[Seminar]MLDS Unit Seminar 2026-2 by Dr. Eshant English

Speaker: Dr. Eshant English, Postodoctoral Researcher, Deep Learning Theory Team, University of Tokyo

Title: Inference from Generative Processes 

Abstract: Generative modelling is a powerful tool for creating new data, but creating new data is not always the end goal. Sometimes we need to draw inferences from the data and applications we have. In this talk, I will show that generative models or, more generally, generative processes, are powerful tools for inference, such as for uncertainty quantification and statistical hypothesis testing. In the end, I will also comment on how these inference methods can reveal inherent fallacies in our generated data.