Software Engineering Seminar

Software Engineering Seminar WS 2024/ 2025

The chair SEDA organizes the joint Software Engineering Seminar for Bachelor and Master students in Winter semester 24/25. The goal of the seminar is to introduce students to the critical reading, understanding, summarizing, and presentation of scientific papers. Contents are selected topics from the field of software and systems engineering, in particular:

Systems Engineering for Cyber-Physical Systems,
Safety, Security, Reliability and Availability,
Risk-Assessment and -Minimization,
Model-Based Safety Analysis.

Registration:

The registration is possible between 30.09.2024 und 17.10.2024, here on the website. Herefore you have to fill the form below. Two topics require some prior knowledge in this area, so please do only select these topics if you have this prerequisites. In this case do also provide an additional document as verification (i.e. transcipt of records, job experience, etc.). Multiple registration will not increase your chances.

Final decisions about acceptance will be send out on 18.10.2024.

Timeline:

Kick-Off Meeting: 23.10.2024, 17:15 c'clock, room 11-220, slides.
Annotated table of content: 20.11.2024
First version of written paper: 20.12.2024
Final version of written papers: 07.02.2025
Presentation: 11.02.2025, 13:00 - 17:30 o'clock, room 32-439.

Links to the Module Handbooks:

Bachelor: https://modhb.uni-kl.de/mhb/modules/INF-01-11-M-4/
Master: https://modhb.uni-kl.de/mhb/modules/INF-33-72-M-7/

Organizers:

Lecturer: Prof. Dr.-Ing. Peter Liggesmeyer
Supervisor and person of contact: Alexander Günther

Topics

Multi Agent LLM-based software engineering platforms

Description: Multi-agent LLM-based software engineering platforms promise to automate most of the software development lifecycle. The goal of this seminar is to gather information about what such platforms can do, how they are structured, and how to evaluate the quality of the software produced.
Literature:
Tools:
- https://chatdev.ai/
  https://github.com/Pythagora-io/gpt-pilot
  https://github.com/geekan/MetaGPT
~~Bachelor/~~Master
Requirements:
- (necessary) Good understanding of SDLC.
- (optional) Understanding of LLM.
Supervisor: Dr. Julien Siebert
Number of students: 1-3

Generating Test Cases with LLMs

Description: LLMs with their generative capabilities are foreseen to impact software engineering deeply. In this seminar the goal is to explore the state of the art of automated test cases generation with LLM.
Literature:
Bachelor/ Master
Supervisor: Dr. Julien Siebert
Number of students: 1-3

Explainable Machine Learning Methods in Safety-Critical Systems

Description: Machine Learning methods generated some stunning results over the last decades and are increasingly influencing our all lives. Nevertheless there black-box character still holds challenges in the the safety assessment. Explainable artificial intelligence is trying to overcome this drawback by enriching the model predictions with additional information. The natural question that arises, are XAI methods capable to overcome the black-box character and allow a safe deployment in safety-critical systems?
Literature:
- Roberto Confalonieri, Ludovik Coba, Benedikt Wagner, and Tarek R. Besold. A historical perspective of explainable artificial intelligence. WIREs Data Mining and Knowledge Discovery, 11(1):e1391, 2021.
- Vinay Chamola, Vikas Hassija, A Razia Sulthana, Debshishu Ghosh, Divyansh Dhingra, and Biplab Sikdar. A review of trustworthy and explainable artificial intelligence (xai). IEEE Access, 11:78994–79015, 2023.
- Jos´e Faria. Machine learning safety: An overview. Safety-critical Systems Symposium 2018 (SSS’18), 02 2018.
- Timo Speith. A review of taxonomies of explainable artificial intelligence (xai) methods. In Proceedings of the 2022 ACM Conference on Fairness, Accountability, and Transparency, FAccT ’22, page 2239–2250, New York, NY, USA, 2022. Association for Computing Machinery.
~~Bachelor/~~Master
Requirements:
- (necessary) Machine Learning 1 (or equivalent).
- (optinal) Safety and Reliability of Embedded System (or equivalent).
Supervisor:Alexander Günther
Number of students: 1

Statistical Model Checking for Risk Assessment of Machine Learning Models

Description: Machine Learning methods generated some stunning results over the last decades and are increasingly influencing our all lives. Nevertheless there black box character still holds challenges in the the safety assessment. The techniques of statistical model checking offers a mathematical sound way to assess (software) systems. Therefore the question arises if these techniques can also be used in case of machine learning models.
Literature:
- Axel Legay, Anna Lukina, Louis Marie Traonouez, Junxing Yang, Scott A. Smolka, and Radu Grosu. Statistical Model Checking, pages 478–504. Springer International Publishing, Cham, 2019.
- Axel Legay, Benoˆıt Delahaye, and Saddek Bensalem. Statistical model checking: An overview. In Howard Barringer, Ylies Falcone, Bernd Finkbeiner, Klaus Havelund, Insup Lee, Gordon Pace, Grigore Ro¸su, Oleg Sokolsky, and Nikolai Tillmann, editors, Runtime Verification, pages 122–135, Berlin, Heidelberg, 2010. Springer Berlin Heidelberg.
- Jos´e Faria. Machine learning safety: An overview. Safety-critical Systems Symposium 2018 (SSS’18), 02 2018.
Bachelor/ Master
Supervisor:Alexander Günther
Number of students: 1

(The chair of professor Hinze did the Seminar https://pl.cs.uni-kl.de/homepage/en/teaching/ss21/seminar/ )

Main navigation