Research topics
The Graph Analytics and Network Learning (GALE) group focuses on bridging the gap between Machine Learning and Distributed Systems to address the challenges posed by centralized AI services. These traditional services often gather and centralize user data, leading to privacy concerns, loss of control, and scalability issues. GALE develops federated and decentralized algorithms that enable advanced data mining and machine learning on distributed, dynamic datasets, aiming to create scalable and secure AI solutions.
Furthermore, GALE is actively involved in research on Information Network Analytics and Graph Learning. This is particularly valuable in domains where data exhibits a linked (graph) nature, such as in Social, Financial, and Transportation Networks, as well as in Fraud Detection and Recommendation Systems. The GALE group utilizes decentralized machine learning techniques, including Gossip Learning and Decentralized Graph Neural Networks, to extract insights and actionable knowledge from distributed graph-structured data.
