SenPAI
Sicherheit und Da­ten­schutz in KI

Künstliche Intelligenz wird immer häufiger in IT-Sicherheitsanwendungen eingesetzt. Doch die Sicherheit der verwendeten Algorithmen ist oft begrenzt – wie gezielte Angriffe und Privatsphäre-Risiken zeigen.

Der Forschungsbereich SenPAI in ATHENE befasst sich mit dem Thema IT-Sicherheit in der KI in Bezug auf Algorithmen und Systeme sowie in Bezug auf Sicherheits-Anwendungen, die auf ML basieren.

Mehr Lesen

Best Paper Award for SenPAI experts

Text forensics experts from SenPAI present an authorship verification method in their award-winning paper "TAVeer - An Interpretable Topic-Agnostic Authorship Verification Method". With the help of AV, one can check whether a specific author actually wrote a piece of text or not.

Mehr Info

How secure is Machine Learning?

Study investigates the application of security aspects in Machine Learning techniques. Practitioners are needed for the survey.

Mehr Info

More Cyber Security for AI

Many companies and public authorities are still hesitant about using AI in the field of cyber security, despite the many opportunities for innovation. One major reason: the performance of systems is often difficult to assess. That is why experts from science and industry have drawn up recommendations on how to overcome obstacles to the use of AI.

Mehr Info

Team

Prof. Dr. Martin Steinebach
Fraunhofer SIT | TU Darmstadt

Koordinator: SenPAI
Projektkoordinator: 

  • Forensic and OSINT Technology with Machine Learning - FROST
  • Robustness in Machine Learning - RoMa

E-Mail

Prof. Dr. Iryna Gurevych
TU Darmstadt

Projektkoordinatorin: 

  • Adversarial Attacks on NLP systems
  • Protecting Privacy and Sensitive Information in Texts
  • Human-AI Collaboration for Cyber­security

E-Mail

Dr.-Ing. Oren Halvani
Fraunhofer SIT


Projektkoordinator: 

  • LLM-Aided and Affected Authorship Verification/Attribution - LAVA
  • Detecting CSAM Without the Need for CSAM Training Data - DecNec
  • Visual Forensic Person Verification - VISPER

E-Mail

Prof. Dr.-Ing. Jörn Kohlhammer
Fraunhofer IGD


Projektkoordinator: 

  • Interactive Visual Cyber Analytics for Trust and Explainability in Artificial Intelligence for Sensitive Data - VCAXAI

E-Mail

Prof. Dr. Carsten Binnig
TU Darmstadt


Projekt:

  • Human-AI Collaboration for Cyber­security

E-Mail

Subhabrata Dutta, PhD
TU Darmstadt


Projekt:

  • Human-AI Collaboration for Cyber­security

E-Mail

Prof. Dr. Kristian Kersting

TU Darmstadt


Projekt:

  • Human-AI Collaboration for Cyber­security

E-Mail

Prof. Dr. Florian Müller

TU Darmstadt


Projekt:

  • Human-AI Collaboration for Cyber­security

E-Mail

Prof. Dr. Dr. Christian Reuter

TU Darmstadt


Projekt:

  • Human-AI Collaboration for Cyber­security

E-Mail

Prof. Dr. Anna Rohrbach

TU Darmstadt


Projekt:

  • Human-AI Collaboration for Cyber­security

E-Mail

Prof. Dr. Marcus Rohrbach

TU Darmstadt


Projekt:

  • Human-AI Collaboration for Cyber­security

E-Mail

Prof. Stefan Roth, Ph.D.

TU Darmstadt


Projekt:

  • Human-AI Collaboration for Cyber­security

E-Mail

Prof. Dr. Haya Schulmann.

TU Darmstadt


Projekt:

  • Human-AI Collaboration for Cyber­security

E-Mail

Prof. Dr. Michael Waidner

TU Darmstadt


Projekt:

  • Human-AI Collaboration for Cyber­security

E-Mail

Niklas Bunzel
Fraunhofer SIT

wiss. MA: RoMa

Igor Cherepanov
Fraunhofer IGD

wiss. MA: VCAXAI

 

Quentin Delfosse
TU Darmstadt

wiss. MA: XReLeaS

Tilo-Lars Flasche
Fraunhofer SIT

wiss. MA: ViSper

Haritz Puerto
TU Darmstadt

wiss. MA: SecLLM

David Sessler 
Fraunhofer IGD

wiss. MA: VCAXAI

York Yannikos
Fraunhofer SIT

wiss. MA: FROST + ML, SePIA