Welcome, Professor Florian Tramèr
Florian Tramèr joined the Department of Computer Science at ETH Zurich in August 2022 as a Tenure Track Assistant Professor of Computer Science. Get to know him in this short interview.
Professor Tramèr, welcome to ETH Zurich. What are your current research interests?
My research is in computer security and privacy, with a focus on machine learning. Lately, I'm particularly interested in understanding how machine learning models memorise training data, and the privacy risks that this poses.
What is the impact of your research on society?
Machine learning technology is being developed and deployed at a very fast pace. With the "move fast and break things" attitude, security and privacy often become afterthoughts. My research aims to remedy this, by proactively demonstrating new vulnerabilities to inform technical as well as legislative protections.
Where were you working before you came to ETH?
I'm originally from Switzerland and did my Bachelor’s and Master’s studies at EPFL. I then did a PhD at Stanford, and spent the past year at Google before joining ETH Zurich.
Which courses will you be teaching at ETH?
I'm not sure yet! I'd really like to teach a class focused on privacy-enhancing technologies, combining beautiful ideas from cryptography, statistics and computer systems.
What do people often get wrong about your field?
Not many people realise how brittle machine learning models actually are. Despite tremendous progress and spectacular results in recent years, we still have only a very superficial understanding of how and what these models actually learn, and why they make "silly" mistakes.
What advice would you give to students who are just starting out in computer science?
Be curious, bold and aim to learn about a wide range of topics! I had no particular interest in machine learning when I started my graduate studies – I was mainly interested in theoretical computer science and cryptography. But my advisor encouraged me to explore this new field, although he didn't know much about it either. It worked out amazingly well!
More information
- external page Prof. Florian Tramèr
- Institute of Information Security
- external page Stanford University
- external page EPFL