Rice CS’ DATA Lab won Outstanding Paper at the 2022 International Conference on Machine Learning (ICML). DATA Lab was founded by Xia “Ben” Hu, CS associate professor and director of Rice’s Data to Knowledge Lab (D2K). Hu specializes in interpretable machine learning, automated machine learning and network analytics.
“This is wonderful and such a pleasant surprise to us,” says Hu about winning Outstanding Paper. “It’s hard to expect an award out of like 6,000 submissions.”
The popularity of machine learning and AI has fueled the growth of ICML, one of the flagship machine learning conferences. This year’s event was held in-person in July in Baltimore, Maryland.
DATA Lab’s winning paper, “G-Mixup: Graph Data Augmentation for Graph Classification,” presents G-Mixup to augment graphs. Extensive experiments show that G-Mixup substantially improves the generalization and robustness of graph neural networks.
“This is a milestone in this research area, to make use of a very specific technology called a graphon for graph augmentation,” says Hu. “It’s a major achievement for network analytics and an accumulation of a decade of research.”
“We started from traditional network analytics and then went into machine learning-based network analytics and now it’s deep learning-based network analytics.”
This line of research has very wide applications to many different areas such as social network analysis, healthcare, disaster relief, biology, and many science and engineering domains that contain graph data. “Hopefully this will transform how the community is using graph data augmentation for graph classification,” says Hu.
The paper’s other authors include lab members, current doctoral students and DATA Lab alumni including Xiaotian Han, Zhimeng Jiang, and Ninghao Liu.
Hu has published more than 100 papers in major academic venues, and his work has been cited more than 14,700 times, with an h-index of 50. He has received the 2021 ACM SIGKDD Rising Star Award as well as an NSF CAREER Award in 2018.
Prior to joining Rice, Hu worked as an associate professor of computer science and engineering at Texas A&M University. He received his Ph.D. in Computer Science from Arizona State University in 2015.
Read the complete paper for more details.