题　目：LLM and FM for Time Series Analysis and Applications: from AIOps to EdTech

嘉　宾：文青松 首席科学家兼AI研究院负责人 松鼠AI

主持人：许欢 教授 上海交通大学安泰经济与管理学院

时　间：2025年7月8日（周二）10:00-11:30

地　点：安泰楼B207室

内容简介：

Time series data underpin critical decisions across cloud operations, energy management, finance, and education, and recent breakthroughs in large language models (LLMs) and foundation models (FMs) are reshaping how we model temporal signals, setting new benchmarks in forecasting, anomaly detection, classification, and causal reasoning. This talk first provides an up-to-date overview of the fast-moving LLM/FM landscape for time-series analysis, then highlights our recent research contributions—Time-LLM, Time-VLM, Time-MoE, Time-MQA, and more—demonstrating their impact on key tasks and real-world applications from AIOps to EdTech. Finllay, we conclude by outlining the transformative potential of these models and pointing to the promising directions for future research.

演讲人简介：

Qingsong Wen is currently the Head of AI & Chief Scientist at Squirrel Ai Learning. Before that, he worked at Alibaba, Qualcomm, Marvell, etc., and received his M.S. and Ph.D. degrees in ECE from Georgia Institute of Technology, USA. His research interests include AI for Time Series, AI for Education, LLM & AI Agent. He has published around 150 top-ranked AI conference and journal papers, had multiple Oral/Spotlight Papers at NeurIPS, ICML, ICLR, ACL, AAAI, had multiple Most Influential Papers at IJCAI, received multiple Innovative Application Awards at AAAI, and won First Place of SP Grand Challenge at ICASSP. Currently, he serves as Chair of IEEE Task Force on AI for Time Series and Spatio-Temporal Data, and Vice Chair of INNS AI for Education Section. He also serves as Associate Editor for IEEE Transactions on Pattern Analysis and Machine Intelligence, IEEE Signal Processing Letters, and Neurocomputing. In addition, he regularly serves as Area Chair of the top conferences including NeurIPS, ICML, KDD, IJCAI, ICASSP, etc.

欢迎广大师生参加！