Details - Tematika

Deep Learning (GAINBAN-MELYTANU-1)

Basic data

Name and type of the study programme

Computer Science Engineering, undergraduate program

Curriculum

2026

Classes / consultation hours

2 + 0 + 2 (L+S+Labs)

Credits

5 credits

Theory – Practice

Theory: 50%, Practice: 50%

Recommended semester

Semester 5

Study mode

full-time

Prerequisites

Evaluation type

Mid-term evaluation

Course category

Compulsory elective in the specialization

Language

English

Instructors

Responsible instructor

Dr. Drenyovszki Rajmund

Responsible department

Department of Information Technologies

Instructor(s)

Dr. Drenyovszki Rajmund

Checked by

Kovács Márk

Course objectives

Students learn the basics and theoretical background of Deep Learning and its applications.

Course content

Lectures

1. Supervised learning. Perceptron, Neural network architecture. Deep learning. 2. Activation functions, backpropagation algorithm. 3. Training strategies, optimization algorithms (Gradient Descent, SGD, Adam), initialization. Hyperparameters. 4. Model evaluation: metrics, validation strategies, bias-variance tradeoff. 5. Regularization techniques (dropout, batch normalization). 6. Convolutional Neural Networks (CNN) and their applications in image processing. 7. Transfer learning techniques and the use of pre-trained models. 8. Recurrent Neural Networks (RNN), LSTM and GRU architectures for sequential data processing. 9. Transformers and the attention mechanism. 10. Unsupervised learning. GAN, Autoencoders (VAE), diffusion models. 11. Vision Language Models (VLM) and multimodal applications. 12. Interpretability and responsible AI. 13. Midterm exam 14. Make-up / Retake midterm exam

Labs

Writing, exploring, and completing Python code snippets related to the lectures, utilizing TensorFlow, PyTorch, JAX, and Keras libraries.

Acquired competences

Knowledge

–

Skills

–

Attitude

–

Autonomy and responsibilities

–

Requirements, evaluation and grading

Mid-term study requirements

Theoretical and practical midterm exam. Completing assigned short tasks / short assignments.

Generative AI usage

Use of GAI tools is permitted in a limited manner (e.g., for literature search support or specific tools). In this case, the course instructor is responsible for defining where and how GAI tools may be used in assignments. The course description must specify in detail how GAI tools may be used during the course.