Bachelor-/ Master Thesis: »Data privacy for machine learning model in manufacturing«

Webseite Fraunhofer Institute for Production Technology IPT

The Fraunhofer-Gesellschaft (www.fraunhofer.com) currently operates 75 institutes and research institutions throughout Germany and is the world’s leading applied research organization. Around 29,000 employees work with an annual research budget of 2.8 billion euros.

The subject data privacy is becoming increasingly important in artificial intelligence (AI) applications in manufacturing. Although AI is increasingly being applied for process and product optimization in industry through several machine learning methods, privacy concerns regarding model transparency and data reconstruction potential represent risks for the company. This thesis will develop a method for assessing model transparency and apply it to manufacturing use cases to uncover and evaluate the risks of a potential privacy attack.

What you will do

  • Understanding the background of privacy attacks on ML models
  • Support through benchmarking model transparency assessment methods
  • Support by evaluating the transparency of models
  • Conducting data regeneration experiments

What you bring to the table

  • You are studying engineering, data science, computer science, CES or a comparable subject
  • You have first experience with object-oriented programming languages, machine learning, TensorFlow, PyTorch, scikit-learn
  • A high degree of initiative and team spirit
  • Good language skills in German and/or English

What you can expect

  • Ideal conditions for practical experience alongside your studies
  • Professional supervision and collaboration in a dedicated team
  • A state-of-the-art machine park equipped with edge cloud systems and 5G infrastructure
  • Flexible working to combine study and job in the best possible way

Questions according to this position will be answered by:
Gustavo Laydner de Melo Rosa, Eng.
Research Fellow Production quality
Phone: +49 241 8904-256

Um dich für diesen Job zu bewerben, besuche bitte s.fhg.de.