Machine Learning and Knowledge Extraction
Code
9229
Academic unit
Faculdade de Ciências e Tecnologia
Department
Departamento de Informática
Credits
6.0
Teacher in charge
Nuno Miguel Cavalheiro Marques, Pedro Manuel Corrêa Calvente Barahona
Weekly hours
2
Total hours
3
Teaching language
Inglês
Objectives
1. Learn how do active research subject matters in MLKE.
2. Develop innovative research work in one of those subject matters.
3. Construct new algorithms by adapting the studied techniques.
4. Understand and change learned models for specific TDM problems.
5. Apply the methods to new problems and in the assessment of results.
Prerequisites
Requires knowledge on :
. Machine Learning and Neural Networks.
. Data and Text Mining techniques.
. Statistical and Computational Methods for Data Management and Text Processing.
Complementary reading materials can be provided in these topics (if requested by the student).
Subject matter
Course topics:
. Knowledge Extraction from Text and Time Series.
. Neural Networks and Deep Learning Methods.
. Sense disambiguation and Machine Translation.
. Final research project with particular focus on the applications of (at least one) of topics to real world data.
Bibliography
Tom Mitchell. 1997. "Machine Learning", McGraw Hill.
Simon Haykin (1998). Neural Networks: A Comprehensive Foundation (2nd
Goodfellow, et. al. 2016. "Deep Learning", Book in preparation for MIT Press. http://www.deeplearningbook.org/ (acessed Sep 2016).
Michael Nielsen. 2016. "Neural Networks and Deep Learning", on line book: http://neuralnetworksanddeeplearning.com/ (acessed Sep 2016).
Robert Dale, Hermann Moisl and Harold Sommers (eds.). 2000. "Handbook of Natural Language Processing”, Marcel Dekker, Inc., New York. Edition). Prentice Hall.
"Survey of Text Mining. Clustering, Classification and Retrieval". Michael W. Berry, editor. Springer, 2008.
More specific articles overviewing the state of the art or supporting the presentation of each problem will be used.
Teaching method
Lectures will present fundamentals on each module. A reading list and some exercises for each module will introduce topics for final exam.
Tutorial supervision will help the student to conclude the final project.
(Students with very good projects will be persuaded to submit a revised version of their work to an appropriate conference).
Evaluation method
Each module will present several base references.
Each module will present several working topics and recommended reading.
Final Exam written exam (3 parts), each part covering a course module (3 x 17% ).
Selected final project (in one or more modules):
. Early presentation of project results (15%).
. Project written report (20%).
. MLKE Final Workshop (15%).
Final course grade is only known after final presentation. Acceptance information will be given after the last part of exam.