FAI: BRIMI - Bias Reduction In Medical Information

FAI：BRIMI - 减少医疗信息中的偏差

• 批准号: 2147305
• 负责人: Shiri Dori-Hacohen
• 金额: $ 39.3万
• 依托单位: University of Connecticut
• 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: 2022-03-15 至 2025-02-28
• 项目状态: 未结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2147305&HistoricalAwards=false
• 关键词: FAI; BRIMI; Bias; Reduction; Medical

This award, Bias Reduction In Medical Information (BRIMI), focuses on using artificial intelligence (AI) to detect and mitigate biased, harmful, and/or false health information that disproportionately hurts minority groups in society. BRIMI offers outsized promise for increased equity in health information, improving fairness in AI, medicine, and in the information ecosystem online (e.g., health websites and social media content). BRIMI's novel study of biases stands to greatly advance the understanding of the challenges that minority groups and individuals face when seeking health information. By including specific interventions for both patients and doctors and advancing the state-of-the-art in public health and fact checking organizations, BRIMI aims to inform public policy, increase the public's critical literacy, and improve the well-being of historically under-served patients. The award includes significant outreach efforts, which will engage minority communities directly in our scientific process; broad stakeholder engagement will ensure that the research approach to the groups studied is respectful, ethical, and patient-centered. The BRIMI team is composed of academics, non-profits, and industry partners, thus improving collaboration and partnerships across different sectors and multiple disciplines. The BRIMI project will lead to fundamental research advances in computer science, while integrating deep expertise in medical training, public health interventions, and fact checking. BRIMI is the first large scale computational study of biased health information of any kind. This award specifically focuses on bias reduction in the health domain; its foundational computer science advances and contributions may generalize to other domains, and it will likely pave the way for studying bias in other areas such as politics and finances. BRIMI has the following objectives: (a) identifying and analyzing bias and language misuse online; (b) advancing the understanding of how misinformation spreads amongst different populations; and (c) triaging health topics with the biggest harms, and creating and disseminating triage guidelines to public health officials and practitioners. BRIMI will develop novel artificial intelligence approaches both to establish health information inequities empirically, and to reduce them. The methods used include large-scale online and social network data collection and a content analysis approach to annotating complex health data; supervised, semi-supervised and transfer learning to detect biased and false health information; controversy and misinformation analysis using community detection, stance detection and claim detection; and intervention design methods based on best practices in public health. The award’s research contributions will include: (a) novel metrics to computationally define biased health information and characterize its dissemination online and in social media, including specifically within divergent population groups; (b) utilizing transfer learning and semi-supervised approaches, in order to generalize solutions developed on and for medical language to lay language; (c) analyzing disagreement within and across populations on health information, which in turn requires improvement in stance detection and claim matching approaches; and (d) novel computational approaches to triage and prioritize misinformation for the purposes of mitigation.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.

该奖项名为“减少医疗信息偏见”（Bias Reduction In Medical Information，BRIMI），重点关注使用人工智能（AI）来检测和减轻对社会中少数群体造成不成比例伤害的偏见、有害和/或虚假健康信息。BRIMI为提高健康信息的公平性、改善人工智能、医学和在线信息生态系统的公平性（例如，健康网站和社交媒体内容）。BRIMI对偏见的新研究将大大促进对少数群体和个人在寻求健康信息时面临的挑战的理解。通过包括针对患者和医生的具体干预措施，并推进公共卫生和事实核查组织的最新发展，BRIMI旨在为公共政策提供信息，提高公众的批判性素养，并改善历史上服务不足的患者的福祉。该奖项包括重要的外联工作，这将使少数群体直接参与我们的科学过程;广泛的利益相关者参与将确保对所研究群体的研究方法是尊重的，道德的和以患者为中心的。BRIMI团队由学术界、非营利组织和行业合作伙伴组成，从而改善了不同部门和多个学科之间的合作和伙伴关系。BRIMI项目将推动计算机科学的基础研究进展，同时整合医疗培训、公共卫生干预和事实核查方面的深厚专业知识。BRIMI是第一个对任何类型的有偏见的健康信息进行大规模计算研究的项目。该奖项特别关注健康领域的偏见减少;其基础计算机科学的进步和贡献可能会推广到其他领域，并可能为研究政治和金融等其他领域的偏见铺平道路。BRIMI有以下目标：（a）识别和分析网上的偏见和语言滥用;（B）增进对错误信息如何在不同人群中传播的理解;（c）对危害最大的健康主题进行分类，并为公共卫生官员和从业人员制定和传播分类指南。BRIMI将开发新的人工智能方法，以经验性地建立健康信息不平等，并减少这些不平等。使用的方法包括大规模在线和社交网络数据收集和内容分析方法，以注释复杂的健康数据;监督，半监督和转移学习，以检测有偏见和虚假的健康信息;使用社区检测，立场检测和索赔检测的争议和错误信息分析;以及基于公共卫生最佳实践的干预设计方法。该奖项的研究贡献将包括：（a）新的指标，以计算方式定义有偏见的健康信息，并描述其在线和社交媒体的传播，特别是在不同的人群中;（B）利用迁移学习和半监督方法，以便将医学语言开发的解决方案推广到非专业语言;（c）分析人口内部和人口之间在健康信息方面的分歧，这反过来又要求改进立场检测和主张匹配方法;以及（d）新的计算方法来分类和优先考虑错误信息，以减轻损失。该奖项反映了NSF的法定使命，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。

项目成果

Information Ecosystem Threats in Minoritized Communities: Challenges, Open Problems and Research Directions

少数民族社区的信息生态系统威胁：挑战、悬而未决的问题和研究方向

• DOI: 10.1145/3477495.3536327
• 发表时间: 2022
• 期刊: SIGIR '22: Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval
• 作者: Dori-Hacohen, Shiri;Hale, Scott A.
• 通讯作者: Hale, Scott A.

Quantifying Misalignment Between Agents

量化代理之间的偏差

• 发表时间: 2022
• 期刊: ML Safety @ NeurIPS 2022
• 作者: Kierans, A.;Hazan, H.;Dori-Hacohen, S.
• 通讯作者: Dori-Hacohen, S.

Fairness via AI: Bias Reduction in Medical Information

通过人工智能实现公平：减少医疗信息的偏见

• 发表时间: 2021
• 期刊: The 4th FAccTRec Workshop on Responsible Recommendation at RecSys 2021
• 作者: Dori-Hacohen, S.;Montenegro, R.;Murai, F.;Hale, S. A.;Sung, K.;Blain, M.;Edwards-Johnson, J.
• 通讯作者: Edwards-Johnson, J.

Benchmarked Ethics: A Roadmap to AI Alignment, Moral Knowledge, and Control

道德基准：人工智能联盟、道德知识和控制的路线图

• DOI: 10.1145/3600211.3604764
• 发表时间: 2023
• 期刊: and Society
• 作者: Kierans, Aidan