Master Seminar Retrieval Augmented Generation (DS-SemRAG-M)

Large Language Models are the basis of numerous applications in natural language processing and dialogue systems such as ChatGPT. In many applications requiring task-specific and other model-external world knowledge, these systems rely on a technique for augmenting the model's knowledge known as Retrieval Augmented Generation (RAG). In this module, participants will thoroughly explore the RAG literature and present a state of the art topic in RAG.

The learning goals for this course are the following:

Participants will:

• Learn to familiarize themselves individually and independently with their respective topic
• Develop skills and understanding of scientific writing and oral communication
• Learn to discuss and evaluate current approaches to RAG with language models
• Gain a deep understanding of the topic in the RAG literature they present.

This 3-ECTS seminar runs each semester.

Participants must be enrolled in one of the Master programms Applied Computer Science or Computing in the Humanities.

The course language will be English. The language of the term paper may be German or English.

The content of the colloquium and the term paper consists of the implementation work done during the course of the semester.

The terms and conditions (e.g., deadline) of the term paper and of the presentation will be announced at the beginning of each course