报告人:庞观松 博士
报告时间:2023年10月30日(星期一)10:00-12:00
报告地点:琶洲实验室一楼会议室
报告名称:Deep Learning of the Unknowns:Anomaly and Out-of-distribution Detection Perspectives
腾讯会议ID: 796-230-243
报告人简介:庞观松博士,新加坡管理大学计算与信息系统学院助理教授、博导。在此之前,他曾在阿德莱德大学的澳大利亚机器学习研究所担任研究员。主要研究数据挖掘、机器学习、计算机视觉算法及其在信息安全、金融科技、医疗健康、生物信息学、生物识别等领域的应用。他在国际顶级数据科学相关会议和期刊上发表50多篇论文,包括KDD、NeurIPS、CVPR、ICCV、ECCV、AAAI、IJCAI、ACM MM、WSDM、IEEE TKDE、IEEE TMM、IEEE TMI、ACM CSUR、ACM TKDD和DMKDJ等。此外,他还担任这些会议的高级程序委员会成员、领域主席或审稿人。他的个人Google Scholar引用次数约5,000次。庞博士曾入选悉尼科技大学2020年校长奖名单,并于2022年和2023年入选斯坦福大学公布的全球顶尖2%科学家。此外,他还获得了DSAA2023应用领域最佳论文奖。他还是IEEE Intelligent Systems 和 International Journal of Data Science and Analytics期刊编委。
报告摘要:Being able to say NO to abnormal/unknown instances is one key capability that machine learning systems should master in broad real-world application domains, especially for security/safety-critical domains, e.g., detecting and rejecting suspicious financial crimes in banking, handling unknown objects in autonomous driving or medical AI systems, etc. Anomaly and out-of-distribution detection are two lines of research to achieve this capability. Despite current approaches who have obtained substantially improved detection accuracy in both research lines, learning of the unknowns is still an open problem due to the difficulty of obtaining instances of the unknowns and their unbounded distribution.