Unleashing the Power of Graph-Based Machine Learning as a Service

Date: Sep 5, 2:15 – 3:05 pm

Abstract: The exponential growth of graph-structured data has created an unprecedented demand for graph learning capabilities across industries, yet domain experts face formidable barriers: massive computational requirements, prohibitive model costs, and complex infrastructure management. Graph-based Machine Learning as a Service (GMLaaS) has emerged as a revolutionary solution, transforming how organizations harness the power of graph machine learning.

This talk unveils the complete GMLaaS ecosystem through three crucial dimensions. First, this talk reveals how GMLaaS users can achieve breakthrough results through unbiased predictions, enhanced privacy preservation, and in-depth explainability that builds trust. Second, for GMLaaS model owners, this talk exposes critical strategies for advanced risk understanding, graph learning model ownership verification, and real-time risk monitoring that safeguard investments for graph machine learning model development. Finally, this talk showcases transformative GMLaaS applications that are revolutionizing three pivotal domains: AI4Science accelerating scientific discovery, AI4Geo powering next-generation geospatial intelligence, and AI4Healthcare saving lives through policy analysis.

Biographical Sketch: Yushun Dong is a tenure-track Assistant Professor in the Department of Computer Science at Florida State University (FSU), where he leads the Responsible AI (RAI) Lab. His research focuses on trustworthy AI with emphases on security, integrity, and explainability, and further spans interdisciplinary applications including political science, neuroscience, and social computing. Dr. Yushun Dong serves as the lead Principal Investigator (PI) on multiple research grants by prestigious institutions, including the National Science Foundation (NSF) and internal awards from Florida State University. As output, his pioneering work has resulted in over 40 peer-reviewed publications in premier venues such as SIGKDD, NeurIPS, ICML, ICLR, and AAAI, with more than 1,600 citations and an h-index of 20. His open-source toolkits?such as PyGDebias and PyGIP?serve the research community in graph learning and trustworthy AI. Dr. Dong serves as an Area Chair and reviewer for major conferences including SIGKDD, ICLR, and ACL. He is committed to advancing the responsible development and deployment of AI technologies that are transparent and secure.

Location and Zoom link: LOV 307 and ZOOM https://fsu.zoom.us/j/7153751215