Dominik Hintersdorf
Artificial Intelligence and Machine Learning Lab TU Darmstadt
My research is centered around the privacy and security of deep learning systems. As deep learning is increasingly used in real-world products and tasks, the data to train such systems is getting more and more relevant and important. In many of those tasks, the models have to be reliable and secure since during training, sensitive data might have been used, such as medical information or other personal data like for example images. In my work, I study possible threats and mitigation techniques to the security and privacy of deep learning models.
I received by Masters from TU Darmstadt and am a PhD student under the supervision of Prof. Kristian Kersting since 2021.
news
Oct 27, 2023 | Our papers Defending Our Privacy With Backdoors and Leveraging Diffusion-Based Image Variations for Robust Training on Poisoned Data got accepted at the NeurIPS 2023 Workshop on Backdoors in Deep Learning. |
---|---|
Sep 11, 2023 | We gave a talk at the AISola conference titled “Balancing Transparency and Risk: The Security and Privacy Risks of Open-Source Machine Learning Models”. |
Sep 11, 2023 | Our paper SEGA: Instructing Text-to-Image Models using Semantic Guidance got accepted at NeurIPS 2023. |