题目：Evidential Reasoning and Data-Driven Decision Modelling（证据推理和数据驱动的决策建模）

时间：2026年9月2日 14:00-17:00

地点：数学院南楼402会议室

报告摘要：The evidential reasoning (ER) rule has been established from the seminal Dempster-Shafer (D-S) theory of evidence to combine multiple pieces of information conjunctively. Through implementing the orthogonal sum operation on weighted belief distributions with reliabilities, the ER rule takes into account both individual and collective support from multiple pieces of evidence reasonably, and it constitutes a generalised Bayesian inference process. The ER rule can be extended to model the causal relationship between antecedent attributes and the consequent as well as to aggregate information for multiple criteria decision analysis. Optimal learning can also be constructed to train prior parameters in the ER rule-based approximate reasoning model when data are available.

报告人简介：陈玉旺，英国曼彻斯特大学教授，人工智能与决策科学研究中心联合主任。2008年于上海交通大学控制科学与工程专业获工学博士学位，2021年至2023年兼任阿兰·图灵研究所Turing Fellow。主要从事智能推理与决策、数据科学和系统科学等相关领域的理论和应用研究。近年来作为项目负责人或主要完成人承担完成Innovate UK, EPSRC, NSFC等资助的科研项目十余项。并在EJOR, JORS, Omega, DSS, IS, IEEE T-SMC等期刊和国际会议上发表了学术论文100多篇。