Prof. Emo Welzl and Prof. Bernd Gärtner
|Mittagsseminar Talk Information|
Date and Time: Thursday, September 25, 2014, 12:15 pm
Duration: 30 minutes
Location: CAB G51
Speaker: Florian Meier
Hinton et al. found an efficient way to train Deep Believe Networks in the mid 2000's. Now they are the state of the art in tasks like feature detection and object recognition. First I will explain the basic model and learning algorithms. The aim of the master thesis was making the model biologically more plausibly, especially changing the network architecture such that it is similar to the one in the human brain. I will present some of the made progress and some of the encountered difficulties.
Automatic MiSe System Software Version 1.4803M | admin login