Machine Learning Meetup in Prague, Czech Republic
Google DeepMind recently published AlphaGo Zero, a new version of its super-human computer player of the Go board game that does not need any training data anymore! Conceptually, this is a much simpler system than the original AlphaGo, while it scales to even higher playing strength using only a fraction of the hardware. After a short intro to Go, we will go through the system’s architecture and neural networks, talk about our experience of reproducing the system, and try to draw some more general lessons from DeepMind’s progress on AlphaGo.
Slides: https://goo.gl/qoAqko
Location / Místo: Paralelni Polis, Prague, Czech Republic
http://www.mlmu.cz
https://twitter.com/mlmucz
https://www.facebook.com/machinelearningmeetups
https://www.meetup.com/Prague-Machine-Learning/
https://www.meetup.com/Machine-Learning-Meetup-Brno/
