Federated Forensics architecture supports multi-agency learning with jurisdictional controls
Original title: Federated Forensics: Jurisdictional constraints as a design driver for multi-agency federated learning
New architecture blocks restricted model updates at jurisdictional borders while maintaining 93% accuracy.
Digital forensics investigations cross borders, requiring agencies to collaborate without centralizing data. The paper derives seven requirements for forensic federated learning and proposes Federated Forensics, a three-layer architecture with a jurisdictional middle layer. Prototype testing on malware classification shows it blocks every restricted update with zero violations and improves restricted-class accuracy from 0% to 93.0%.
- Seven requirements derived for forensic FL systems
- Prototype blocks restricted updates at border
- Accuracy rises from 0% to 93.0% on malware task
From the original article
1 min read at arxiv.orgDigital forensic investigations increasingly cross borders, and law enforcement agencies need to build analytical models together without centralizing their data. Federated Learning (FL) keeps data local, yet neither FL nor its frameworks enforce the jurisdictional, governance, and accountability rules that multi-agency forensic work requires. This paper takes a design-first approach to that gap.
By Aruna Withanage · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.