Andreas Stergiou (King's College): Neural Spectral Bias and Conformal Correlators

Thermal Seminars 23/06/26 ➤ Andreas Stergiou (King's College): Neural Spectral Bias and Conformal Correlators Can a simple neural network deduce the full functional form of a conformal field theory (CFT) correlation function from almost nothing? In this talk, I will demonstrate that by training a neural network using only crossing symmetry, correlator asymptotics, and a single “anchor point”, one can successfully recover correlators across a wide range of theories. I will argue that this success is driven by the spectral bias of neural networks, which inherently focuses on maximally smooth functions. This suggests a potential new aspect of CFTs: their correlators are among the smoothest crossing-symmetric functions. Based on arXiv:2604.18673, 2604.18686 [hep-th] ➤ The Thermal Seminars are a series of webinars on finite temperature quantum field theories, conformal field theories, and related topics, from holography, to black holes and bootstrap approaches. ■ Website: sites.google.com/view/thermalseminars ■ Email: [email protected] ✎Organized by: Julien Barrat (DESY), Deniz Bozkurt (U. Hamburg), Enrico Marchetto (DESY), Alessio Miscioscia (YITP)