SoCS: Achieving the Interpersonal Function of Affect in Human-Machine Collaboration

愤怒或喜悦的表现可以深刻地塑造我们周围人的行为。这个项目从功能主义的角度研究了这些对人类决策的影响。与其将情感展示仅仅视为内部经验的表现，它们可以被视为促进社会目标的战略性他者定向信号，例如通过传达特定的内在信念、愿望和未来意图。该项目探索了“逆向评估”理论，作为解释和计算模拟这一社会过程的框架。具体地说，这一理论将被用来揭示情感展示在社会困境和谈判中影响结果的机制。这一功能分析也为人机和计算机协调协作系统的设计提供了依据。该项目使用了一个基于“虚拟同盟”--即拟人化的计算机角色--的框架来研究社会影响。它建立在逆向评估的洞察力之上，即重要的是信息(功能)，而不是本身(形式)的表现，并研究情感的社会功能如何应用于超越虚拟人类的系统。该项目促进了社会科学和计算科学之间的协同作用：它有助于理解情感表达如何影响人们在社会、经济和日常生活中的决策的心理学；它还促进了人机系统在谈判等实用领域的协作状态，并对设计协作计算机中介系统具有影响。这项研究产生了新的工具--虚拟联盟--或研究社会过程的工具，这些工具将免费提供给科学界。最后，该项目支持研究生和博士后学者的研究和指导活动，以实现情感和决策的多学科专业化。

Displays of anger or joy can profoundly shape the behavior of those around us. This project brings a functionalist perspective to the study of such influences on human decision-making. Rather than treating affective displays as mere manifestation of internal experience, they can be viewed as strategic other-directed signals that facilitate social goals, for example by conveying specific internal beliefs, desires and future intentions. The project explores the theory of "reverse appraisal" as a framework for explaining and computationally modeling this social process. Specifically, this theory will be used to unpack the mechanisms by which affective displays influence outcomes in social dilemmas and negotiations. This functional analysis also informs the design of human-computer and computer-mediated collaboration systems. The project uses a framework based on "virtual confederates" - i.e., anthropomorphic computer characters - to study social affect. It builds on reverse appraisal's insight that what matters is the information (function) not the display in itself (form) and studies how the social functions of affect apply to systems that extend beyond virtual humans.Broader Impact. The project promotes synergies between social and computational sciences: it contributes to understanding the psychology of how affective expressions impact people's decisions in social, economic and daily life; and, it advances the state of human-computer systems for collaboration in practical domains such as negotiation, and has implications for designing collaborative computer-mediated systems. The research results in new tools -- virtual confederates -- or studying social processes, which will be offered freely to the scientific community. Finally, the project supports research and mentoring activities for graduate students and postdoctoral scholars towards a multidisciplinary specialization in affect and decision making.