Deep Learning

You are here

Credits
4.5
Types
Elective
Requirements
This subject has not requirements, but it has got previous capacities
Department
BSC;CS
This subject aims to familiarize the student with theoretical and practical aspects of Deep Learning (DL) techniques. The lectures will refresh the basic concepts of DL (CNNs, etc.), assuming some prior knowledge. Some the most popular architectures will be introduced, as well as neural network configurations that have proved useful for specific problems. In the practical part, the student will have to carry out experiments using DL libraries, and experiment with the various components that have been proposed in the field.

Teachers

Person in charge

  • Enrique Romero Merino ( )
  • Luis Antonio Belanche Muñoz ( )

Others

  • Caroline König ( )

Weekly hours

Theory
2
Problems
0
Laboratory
0.3
Guided learning
0.3
Autonomous learning
4.93

Competences

Technical Competences of each Specialization

Professional

  • CEP3 - Capacity for applying Artificial Intelligence techniques in technological and industrial environments to improve quality and productivity.
  • CEP4 - Capability to design, write and report about computer science projects in the specific area of ??Artificial Intelligence.

Transversal Competences

Reasoning

  • CT6 - Capability to evaluate and analyze on a reasoned and critical way about situations, projects, proposals, reports and scientific-technical surveys. Capability to argue the reasons that explain or justify such situations, proposals, etc..

Analisis y sintesis

  • CT7 - Capability to analyze and solve complex technical problems.

Objectives

  1. Understand the various techniques that can be integrated into a deep learning system, and know how to experiment with them coherently in a realistic production environment through the use of third-party libraries.
    Related competences: CT7, CEP3,
  2. Be able to understand scientific articles from the area of deep learning, to extract the most relevant conclusions, and to derive possible applications or limitations.
    Related competences: CT6, CEP4,

Contents

  1. Introduction and refresher
    Fundamental necessary concepts of ML will be reviewed, with special focus on those most influential to deep learning.
  2. Convolutional Neural Networks
    We will review the main aspects of Convolutional Neural Networks. How they work, why, and how can they be improved.
  3. Embeddings and autoencoders
    Autoencoders, Variational autoencoders, their modifications and uses. Embeddings. Transfer learning.
  4. Few-shot and zero-shot learning
    Zero-shot learning (ZSL), One-Shot Learning (OSL), and Few-Shot Learning (FSL) as approaches to handling limited training data in machine learning
  5. Diffusion Networks
    We will review the main aspects of Diffusion Networks. How they work, why, and how can they be used and trained.
  6. Transformer Networks
    We will review the main aspects of Transformer Networks. How they work, why, and how can they be used and trained.
  7. Graph Neural Networks
    We will review the main aspects of Graph Neural Networks. How they work, why, and how can they be used and trained.
  8. Interpretability and related issues. Negative deep learning
    We discuss interpretability aspects of DL and related issues. How negative deep learning can affect performance.

Activities

Activity Evaluation act


Practical experimentation

Experimentation using deep learning libraries, and reporting of the relevant conclusions.
Objectives: 1
Week: 13 (Outside class hours)
Theory
0h
Problems
0h
Laboratory
0h
Guided learning
0h
Autonomous learning
34h

Exam

Exam
Objectives: 1
Week: 13 (Outside class hours)
Theory
2h
Problems
0h
Laboratory
0h
Guided learning
0h
Autonomous learning
10h

Introduction and Refresher


Objectives: 1 2
Theory
4h
Problems
0h
Laboratory
0h
Guided learning
0h
Autonomous learning
4h

Convolutional Neural Networks


Objectives: 1 2
Theory
6h
Problems
0h
Laboratory
0h
Guided learning
0h
Autonomous learning
6h

Lab on Convolutional Neural Networks


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

Embeddings and autoencoders


Objectives: 1 2
Theory
4h
Problems
0h
Laboratory
0h
Guided learning
0h
Autonomous learning
4h

Few-shot and zero-shot learning


Objectives: 1 2
Theory
4h
Problems
0h
Laboratory
0h
Guided learning
0h
Autonomous learning
4h

Diffusion Networks


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

Transformer Networks


Objectives: 1 2
Theory
4h
Problems
0h
Laboratory
0h
Guided learning
0h
Autonomous learning
4h

Lab on embeddings, autoencoders, few-shot and zero-shot learning


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

Graph Neural Networks


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

Interpretability and related issues. Negative deep learning


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

Teaching methodology

This subject has a theoretical component and a practice.

The theoretical component consists of face-to-face classes where the teacher will review concepts of Deep Learning, present applications, and other recent trends in the field.

The practical component is composed by two group practicals, where students will have to experiment with the various techniques of Deep Learning. Based on simple experiments, and using popular Deep Learning libraries (e.g., Keras, TensorFlow, Pytorch, ...), the students will test the effects of the various available techniques.
At the end of the course, there will be an exam.

Evaluation methodology

This subject will be evaluated taking into account the theoretical and practical aspects:

P1: grade of practical 1
P2: grade of practical 2
E: grade of the exam

The final grade will be computed as: 0.4 * P1 + 0.4 * P2 + 0.2 * E

Bibliography

Basic:

Previous capacities

Basic concepts of neural networks (SGD, back-propagation, loss functions) and machine learning (classification, regression, evaluation methodologies) are required.

Students must be able to program autonomously (Python), to work on a remote server through a terminal (ssh, bash), and to interact with third-party libraries.