6th Deep Learning Meetup in Vienna

Hi Deep Learners,

Our next meetup is very soon: Wednesday, 12. Oct, at Sektor 5 again.

As we know we have both beginners and advanced people in our meetup, we try to balance for this, having both an entry level and an advanced talk. This time, we are going to have the following two talks and speakers:

• Benjamin Freundorfer: An Intro to Neural Networks

This is an intro to the building blocks of neural networks which are also the building blocks of deep neural nets. We'll try to get a deep intuitive understanding of basic questions like: How can neural nets be so powerful? What does a single neuron actually do? What are the problems with deeper neural nets?

• Kornél Kis: Deep learning in practice - a Text-to-Speech system built with neural networks

This talk will present a deep neural network based Text-to-Speech system. Through this use-case, we could see the full 'life-cycle' of a deep learning project: the preprocessing, the initial parameter settings, the learning phase, then some hyperparameter optimization. The techniques discussed here can hopefully be adopted to several other, very different projects. There will also be slides about ensemble architectures/ ensemble learning. This part is especially useful for more experienced machine learning practicioners, who have yet to add neural networks to their toolbox.

As usual we are also going to have a 'hot topics' section and then open the floor then to discussions about above and other topics.

If you have some 'hot topics' to share, let us know please.

Looking forward to the next meetup!

Tom