CyberTraining: Implementation: Small: Promoting AI Readiness for Machine-Assisted Secure Data Analysis (PAIR4MASDA)

网络培训：实施：小型：促进人工智能为机器辅助安全数据分析做好准备 (PAIR4MASDA)

• 批准号: 2320951
• 负责人: Alvis Fong
• 金额: $ 49.99万
• 依托单位: Western Michigan University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2023
• 资助国家: 美国
• 起止时间: 2023-09-01 至 2027-08-31
• 项目状态: 未结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2320951&HistoricalAwards=false
• 关键词: CyberTraining; Implementation; Small; Promoting; AI

Democratization of artificial intelligence (AI) technologies, such as large language models and chatbots, has accelerated the need for an AI ready workforce. To meet this challenge, the project aims to instill AI readiness in a broad spectrum of users and researchers of advanced cyberinfrastructure (CI), so they can productively and harmoniously use AI for secure big data analysis. The project directly involves eighteen science and engineering faculty members, ten industry experts, and over a thousand students across partner universities and colleges, including several minority serving institutions. By promoting AI readiness across diverse populations with accessible and relatable learning materials and holistic professional development, the project aims to level the playing field for those who feel left behind via bottom-up empowerment. Collective impact guides the team's concerted effort to broaden participation from underrepresented groups such as women, minorities, and veterans. All training materials have reproducibility built in by design to facilitate adoption by a broad range of learners, including those not currently attending college or otherwise having limited access to computational or traditional educational resources.Advanced cyberinfrastructure (CI) enabled artificial intelligence (AI) with big data analysis is an important enabler for science and engineering (S&E) research. However, as recent mishaps involving chatbots show, AI can produce convincing but incorrect or even harmful results. AI cannot yet be trusted with full autonomy, especially in mission critical applications. There is an emphasis on machine assisted secure data analysis, with human in the loop to enhance safety, security, and reliability. Responding to repeated calls for a multi-faceted approach to AI training, the project studies AI readiness along three dimensions: technical, psychological, and behavioral. Expectation confirmation theory is extended with decision quality and coupled with pervasive technology strategies. This development both advances knowledge and provides a theoretical basis for precise and measurable outcomes. A comprehensive suite of experiential learning modules forms the core of a set of customizable training materials. Critical knowledge and skills are distilled into bitesize units known as flexible micro modules (FMMs) to rapidly upskill learners, who in turn create their own personalized toolkits. An additional layer is built on top to give learners multimodal immersive experiences using extended reality (XR) technologies. These immersive experiences help learners better comprehend difficult concepts and internalize complex sequences of tasks. This award by the NSF Office of Advanced Cyberinfrastructure is jointly supported by the Division of Graduate Education (DGE) within the NSF Directorate for STEM Education (EDU) and Information and Intelligent Systems (IIS) division within the Computer and Information Science and Engineering (CISE) directorate.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.

人工智能(AI)技术的民主化，如大型语言模型和聊天机器人，加速了对准备好AI的劳动力的需求。为了应对这一挑战，该项目旨在向先进网络基础设施(CI)的广泛用户和研究人员灌输人工智能准备，以便他们能够高效、和谐地使用人工智能进行安全的大数据分析。该项目直接涉及18名科学和工程教职员工，10名行业专家，以及合作大学和学院的1000多名学生，其中包括几个少数族裔服务机构。通过利用可获得和相关的学习材料和全面的专业发展，促进不同人群的人工智能准备，该项目旨在通过自下而上的赋权，为那些感到被落在后面的人创造公平的竞争环境。集体影响指导团队共同努力，扩大妇女、少数群体和退伍军人等代表性不足群体的参与。所有培训材料在设计上都具有可重复性，以便于广泛的学习者采用，包括那些目前没有上大学或以其他方式有限地获得计算或传统教育资源的学习者。先进的网络基础设施(CI)支持的人工智能(AI)和大数据分析是科学和工程(S和安培；E)研究的重要推动因素。然而，正如最近涉及聊天机器人的不幸事件所表明的那样，人工智能可以产生令人信服但不正确甚至有害的结果。人工智能还不能完全自主，特别是在任务关键型应用程序中。重点是机器辅助的安全数据分析，人在循环中，以提高安全性、安全性和可靠性。为了响应人们对人工智能培训多方面方法的反复呼吁，该项目从三个维度研究了人工智能的准备情况：技术、心理和行为。预期确认理论被扩展到决策质量，并与普适技术策略相结合。这一发展既促进了知识的进步，又为精确和可衡量的结果提供了理论基础。一套全面的体验式学习模块构成了一套可定制培训材料的核心。关键知识和技能被提炼成称为柔性微模块(FMM)的单位，以快速提升技能学习者，他们反过来又创建自己的个性化工具包。在顶部构建了一个附加层，使用扩展现实(XR)技术为学习者提供多模式身临其境的体验。这些身临其境的体验帮助学习者更好地理解困难的概念并将复杂的任务序列内在化。这项由NSF高级网络基础设施办公室颁发的奖项由NSF STEM教育局(EDU)内的研究生教育司(DGE)和计算机与信息科学与工程(CEISE)局内的信息与智能系统(IIS)司共同支持。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。