December 6-án, csütörtökön 10:15-től Csáji Balázs Csanád Stochastic Approximation témában tart tutorialt Kende utcai épületünk Nagytanácstermében.
A tutorial absztraktja: Stochastic Approximation (SA) is a mathematical framework for recursive (on-line) algorithms working with uncertain data. SA methods are fundamental for many fields, including machine learning, system identification, signal processing and adaptive control. This tutorial-style talk first builds up some intuitions behind SA methods, then presents several important examples, such as the Robbins-Monro algorithm, the Kiefer-Wolfowitz method and a number of other stochastic gradient descent methods, as well as typical SA algorithms from reinforcement learning. Finally, a few theoretical results are surveyed which ensure the strong consistency of these methods.
Az esemény nyilvános, minden érdeklődőt szeretettel várunk!