Smart Embedded Systems Lab · OTH Regensburg

Trustworthy AI: detecting AI-generated content and explaining why.

The Smart Embedded Systems Lab (SES Lab) at OTH Regensburg (Ostbayerische Technische Hochschule Regensburg, Germany) builds AI systems that hold up outside the lab. Our focus is the detection of AI-generated text and images: detectors that generalise across datasets, generators and domains, that survive adversarial manipulation of their input, and whose decisions can be explained to the people who rely on them.

Lab website · GitHub · Real or Fake? AI Challenge

Featured: DeBERTa-ConPara (AACL-IJCNLP 2026)

Attack-aware detection of AI-generated text, built for deployment: unknown domains, unknown generators and adversarially perturbed input. On the official RAID leaderboard: AUROC 99.61 %, TPR 96.57 % at 1 % FPR.


Model	SES-Lab-OTH/deberta-conpara
Live demo	Space: deberta-conpara
Paper	arXiv:2610.00883
Code	github.com/SES-Lab-OTH/deberta-conpara

Datasets

Dataset	Size	Content
Academic-Text-arxiv-gpt-gemini	669,008 paragraphs	Human academic paragraphs from arXiv (papers before 2022) and AI-generated counterparts from GPT-3.5-Turbo and Gemini 2.0 Flash
HC3-Gemini-Flash-Responses	23,463 responses	Gemini 2.0 Flash answers to the HC3 questions, for measuring generator shift against HC3

People

Prof. Dr.-Ing. Johannes Reschke, professor of self-learning and adaptive systems
Mohamed Mady, doctoral researcher (cooperative doctorate with the Technical University of Munich)
Gerald Schickhuber, lab engineer

Publications, contact and more on our GitHub page.