NAG

Abstract

I will explain why the determinants of the period matrices of holonomic D-modules on curves admit a natural epsilon-factorization, as was conjectured by Deligne back in 1984. See http://xxx.lanl.gov/abs/0903.2674