Algebraic Geomtry for the Resolution of Singularities to Singular Models

Feb 01 (Monday) at 1:30 pm GHC 8102

Speaker: Jisu Kim

Abstract: Many statistical models are singular, because they are nonidentifiable or their Fisher information matrices are singular. One way to tackle this is to apply resolution of singularities from the field of algebraic geometry. I will introduce how resolution of singularities can be applied, and I will discuss some examples. This talk is based on following book: Sumio Watanabe. Algebraic Geometry and Statistical Learning Theory.