[Pycon] [new paper] "Alexander Hendorf" - Deep Learning with PyTorch for Fun and Profit (Part III / Italian Edition: Divina Commedia)

[info a pycon.it]

Title: Deep Learning with PyTorch for Fun and Profit (Part III / Italian Edition: Divina Commedia)
Duration: 60 (includes Q&A)
Q&A Session: 15
Language: en
Type: Talk

Abstract: There are all these great articles and blog posts about Deep Learning describing all that awesome stuff. - Is it all that easy? Let's check! 

We'll look into: style transfer (making a picture look like painting), speech generation (like Siri or Alexa) and text generation (writing a story). In this talk I'll describe the whole journey: A fun ride from the idea to the very end including all the struggles, failures and successes.

This is an ongoing talk on how far we can get creating a full [radio drama (Hörspiel)][1] with deep learning and the resources required.
  
Steps, we'll cover:

 - The data challenge: get the data ready
 - Creating a character-level language models with an Recurrent Neural Network
 - Creating a text generator
 - Creating artwork
 - Synthesising speech
 - Making it sound like a real person

In this special PyCon X edition we will also try to recreate text in the style of Dante' Divina Commedia.

  [1]: https://en.wikipedia.org/wiki/Radio_drama


Tags: [u'art', u'deep learning', u'Pytorch', u'nlp', u'Artificial Intelligence']