FW-HTF-RL: Collaborative Research: Future expert work in the age of "black box", data-intensive, and algorithmically augmented healthcare

FW-HTF-RL：协作研究：“黑匣子”、数据密集型和算法增强医疗保健时代的未来专家工作

• 批准号: 1928586
• 负责人: Mark Riedl
• 金额: $ 49.87万
• 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: 2019
• 资助国家: 美国
• 起止时间: 2019-09-01 至 2024-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1928586&HistoricalAwards=false
• 关键词: FW; HTF; RL; Collaborative; Research

The nature of expert work is changing. Technological advances such as artificial intelligence and data science increasingly enable new computerized tools and products that make predictions and recommendations which were previously made by human experts. However, many of these new tools are "black boxes" whose inner workings are often not understood by their users, place demands that create cognitive load, and de-emphasize abstract problem solving. As these technologies are being deployed, there is little understanding of how they affect experts' work practices, perceptions of the value of work, and the expert-client relationship. Foundational research is needed in order to understand and improve work in an age of data-intensive enhanced cognition, especially in healthcare where such new technologies are rapidly changing expert work. This project is expected to transform the future of expert work through a combined redesign of technology, workflow, and interactions. It will lead to: a healthier and better-informed population; efficient deployment of human capabilities in restructured healthcare occupations; healthcare providers reducing the proportion of time spent on repetitive tasks while increasing time devoted to value-adding, meaningful activities; guidelines on design and delivery of cognition-augmenting expert advice; and students who are well versed in cross-disciplinary research on cognition-augmenting technologies in the workplace.The project's goals are: i) to study the relationships between experts, patients, and technologies in a multidisciplinary way; ii) to develop new ways for these technologies to serve experts and clients; and iii) to make expert work more responsive, value-adding, and meaningful. The project includes two strands. In the "Understand" strand, the interactions between experts, clients and cognition-augmenting technologies are examined. In the "shape" strand, the project lays the foundations for technological and organizational interventions that will make the interactions between experts, clients, and technology more effective and empowering. With a multidisciplinary team including researchers in computer science, human-computer interaction, dynamical systems, and organization alongside with medical clinicians, the project will contribute: i) scalable approaches toward quantifying the benefits and drawbacks of cognition-augmented interactions, as well as measuring information flow in relationships between experts, clients, and cognition augmenting technologies; ii) insights into when, why, and how cognition-augmenting technologies are experienced as expertise enhancing, rather than degrading; iii) data-driven methodologies to predict the effects of technical and organizational interventions on experts' work and experts' interaction with patients; iv) novel tools and workflows for experts and clients to interact with black-box cognition-augmenting technologies; v) modeling how representation of problems can be embedded in expert; and vi) systematic exploration of explanation and dialogue interventions with regard to how they affect experts' work and expert-client relationship.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.

专家工作的本质正在发生变化。人工智能和数据科学等技术进步越来越多地实现了新的计算机工具和产品，这些工具和产品可以做出以前由人类专家提出的预测和建议。但是，这些新工具中有许多是“黑匣子”，其用户通常无法理解其内部工作，而是造成认知负载的地方要求，并逐渐强调抽象问题解决。随着这些技术的部署，对它们如何影响专家的工作实践，对工作价值的看法以及专家 - 客户关系的了解几乎没有理解。为了理解和改善数据密集型认知时代的工作，尤其是在此类新技术正在迅速改变专家工作的医疗保健中，需要基础研究。预计该项目将通过技术，工作流和互动的综合重新设计来改变专家工作的未来。它将导致：一个更健康，更有信息的人口；有效地部署人类能力在重组的医疗保健职业中；医疗保健提供者减少了在重复任务上花费的时间的比例，同时增加了专门用于增值，有意义的活动的时间；关于认知提升专家建议的设计和交付指南；精通工作场所认知提升技术的跨学科研究的学生。项目的目标是：i）以多学科的方式研究专家，患者和技术之间的关系； ii）为这些技术提供新的方式为专家和客户提供服务； iii）使专家的工作更加敏感，增值和有意义。该项目包括两个股。在“理解”链中，研究了专家，客户和认知提升技术之间的相互作用。在“形状”链中，该项目为技术和组织干预措施奠定了基础，这将使专家，客户和技术之间的互动更加有效和授权。通过一个多学科的团队，包括计算机科学领域的研究人员，人力计算机互动，动力学系统和组织以及与临床医生一起，该项目将有助于：i）可扩展的方法，以量化认知互动的益处和缺点，并测量专家，客户，客户，和增强技术的关系流动中信息流中的信息流； ii）洞悉何时，为什么以及如何作为增强专业知识而不是退化的专业知识来体验认知提升技术； iii）数据驱动的方法论，以预测技术和组织干预对专家与患者的工作和专家互动的影响； iv）专家和客户与黑盒认知提升技术互动的新颖工具和工作流程； v）建模如何将问题表示可以嵌入专家； vi）关于他们如何影响专家的工作和专家 - 客户关系的解释和对话干预措施的系统探索。该奖项反映了NSF的法定任务，并被认为是通过基金会的知识分子和更广泛影响的评估审查标准的评估来通过评估来获得支持的。

项目成果

Operationalizing Human-Centered Perspectives in Explainable AI

• DOI: 10.1145/3411763.3441342
• 发表时间: 2021-05
• 期刊: Extended Abstracts of the 2021 CHI Conference on Human Factors in Computing Systems
• 作者: Upol Ehsan;Philipp Wintersberger;Q. Liao;Martina Mara;M. Streit;Sandra Wachter;A. Riener;Mark O. Riedl

Expanding Explainability: Towards Social Transparency in AI systems

• DOI: 10.1145/3411764.3445188
• 发表时间: 2021-01-01
• 期刊: CHI '21: PROCEEDINGS OF THE 2021 CHI CONFERENCE ON HUMAN FACTORS IN COMPUTING SYSTEMS
• 作者: Ehsan, Upol;Liao, Q. Vera;Weisz, Justin D.
• 通讯作者: Weisz, Justin D.

Social Construction of XAI: Do We Need One Definition to Rule Them All?

XAI 的社会构建：我们需要一个定义来统治它们吗？

• 发表时间: 2022
• 期刊: NeurIPS workshop on Human-centered AI
• 作者: Ehsan, Upol;Riedl, Mark O.
• 通讯作者: Riedl, Mark O.

Human-centered Explainable AI: Towards a Reflective Sociotechnical Approach

• DOI: 10.1007/978-3-030-60117-1_33
• 发表时间: 2020-02
• 期刊: ArXiv
• 作者: Upol Ehsan;Mark O. Riedl

Beyond Prompts: Exploring the Design Space of Mixed-Initiative Co-Creativity Systems

• DOI: 10.48550/arxiv.2305.07465
• 发表时间: 2023-05
• 期刊: ArXiv
• 作者: Zhiyu Lin;Upol Ehsan;Rohan Agarwal;Samihan Dani;Vidushi Vashishth;Mark O. Riedl