BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:Konstantin Matchev - Machine Learning Symmetries in Physics from F
 irst Principles
DTSTART:20250527T130000Z
DTEND:20250527T150000Z
DTSTAMP:20260512T221700Z
UID:indico-event-46904@agenda.infn.it
DESCRIPTION:Symmetries are the cornerstones of modern theoretical physics\
 , as they imply fundamental conservation laws. The recent boom in AI algor
 ithms and their successful application to high-dimensional large datasets 
 from all aspects of life motivates us to approach the problem of discovery
  and identification of symmetries in physics as a machine-learning task. I
 n a series of papers\, we have developed and tested a deep-learning algori
 thm for the discovery and identification of the continuous group of symmet
 ries present in a labeled dataset. We use fully connected neural network a
 rchitectures to model the symmetry transformations and the corresponding g
 enerators. Our proposed loss functions ensure that the applied transformat
 ions are symmetries and that the corresponding set of generators is orthon
 ormal and forms a closed algebra. One variant of our method is designed to
  discover symmetries in a reduced-dimensionality latent space\, while anot
 her variant is capable of obtaining the generators in the canonical sparse
  representation. Our procedure is completely agnostic and has been validat
 ed with several examples illustrating the discovery of the symmetries behi
 nd the orthogonal\, unitary\, Lorentz\, and exceptional Lie groups.\n\nhtt
 ps://agenda.infn.it/event/46904/
LOCATION:Aula B (Via della Vasca Navale 84)
URL:https://agenda.infn.it/event/46904/
END:VEVENT
END:VCALENDAR