Machine Learning II

Credits
6
Types
Compulsory
Requirements
This subject has not requirements, but it has got previous capacities
Department
CS;TSC
The goal of machine learning is the development of theories, techniques and algorithms that allow a system to modify its behaviour by observing data that represents incomplete information about a process or phenomenon subject to statistical uncertainty. Automatic learning is a meeting point for different disciplines: multivariate statistics, artificial intelligence, algorithmics and mathematical optimization, among others.The course takes advantage of modern learning techniques based on data, including deep artificial neural networks, reinforcement learning and kernel based methods. A side goal is to get familiar with the corresponding computing environments.

Teachers

Person in charge

  • Xavier Giró Nieto ( )

Others

  • Luis Antonio Belanche Muñoz ( )

Weekly hours

Theory
2
Problems
0
Laboratory
2
Guided learning
0.333
Autonomous learning
5

Competences

Technical Competences

Technical competencies

  • CE1 - Skillfully use mathematical concepts and methods that underlie the problems of science and data engineering.
  • CE2 - To be able to program solutions to engineering problems: Design efficient algorithmic solutions to a given computational problem, implement them in the form of a robust, structured and maintainable program, and check the validity of the solution.
  • CE3 - Analyze complex phenomena through probability and statistics, and propose models of these types in specific situations. Formulate and solve mathematical optimization problems.
  • CE4 - Use current computer systems, including high performance systems, for the process of large volumes of data from the knowledge of its structure, operation and particularities.
  • CE5 - Design and apply techniques of signal processing, choosing between different technological tools, including those of Artificial vision, speech recognition and multimedia data processing.
  • CE6 - Build or use systems of processing and comprehension of written language, integrating it into other systems driven by the data. Design systems for searching textual or hypertextual information and analysis of social networks.
  • CE8 - Ability to choose and employ techniques of statistical modeling and data analysis, evaluating the quality of the models, validating and interpreting them.
  • CE9 - Ability to choose and employ a variety of automatic learning techniques and build systems that use them for decision making, even autonomously.

Transversal Competences

Transversals

  • CT3 - Efficient oral and written communication. Communicate in an oral and written way with other people about the results of learning, thinking and decision making; Participate in debates on topics of the specialty itself.
  • CT5 - Solvent use of information resources. Manage the acquisition, structuring, analysis and visualization of data and information in the field of specialty and critically evaluate the results of such management.
  • CT7 - Third language. Know a third language, preferably English, with an adequate oral and written level and in line with the needs of graduates.

Generic Technical Competences

Generic

  • CG1 - To design computer systems that integrate data of provenances and very diverse forms, create with them mathematical models, reason on these models and act accordingly, learning from experience.
  • CG2 - Choose and apply the most appropriate methods and techniques to a problem defined by data that represents a challenge for its volume, speed, variety or heterogeneity, including computer, mathematical, statistical and signal processing methods.
  • CG3 - Work in multidisciplinary teams and projects related to the processing and exploitation of complex data, interacting fluently with engineers and professionals from other disciplines.
  • CG4 - Identify opportunities for innovative data-driven applications in evolving technological environments.
  • CG5 - To be able to draw on fundamental knowledge and sound work methodologies acquired during the studies to adapt to the new technological scenarios of the future.

Objectives

  1. Organize the flow of solution of a machine learning problem, analyzing the possible options and elegint the mismatches to the problem
    Related competences: CE8, CG2,
  2. Decide, defend and criticise a solution for a machine learning problem, arguing the strong and weak points of appropriation.
    Related competences: CE1, CE2, CE3, CE4, CE5, CE6, CE8, CE9, CG1, CG2,
  3. Know and know how to apply advanced techniques to solve non-supervised learning problems, especially clustering.
    Related competences: CE1, CE2, CE4, CT5, CG1, CG2,
  4. Know how to apply deep feed-forward multilayer neuronal network techniques to solve complex supervised learning problems.
    Related competences: CE1, CE3, CE4, CT7, CG1, CG5,
  5. Know and apply advanced techniques of learning methods based on kernel functions, for the resolution of learning problems, both supervised and unsupervised.
    Related competences: CE1, CE2, CE3, CE4, CE8, CT5, CT7, CG1, CG2,
  6. To know and to know how to apply the advanced techniques for the resolution of learning problems by reinforcement, and its relation with techniques of deep learning.
    Related competences: CE1, CE2, CE3, CE4, CE5, CE6, CT5, CG1, CG2, CG4,
  7. Know how to identify problems involving signal processing, such as data in the form of audio, image or video, or a combination of you, and solve them with advanced computer learning techniques.
    Related competences: CE1, CE2, CE3, CE4, CE5, CE8, CE9, CT5, CT7, CG1, CG2, CG3, CG4,
  8. Know how to identify problems involving human or non-human speech processing, such as data in the form of audio or text, or a combination thereof, and solve them with advanced computational learning techniques
    Related competences: CE1, CE2, CE3, CE4, CE6, CT3, CT7, CG1, CG2, CG3, CG4, CG5,

