Deep Variational Information Bottleneck

Feb 21, 2020 - 3:00 pm to 4:00 pm

Campus, PAB 232 

Sebastian Wagner-Carena

Join us Friday Feb 21st at 3pm in PAB 232 on campus (and on zoom ) for the next meeting of the Stats and ML Journal Club. This week, Sebastian Wagner-Carena will be leading an introduction on Variational Information Bottlenecks, an interesting twist on VAEs that has exciting applications. See you then!


Title: Deep Variational Information Bottleneck
Abstract: We present a variational approximation to the information bottleneck of Tishby et al. (1999). This variational approach allows us to parameterize the information bottleneck model using a neural network and leverage the reparameterization trick for efficient training. We call this method "Deep Variational Information Bottleneck", or Deep VIB. We show that models trained with the VIB objective outperform those that are trained with other forms of regularization, in terms of generalization performance and robustness to adversarial attack.