CAREER: A Virtual Laboratory for Studying Long-Term Human-Computer Relationships

随着计算机通过机器人、可穿戴设备、PDA和各种其他无处不在的界面以越来越复杂和人性化的方式与人互动，我们与它们之间关系的心理方面扮演着越来越重要的角色。重要的是不仅要了解这一现象的性质及其在工作和休闲环境中的影响，而且要制定建立和管理这些关系的战略，这些关系直接影响到生产力、享受性、参与度和人机交互的其他重要结果。在这项研究中，PI打算通过关系代理来探索这些问题，关系代理是一种计算人工制品，旨在通过管理预期、态度和意图与用户建立长期的社会情感关系。虽然这样的代理可以采取许多实施例(例如，作为珠宝、服装、手持设备、机器人和各种非类人的物理或非物理形式)，但当前的项目将专注于纯软件类人动画代理的开发。在他之前工作的基础上，PI将开发一个由联网的软件体系结构和实验方法组成的“虚拟实验室”，以支持非常长期的人机交互研究，其中新的实验和代理能力可以动态地集成到一个服务于一组持久的人类受试者的运行系统中。然后，虚拟实验室将被用作一种手段，研究社交界面代理如何与用户进行非常长期的互动-跨越数月或数年的日常使用-以及这些互动可能对用户教育、行为改变和整体福祉产生的影响。虽然社交界面代理具有广泛的适用性，但本项目将针对一个应用领域(身体活动促进)和一个用户群体(城市老年人)，以便集中研究和奠定基础。将特别注意向代理人传授与用户自然互动的能力，并随着时间的推移与他们建立社会-情感关系，以促进他们在保持长期互动和实现积极任务成果方面的有效性。除了促进与人机交互相关的知识外，这项研究还将在计算语言学、心理学和传播学领域做出重大贡献。在计算语言学领域，这项工作将促进我们对如何建立对话规划器的了解，该对话规划器可以产生面向任务的对话和社会对话-包括语言和非语言对话行为-旨在建立和维护与特定目标任务领域内的用户的社会关系；它还将解决如何对跨越多个交互的人机对话进行建模和推理。在心理学和传播学中，这项工作将提供更好的理解，了解人们与计算机代理可以拥有的社会关系的性质和类型，代理可以用来建立和维护这些关系的技术，以及人机关系如何提高某些类型的任务绩效。广泛的影响：这项工作的结果将包括建模人类关系和社会和关系行为的新方法。该项目还将有助于理解人们和计算机人员如何在较长的时间内以最佳方式共同生活和工作，这反过来可能通过提供新的词汇、模型和范式来思考关系，从而提高我们对人与人之间关系的理解。最后，这项工作将针对服务不足的人群(老年人)，同时侧重于一个可能对我们整个社会产生重大积极影响的应用领域(健康生活方式促进)。

As computers interact with people in increasingly complex and human ways through robots, wearable devices, PDAs, and various other ubiquitous interfaces, the psychological aspects of our relationships with them take on an increasingly important role. It is important to not only understand the nature of this phenomenon and its effects in work and leisure contexts, but also to develop strategies for building and managing these relationships, which directly impact productivity, enjoyment, engagement and other important outcomes of human-computer interaction. In this research, the PI intends to explore such issues via relational agents, computational artifacts designed to build long-term, social-emotional relationships with their users by managing expectations, attitudes and intentions. While such agents can take on a number of embodiments (e.g., as jewelry, clothing, handheld, robotic, and various non-humanoid physical or nonphysical forms), the current project will focus on the development of purely software humanoid animated agents. Building on his prior work, the PI will develop a "virtual laboratory" consisting of a networked software architecture and experimental methodology to support very long-term human-computer interaction studies, in which new experiments and agent capabilities can be dynamically integrated into a running system serving a persistent group of human subjects. The virtual laboratory will then be used as a means to studying how social interface agents can conduct very long-term interactions with users - spanning months or years of daily use - and the impacts these interactions can have on user education, behavior change and overall well-being. While social interface agents have broad applicability, this project will target one application domain (physical activity promotion) and one user group (urban older adults) in order to focus and ground the research. Special care will be paid to imparting to the agents the ability to interact naturally with users and to form social-emotional relationships with them over time, in order to promote their effectiveness in both maintaining long-term interaction and achieving positive task outcomes. In addition to advancing knowledge relating to human-computer interaction, this research will make significant contributions to the fields of computational linguistics, psychology, and communication. In the field of computational linguistics, the work will advance our knowledge of how to build dialog planners that can produce both task-oriented and social dialogue - including both verbal and nonverbal conversational behavior - intended to build and maintain social relationships with users within specifically targeted task domains; it will also address how to model and reason about human-machine dialog that spans multiple interactions. In psychology and communication, this work will provide a better understanding of the nature and types of social relationships that people can have with computer agents, techniques that agents can use to build and maintain these relationships, and how human-computer relationships can improve certain kinds of task performance.Broader Impacts: The outcomes of this effort will include new ways of modeling human relationships and social and relational behavior. The project will also lead to an understanding of how people and computer agents can optimally live and work together over extended periods of time, which in turn may improve our understanding of human-human relationships by providing new vocabulary, models and paradigms for thinking about relationships. Finally, this work will target an underserved population (older adults) while focusing on an application domain (healthy lifestyle promotion) that has the potential for significant positive impacts on our overall society.