Abstract
In this talk, Parviainen will give an overview of Bayesian networks: what are they, what can be done with them and where do they come from? They were invented in the 1980s, mainly as a result of research by Judea Pearl. Bayesian networks are probabilistic graphical models that represent relations between random variables. They enable efficient probabilistic inference and causal analysis and are found “under the hood” in many probabilistic models. They have been significant in reasoning with uncertainty in Artificial Intelligence. Baysian networks are now dominating the field of uncertainty reasoning.
Biography
Pekka Parviainen is an associate professor in machine learning at the University of Bergen. His research interests are concentrated on probabilistic models in machine learning. Especially, he has studied structure learning in Bayesian networks. Before Pekka came to Bergen, he was employed at Aalto University, Finland. He wrote his PhD titled “Algorithms for Exact Structure Discovery in Bayesian Networks” at the University of Helsinki in 2012.