Code |
14499
|
Year |
1
|
Semester |
S1
|
ECTS Credits |
6
|
Workload |
OT(15H)
|
Scientific area |
Informatics
|
Entry requirements |
Know how to program and have notions of probability and statistics.
|
Mode of delivery |
Face-to-face
|
Work placements |
Not applicable.
|
Learning outcomes |
Introduce the concepts, models and language adequate to problem solving using neural networks. At the end of this course unit the student should be able to solve problems using neural networks and eventually be able to propose improvements to current methods.
|
Syllabus |
1-Feed-forwd and recurrent networks. 2-Networks for clustering, classification and regression. 3-Reservoir networks. 4-Learning algorithms. 5-Topologies. 6-Cost functions. 7-Performance evaluation. 8-Shallow versus deep networks. 9-Applications to signal and image processing.
|
Main Bibliography |
Main: -Deep Learning, Ian Goodfellow and Yoshua Bengio and Aaron Courville, MIT Press book, 2016 -Neural Networks and Learning Machines (3rd Edition), S. Haykin, Prentice Hall; 3 edition, 2008
Complementary: -Pattern Recognition and Machine Learning, C. Bishop, Springer, 2006 -Several journals in this area.
|
Teaching Methodologies and Assessment Criteria |
Assessment is done through two practical assignments (each worth 50% of the final grade). Each assignment includes a written report and a public presentation. The first work is a review of articles in the area and the second has a practical component linked to the implementation of one of the approaches studied.
|
Language |
Portuguese. Tutorial support is available in English.
|