应我校数学科学学院邀请，复旦大学姚烨副教授将于2026年1月18日为我校数学专业教师和研究生讲学。欢迎数学科学学院及全校相关教师、博士生、硕士生参加！

报告题目：A Coupled Network Framework for Decoding Sentiment-Driven Human Mobility Using LLMs

报告人：姚烨副教授

报告人单位：复旦大学

时间：2026年1月18日（周日）上午9:00

地点：黑龙江大学丽泽雅苑四楼报告厅

主办单位：重点建设与发展工作处、数学科学学院

姚烨副教授简介：复旦大学公共卫生学院姚烨副教授，主要从事医学大数据驱动的人工智能与统计模拟研究，主要研究方向为基于马尔可夫模型的参数估计与模拟以及传染病数学建模。其近几年在医学著名期刊European Respiratory Journal、Journal of Infection、eLife、BMC Medicine、BMJ Global Health、Infectious Disease Modelling、International Journal of Infectious Disease、Infectious Diseases of Poverty等杂志以第一作者/通讯作者身份发表过多篇论文，包括三篇3高被引论文、1篇热点论文。研究成果亦已通过各级内参/专报，以政策建议的形式报送各级政府，获得国家疾控局发函肯定与中央政治局委员批示。另外，其已经主持过多项自然科学基金类项目以及若干横向课题，近五年总经费超过400万。姚烨老师先后获得复旦大学仲英青年学者、上海市科技进步二等奖（2/10）、上海市卫生健康青年人才、复旦大学青年教师教学竞赛一等奖、一健康基金优秀教师二等奖、复旦大学抗击新冠疫情先进个人。

报告摘要：We propose a dual-layer coupled network framework to quantify how online sentiment dynamics act as a leading indicator of human mobility in mosquito-borne disease systems. Using dengue data from Xishuangbanna, China, we integrate a physical mobility layer with an information diffusion layer constructed from social media sources. Large language models are employed to extract time-varying sentiment distributions, and emotional fluctuations are quantified using the 1D Wasserstein distance. Our results show that sentiment shifts driven by official information sources are more strongly correlated with population inflow and outflow than those from individual accounts. Time-lag analysis reveals a consistent decision window of approximately three weeks between sentiment change and mobility response. The strength of sentiment–mobility coupling further varies spatially and is associated with vector-borne exposure in origin–destination regions. This study establishes sentiment as a predictive behavioral signal and provides a mathematically grounded framework for early intervention in mobility-driven epidemic risk.

欢迎各位老师同学届时参加！