After completing this course, students will be able to:explain and discuss the potential of recent research in software engineering of using AI for automating tasks in software developmentexplain the basic concepts of embeddings and large language models and methods to train themapply LLMs for automating a specific task in software development, such as detecting a vulnerability or repairing a small program

Lectures, literature research, discussion, lab assignment

Topics will include recent research articles on:Training and using Embeddings and LLMs Auto-repairing programs and (build) configurationsVulnerability detection and remediationGenerating API documentation and/or commit messages

Participants should have:Basic knowledge of machine learningGood programming skills (in Python)

Various references to related publications will be provided in the course (on the course slides).Parts of the slides are based on:Hands-on Large Language Models: Language Understanding and Generation, Alammar and Grootendorst, O'Reilly, 2024see also: https://jalammar.github.io/illustrated-transformer/

One lab assignment on automating a software engineering task, student presentation and written report

Lab assignment on automating a software engineering task

The following criteria are used:Quality of the solutionQuality of the presentation (content, style, language, slides) Quality of the written report (content, language)Please note, the use of AI tools in this course is generally permitted throughout the work and writing process. However, students agree to disclose any use of AI tools in the work and text product and to mark AI-generated code and text passages accordingly. Furthermore, students assume sole responsibility for the code and text produced.