Home  | News

06.03.2024

Teaser image to Daniel Rückert on how datasecurity in AI in medicine can be realized

Daniel Rückert on How Datasecurity in AI in Medicine Can Be Realized

TUM Magazine “Faszination Forschung”

The current issue of the TUM magazine "Faszination Forschung" takes a closer look at the work of our Director Daniel Rückert.

With his team, he strives to train medical AI in a way that ensures both patient data privacy and the impartiality of the AI in diagnosis. To achieve this, he employs "Differential Privacy", a method that adds a kind of statistical noise to the data.

«The requirements for AI systems are high. They should handle patients' personal data carefully and not store any identifiable information.»


Daniel Rückert

MCML Director

#research #rueckert
Subscribe to RSS News feed

Related

Link to MCML at AAAI 2026

19.01.2026

MCML at AAAI 2026

MCML researchers are represented with 11 papers at AAAI 2026 (8 Main, and 3 Workshops).

Link to MCML PI Frauke Kreuter Featured on ARD alpha on AI

19.01.2026

MCML PI Frauke Kreuter Featured on ARD Alpha on AI

MCML PI Frauke Kreuter featured on ARD alpha discussing AI in daily life, workplace applications, and responsible, future-ready use.

Link to Blind Matching – Aligning Images and Text Without Training or Labels

15.01.2026

Blind Matching – Aligning Images and Text Without Training or Labels

CVPR 2025 research from Daniel Cremers’ group shows how images and text can be aligned without training data, labels, or paired examples.

Link to MCML PIs Featured in Süddeutsche Zeitung

12.01.2026

MCML PIs Featured in Süddeutsche Zeitung

MCML PIs Xiaoxiang Zhu and Felix Dietrich featured in Süddeutsche Zeitung for TU Munich’s 2025 breakthroughs in AI and data science.

Link to High-Res Images, Less Wait: A Simple Flow for Image Generation

08.01.2026

High-Res Images, Less Wait: A Simple Flow for Image Generation

ECCV 2024 research led by Björn Ommer’s team enables faster high-resolution image generation by boosting diffusion models with flow matching.

Back to Top