主 讲 人：李凌风    

活动时间：2026-05-12 14:30:00

地　　点：理科2号楼B409

主办单位：金沙以诚为本赢在信誉

讲座内容：

Particle physics is expecting a precision era with the upcoming experiments and facilities being built. A key aspect of this feature is that the resolution of SM processes will be much greater than the SM predictions. In such scenarios, traditional analysis techniques cannot handle the large amount of information collected by the detector, resulting in the waste of useful information. Machine learning techniques, on the other hand, can extract useful information from high-dimensional and noisy data. The use of machine learning techniques is, therefore, unavoidable in the precision era. Unlike previous efforts in HEP that focused on signal-background classification, there is often no "signal" defined for precision measurements. Instead, the analysis focuses on theoretical parameters, such as EFT Wilson coefficients, and the final states only differ by differential distributions. We will discuss recent progress in various levels of machine learning techniques, from input to network structure, and conclude with strategy and loss functions. We stress new theoretical developments in terms of simulation-based inference (SBI) for measuring EFT coefficients.

主讲人介绍：

李凌风博士，中国科学院大学国际理论物理中心工作。于2013年本科毕业于北京大学，后在加州大学戴维斯分校师从Hsin-Chia Cheng教授，于2018年获理论高能物理博士学位。此后，他先后在香港科技大学和布朗大学从事博士后研究。他的研究方向包括利用新模型、新方法和新思维来精确检验标准模型，并探索新物理。