Graph Learning for Intelligent and Efficient Reasoning
Recent years have seen tremendous progress in modeling graph-structured data through deep learning, transforming models' ability to understand relational structure. In this talk I will demonstrate explorations that leverage graph structure to empower complex and efficient reasoning in various machine learning scenarios, focusing on the use in foundation models. This talk focuses on 3 aspects of foundation models: architecture, objectives, and inference, where we leverage graph and relational learning for intelligent and efficient reasoning. In particular, I will discuss sparse and efficient transformer architecture backbone, efficient self-supervised pre-training pipeline, and the use of relational reasoning in large language models.
Lunch will be served at 11:45 AM.
Rex Ying is an assistant professor in the Department of Computer Science at Yale University. His research focus includes algorithms for graph learning, geometric embeddings, and trustworthy deep learning. He is the author of many widely used graph learning algorithms such as GraphSAGE, PinSAGE and GNNExplainer. Rex worked on a variety of applications of graph learning in physical simulations, social networks, NLP, knowledge graphs and biology. He developed the first billion-scale graph embedding services at Pinterest, and the graph-based anomaly detection algorithm at Amazon. Graduating with highest distinction, he received his B.S. degree at Duke, his Ph.D. at Stanford, and won the dissertation award at KDD 2022.