CASK: A Gauge Covariant Transformer for Lattice Gauge Theory

Add to your list(s) Download to your calendar using vCal

• Akio Tomiya (Tokyo Woman's Christian University)
• Wednesday 26 March 2025, 15:15-15:45
• DAMTP, MR11.

If you have a question about this talk, please contact Sven Krippendorf.

We introduce a Transformer architecture for lattice QCD that is designed to respect the gauge symmetry and the discrete rotational and translational symmetries of the lattice. The core innovation lies in defining the attention matrix via a Frobenius inner product between link variables and extended staples, ensuring gauge covariance. We apply this method to self-learning HMC and find that it surpasses existing gauge covariant neural networks in performance, demonstrating its potential to enhance lattice QCD computations. This talk is based on https://arxiv.org/abs/2501.16955

This talk is part of the DAMTP Data Intensive Science Seminar series.

This talk is included in these lists:

• DAMTP Data Intensive Science Seminar
• DAMTP, MR11

Note that ex-directory lists are not shown.