CAREER: Leveraging Recommendations for Self-Actualization

该项目的主要研究目标是利用推荐技术来支持用户根据他们的长期目标和抱负来开发，探索和理解他们的偏好。推荐系统根据用户表达的偏好为用户提供推荐，但决策科学家已经证明，用户的偏好通常是动态构建的，因此专注于他们的即时需求，而不是他们的长期目标。反过来，这些推荐往往反映了这些短期的喜好，忽略了用户的抱负和长期目标。该提案将通过开发和测试两种新颖的用户自适应交互机制来摆脱这种恶性循环，这两种机制利用推荐算法来支持用户开发，发现和理解自己的偏好的独特新方法。这将使这些系统的用户成为自信的消费者，他们能够构建专注于更广泛的长期目标的偏好。 决策支持平台和建议的交互机制将提供给学术界，以支持他们对推荐系统和决策的研究。两种创新的交互机制可以通过将个性化决策支持工具的范围扩展到“选择”之外，从而推进推荐系统研究领域和决策科学领域：（1）个性化偏好配置文件，使用推荐算法来可视化用户的偏好并突出令人惊讶的偏好动态。这些配置文件将通过帮助用户了解他们的偏好，反思他们的长期目标，并做出更明智的决定来诱导偏好构建。(2)基于偏好的社区，使用推荐算法自动创建特定的用户组，这些用户组的兴趣重叠足以产生有趣的讨论，但不会太多，以至于它们变成认识泡沫。这些社区试图通过讨论来引导偏好的构建，从而恢复积极提供咨询的做法的社会和知识效益。该项目将把这些互动机制整合到实时系统中，支持教育和推广活动，所有年龄段的学生都将使用这些系统根据他们的偏好和生活目标做出决定。该奖项反映了NSF的法定使命，并通过使用基金会的智力价值和更广泛的影响审查标准进行评估，被认为值得支持。

The primary research goal of this project is to leverage recommendation technology to support users in developing, exploring, and understanding their preferences based on their long-term goals and ambitions. Recommender systems provide their users with recommendations based on their expressed preferences, but decision scientists have demonstrated that users' preferences are often constructed on the fly and thereby focused on their immediate desires rather than their longer-term goals. The recommendations, in turn, tend to reflect these short-term likes, ignoring users' ambitions and long-term goals. This proposal will take an important step towards escaping this vicious circle by developing and testing two novel user-adaptive interaction mechanisms that leverage recommendation algorithms to support unique new ways for users to develop, discover, and understand their own preferences. This will turn the users of these systems into confident consumers who are able to construct preferences that focus on their broader long-term objectives. The decision-support platform and the proposed interaction mechanisms will be made available to academics to support their research on recommender systems and decision-making.Two innovative interaction mechanisms can advance the field of recommender systems research as well as the field of decision science by extending the reach of personalized decision-support tools beyond "choice" towards preference construction: (1) Personalized preference profiles that use recommendation algorithms to visualize users' preferences and highlight surprising preference dynamics. These profiles will induce preference construction by helping users understand their preferences, reflect upon their long-term goals, and make better-informed decisions. (2) Preference-based communities that use recommendation algorithms to automatically create ad-hoc groups of users whose interests overlap enough to generate interesting discussion, yet not so much that they turn into epistemic bubbles. These communities seek to induce preference construction through discussion, thereby reinstating the social and intellectual benefits of active advice-giving practices. This project will integrate these interaction mechanisms into live systems, supporting education and outreach activities where students of all ages will use these systems to make decisions based on their preferences and life goals.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.