---
date: 2012-07-13
excerpts: 2
---

# Studying model asymptotics with singular learning theory

## Abstract
Singular statistical models occur frequently in machine learning and computational biology. An important prob-lem in the learning theory of singular models is determin-ing their asymptotic behavior for massive data sets. In this talk, we give a brief introduction to Sumio Watanabe’s Singular Learning Theory, as outlined in his book “Algebraic Geometry and Statistical Learning Theory.” We will also explore the rich algebraic geometry and combinatorics that arise from studying the asymptotics of Bayesian integrals.

## Details
[Modern Massive Data Sets](https://www.youtube.com/playlist?list=PLl_tHArJvfnBS-OPQa2jFYhb1ZrCAnyyL)

[Slides](https://w3id.org/people/shaoweilin/public/mmds.pdf)

[YouTube](https://www.youtube.com/watch?v=T6fL8uAao0A)