• Intelligent Software Systems

Chair of Intelligent Software Systems


  • Norbert Siegmund is co-founder of the Java User Group Thuringia (follow us on Twitter: @jug_th)
  • The German Research Foundation (DFG) invest 289,000€ in funding a cooperative research project together with the University of Passau. The Project Pervolution will investigate how performance changes during the evolution of highly configurable software systems.
  • New master thesis started: "Deep Code Generation via Hierarchical Neuronal Networks"
  • Joined the programm committee at NIERS@ICSE
  • Great presentations and discussions at VaMoS Workshops!
  • Update: Deadline Extension until March 31st, 2017: 2 open positions available; details >here<:

    • PhD candidate in the area of deep learning for automated code generation (duration 3+3 years, 100% TVL - E13).
    • PhD candidate / post-doc to support the introduction and establishment  of the new master curriculum Digital Engineering (duration 1 year, 100% TVL - E13; in case of successfull third party funding, the appointment might be extended).

  • New bachelor thesis started: "Code Generation via Generative Adversarial Networks"
  • New master thesis started: "Quantifying the Impact of Web Server Features on Energy Consumption"



Research Topics:

The Chair of Intelligent Software Systems develops methods, tools, and theories to improve the construction, maintenance, usability, and optimization of complex, configurable software systems. Our main focus is on the non-functional properties of software systems, such as performance and energy consumption. We combine classical and novel machine learning techniques to solve challenges in the area of software engineering. This relative young branch of software engineering, called "Search-Based Software Engineering", explores the synergies of machine learning and software-engineering methods:

  • Software Engineering fundamentally aims at automatization and optimization in software development and usage. Machine learning techniques can give new insights and solve tasks that are too complex and too big to for the human brain. 
  • Vice versa, also techniques from software engineering, especially from domain of configurable software systems and variant management, can help to maintain and control the ever growing number of increasingly complex machine learning techniques. In this line of research, we also strive for automatization of complex systems and to make them accessible and usable for non-experts in machine learning.