Seminar: Squared Neural Probabilistic Models by Visiting Professor Dino Sejdinovic

02 Jul 2025 10.00 AM - 11.00 AM LT5 Current Students, Industry/Academic Partners

Abstract:
We describe a new class of probabilistic models, squared families, where densities are defined by squaring alinear transformation of a statistic and normalising with respect to a base measure. Key quantities, such as the normalisingconstant and certain statistical divergences, admit a helpful parameter-integral decomposition giving a closed formnormalising constant in many cases of interest. Parametrising the statistic using neural networks results in highly expressive yet tractable models, with universal approximation properties. This approach naturally extends to otherprobabilistic settings, such as modelling point processes. We illustrate the effectiveness of squared neural probabilisticmodels on a variety of tasks, demonstrating their ability to represent complex distributions while maintaining analyticaland computational advantages. Joint work with Russell Tsuchida, Jiawei Liu, and Cheng Soon Ong.

 

Speaker profile:

Dino Sejdinovic is a Professor of Statistical Machine Learning at Adelaide University (since 2022),where he is affiliated with the Australian Institute for Machine Learning (AIML) and the Responsible AI Research Centre(RAIR). He also holds Visiting Professor appointments with the Nanyang Technological University, Singapore and the Institute of Statistical Mathematics, Tokyo. He was previously an Associate Professor at the Department of Statistics,University of Oxford and a Turing Faculty Fellow of the Alan Turing Institute. He held postdoctoral positions at theUniversity College London and the University of Bristol and received a PhD in Electrical and Electronic Engineering fromthe University of Bristol (2009). His research spans a wide variety of topics at the interface between machine learning andstatistical methodology, including large-scale nonparametric and kernel methods, robust and trustworthy machine learning,causal inference, and uncertainty quantification.

 

Categories