您当前位置: 首页  >  新闻中心  >  学院通知  >  正文

学院通知

计算机科学技术名家讲座(十一)——胡侠

发布日期:2018-06-07 发布人:科研办 点击量:

报告题目:Human-centric Machine Learning in the Big Data Era

报告时间:2018615 周五 900 - 1000

报告地点:吉林大学 计算机楼 A521

报告人:   胡侠

报告人简介:胡侠博士,博士生导师,现任美国德州农工大学数据挖掘实验室主任、计算机学院终身教职系列助理教授。长期从事人工智能,机器学习,社交媒体和医学数据挖掘研究,在相关顶级国际会议及期刊(CCF A/B类,包括KDD, WWW, SIGIR, IJCAI, AAAI, TKDD, TKDE等)发表论文100余篇,他引超过3000次。获美国国家科学基金委杰出青年科学基金资助(NSF CAREER Award, IJCAI 2016 BOOM 最佳论文奖,WSDM'13 国际会议最佳学生论文提名,亚利桑那州立大学校长奖,IEEE Atluri学者奖。长期担任国际顶级期刊编委(包括TKDE, TKDD, TOIS, TIST等),及顶级国际学术会议程序委员会委员(包括NIPS, KDD, IJCAI, AAAI等)。现作为负责人承担美国国家自然科学基金、各国家部委多项人工智能相关科研课题,与工业界合作紧密,承担多项包括阿里巴巴,Ingersoll RandApple等公司相关科研课题

报告摘要: This talk will cover the recent progress of several ongoing projects directed by Dr. Xia "Ben" Hu on human-centric machine learning. Though machine learning has achieved a lot of success in different applications, it is still challenging for domain experts to easily make use of machine learning systems and algorithms in applications that matter, especially in the big data era. To bridge the gap, we propose to develop interpretable and automated machine learning systems to handle big data through the following efforts. First, we will present a specific example using social spammer detection to show how we can develop machine learning algorithms to handle large-scale, heterogeneous and dynamic big data? Second, we will discuss the system architecture and main algorithms, as well as our current progress, to bridge the gap between powerful deep learning algorithms and interpretable shallow models. This is to answer the question how we could enable interpretable machine learning. Third, we will briefly introduce the system architecture and main algorithms, as well as our current progress, to develop an end-to-end automated machine learning system. This is to show that how we could enable automated machine learning for domain experts with limited data science background.


主办单位:

吉林大学计算机科学与技术学院

吉林大学软件学院

吉林大学计算机科学技术研究所

符号计算与知识工程教育部重点实验室