Provably Powerful Graph Neural Networks for Directed Multigraphs
- Beni Egressy
- Luc Von Niederhäusern
- et al.
- 2024
- AAAI 2024
Jovan Blanuša is a Postdoctoral Researcher at IBM Research Zurich. His main research interests are graph processing, graph machine learning, and the acceleration of graph-based algorithms with applications in financial crime detection and knowledge extraction. Jovan first joined IBM Research Zurich in 2019 as a Predoctoral Researchers, where he worked on real-time detection of suspicious financial transactions in financial transaction graphs. For this purpose, he developed a Graph Feature Preprocessor library, which has been integrated into IBM mainframe software products, namely IBM Cloud Pak for Data on Z and AI Toolkit for IBM Z and LinuxONE. As a result of this work, Jovan received an Outstanding Research Accomplishment for contributions to IBM's System Z AI offerings.
In 2023, Jovan obtained his Ph.D. in Computer Science from EPFL with the thesis titled Acceleration of graph pattern mining and applications to financial crime. The same year, he obtained the Fritz Kutter Award for the best Swiss-based industry-related PhD thesis in CS. Prior to joining IBM Research, Jovan obtained his Masters degree in Electrical Engineering from EPFL and his Bachelors degree in Electrical Engineering from the University of Belgrade, and he carried out several internships with NVIDIA and Microsoft. He has been a member of ACM since 2022.