[AIP AI4S Seminar Series] 'Breaking of Replica Symmetry on the Nishimori Line and its Implications for Bayesian Inference' Talk by Ayaka Sakata (Ochanomizu University and RIKEN AIP)

An AIP AI4S (AI for Science) Seminar Series talk by Ayaka Sakata (Ochanomizu University / RIKEN AIP) on the breaking of replica symmetry on the Nishimori line and its implications for Bayesian inference. The talk uses mean-field theory to explain the thermodynamic behavior of the Nishimori line and the associated phase transitions revealed by the recently proposed 'Nishimori Model,' going beyond the conventional paradigm of RS stability. The speaker discusses what these insights mean for the fundamental limits of statistical estimation in Bayesian inference. Hybrid format; in-person attendance is limited to AIP researchers, online open more broadly.