当前位置: 首 页 > 学术科研 > 讲座报告 > 正文

讲座 | 大数据与管理科学系列讲座(第21期)

发布时间:2022-05-10来源:管理科学系 浏览次数:

讲座题目:《数据科学的三个神话》

主讲嘉宾:Ryan Shaw

主 持 人:韩璐

讲座时间:2022年5月27日8:00-9:40(北京时间)

讲座地点:腾讯会议(会议号:576 212 574)

讲座摘要:在现今的研究中,数据科学被广泛的应用于决策优化,普遍认为数据科学的基础特征使之不仅可以优化决策,还可以通过分析技术进行稳定的可实践的概念浮现。但这些数据科学的基本假设建立在外部化表征的错误之上。在本讲座中将通过例子讨论三个这些错误假设的神话:the conduit metaphor, the knowledge pyramid, and the common substrate。这些神话需要数据科学正视。

嘉宾简介:Ryan Shaw,副教授,美国北卡罗莱纳大学教堂山分校信息与图书馆科学学院本科项目主管。2010年他于美国加利福尼亚大学,伯克利信息科学学院获得博士学位。2012年,他获得了博物馆和图书馆服务研究所(Institute of Museum and Library Services)的一项为期三年的发展资助,以发明新工具计算文本处理技术来组织历史收藏。他还是编辑笔记项目(Editors’NotesProject)的联合首席执行官,该项目由梅隆基金会(Mellon Foundation)资助,旨在为人文主义者开发开放、协作的软件记录工作。他还是PeriodO项目的联合首席执行官,该项目由国家环境卫生研究院(NEH)资助,对历史、艺术历史和考古学时期的文本进行学术断言和预测。他在研究领域发表了50余篇论文,主办了10余场国际会议。详见:https://sils.unc.edu/people/faculty/profiles/Ryan-Shaw

 

主办单位:新葡的京集团8814登录入口


Topic:Three Foundational Myths of Data Science

Lecturer:Ryan Shaw(Associate Professor, Undergraduate Program Coordinator, UNC School of Information and Library Science)

Host:Lu Han

Time:2022 May 27,08:00-09:40 (Beijing, China)

Software:Tencent VooV Meeting (Num:576 212 574)

Abstract:A wide variety of social, economic, and cultural practices are mediated by networked information and communication systems, which generate fine-grained documentary traces of these practices: data. The analysis of that data to optimize decision-making has been dubbed “data science.” Data science techniques have undoubtedly proven to be effective for optimizing well-defined decision procedures. Yet the characterization of data science as a “science” suggests that it can produce not only optimized decisions, but also stable, quantifiable concepts for understanding the practices that generated the data. Some even argue that, given enough data, such concepts will simply “emerge” from analysis. This view of data science rests on a foundation of faulty assumptions about the role of external representations in facilitating communication. In my talk I will discuss three “myths” that exemplify these faulty assumptions: the conduit metaphor, the knowledge pyramid, and the common substrate. A data science worthy of the name will need to abandon these myths.