Programme

• 16:00-17:00
  Keynote Presentation by Christian A. Naesseth, Machine Learning Assistant Professor at University of Amsterdam, ELLIS Member (venue: L3.33 & L3.35)
• 17:00-19:30
  Poster session with bites and drinks (venue: Ground floor)

For logistic purposes, kindly register here.

Keynote Presentation details

Title: ‘Diffusions, flows, and other stories’

Generative models have taken the world by storm. Using generative modeling, a.k.a. generative AI, we can construct probabilistic approximations to any data-generating process. 

In the context of text, large language models place distributions over the next token, for images it is often a distribution over pixel color values, whereas for molecules it can be a combination of atom types, positions, and various chemical features. 

This talk will explore some of the dominant paradigms, applications, and recent developments in generative modeling.

Bio

Christian A. Naesseth is an Assistant Professor of Machine Learning at the University of Amsterdam, a member of the Amsterdam Machine Learning Lab, the lab manager of the UvA-Bosch Delta Lab 2, and an ELLIS member.

His research interests span statistical inference, uncertainty quantification, reasoning, and machine learning, as well as their application to the sciences. He is currently working on generative modelling (diffusions, flows, AI4Science), approximate inference (variational and Monte Carlo methods), uncertainty quantification and hypothesis testing (E-values, conformal prediction). 

Previously, he was a postdoctoral research scientist with David Blei at the Data Science Institute, Columbia University. He completed his PhD in Electrical Engineering at Linköping University, advised by Fredrik Lindsten and Thomas Schön.

At the 2024 NeurIPS Conference, his lab and collaborators will present 5 accepted papers:

1. Neural Flow Diffusion Models

2. Equivariant Neural Diffusion for Molecule Generation

3. Fast yet Safe: Early-Exiting with Risk Control

4. VISA: Variational Inference with Sequential Sample-Average Approximations

5. Variational Flow Matching for Graph Generation