Machine Learning


The main objective of the Seminar on Machine Learning is to give PhD students a broad understanding of what Machine Learning is, and a framework to understand both the overarching features and problems of Machine Learning, the various approaches and their underlying principles.

The other objectives are to:

  • Give a structured view of various particular approaches to Machine Learning.
  • Show the various machine learning research areas at NOVA IMS.
  • Provide into the state-of-the-art of particular topics in Machine Learning
  • Discuss the ethical, legal, and societal issues of the use of Machine Learning.
  • Help the students write a scientific paper in the area of Machine Learning, preferably with a practical application of a Machine Learning algorithm to the problem of their thesis, although other themes are possible

General characterization





Responsible teacher

Victor José de Almeida e Sousa Lobo


Weekly - Available soon

Total - Available soon

Teaching language

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


There are no requirements beyond those of admission to the PhD program.


Mitchell T.M. (1997) ¿Machine Learning¿, ISBN 0-07-115467-1, McGraw-Hill, 1997, ISBN 0-07-115467-1

Pedro Domingos (2015), ¿The Master Algorithm: How the Quest for the Ultimate Learning Machine Will Remake Our World¿, ISBN:978-0465065707

Russel & Norvig (2016), ¿Artificial Intelligence¿, Pearson Education, 2016. ISBN 9781292153964.

Bishop, C. M. (2011), ¿Pattern Recognition and Machine Learning¿ (Information Science and Statistics). Berlin, Heidelberg: Springer-Verlag. ISBN:978-0387310732.

Bishop C., (1995) ¿Neural Networks for Pattern Recognition¿, Oxford University Press

Géron, A., (2017) ¿Hands-On Machine Learning with Scikit-Learn and TensorFlow: Concepts, Tools, and Techniques to Build Intelligent Systems¿ O¿Reilly Media Inc.

Dem¿ar, J. (2006). Statistical Comparisons of Classifiers over Multiple Data Sets. J. Mach. Learn. Res., 7, 1¿30.

Guyon, I. (2003). Design of experiments of the NIPS 2003 variable selection benchmark. In Advances in Neural Information Processing Systems 16 (NIPS 2003). Vancouver, British Columbia, Canada.

Langley, P. (2000). Crafting Papers on Machine Learning. In ICML.

Teaching method

This course is based on theoretical sessions, invited lectures, a hands-on workshop, and student presentations. The theoretical sessions include the presentation of the main paradigms and principles of machine learning. The invited lectures address cutting edge issues in machine learning. The hands-on workshop is intended to familiarize students with the development and implementations of prototypes and experiences in the area of machine learning. In addition, in the hands-on workshop, students will be assisted in the development of their self-learning projects that will form the basis of their research articles.

Evaluation method

Evaluation of a paper written by the student (100%).

That paper should be submitted to a Conference or Scientific Journal, and will have a better grade if it is. However, if the student¿s thesis is unrelated to Machine Learning, this paper may be a review or simple application paper with a format compatible with a conference or journal paper, without actually being submitted.


Subject matter

1. History, definitions, and philosophical issues of Machine Learning

2. Review of basic concepts concerning Data, Curve Fitting, and types of problems.

3. Fundamental and general issues in Machine Leaning: Generalization, Error measures, Model Selection, available data and degrees of freedom of models, explainability and causality.

4. Overview of the ¿5 tribes¿ of Machine Learning: Bayesians, Analogisers, Connectionists, Evolutionaries, and Symbolists.

5. Bayesian based approaches. Decision Theory. MAP, ML, Naïve bayes, Belief Networks, etc.

6. Analogy based approaches. Nearest Neighbors, Case Based Reasoning, Lazy Learning, etc.

7. Connectionist based approaches. Neural Networks, Deep Learning, etc.

8. Evolution (and other biology) based approaches. Genetic Algorithms, Genetic Programming, ant colonies and swarms.

9. Symbolist based approaches. Programming in Logic, Expert Systems, Classification and Regression Trees.

10. Hands-on workshop with a machine learning software package.

11. Invited lectures by experts in specific topics of machine learning.


Programs where the course is taught: