CRII: CPS: Society-in-the-Loop Personalized Computing

该项目探索支持公平感知、隐私感知、社会在环、个性化物联网(IoT)系统的框架、工具和方法。由于移动计算和可穿戴技术的快速增长，监控复杂的人类环境变得可行，这为开发自然进化以适应人类和环境状态的人在环物联网系统铺平了道路。然而，随着我们朝着建立有效的、普遍的、既自主又个性化的计算系统的长期愿望前进，我们可以推动这些系统的极限，对社会产生积极的附带影响。我们可以设计出能够感知并相互适应其周围环境的个性化系统，以实现用户的集体利益。这需要新的计算范式和算法来实现我们对社会在环中的个性化计算的愿景。设计这种个性化的物联网应用面临着许多挑战，这些挑战源于“人类的变异性”。这种可变性源于不同的人，在与物联网应用交互时(即人与人之间的可变性)、在与同一物联网应用交互时(即人内可变性)以及在与同一环境中的一组人交互时(即多人可变性)表现出不同的行为。为了利用这些变异性，该项目开发了：(I)当多个人类共存于同一环境时的“适应公平”框架；(Ii)一套“适应代理人”利用人类变异性的算法；(Iii)“适应代理人”之间关于协调战略的协作协议；以及(Iv)保护适应代理人持有的私人信息的方法。经验验证是在两个截然不同的交互领域进行的，UCI试验台用于汽车高级驾驶辅助系统和智能家居自动化。该项目邀请研究生、本科生和高中生参与研究、测试和评估。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。

This project explores frameworks, tools, and methodologies that enable fairness-aware, privacy-aware, society-in-the-loop, personalized Internet-of-Things (IoT) systems. Thanks to the rapid growth in mobile computing and wearable technologies, monitoring complex human context becomes feasible, which paves the way to develop human-in-the-loop IoT systems that naturally evolve to adapt to the human and environment state autonomously. Nevertheless, as we move forward towards a long-standing desire to build effective, pervasive computing systems that are both autonomous and personalized, we can push the envelope of these systems to have a positive collateral effect on society. We can design the personalized systems that are aware of and can mutually adapt to the context of their surroundings for the collective benefit of the users. This requires new computation paradigms and algorithms to achieve our vision for society-in-the-loop personalized computing.Designing such personalized IoT applications faces many challenges that arise from "human variability." Such variability stems from different humans, exhibiting different behaviors when interacting with IoT applications (i.e., inter-human variability), when interacting with the same IoT application (i.e., intra-human variability), and when interacting with a group of humans in the same environment (i.e., multi-human variability). To harness these variabilities, this project develops: (i) a framework for "adaptation fairness" when multiple humans co-located in the same environment; (ii) a suite of algorithms for "adaptation agents" to harness human variability, (iii) protocols for collaboration among "adaptation agents" on coordination strategies, and (iv) methods to protect private information held by adaptation agents. Empirical validation is conducted in two very different domains of interaction, the UCI testbeds for automotive advanced driver assistance systems and for smart home automation. The project engages graduate, undergraduate, and high-school students in the research, test, and evaluation.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.