Hi, I’m Jonas. I am starting as a research group leader in Tübingen, where I’m building a group for safety- & efficiency- aligned learning (🦭). Before this, I’ve spent time at the University of Maryland and the University of Siegen.
I am mostly interested in questions of safety and efficiency in modern machine learning. There are a number of fundamental machine learning questions that come up in these topics that we still do not understand well. In safety, examples are questions about the principles of data poisoning, the subtleties of water-marking for generative models, privacy questions in federated learning, or adversarial attacks against large language models. Can we ever make these models “safe”, and how do we define this? Are there feasible technical solutions that reduce harm?
Further, I am interested in questions about the efficiency of modern AI systems, especially for large language models. How efficient can we make these systems, can we train strong models with little compute? Can we extend the capabilities of language models with recursive computation? How do efficiency modifications impact the safety of these models?
- Safety, Security and Privacy in Machine Learning
- Efficient Machine Learning (especially for NLP)
- Trustworthy AI
- Deep Learning as-a-Science
Incoming PhD Students:
If you are interested in these topics, feel free to reach out for more information! I’m currently hiring through the following PhD programs:
- ELLIS PhD program
- Max Planck & ETH Center for Learning Systems (CLS)
- International Max Planck Research School for Intelligent Systems (IMPRS-IS)
For more details, make sure to also check out the openings page.
- Decepticons: Corrupted Transformers Breach Privacy in Federated Learning for Language ModelsIn International Conference on Learning Representations, Feb 2023
- Cramming: Training a Language Model on a Single GPU in One Day.In Proceedings of the 40th International Conference on Machine Learning, Jul 2023
- Baseline Defenses for Adversarial Attacks Against Aligned Language Modelsarxiv:2309.00614[cs], Sep 2023
- A Watermark for Large Language ModelsIn Proceedings of the 40th International Conference on Machine Learning, Jul 2023
- Diffusion Art or Digital Forgery? Investigating Data Replication in Diffusion ModelsIn Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Jul 2023
- Tree-Ring Watermarks: Fingerprints for Diffusion Images That Are Invisible and Robustarxiv:2305.20030[cs], May 2023