Skip to main content Skip to main navigation

Publication

AI combats Cybercrime - Deep Learning for Forensics

Tobias Wirth; Andrey Guzhov; Matthias Müller; Tobias Fischer; Lucas Howes; Dominik Ospelt
In: Future Insight e.V. (Hrsg.). Curious 2024. Curious - Future Insight Conference (Curious-2024), July 10-11, Mainz, Germany, www.curiousfutureinsight.org, 7/2024.

Abstract

The German Research Center for Artificial Intelligence (DFKI), The State Criminal Police of Rhineland-Palatinate and the software developer Binary Impact join forces to develop an AI-enabled assistant to support forensic experts. In this presentation we show how the AI-assistant (i) helps to classify a sheer mass of fragments of deleted files, (ii) clusters all fragments that belong to individual files, and (iii) in the final step puts together the "puzzle" of fragments to restore the digital evidence. We developed an AI pipeline built on top of state-of-the-art transformer architectures combined with a UI/UX frontend that is inspired by the best of gaming standards. The yet to be answered question will be "Can the re-assembled digital evidence stand up in court?" This research is funded by the project Carve-DL; sponsored by The German Federal Ministry of Education and Research (BMBF).

Projects