Collaborative Research: SBP: Scientific topics and careers at the intersection: an algorithmic approach

合作研究：SBP：交叉点上的科学主题和职业：算法方法

• 批准号: 2152303
• 负责人: Cassidy Sugimoto
• 金额: $ 21.9万
• 依托单位: Georgia Tech Research Corporation
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2022
• 资助国家: 美国
• 起止时间: 2022-04-01 至 2025-03-31
• 项目状态: 未结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2152303&HistoricalAwards=false
• 关键词: Collaborative; Research; SBP; Scientific; topics

The goal of this project is to investigate the direct harms to science wrought by structural racism and the benefits derived by the inclusion of people of color and other historically marginalized groups in the scientific workforce. Specifically, this work seeks to (a) quantify the participation of people of color and members of historically marginalized populations in the production of science, (b) elucidate their role in propelling intellectual innovations, and (c) understand how the distribution of labor and composition of scientific teams creates barriers and pathways to their scientific success. The project will support the mission of open science, by making the algorithms and publications openly available to propel this area of research. Finally, the PI team will recruit a cohort of a dozen student Fellows from a variety of disciplines and countries to discuss the ways in which they incorporate their lived experiences into research design and the challenges and barriers to this process. Priority will be given to doctoral students of color, or who identify as a member of a historically marginalized population within their country of affiliation. The goal of the fellowship is to empower students to navigate academic spaces by suggesting new topical directions with advisors, to cultivate change in terms of how authors are distributed in scientific publications, and to examine what and how science is conducted. Our research aims to empirically examine the degree to which diversity in the scientific workforce creates a more innovative and robust scientific system. The research has strong implications for all sectors of society.This research builds upon previous quantitative analyses to construct more robust and equitable algorithms that take into consideration contextual factors that influence the performance of the algorithm. To address our primary aim we use articles’ abstract, title, and keywords to train a Latent Dirichlet Allocation (LDA) model to infer the topics within a corpus of papers and the distribution of topics within each article. Data sources include millions of articles and distinct authors indexed in the Web of Science (WoS) database. To address our primary aim we will use articles’ abstract, title, and keywords to train a Latent Dirichlet Allocation (LDA) model and to extend our work on intersecting race, ethnicity, and gender inequalities in the US research landscape to citation and collaboration patterns, the role of institutional affiliation and changes over time; infer the topics within a corpus of papers, and the distribution of topics within each article. Our second aim is to determine if variation by race, ethnicity and gender identified in the US context translates to other national contexts. To address this second aim we will replicate and expand our methodology to two other scientifically productive, diverse societies. Comparison across all three nation states will allow for the identification of potentially generalizable characteristics, mechanisms that can be used to improve equity in science across the globe, and knowledge of how topicality of research in different countries is affected by the racial composition of teams. This research will provide a scalable methodological contribution that extends beyond the confines of this single research project and will allow other researchers to analyze race, ethnicity, and gender in any dataset that includes individual names.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

该项目的目标是调查结构性种族主义对科学造成的直接伤害，以及将有色人种和其他历史上被边缘化的群体纳入科学队伍所带来的好处。具体来说，这项工作旨在(a)量化有色人种和历史上被边缘化的人群在科学生产中的参与，(b)阐明他们在推动智力创新方面的作用，以及(c)了解科学团队的劳动力分配和组成如何为他们的科学成功创造障碍和途径。该项目将通过公开算法和出版物来推动这一领域的研究，从而支持开放科学的使命。最后，PI团队将从不同学科和国家招募十几名学生研究员，讨论他们如何将自己的生活经历纳入研究设计，以及这一过程中的挑战和障碍。优先考虑有色人种的博士生，或在其所属国家中被历史边缘化的群体的博士生。该奖学金的目标是通过与导师提出新的主题方向，使学生能够驾驭学术空间，培养科学出版物中作者分布方式的变化，并研究科学是什么以及如何进行的。我们的研究旨在实证检验科学劳动力的多样性在多大程度上创造了一个更具创新性和更强大的科学体系。这项研究对社会各阶层都有重大影响。本研究建立在以前的定量分析的基础上，构建更稳健和公平的算法，考虑到影响算法性能的上下文因素。为了实现我们的主要目标，我们使用文章的摘要、标题和关键词来训练潜在狄利克雷分配（LDA）模型，以推断论文语料库中的主题和每篇文章中的主题分布。数据源包括在Web of Science （WoS）数据库中索引的数百万篇文章和不同的作者。为了实现我们的主要目标，我们将使用文章的摘要、标题和关键词来训练潜在狄利克雷分配（LDA）模型，并将我们对美国研究领域中交叉种族、民族和性别不平等的研究扩展到引用和合作模式、机构隶属关系的作用和随时间的变化；推断论文语料库中的主题，以及每篇文章中主题的分布。我们的第二个目标是确定美国背景下的种族、民族和性别差异是否可以转化为其他国家的背景。为了实现第二个目标，我们将把我们的方法复制并扩展到另外两个科学上富有成效的、多样化的社会。在这三个民族国家之间的比较将允许识别潜在的可概括的特征，可用于改善全球科学公平的机制，以及不同国家研究的话题性如何受到团队种族组成的影响的知识。这项研究将提供一个可扩展的方法贡献，超出了这个单一研究项目的范围，并将允许其他研究人员在任何包含个人姓名的数据集中分析种族、民族和性别。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。