Detail of the student project

Topic:[Deep Learning] Binary Neural Architecture Search
Department:Katedra kybernetiky
Supervisor:Mgr. Oleksandr Shekhovtsov, Ph.D.
Announce as:Diplomová práce, Bakalářská práce, Semestrální projekt
Description:Apply techniques of Bayesian Optimization to search for the configuration of a neural network giving the best validation performance, along the lines of "DARTS: Differentiable Neural Architecture Search". Of particular focus are neural networks with binary weights and activations. Experimental verification of methods developed at the department, research on the problem formulation, connection to hypergradient techniques.
Responsible person: Petr Pošík