Double Debut – Antrittsvorlesungen von Flavio Toffalini und Yannic Noller

Seit Juli 2024 ist Yannic Noller Professor für Software Quality an der Fakultät für Informatik. Er entwickelt Methoden für die automatisierte Fehlersuche und -reperatur in Software Code. Bislang geschieht dies oft in zwei separaten Prozessen. Noller forscht an Methoden, die Entwicklerinnen und Entwickler auf Fehler in ihrem Code aufmerksam machen und zugleich Lösungsvorschläge unterbreiten, die zur bestehenden Logik des restlichen Programmcodes passen.

Abstract zur Vorlesung:

In this lecture, I will reflect on my journey toward making software „better“ (i.e., more resilient) through automated software engineering. We will explore key approaches and techniques for automatically detecting bugs and vulnerabilities, and generating corresponding patches. I will begin by discussing the roles of fuzzing and symbolic execution in automated testing and repair, including their integration for differential analysis. This will lead into a case study on the detection, quantification, and repair of software side-channel vulnerabilities. Next, I will delve into automated program repair using semantic analysis and hybrid techniques, addressing both traditional software and machine learning models. I will also highlight our recent initiatives to enhance computer science education through automation. Finally, I will present my latest work at the intersection of repair, formal methods, and large language models, and share a vision for the future of software engineering.