Education DCL: EAGER: Generative AI-based Personalized Cybersecurity Tutor for Fourth Industrial Revolution

第四次工业革命(4IR)系统的日益复杂使其极易受到网络攻击，这证明了对4IR员工进行网络安全培训是合理的。虽然在线课程具有高度的可扩展性，并且是重新培训和提升现有劳动力技能的理想选择，但目前它们不适合4IR网络安全培训，因为4IR网络安全培训需要访问专业硬件。此外，来自不同背景的学生有不同的学习风格，这取决于他们的成长背景、教育背景和动机。成功完成课程对来自边缘化、代表性不足的社区的学生来说尤其具有挑战性，因为在他们的成长期(学前/初中/高中)，历史上缺乏获得高质量教育的机会可能会阻碍他们的成功。为了应对这些挑战，该项目利用了基于生成性人工智能的个性化网络安全导师(GAI-PCT)，这是一个可扩展的在线学习框架，通过视频讲座和基于虚拟现实(VR)的网络安全实验室培训4IR安全工作人员，通过生成性AI根据学生的学习需求进行个性化，使用情感分析来评估学生的兴趣和知识理解。这项研究是由亚利桑那大学和桑迪亚国立大学的网络安全和教育研究人员合作完成的。研究团队将探索：1)基于生成式人工智能的交互式教师，帮助和指导学生完成他们的在线课程；2)基于虚拟现实的网络安全培训实验室，允许游戏化的4IR网络安全培训，包括在不冒昂贵设备或生命危险的情况下进行极端场景的培训；3)基于转换器的机器学习方法，在学习活动中测量学生的情感和知识理解；4)基于生成性人工智能的方法，结合布鲁姆的分类，根据学生的学习需求个性化课程工作；5)一种衡量学生学习干预措施有效性的方法，同时服务于边缘学生；6)一种增加学生归属感、学术成就和工程认同感的方法，这将导致通过学位提高学术坚持性。该奖项反映了NSF的法定使命，并通过使用基金会的智力价值和更广泛的影响审查标准进行评估，被认为值得支持。

The increasing complexity of Fourth Industrial Revolution (4IR) systems makes them highly vulnerable to cyberattacks, justifying cybersecurity training for the 4IR workforce. Although online programs are highly scalable and ideal for reskilling and upskilling an existing workforce, currently, they are unsuitable for 4IR cybersecurity training which requires access to specialized hardware. In addition, students from different backgrounds have different learning styles depending on their upbringing, educational backgrounds, and motivations. Successful program completion is especially challenging for students from marginalized, underrepresented communities, as a historical lack of access to high-quality education during their formative years (pre/middle/high schools) can impede their success. To address these challenges, this project leverages a Generative AI based Personalized Cybersecurity Tutor (gAI-PCT), a scalable online learning framework that trains a 4IR security workforce through video lectures and virtual reality (VR) based cybersecurity labs, personalized to the students learning needs through generative AI, using sentiment analysis to evaluate student interest and knowledge comprehension. This research is a collaboration between cybersecurity and education researchers at the University of Arizona and Sandia National Labs.The research team will explore: 1) a Generative AI-based interactive instructor to assist and guide students through their online coursework; 2) Virtual Reality based cybersecurity training labs that allow gamified 4IR cybersecurity training, including training on extreme scenarios without risking expensive equipment or human life; 3) Transformer-based machine learning approach to gauge student sentiments and knowledge comprehension during learning activities; 4) Generative AI-based approach to use student sentiment, and comprehension, in combination with Bloom's Taxonomy, to personalize coursework according to student's learning needs; 5) An approach to gauge non-normative outcomes measuring the effectiveness of students learning interventions while serving marginalized students; 6) An approach to increasing student's sense of belonging, academic attainment, and engineering identity that will result in increased academic persistence through the degrees.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.