Berlin 2024 – scientific programme
Parts | Days | Selection | Search | Updates | Downloads | Help
TT: Fachverband Tiefe Temperaturen
TT 2: Focus Session: Artificial Intelligence in Condensed Matter Physics I (joint session TT/DY)
TT 2.8: Talk
Monday, March 18, 2024, 13:00–13:15, H 0104
Machine determination of a phase diagram with and without deep learning — •Burak Çivitcioğlu1, Rudolf A. Römer2, and Andreas J. Honecker1 — 1Laboratoire de Physique Théorique et Modélisation, CNRS UMR 8089, CY Cergy Paris Université, France — 2University of Warwick, Coventry, UK
We study the performance of unsupervised learning in detecting phase transitions in the J1-J2 Ising model on the square lattice. We use variational auto encoders (VAE) and the reconstruction error, defined as the mean-squared error between two configurations, to explore the phase diagram of the system. Moreover, we propose as simple alternative method a direct spin comparison. The results of the spin comparison are contrasted with that of the VAEs. Our findings highlight that for certain systems, the simpler method can yield results comparable to a much more complex model, namely the VAE. This work contributes to the broader understanding of machine-learning applications in statistical physics and introduces an efficient approach to the detection of phase transitions using machine determination techniques.
Keywords: machine learning; variational autoencoders; Ising model; deep learning; next nearest neighbour interaction