Temporal Link Prediction in Heterogeneous Social Networks with Attention-Based Graph Encoders
K. Selvarajah, A. Patel, M. Kobti
IEEE Transactions on Computational Social Systems, vol. 12, no. 2, pp. 410–423, 2025
Professor of Computer Science
School of Computer Science, University of Windsor
Understanding how people, information and influence move through networks.
Dr. Kalyani Selvarajah is a computer scientist whose research sits at the intersection of artificial intelligence and social network analysis. Her work develops graph-based machine learning methods for link prediction, community detection and influence modelling, and applies them to real problems with industry and community partners.
Research focus
Research
Recent and representative work on social network analysis and graph machine learning.
K. Selvarajah, A. Patel, M. Kobti
IEEE Transactions on Computational Social Systems, vol. 12, no. 2, pp. 410–423, 2025
K. Selvarajah, R. Nguyen
Proceedings of the IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining, pp. 133–140, 2024
K. Selvarajah, J. Rivera, P. Osei
ACM Computing Surveys, vol. 55, no. 4, pp. 1–38, 2022
Partnerships
Applied research projects developed with industry, healthcare and public-sector partners.
Modelled a multi tier supplier network as a temporal graph and trained link prediction models to flag suppliers at risk of disruption weeks before delays appeared in production data.
Analysed anonymised referral data to surface care communities and bottlenecks, helping administrators redesign referral pathways and reduce wait times for specialist appointments.
Identified high influence community nodes on social platforms to improve the reach of vaccination and wellness campaigns while respecting privacy constraints.
Academic & professional activities
Events coordinated or delivered as part of research, teaching and community engagement.
A one-day workshop bringing together researchers, industry practitioners and policy makers on ethical and privacy-preserving network analysis.
End-of-term exhibition where capstone teams demonstrate AI and network-analysis projects developed with community and industry partners.
In the media