Speaker: Csaba Szepesvari (University of Alberta)
Research Centre: Research Centre for Machine Learning
Most AI researchers would agree that machine learning is key to "solving AI" and most of what we do on machine learning hinges upon a statistical approach to learning. In this talk, after shortly clarifying what do I mean by statistical learning and why I say that most of machine learning hinges upon statistical learning, I will explore evidence in support and also evidence against that statistical learning can "solve" various AI tasks. I will conclude with discussing approaches that go beyond statistical learning.
Csaba Szepesvári is the team-lead for the "Foundations" team at Deepmind, UK.
He is currently on leave from the University of Alberta, where he is a Professor of Computing Science. His interest is artificial intelligence (AI) and, in particular, principled approaches to AI that use machine learning. He has extensive industrial and academic experience, including various positions at leading journals and conferences. He is the co-inventor of UCT, a widely successful Monte-Carlo tree search algorithm, which ignited much work in AI, contributing significantly to the leap in performance of computer Go programs and eventually leading to Deepmind's AlphaGo which the top Go professional Lee Sedol in a landmark game. This work on UCT won the 2016 test-of-time award at ECML/PKDD where it was originally published in 2006.
Slides from this presentation can be found here.
Share this event
When and where
5.00pm - 6.00pmFriday 16th March 2018