EAGER: SaTC-EDU: Training Mid-Career Security Professionals in Machine Learning and Data-Driven Cybersecurity

EAGER：SaTC-EDU：在机器学习和数据驱动的网络安全方面培训职业中期安全专业人员

• 批准号: 2041970
• 负责人: Nicholas Feamster
• 金额: $ 29.99万
• 依托单位: University of Chicago
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2020
• 资助国家: 美国
• 起止时间: 2020-09-01 至 2023-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2041970&HistoricalAwards=false
• 关键词: EAGER; SaTC; EDU; Training; Mid

The cybersecurity and machine learning (ML) fields have evolved relatively independently. The occasional overlap between the two fields generally takes the form of either (1) applications of ML to statistical anomaly detection (e.g., malware detection); or (2) adversarial attacks on ML detection algorithms (e.g., adversarial ML). The cybersecurity and ML fields are also rapidly advancing, which makes education both in these respective fields and at their intersection critical. Advancement and re-skilling the United States cybersecurity workforce through large-scale, online training in data-driven and ML methods is critical for keeping the country secure and the workforce competitive. The project team will address this critical need by developing curricula for large-scale, online training of mid-career security professionals who aim to develop the skills to apply both conventional and cutting-edge ML tools to cybersecurity. This project will develop curricula at the intersection of ML and cybersecurity with a focus on applications of ML to practical, real-world security use cases. In addition, the project will establish a pedagogical foundation for security researchers to evaluate and apply various potential ML-based approaches to cybersecurity. The project is focused, in particular, on training mid-career professionals who have a classical training in cybersecurity (and thus an understanding of practical concepts), but need to gain a stronger foundation in data-driven methods that have become the basis for most applied cybersecurity in the past decade. The project outcomes will include: (1) online curricular development in data-driven security, to provide mid-career professionals foundations and practical tools for applying these methods to practical problems in network security; (2) formative research to elicit desired skills and use cases from the workforce; (3) modular public toolkits and datasets for use in both courses and as resources for professionals to apply in practical settings; and (4) augmented teaching materials, tailored to individual students, based on intelligent tutoring systems.This project is supported by a special initiative of the Secure and Trustworthy Cyberspace (SaTC) program to foster new, previously unexplored, collaborations between the fields of cybersecurity, artificial intelligence, and education. The SaTC program aligns with the Federal Cybersecurity Research and Development Strategic Plan and the National Privacy Research Strategy to protect and preserve the growing social and economic benefits of cyber systems while ensuring security and privacy.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.

网络安全和机器学习（ML）领域相对独立地发展。这两个领域之间的偶尔重叠通常采取以下形式：(1)ML应用于统计异常检测（例如，恶意软件检测）；(2)对机器学习检测算法的对抗性攻击（例如，对抗性机器学习）。网络安全和机器学习领域也在迅速发展，这使得这些领域及其交叉领域的教育至关重要。通过数据驱动和机器学习方法的大规模在线培训，提升和重新培训美国网络安全劳动力，对于保持国家安全和劳动力竞争力至关重要。项目团队将通过为职业中期安全专业人员开发大规模在线培训课程来解决这一关键需求，这些专业人员旨在培养将传统和尖端机器学习工具应用于网络安全的技能。该项目将开发机器学习和网络安全交叉的课程，重点是将机器学习应用于实际的、现实世界的安全用例。此外，该项目将为安全研究人员建立一个教学基础，以评估和应用各种潜在的基于ml的网络安全方法。该项目特别侧重于培训具有网络安全经典培训（从而理解实际概念）的职业中期专业人员，但需要在数据驱动方法方面获得更坚实的基础，这些方法在过去十年中已成为大多数应用网络安全的基础。项目成果将包括：(1)数据驱动安全的在线课程开发，为职业生涯中期的专业人员提供基础和实用工具，将这些方法应用于网络安全的实际问题；(2)形成性研究，从劳动力中引出所需的技能和用例；(3)模块化公共工具包和数据集，供课程使用，并作为专业人员在实际环境中应用的资源；(4)基于智能辅导系统的个性化增强型教材。该项目由安全与可信网络空间（SaTC）计划的一项特别倡议支持，旨在促进网络安全、人工智能和教育领域之间前所未有的合作。SaTC项目与《联邦网络安全研究与发展战略计划》和《国家隐私研究战略》保持一致，旨在保护和维护网络系统日益增长的社会和经济效益，同时确保安全和隐私。该奖项反映了美国国家科学基金会的法定使命，并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

Contextual Active Online Model Selection with Expert Advice

根据专家建议进行上下文主动在线模型选择

• 发表时间: 2022
• 期刊: ICML2022 Workshop on Adaptive Experimental Design and Active Learning in the Real World
• 作者: Liu, Xuefeng;Xia, Fangfang;Stevens, Rick L.;Chen, Yuxin
• 通讯作者: Chen, Yuxin

Iterative Machine Teaching for Black-box Markov Learners

黑盒马尔可夫学习者的迭代机器教学

• 发表时间: 2023
• 期刊: Workshop on Theory of Mind in Communicating Agents
• 作者: Wang, Chaoqi;Zilles, Sandra;Singla, Adish;Chen, Yuxin
• 通讯作者: Chen, Yuxin