On June 11th, 2026, the 10th Peking-Tsinghua Joint Statistics Colloquium took place at the Lyu Dalong Building, Tsinghua University. Established jointly by the Center for Statistical Science at Peking University and the Department of Statistics and Data Science at Tsinghua University, the colloquium has reached its tenth consecutive session, with this year's edition organized by Tsinghua's Department of Statistics and Data Science. Dedicated to the latest developments in statistics and data science, the colloquium drew a diverse audience, including faculty and students from the two host institutions, as well as participants from other universities and researchers from industry. The event was co-chaired by Associate Professor Hanzhong Liu of Tsinghua University and Associate Professor Ruixun Zhang of Peking University.
Associate Professor Hanzhong Liu
Associate Professor Ruixun Zhang
Professor Jun Liu, Head of the Department of Statistics and Data Science at Tsinghua University and Member of the U.S. National Academy of Sciences, and Professor Fang Yao, Director of the Center for Statistical Science at Peking University, each delivered opening remarks. Professor Jun Liu drew on his own academic journey to reflect on the long-standing and deep intellectual ties between the two universities. Professor Fang Yao reviewed the decade-long trajectory of collaborative development between the two statistics programs and looked back on the history of the Peking-Tsinghua Statistics Colloquium.
Professor Jun Liu
Professor Fang Yao
Professor Ke Deng from Tsinghua University delivered a keynote presentation titled AI × Statistics: Integrating AI and Statistics for a Brighter Future of Data Science. The talk centered on the opportunities and challenges facing statistics in the era of artificial intelligence, highlighting the significant complementarity between statistics and AI in areas such as data types, model architecture, interpretability, and uncertainty quantification. Drawing on case studies including surgical video analysis, multimodal learning, and nonparametric density estimation, Professor Deng demonstrated how statistical modeling principles can be integrated with deep learning methods, and how statistical tools can enhance the recognition accuracy and data efficiency of AI models. He also explored the potential of the pre-training and fine-tuning framework for statistical innovation. The talk underscored that the fundamental principles of statistics remain highly valuable in the age of AI, and that the deep integration of AI and statistics will open new directions for the future of data science.
Professor Ke Deng
Professor Yijuan Hu from Peking University delivered a keynote presentation titled SMS: Symmetric Mediation Statistics for Powerful High-Dimensional Mediation Analysis. The talk focused on high-dimensional mediation analysis in the context of omics data, highlighting its critical role in uncovering complex biological mechanisms and the statistical challenges it entails. In response to the limitations of existing methods in handling composite null hypotheses and achieving adequate statistical power, Professor Hu introduced a novel framework—Symmetric Mediation Statistics (SMS)—which addresses composite null hypotheses holistically while accommodating imbalances in association strength across different mediators, thereby enabling more effective identification of mediating variables. Through simulations and real-world case studies, including metabolomics research on childhood obesity, the presentation demonstrated the strong performance and broad application potential of the SMS method in high-dimensional biomedical data analysis, offering a new statistical tool for omics-driven causal mechanism research.
Professor Yijuan Hu
Poster Presentation Session
The colloquium then moved to a poster presentation and evaluation session, featuring poster displays by students from both Tsinghua University and Peking University, showcasing their respective research findings. Participating faculty and students engaged in in-depth discussions, fully demonstrating the research innovation capabilities and academic excellence of the young scholars from both institutions. Faculty members from both universities carefully evaluated the presented posters. After comprehensive review, four Outstanding Poster Awards were selected. Kun Qian and Zhifei Wang from Peking University, along with Dong Huang and the collaborative team of Yaqi Zhou and Zihan Wang (who co-presented a single poster) from Tsinghua University, received the "10th Peking-Tsinghua Joint Statistics Colloquium Outstanding Poster Award." Additionally, Jingkun Qiu from Peking University and Yicheng Li from Tsinghua University were honored with the "10th Peking-Tsinghua Joint Statistics Colloquium Outstanding Graduate" title.
Scenes from the Colloquium
Award Presentation Session
The 10th Peking-Tsinghua Joint Statistics Colloquium thus drew to a successful conclusion. We eagerly anticipate our reunion at next year's edition!
Edited by Haoyang Yu
Reviewed by Hanzhong Liu, Lin Hou, Lefei Li
