Detail of the student project

Topic:[Deep Learning] Stochastic Binary Neural Networks
Department:Katedra kybernetiky
Supervisor:Mgr. Oleksandr Shekhovtsov, Ph.D.
Announce as:Diplomová práce, Bakalářská práce, Semestrální projekt
Description:Binary networks use only binary weights and activations. A forward computation is extremely efficient and the hardware support is growing. This makes binary networks very computation and energy efficient and attractive for mobile and autonomous robotics applications.
Training such models is feasible with stochastic relaxation methods.
You will work on the efficient implementation of the methods developed at the department. Experimental comparison of different methods. Theoretical improvement of methods.
Responsible person: Petr Pošík