Personal tools
You are here: Home Events ANC Seminar: Christopher Yau

ANC Seminar: Christopher Yau

— filed under:

Embedding additive models within nonlinear latent variable models

What
  • ANC/DTC Seminar
When May 28, 2019
from 11:00 AM to 12:00 PM
Where IF 4.31/4.33
Add event to calendar vCal
iCal

Abstract

Nonlinear latent variable models, such as Gaussian Process Latent Variable Models or Variational Autoencoders, provide a powerful means of identifying latent substructures in high-dimensional data problems. Recently, we have developed variants of such models that embed Generalise Additive Model type constraints in the way latent and fixed inputs contribute to feature-level variability. I will show that this requires modifications in inference procedures for these models to ensure identifiability. I will demonstrate the utility of such variants with examples from genomics.

 

Dr Christopher Yau

Reader in Computational Biology, University of Birmingham Turing Fellow, The Alan Turing Institute