Module Artificial Intelligence, Computer Science (Master) (ER 8)

This lecture covers current developments and recent research findings in the field of Artificial Intelligence, with a particular focus on Deep Learning.

After completing the lecture, students are able to assess the fundamentals of neural networks and evaluate current research methods in the domains of Computer Vision and Natural Language Processing.

Students will be capable of evaluating basic architectures such as Convolutional Neural Networks and Recurrent Neural Networks, as well as analyzing advanced architectures like Transformer models. Furthermore, students are able to assess appropriate architectures for application areas in Computer Vision and Natural Language Processing, comprehend scientific literature such as recent publications, and analyze specific aspects. Additionally, they will understand current methods from subfields such as Continual Learning.

The lecture serves as a theoretical foundation for the overall module “Artificial Intelligence” and complements the practical course “AI Lab.” Practical exercises are therefore not part of this lecture. Group presentations are an option for a deeper exploration of specific topics.

The lecture covers the following topics:

• Review of the fundamentals of neural networks and Deep Learning
• Various architectures of CNNs
• Application examples in Computer Vision, such as Object Detection and Instance Segmentation
• Models for processing sequential data, such as RNNs and (Bi-)LSTMs
• Application examples in Natural Language Processing, such as Machine Translation and Information Extraction
• Language models and word embeddings
• Attention mechanism and Transformer models
• Large language models (LLMs)
• Core cross-domain methods like Transfer Learning
• Introductions to specialized areas such as Explainable AI, Continual Learning (e.g., for improving LLMs), or Diffusion Models

Students' understanding of the fundamental concepts and architectures of Deep Learning, as well as their ability to evaluate current research findings, will be assessed through an exam. Answering the questions will require both individual formulations and performing calculations based on examples from the application domains.

In addition, students have the option to demonstrate their ability to evaluate and explain specific topics in detail through a voluntary group presentation on a specialized topic. A presentation that is convincing due to thorough analysis and clear explanation will result in an improvement of the exam grade by one grade level.

This lab implements the theoretical foundations from the lecture into practical tasks.

For this, tasks from the following three domains are tackled:

* Computer Vision

* Natural Language Processing

* Reinforcement Learning

Requirements:

• Basic knowledge in Python
• Basic knowledge in Machine Learning