Developing core–technologies for tree-based models

Krzysztof Bartoszek, LiU and Niklas Wahlberg, LU

Graph learning is an important topic in science and technology. In this project we will develop new methods for tree-structured graphs, motivated by phylogenetics but with potential use in other network-based applications e.g. network routing, and discovery and analysis of gene-regulatory networks. We will develop probabilistic models for tree-structured graphs that enable learning these graphs from data while reasoning about the uncertainties in the learnt structures. We will also contribute to the development of probabilistic programming languages (PPLs), on which there is currently a lot of focus in the AI and ML communities, for automatic and efficient inference in these models.

Project number: C02