Prof. Emo Welzl and Prof. Bernd Gärtner
|Mittagsseminar Talk Information|
Date and Time: Thursday, November 16, 2017, 12:15 pm
Duration: 30 minutes
Location: CAB G51
Speaker: Asier Mujika
Processing sequential data of variable length is a major challenge in a wide range of applications, such as speech recognition, language modeling, generative image modeling and machine translation. We address this challenge by proposing a novel recurrent neural network (RNN) architecture, the Fast-Slow RNN (FS-RNN). The FS-RNN incorporates the strengths of both multiscale RNNs and deep transition RNNs as it processes sequential data on different timescales and learns complex transition functions from one time step to the next. We evaluate the FS-RNN on character level language modeling data sets. Our approach outperforms the best known compression algorithms on Wikipedia data. We also present an empirical investigation of the learning and network dynamics of the FS-RNN, which explains the improved performance compared to other RNN architectures. The first part of this talk will cover the basics of recurrent networks, such that anyone should be able to follow the rest. Joint work with Florian Meier and Angelika Steger.
Automatic MiSe System Software Version 1.4803M | admin login