PI（July 2020 to present）
Special-term Research Fellow、Assistant Professor
Yang Yuan is now an assistant professor at IIIS, Tsinghua. He finished his undergraduate study at Peking University in 2012. Afterwards, he received his PhD at Cornell University in 2018, advised by Professor Robert Kleinberg. During his PhD, he was a visiting student at MIT/Microsoft New England (2014-2015) and Princeton University (2016 Fall). Before joining Tsinghua, he spent one year at MIT Institute for Foundations of Data Science (MIFODS) as a postdoc researcher. He works on AI+Healthcare, AI Theory and Applied Category Theory.
Beijing Zhiyuan Young Scientists (2020)
2019 Forbes China 30 under 30 (Science List)
AI+healthcare, AI theory and applied category theory
Category theory is the theoretical foundation of our system.
We start from traditional Chinese medicine (TCM), gradually add elements from modern medicine, and will finally build a new healthcare system that is purely data driven.
The symptoms and medicines in TCM, are exactly two categories. The treatment of a disease is a functor from the symptom category to the medicine category.
Use category theory to analyze the power of foundation models, learning of concept, and present a framework of general intelligence
AI-Interpretability in Machine Learning
The rapid development of deep learning technology has brought revolutionary progress to many fields, and the performance of deep learning models even surpasses that of humans in many tasks, especially in fields like image recognition and natural language processing. Our research group is committed to exploring cutting-edge AI interpretability, striving to not only understand the results of deep learning, but also why these results are achieved. We also plan to deploy our research findings in medical scenarios, developing more reliable deep learning assisted diagnosis systems.
Yang Yuan's single authored paper On the Power of Foundation Models has been accepted by ICML’2023. This paper introduces category theory to the analysis of foundation models for the first time, and theoretically characterizes the power of foundation models.
•On the Power of Foundation Models, Yang Yuan, ICML 2023.
•On Uni-Modal Feature Learning in Supervised Multi-Modal Learning, Chenzhuang Du, Jiaye Teng, Tingle Li, Yichen Liu, Tianyuan Yuan, Yue Wang, Yang Yuan, Hang Zhao, ICML 2023.
•Finding Generalization Measures by Contrasting Signal and Noise, Jiaye Teng, Bohang Zhang, Ruichen Li, Haowei He, Yequan Wang, Yan Tian, Yang Yuan, ICML 2023.
•Predictive Inference with Feature Conformal Prediction, Jiaye Teng*, Chuan Wen*, Dinghuai Zhang*, Yoshua Bengio, Yang Gao, Yang Yuan,ICLR 2023. •Jiaye Teng, Jianhao Ma, Yang Yuan，Towards Understanding Generalization Via Decomposing Excess Risk Dynamics，2022 International Conference on Learning Representations（ICLR), 2022.
•Jiaye Teng , Zeren Tan , Yang Yuan，T-SCI: A Two-Stage Conformal Inference Algorithm with Guaranteed Coverage for Cox-MLP，2021 International Conference on Machine Learning（ICML), 2021