NRT-HDR: Detecting and Addressing Bias in Data, Humans, and Institutions

数据科学已经成为产生新知识、推动创新和处理社会最紧迫问题的强大工具。利用这场数据革命需要扩大美国公民在数据科学劳动力中的参与，特别是女性和其他代表性不足的群体。国家科学基金会研究培训(NRT)奖授予石溪大学(纽约州立大学石溪分校)，将培养一批不同的学生，使他们在数据科学和以人为中心的科学方面都能熟练掌握。该项目将使学生能够在将数据科学与深厚的学科知识相结合的问题领域进行趋同研究。该项目预计将培训六十八(68)名博士生，其中包括三十四(34)名由NRT资助的受训人员，他们来自以人为中心的科学(包括心理学、语言学、经济学、神经生物学和行为学、政治学和社会学)和数据科学(包括计算机科学和应用数学与统计学)。在美国，以科学为导向的国内学生倾向于聚集在不同的竖井中，HCS领域的女性比例明显高于DS领域。HCS学生在其所在领域的传统经验方法方面接受了良好的培训。然而，许多人并不认为自己是“数据科学家”。这些HCS学生可能没有准备好潜入计算机科学的研究生课程，以获得人工智能、机器学习和大数据等领域的尖端研究技能，从而支持他们的研究。许多DS学生对如何收集数据集及其局限性缺乏深入的理论和实践理解。此外，DS学生可能缺乏评估基于数据的技术对人类和社会的影响的技能。在这种双途径培训模式中，DS学员将参加HCS领域的研究生课程，获得以人为中心的科学证书。与此同时，HCS学员将参加任何必要的强化桥牌课程，为他们在DS中的研究生课程做准备，并获得人工智能证书。受训者将在三个团队授课的研究实践中聚集在一起，围绕一个相关和包容的主题：对数据、对人类和对机构的偏见。这些实践将侧重于发现如何使用强大的量化方法来掩盖偏见(无论是隐性的还是显性的)，在给定的背景下，这些偏见对一些人有利，对另一些人不利。受训者和培训者都将练习在他们自己和他们的研究中以及在社会和机构环境中发现和解决偏见的技术。他们将共同加入针对普遍存在的不平等和技术伦理使用问题的研究合作。NSF研究培训(NRT)计划旨在鼓励开发和实施大胆的、具有潜在变革意义的STEM研究生教育培训模式。该计划致力于通过创新的、基于证据的、与不断变化的劳动力和研究需求保持一致的综合实习生模式，在高度优先的跨学科或趋同研究领域对STEM研究生进行有效培训。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Data science has become a powerful tool for generating new knowledge, fueling innovation, and dealing with society's most pressing problems. Harnessing this data revolution requires broadening the participation of U.S. citizens in the data science workforce, especially of women and other under-represented groups. This National Science Foundation Research Traineeship (NRT) award to Stony Brook University (the State University of New York at Stony Brook) will prepare a diverse pool of students to become fluent in both data sciences and human-centered sciences. The program will enable students to conduct convergent research in problem domains that blend data science with deep disciplinary knowledge. The project anticipates training sixty-eight (68) Ph.D. students, including thirty-four (34) NRT-funded trainees, recruited from both the human-centered sciences (HCS, consisting of Psychology, Linguistics, Economics, Neurobiology & Behavior, Political Science, and Sociology) and the data sciences (DS, consisting of Computer Science and Applied Mathematics & Statistics). In the U.S., scientifically oriented domestic students tend to gather in distinct siloes, with significantly higher proportions of women in HCS than in DS fields. HCS students are well-trained in the traditional empirical methods of their fields. However, many do not identify as "data scientists." These HCS students may lack the preparation to dive into graduate coursework in computer science to acquire cutting-edge research skills in areas such as artificial intelligence, machine learning, and big data that could empower their research. And many DS students lack a deep theoretical and practical understanding of how datasets are collected and their limitations. Moreover, DS students may lack the skills to assess the impacts of data-based technology upon human beings and society. In this two-pathway training model, DS trainees will take graduate coursework in an HCS domain, earning a certificate in the Human-Centered Sciences. Meanwhile, HCS trainees will take any needed intensive bridge courses to prepare them for graduate coursework in DS and earn an Artificial Intelligence certificate. Trainees will come together in three team-taught research practica centered on a relevant and inclusive theme: bias—in data, in humans, and in institutions. These practica will focus on discovering how the use of powerful quantitative methods can camouflage biases (both implicit and explicit) that, in a given context, advantage some human beings and disadvantage others. Both trainees and trainers will practice techniques for detecting and addressing biases in themselves and their research, as well as within social and institutional settings. Together, they will join convergent research collaborations that target pervasive problems of inequity and the ethical uses of technology. The NSF Research Traineeship (NRT) Program is designed to encourage the development and implementation of bold, new potentially transformative models for STEM graduate education training. The program is dedicated to effective training of STEM graduate students in high priority interdisciplinary or convergent research areas through comprehensive traineeship models that are innovative, evidence-based, and aligned with changing workforce and research needs.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.