Taller 1: Introduction and Limitation of Deep Learning

 
"A little learning is a dangerous thing; drink deep, or taste not the Pierian spring: there shallow draughts intoxicate the brain, and drinking largely sobers us again."

Alexander Pope 
Instructores:

• Ulises Moya Sánchez, BSC/UAG
• Ulises Cortés, BSC

Resumen
This course presents an introduction of the bases and practices of Deep Learning (DL). We will focus in  Convolutional Neural Networks for  image classification in large datasets in  the hands-on session.
 
Contenido

1. Introduction to Neural networks and Deep Learning.
2. Understand the relations and differences between shallow and deep nets.
3. Main characteristics  and  DL  net selection (Boltzmann machines, convolutional nets,  recurrent nets).
4. Application and use cases  of DL methods.
5. Convolutional neural networks application: CNN architectures, Training, fine tuning and regularization (Hands on session)
6. Parallel computing tools for DL (Data augmentation)
7. DL limitations (noise, size, overfitting).

Referencias:
Textbook
Ian Goodfellow and Yoshua Bengio and Aaron Courville
Deep Learning
MIT Press 2016
http://www.deeplearningbook.org

Software
Keras (as wrapper) on TensorFlow
https://keras.io/