- This event has passed.
Deep Reinforcement Learning Series – Part 3: function approximators
May 13, 2017 @ 10:00 am - 11:00 am
Metis is hosting Chi-Town Machine Learning for their Deep Reinforcement Learning Series – Part 3: Function Approximators with Jeremy Watt.
Breakfast will be provided!
In this series of tutorial talks we will be Deep Reinforcement Learning from start to finish – the tech powering self-playing Atari games, Alpha Go, problems in automatic control and more.
In Part 3 of the series we will be covering extensions of Q-Learning to problems – like chess, control, and video games – where the enormous size of the state space makes resolving Q – directly – impossible. As cycling through even a reasonable portion of the states is computationally impossible, function approximators are introduced to greatly generalize the effect of a Q-Learner.
This talk will be highly interactive with a number of live code demonstrations and a fully featured Jupyter Notebook.
If you did not attend previous events in this series be sure to check out the resources listed in those events – including slides, blog posts, etc.,