CCRI: Planning: RecommendNews: Community Research Infrastructure for Online Field Experiments

这项规划赠款将发展一个共同的愿景和建立创新研究基础设施所需的伙伴关系，其中不仅包括一个计算机系统，还包括一个用户社区，以开发适用于在线新闻的新形式的推荐系统。 这项规划赠款以及随后基于该计划的基础设施有三个关键的更广泛的影响。首先，新闻推荐具有明显的社会重要性。组织研究界支持多样性、透明度、可信度和有效的新闻推荐工作，不仅有利于研究，也有利于社会。其次，计划中的基础设施不仅有可能使目标核心研究团体受益，而且还可能使全国和世界各地的其他研究团体受益，包括新闻和大众传播领域的研究团体。第三，规划过程将吸引各种研究人员，包括许多来自传统上服务不足的地区的研究人员，以尽可能扩大这项工作的影响。这种基础设施的设计、建设和治理具有重大的研究挑战。 在这种情况下，系统是一个新闻推荐应用程序（加上支持它的后端），用户将是通过应用程序阅读新闻的人。该基础设施还包括实验控制，允许研究人员通过改变应用程序的推荐算法和界面以及测量结果来运行实验。 这样的基础设施可以通过启用新的应用程序、研究设计和成功指标来重振推荐系统的研究。该项目的核心是为社区举办一系列外联活动和规划会议。这些活动包括广泛的调查和为100名研究人员举办的社区论坛（与会议同时举行），以及为目标用户和执行伙伴举办的小型会议。这些会议将确定社区的需求和优先事项，设计基础设施本身及其运营所需的治理模式，并制定确保可持续性的模式。 这项计划将吸引研究界的不同子集进行深入和深思熟虑的讨论，讨论我们今天无法在学术环境中开展哪些研究，我们重视哪些研究，以及我们需要哪些基础设施来推进该领域的研究。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。

This planning grant will develop a shared vision and the partnerships required for an innovative research infrastructure that includes not only a computer system but also a community of users, for the development of new forms of recommender system applied to online news. There are three key broader impacts for this planning grant, and the subsequent infrastructure based on the plan. First, news recommendation has clear societal importance. Organizing the research community to support work on diversity, transparency, credibility, and effective news recommendation is a benefit not only for research, but for society. Second, the planned infrastructure has the potential to benefit not only the core research communities targeted, but also other research communities across the nation and around the world, including those in journalism and mass communication. Third, the planning process will engage a diverse set of researchers, including many from areas traditionally underserved, to extend the impact of this work as far as possible. The design, construction, and governance of this infrastructure has significant research challenges. In this case, the system is a news recommender application (plus the back-end to support it) and the users will be people who read news through the application. The infrastructure also includes experimental controls to allow researchers to run experiments by varying the recommendation algorithms and interfaces of the application and measuring results. Such infrastructure can reinvigorate research in recommender systems by enabling new applications, study designs, and success metrics. Central to this project is a series of outreach activities and planning meetings for the community. These range from a broad survey and a community forum for 100 researchers (held in conjunction with a conference) to smaller meetings of intended users and partners in implementation. These meetings will identify community needs and priorities, design both the infrastructure itself and the governance models needed to operate it, and develop models to ensure sustainability. This planning effort will engage a diverse subset of the research community in deep and thoughtful discussions about what research we are unable to carry out in academic settings today, what research we value, and what infrastructure we need to advance research in the field.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.