Collaborative Research: Closing the Affective Gap

我们的情绪和情绪在我们的日常体验中起着核心作用。为了设计更充分地利用这一全方位人类体验的界面，一些研究人员现在倡导情感计算，即计算机自动感知、处理和回应人类情感，但在用户没有明确意识和同意的情况下，让计算机测量和报告隐私数据引发了对谁在控制的重大担忧。此外，准确地感知受许多生理、个体和文化因素影响的复杂和难以捉摸的情绪是具有挑战性的，这些因素是计算传感器无法获得的，需要简化情绪以对其进行计算建模，这可能会导致接口使我们的情绪体验扁平化，而不是让我们的情绪充分参与到日常生活中。这类问题是基于一种基本的情感鸿沟，即人们思考和体验情感的主观、情境、文化和社会方式与计算机模拟情感的客观、形式化方式之间的脱节。在这个项目中，PI将试图通过分担人和机器之间的情感解释负担来缩小情感鸿沟。在这个模型中，人和机器协作构建情感体验；机器跟踪用户行为并提供反馈，用户解释这些反馈以培养情感意识。将解释和反思的中心从系统转移到用户身上，让用户负责情感意义的产生，解决仅靠计算系统无法真正理解的情感，并将我们的设计努力集中在丰富、复杂、特殊和潜在的神秘情感体验上，而不是情感的形式化。需要用户主动解释的系统对人机界面的设计和评估提出了严重的挑战，因为我们不再直接控制系统的意义，而是为了意外的挪用而进行设计和评估。从艺术、人文和批判性设计中汲取，个人投资指数的系统将通过反映情感的复杂、模棱两可的性质来鼓励解释，使人们能够在他们的日常生活和关系的背景下积极地理解系统的输出。广泛的影响：个人投资指数将制定系统的设计和评估策略、理论基础和详细的案例研究，使情感存在对人机交互的贡献是可持续的、可重复的。这项工作将带来新的技术，让用户负责解释他们的情感，同时允许构建涉及隐私问题而不具侵入性的系统，支持用户对技术中情感的作用进行反思，并使设计师能够创造一种新的、广泛的情感体验，尽管计算机很难对这些体验进行建模。

Our emotions and moods play a central role in our everyday experiences. To design interfaces that more adequately engage this full range of human experience, some researchers now advocate affective computing, in which computers automatically sense, process, and respond to human emotions, but having computers measure and report on intimate data without users' explicit awareness and consent raises substantial concerns around who is in control. Furthermore, it is challenging to accurately sense complex and elusive emotions influenced by many physiological, individual, and cultural factors that are unavailable to computational sensors, and the need to simplify emotions in order to computationally model them may lead to interfaces that flatten our emotional experiences rather than engaging our emotions in their full everyday richness. Such problems are based on a fundamental affective gap, a disjunction between the subjective, situated, cultural, and social ways in which people think about and experience emotions and the objective, formalized ways in which computers model them. In this project the PIs will attempt to close the affective gap by sharing the burden of affective interpretation between people and machines. In this model, people and machines collaborate to construct affective experiences; machines track user behavior and supply feedback, which users interpret to develop affective awareness. Shifting the center of interpretation and reflection from the system towards the user keeps users in charge of emotional meaning-making, addresses emotions that computational systems alone cannot truly understand, and focuses our design efforts not on formalizations of affect but on rich, complex, idiosyncratic, and potentially enigmatic emotional experiences. Systems requiring active user interpretation pose serious challenges for design and evaluation in HCI, since we no longer have direct control of system meaning but instead design and evaluate for unexpected appropriation. Drawing from the arts, humanities, and critical design, the PIs' systems will encourage interpretation by reflecting the complex, ambiguous nature of affect, allowing people to actively make sense of the system's output in the context of their everyday lives and relationships.Broader Impacts: The PIs will develop systematic design and evaluation strategies, a theoretical foundation, and detailed case studies that make affective presence a sustainable, reproducible contribution to HCI. The work will lead to new technology that leaves users in charge of the interpretation of their emotions yet allows construction of systems that touch on intimate issues without being intrusive, supports user reflection on the role of affect in technology, and enables designers to create a new, wide range of emotional experiences even though computers have difficulty modeling those experiences.