Fáth, Gábor (Morgan Stanley)
Machine Learning Techniques in Chess
and Go
There have been significant breakthrough developments in Artificial
Intelligence for board games recently. Tabula Rasa Learning for Deep Neural Networks,
combined with Monte Carlo Tree Search, is able to
achieve super-human performance levels in the games of Go and Chess, and
surpass thousands of years of meticulously assembled human knowledge within
hours. Can this be extended to other areas of general interest for humankind? A
great question that remains to be answered in the coming years. In the talk we
review how recent machine learning techniques are implemented for board games,
how far AI got in this domain, and what lessons the Go and Chess communities are able to learn from these new developments.
The talk is held in English!
Az előadás nyelve angol!
Date: Sep 11, Tuesday 4:15pm
Place: BME, Building „Q”, Room QBF13