Deep Learning

Objectives

Present the area of deep learning. Discuss different architectures for addressing problems with images, time series, simple observations. Be able to implement and train a neural network in Keras.

General characterization

Code

200180

Credits

3.5

Responsible teacher

Mauro Castelli

Hours

Weekly - Available soon

Total - Available soon

Teaching language

Portuguese. If there are Erasmus students, classes will be taught in English

Prerequisites

NA

Bibliography

Deep Learning. Ian Goodfellow, Yoshua Bengio, Aaron Courville. MIT Press, 2016.

Teaching method

Theoretical and practical classes.

Evaluation method

First epoch: deep learning project.

Second epoch: deep learning project.

More details will be provided during the course.

Subject matter

Introduction to deep learning.

History and cognitive basis of neural computation.

The perceptron / multi-layer perceptron.

The neural net as a universal approximator.

Optimization by gradient descent and Back propagation. 

Overfitting and regularization (Dropout). 

Convolutional Neural Networks (CNNs). 

Training with shared parameters: the convlutional model. 

Recurrent Neural Networks (RNNs). 

Exploding/vanishing gradients Long Short-Term Memory Units (LSTMs).

Programs

Programs where the course is taught: