Transforming the Bootstrap: Using Transformers to Compute Scattering Amplitudes in Planar N = 4 Super Yang-Mills Theory

We pursue the use of deep learning methods to improve state-of-the-artcomputations in theoretical high-energy physics. Planar N = 4 Super Yang-Millstheory is a close cousin to the theory that describes Higgs boson production atthe Large Hadron Collider; its scattering amplitudes are large mathematicalexpressions containing integer coefficients. In this paper, we applyTransformers to predict these coefficients. The problem can be formulated in alanguage-like representation amenable to standard cross-entropy trainingobjectives. We design two related experiments and show that the model achieveshigh accuracy (> 98applied successfully to problems in theoretical physics that require exactsolutions.