Neural Networks and Deep
Learning
António Ruano
[email protected]
Timetable, Objectives
• Timetable, Presential:
• T– Fridays, 18:00 – 20:00;
• PL – Fridays, 20:00-22:00
• Objectives: At the end of this lecture course the student should:
• Understand the biological inspiration of neural networks, their main characteristics and applications;
• To know multi-layer perceptrons and radial base function networks, as well as the most common
techniques of supervised training and adaptation; know how to use them in the context of systems
approximation and classification;
• To know the most common unsupervised training methods, based on Hebbian rules and on competition,
and the Kohonen networks; know how to use them in data clustering and classification problems.
• To know the main recurrent neural networks, and their training algorithms; know how to use them in
practical applications;
• To know the main mechanisms of deep learning; know how to use deep belief and convolutional
networks in practical applications;
Contents
• 1. Introduction to Artificial Neural Networks
• 2. Supervised Learning for Feedforward Neural Networks
• Models
• Learning Algorithms
• Applications to function approximation and pattern recognition
• 3. Unsupervised Learning
• Non-Competitive Learning
• Competitive Learning
• 4. Recurrent Neural Networks
• Introduction to nonlinear dynamic systems
• Models
• Training and Operation
• Applications of Recurrent Neural Networks
• 5. Deep Learning
• Revisiting Gradient Descent Algorithms
• Restricted Boltzmann Machines and Deep Belief Networks
• Convolutional Neural Networks
Assessment and Bibliography
• Assessment
• In this discipline there will be a group mini-project, with a weight of 50% of the grade, and an individual
project, contributing also with 50% to the final grade. These will be assessed during the classes. If the final
grade is equal to or higher than 10, and both components are graded with a mark higher than 8, the
student passes.
• Bibliography
• Apontamentos on-line
• Ruano, A. E. (Ed.). Intelligent Control using Intelligent Computational Techniques. 2005: The Institution of
Electrical Engineers.
• Gori, M., Machine Learning. A Constraint-based Approach. 2018: Morgan Kaufmann.
• Haykin, S.O., Neural Networks and Learning Machines. 3 ed. 2008: Prentice Hall.
• Aggarwal, C.C., Neural Networks and Deep Learning. A Textbook. 2018: Springer.
• Buduma, N. and N. Locascio, Fundamentals of Deep Learning: Designing Next-Generation Machine
Intelligence Algorithms. 2017: O¿Reilly Media.
• Skansi, S., Introduction to Deep Learning: From Logical Calculus to Artificial Intelligence. 2018: Springer
International Publishing.
