5 papers
Hybrid Inspection and Task-Based Access Control in Zero-Trust Agentic AI
Majed El Helou, Benjamin Ryder, Chiara Troiani +3
Authorizing Large Language Model (LLM)-driven agents to dynamically invoke tools and access protected resources introduces significant security risks, and the risks grow dramatical…
Delegated Authorization for Agents Constrained to Semantic Task-to-Scope Matching
Majed El Helou, Chiara Troiani, Benjamin Ryder +3
Authorizing Large Language Model driven agents to dynamically invoke tools and access protected resources introduces significant risks, since current methods for delegating authori…
VerA: Versatile Anonymization Applicable to Clinical Facial Photographs
Majed El Helou, Doruk Cetin, Petar Stamenkovic +2
The demand for privacy in facial image dissemination is gaining ground internationally, echoed by the proliferation of regulations such as GDPR, DPDPA, CCPA, PIPL, and APPI. While…
Facial Image Feature Analysis and its Specialization for Fréchet Distance and Neighborhoods
Doruk Cetin, Benedikt Schesch, Petar Stamenkovic +3
Assessing distances between images and image datasets is a fundamental task in vision-based research. It is a challenging open problem in the literature and despite the criticism i…
Which Information Matters? Dissecting Human-written Multi-document Summaries with Partial Information Decomposition
Laura Mascarell, Yan L'Homme, Majed El Helou
Understanding the nature of high-quality summaries is crucial to further improve the performance of multi-document summarization. We propose an approach to characterize human-writt…