In today's rapidly evolving technological landscape, artificial intelligence, specifically deep learning, becomes able to solve increasingly complex tasks. Their increasing capabilities make them useful for real-world problems and deployed even for critical tasks such as making hiring decisions, medical diagnostics, or safety-critical operations. However, deep learning algorithms face new challenges, constraints, and threats when deployed outside well-controlled test environments to solve real-world problems.
In this seminar, students will summarize and critically analyze the recent literature on the assigned topic in the form of a report. Additionally, each student will present his work in front of the group at the end of the semester.
Possible topics include:

  • Fair and non-biased deep learning
  • Explainable and trustworthy decisions
  • Privacy and the right to be forgotten for deep neural networks
  • Robust and trustworthy deep learning for security-critical applications
  • Distributed learning schemes
  • Deep neural networks in environments with runtime constraints
  • Detecting AI-generated content
  • Responsible deep fake generation