Send email Copy Email Address
Research Group

Backes

My group is working on various aspects of information security, with a current main focus on the intersection of information security with AI / Machine Learning. Further topics of interest include: trustworthy information processing of medical information; design, analysis and verification for security-critical systems and services; and universal solutions in system and software security.

Head of Group

Michael Backes

Email

Address

Stuhlsatzenhaus 5
66123 Saarbrücken (Germany)

Members

Placeholder

Yihan Ma

Xinyue Shen

Yiting Qu
Placeholder

Wai Man Si
Placeholder

Yuan Xin

Junjie Chu

Hai Huang

Yugeng Liu
Placeholder

Yiyong Liu

Minxing Zhang
Placeholder

Yixin Wu
Placeholder

Ziqing Yang
Placeholder

Wenhao Wang

Vincent Hanke

Mingjie Li

Yukun Jiang
Placeholder

Xun Wang

Yage Zhang

Most Recent Publications

Year 2026

2026-04-24

Benchmarking Empirical Privacy Protection for Adaptations of Large Language Models

Conference / Medium

International Conference on Learning Representations (ICLR)

Tags
Authors
Visit Detail Page

Year 2025

2025-12-05

Adjacent Words, Divergent Intents: Jailbreaking Large Language Models via Task Concurrency

Conference / Medium

Conference on Neural Information Processing Systems (NeurIPS)

Tags
Trustworthy Information Processing
Authors
Full Paper Visit Detail Page
2025-12-03

Finding and Reactivating Post-Trained LLMs’ Hidden Safety Mechanisms

Conference / Medium

Conference on Neural Information Processing Systems (NeurIPS)

Tags
Trustworthy Information Processing
Authors
Full Paper Visit Detail Page
2025-11-04

Breaking Agents: Compromising Autonomous LLM Agents Through Malfunction Amplification

Conference / Medium

Conference on Empirical Methods in Natural Language Processing (EMNLP)

Tags
Trustworthy Information Processing
Authors
Full Paper Visit Detail Page
2025-10-21

Hate in Plain Sight: On the Risks of Moderating AI-Generated Hateful Illusions

Conference / Medium

IEEE International Conference on Computer Vision (ICCV)

Tags
Trustworthy Information Processing
Authors
Full Paper Visit Detail Page