With a Podcast Series entirely focused on the Principal Investigators of the ELLIS Institute Tübingen, Cyber Valley continues to explore the profound implications of artificial intelligence with experts in the field.
The second episode features Jonas Geiping, our PI expert on safety and efficiency of modern machine learning. With a background in mathematical optimization and a strong interest in safety and efficiency questions in deep learning, Jonas' research agenda is based on constructively understanding machine learning, directly testing theory, privacy and safety, and designing more secure and private ML systems to advance our fundamental understanding of trustworthy AI.
In this episode, Jonas, who currently leads the Safety- and Efficiency- aligned Learning group, talks about his vision of AI, his research, and the challenges he is now facing. He also discusses AI misconceptions, biases, ethics, and privacy issues:
-
What is his definition of AI?
-
What are the challenges in making a system both safe and efficient?
-
What is the relationship between academia and industry in advancing AI research?
-
What does trustworthy AI really mean?
For this and more, watch the full episode below, or listen to it on Spotify here.
More Information
• Learn more about Jonas’ Safety- and Efficiency- aligned Learning group here.
• To keep up with the latest news, follow us on X and LinkedIn.
Video