CAREER: Algorithms, Incentives, and Policy for Data Privacy

隐私问题正在成为使用数据的主要障碍，目前的法规应该如何转化为技术往往并不清楚。随着差异隐私和其他隐私工具在实践中的应用，在确保这些应用程序保持理论上预期的隐私保障方面出现了新的挑战。那么，组织应该如何利用隐私技术的进步来利用潜在的敏感数据？这个项目通过解决理论隐私技术在现实世界中实施时出现的新的技术挑战，在正式隐私研究的理论和实践之间架起了一座桥梁。这个项目解决了三个主要的技术问题：首先，数据管理员应该如何在多个分析任务之间最优地分配固定的差异隐私预算，以权衡准确分析的价值和隐私预算？其次，在实践中，人们如何对自己的隐私进行推理和评估，这些评估是如何根据数据或分析任务的背景而变化的？第三，应该如何监管个人数据市场，以保护个人隐私和人口层面的公平，同时仍然能够进行有价值的数据驱动的决策？回答这些问题需要一种跨学科的方法，将计算机科学的工具整合到不同的私人算法设计中，从经济学角度理解组织和个人的激励，以及监管隐私技术的公共政策。为了确保这项研究的广泛影响，该项目还包括一个重要的教育和推广部分，包括课程开发、学生指导和旨在改善研究生教育多样性的研讨会组织。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。

Privacy concerns are becoming a major obstacle to using data, and it is often unclear how current regulations should translate into technology. As differential privacy and other privacy tools are brought to bear in practice, new challenges arise in ensuring these applications maintain the privacy guarantees intended by theory. How, then, should organizations utilize advances in privacy technologies to make use of potentially sensitive data? This project bridges the gap between theory and practice in the formal study of privacy by addressing new technical challenges that arise when theoretical privacy technologies are implemented in real-world settings.This project addresses three main technical questions: Firstly, how should a data curator optimally allocate a fixed differential privacy budget across multiple analysis tasks, to trade off the value of accurate analysis with the privacy budget? Secondly, how do people reason about and value their privacy in practice, and how do these valuations change based on context of the data or analysis task? Thirdly, how should markets for personal data be regulated to protect individual privacy and population-level fairness, while still enabling valuable data-driven decision making? Answering these questions requires an interdisciplinary approach that integrates tools from computer science for differentially private algorithm design, economics to understand incentives of organizations and individuals, and public policy for regulation of privacy technologies. To ensure the broad impact of this research, this project also includes a significant educational and outreach component, including curriculum development, mentorship of students, and workshop organization aimed at improving diversity in graduate education.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.