Graduate School and Research Center in Digital Sciences

Seminar: Overview of Deep Learning Models

Grigory Antipov - PhD student at Eurecom and Orange Labs

Multimedia Communications

Date: December 11, 2014

The presentation is an overview of the main models and methods which are known by the common notion of "Deep Architectures". The presented models are split into supervised (Deep Feedforward Networks, Convolutional Neural Networks, Recurrent Neural Networks, LSTM) and unsupervised (Energy-Based Models, Deep Autoencoders) parts. Each deep architecture is considered both from theoretical and applicative perspectives.

Overview of Deep Learning Models