报告主题：AI Interpretation and its Applications in Smart Grid

报告人：Professor Maozhen Li,Dept. of Electronic and Electrical Engineering, Brunel University of London

时间：2025年11月28日 16:00

腾讯会议：795-488-601

报告摘要：AI models normally work in a black-box style, making it challenging for deployment in situations where the decision made by an AI model should be interpretable to users. This talk starts with a brief review on interpretation methods of AI models based on which it presents SA-CAM, a self-attention CAM (class activation map) method for visual interpretations of AI models. Further down the line, this talk presents CENet, a computation efficient lightweight AI model which can potentially be deployed on resource constrained mobile devices like drones. Towards the end, this talk touches upon the applications of AI interpretation in critical infrastructures such as smart grids and discusses how Causal AI would serve as a better solution for AI interpretation.

报告人简介：李茂贞教授现任英国布鲁内尔大学电子与电气工程系教授。他于1997年获得中国科学院软件研究所博士学位，1999-2002年在英国卡迪夫大学计算机科学系从事博士后研究工作。主要研究方向包括高性能计算、大数据分析及智能系统，并将其应用于智能电网、智能制造与智慧城市领域。在上述研究领域发表学术成果约240项，含专著4部。其中，2005年由Wiley出版社出版的专著《The Grid: Core Technologies》广受学术界认可。李教授是英国计算机学会会士、英国工程技术学会会士，曾参与30余个IEEE国际会议的组织工作，并担任多个学术期刊的编委会成员。其主持的“Big Data Modelling on Large Scale Road Networks”研究项目，于2018年5月入选英国《Computing UK》“Most Innovative Big Data Solution”类别“BIG DATA EXCELLENCE AWARDS”奖项短名单。