Associate Professor Jingrui He will present her research at the International Workshop on Deep Learning on Graphs, which will be held in conjunction with the 35th AAAI Conference on Artificial Intelligence (AAAI'21) on February 2-9. The one-day workshop brings together academic researchers and industrial practitioners from different backgrounds to investigate new approaches and methods at the intersection of Graph Neural Networks and real-world applications.
Her talk, "Exploring Rare Categories on Graphs: Local vs. Global," will address "rare categories," or underrepresented classes in data sets where the proportion of examples belonging to different categories varies greatly.
"For example, in a data set consisting of a large number of financial transactions, only a small number of them are marked as fraudulent, and the remaining are legitimate," she said.
According to Associate Professor He, rare categories are prevalent across many high-impact applications in the security domain where the input data can be represented as graphs. Her talk will focus on two strategies for exploring rare categories—local and global.
"With the local strategy, the goal is to explore a small neighborhood around a seed node (a known example from the rare category, such as a fraudster in my previous example) to identify additional rare examples. The global strategy explores the entire graph, in order to identify rare category-oriented representations," she said.
Jingrui He's general research theme is to design, build, and test a suite of automated and semi-automated methods to explore, understand, characterize, and predict real-world data by means of statistical machine learning. She received her PhD in machine learning from Carnegie Mellon University.