was learned by the neural networks.