Contents

  1. Introduction to Machine learning
    Description and approach of the problems attacked by machine learning. Identification of modern areas and problems (data, scalability, heterogeneity, etc). Introduction to modern and advanced techniques covered in progress. Advanced application examples.
  2. Introduction to kernel based methods
    Kernel ridge regression. Feature maps. Introducció als espais de Hilbert. Representer Theorem. Spectral clustering.
  3. Non-standard kernel functions
    Non-standard kernel functions: refresh and expand the kernel function definition. Learning in RKHS. Kernel functions for text analysis, graphs, biomedical data (-OMIC).
  4. Advanced kerned methods
    Advanced kernel methods: kPCA, KFDA, Relevance Vector Machines. Advanced applications.
  5. Neural layers
    - Perceptron and multilayer perceptron.
    - Convolutional layers.
    - Recurrent layers.
    - Residual layers
    - Mechanisms of attentio
  6. Training of deep neural networks
    - Backpropagation.
    - Loss functions.
    - Optimizers.
    - Methodology.
    - Data augmentation.
    - Batch normalization

Activities

Activity Evaluation act


Development of subject 1


Objectives: 1
Contents:
Theory
1h
Problems
0h
Laboratory
0h
Guided learning
0h
Autonomous learning
2h

Development of subject 2


Objectives: 3
Contents:
Theory
4h
Problems
0h
Laboratory
4h
Guided learning
0h
Autonomous learning
10h

Development of subject 3


Objectives: 5
Contents:
Theory
4h
Problems
0h
Laboratory
4h
Guided learning
0h
Autonomous learning
10h

Development of subject 4


Objectives: 2 5 7 8
Contents:
Theory
6h
Problems
0h
Laboratory
6h
Guided learning
0h
Autonomous learning
15h

Development of subject 5


Objectives: 2 4 7 8
Contents:
Theory
6h
Problems
0h
Laboratory
6h
Guided learning
0h
Autonomous learning
15h

Development of subject 6


Objectives: 1 2 4 7 8
Contents:
Theory
6h
Problems
0h
Laboratory
6h
Guided learning
0h
Autonomous learning
15h

Development of subject 7


Objectives: 7 8 1 2 3 6
Theory
3h
Problems
0h
Laboratory
4h
Guided learning
0h
Autonomous learning
8h

Practicum evaluatio


Objectives: 1 2
Week: 10
Type: assigment
Theory
0h
Problems
0h
Laboratory
0h
Guided learning
2h
Autonomous learning
0h

Practicum report


Objectives: 1 2 3 4 5 6 7 8
Week: 15
Type: assigment
Theory
0h
Problems
0h
Laboratory
0h
Guided learning
3h
Autonomous learning
0h

Teaching methodology

The theory classes introduce all the knowledge, techniques, concepts and results necessary to reach a well-founded level. These concepts are put into practice in the laboratory classes, in which the student learns to develop machine learning solutions to real problems of a certain complexity.

In the problem classes we will delve deeper to understand the theory by problem solving or by extending the concepts already seen.

In the laboratory classes, code is provided in various computing environments that allow solving a problem completely with the technique or techniques corresponding to the current topic. This laboratory also serves as a guide for the corresponding part of the practical works developed by the students.

Evaluation methodology

The subject is evaluated through a partial exam, a final exam and practical work in which real problems are addressed, writing the corresponding reports.

The partial exam will correspond to the kernel-based methods part and will discard this part of the syllabus.

The final grade is calculated as:

Grade = 0,4 * Jobs + 0,6 * (Final + Partial) / 2

For the students who attend the re-evaluation, the re-evaluation exam note will include the two parts of the syllabus and will substitute 0.6 * (Final + Partial) / 2.

Bibliography

Basic:

Previous capacities

Programming in Python.

Addendum

Contents

NO HI HA CANVIS RESPECTE LA INFORMACIÓ PUBLICADA A LA GUIA DOCENT.

Teaching methodology

Les sessions de teoria es realitzaran de forma no presencial. Les sessions de problemes/laboratori es realitzaran amb els ordinadors personals de cada estudiant, que hauran de portar, en aules electrificades i amb accés sense fils a Internet mitjançant la xarxa Eduroam.

Evaluation methodology

NO HI HA CANVIS RESPECTE LA INFORMACIÓ PUBLICADA A LA GUIA DOCENT.

Contingency plan

Les sessions de problemes/laboratori es passaran a un format en línia síncron en cas que es cancel·lés tota activitat lectiva físicament presencial. Els exàmens (parcial, final, reavaluació), en cas de no poder fer-se presencials, es farien no presencials, via "Pràctiques" del racó o via Atenea.