Doctoral thesis in applied mathematics: Artificial fingerprints train safer AI
Donglin Liu, a doctoral student in applied mathematics, has used several different mathematical models in order to improve artificial intelligence at a fundamental level.
– Published 6 October 2025
He has, among other things, produced synthetic but realistic fingerprints which, without violating anyone's privacy, can be used to train improved fingerprint readers. He has also worked with mixed mathematical models of artificial intelligence, which combine several different types of algorithms. This increased the reliability and realism of the models.
Donglin Liu will defend his thesis “System Identification and Data-Driven Modeling” on October 17, 2025.
Read more about Donglin Liu’s research in Lund University's research portal:
https://portal.research.lu.se/en/persons/donglin-liu