Course Title

Deep Learning

Course Code

SDS 422

Course Type

Elective

Level

Master’s

Year / Semester

2^nd Semester

Instructor’s Name

Assoc. Prof. Mihalis A. Nicolaou (Lead Instructor), Dr Charalambos Chrysostomou

ECTS

5

Lectures / week

2

Laboratories / week

1

Course Purpose and Objectives

To teach students the fundamentals of neural networks and deep learning, including concepts such as activation functions, optimization algorithms, and different types of layers (E.g., recurrent, attention) and architectures, addressing both generative and discriminative learning. Developed knowledge will be linked to real-world applications and datasets, using modern deep learning frameworks.

Learning Outcomes

By the end of the course, students will be able to:

• Explain basic principles behind deep learning and neural networks
• Explain particular architectures, identifying their properties and being able to articulate suitability for a particular problem
• Select, implement, and apply the appropriate deep learning algorithms for given tasks, and datasets of different properties
• Rigorously evaluate the performance of deep learning algorithms on target datasets and applications
• Use modern deep learning frameworks such as pytorch and tensorflow.
• Explain and develop solutions for state-of-the-art applications of deep learning

Prerequisites

Course Content

Week 1. Introduction to Neural Networks and Deep Learning and modern GPU- accelerated deep learning frameworks.

Week 2. Neural Network Optimization. Gradient Descent, Backpropagation, Regularization, momentum, normalization.

Week 3. Convolutional Neural Networks. Architecture design (e.g., layers, normalization, pooling) loss functions, properties

Week 4. Deep Learning for Sequential Data. Recurrent Neural Networks, attention- based architectures

Week 5. Generative Models: Introduction to generative models in deep learning covering current algorithms such as Generative Adversarial Networks and Diffusion Models.

Week 6. Generative Models II: Transformer-based Architectures and applications.

Teaching Methodology

Lectures, Labs

Bibliography

• Goodfellow, Bengio A. Courville, “Deep Learning”, MIT press
• Aston Zhang, Zachary Lipton, Mu Li, Alexander Smola - Dive into Deep Learning
• Michael Nielsen- Neural Networks and Deep Learning

Assessment

Combination of coursework and exam

Language

English