Dataset Similarity Detection for Reuse Protection in Federated
Data Spaces with Privacy Considerations

Today, organizations use shared networks to safely exchange and sell data under strict agreements. However, a major problem remains: once data leaves its owner, it is incredibly hard to track. If someone reuses, leaks, or changes that data without permission, it usually goes unnoticed.

To fix this, this paper introduces the Off-Platform Contract Inspector. Think of it as a digital detective designed to catch unauthorized data use. It scans files, tables, and text to ensure everyone plays by the rules, offering three main breakthroughs:

• Smart Filtering: It doesn’t get confused by common industry buzzwords, allowing it to accurately spot actual data theft without triggering false alarms.

• Privacy-First Checking: It can compare datasets to find hidden copies without ever looking at the sensitive, raw information itself.

• Real-World Proof: The system successfully outsmarted realistic data-disguising tricks when tested on real city data from the transportation, energy, and automotive sectors.