BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:BART-L: Towards a Foundation Model for Theoretical High-Energy Phy
 sics - Yong Sheng Koay (Uppsala University)
DTSTART:20251014T130000Z
DTEND:20251014T140000Z
UID:TALK238123@talks.cam.ac.uk
CONTACT:Sven Krippendorf
DESCRIPTION:Transformers excel at natural language tasks\, which naturally
  raises the question of whether they can also learn the mathematical langu
 age of particle physics. In this talk\, I will introduce BART-L\, a transf
 ormer-based model designed to generate particle physics Lagrangians from t
 he field content and their symmetry information. Trained on Lagrangians re
 specting the Standard Model gauge group SU(3)×SU(2)×U(1)\, BART-L achiev
 es over 90 % accuracy for Lagrangians involving up to six matter fields.
  Embedding analyses show that the model internalizes concepts such as grou
 p representations and conjugation operations\, despite not being explicitl
 y trained for them.  We further examine its out-of-distribution behavior t
 o identify architectural constraints that limit generalization. Finally\, 
 we discuss how this framework provides an early indication of what a found
 ation model for theoretical high-energy physics might look like\, along wi
 th its potential capabilities and inherent limitations.
LOCATION:DAMTP\, MR11
END:VEVENT
END:VCALENDAR