Social Information Retrieval

该计划将战略性地解决错误信息日益增加的挑战，同时为社会信息检索提供基础贡献。从实践的角度来看，它努力(1)对错误信息进行语义建模，并识别网络平台上真实事件与错误信息之间的因果关系；(2)有效检索与错误信息相反的事实信息，并有选择地呈现以优化吸收。从理论角度来看，它将开发以下技术：(1)在知识图的背景下为主题集中的时间文本语料库建模，以测量语义关联和因果关系；(2)通过考虑覆盖、信任、说服和不和谐等方面来有效检索内容。

This proposed program will strategically address the increasing challenge of misinformation while providing foundational contributions in social information retrieval. From a practical perspective, it endeavors to (1) semantically model misinformation and identify causal relations between real-world events and misinformation in online platforms, and (2) effectively retrieve factual information counter to misinformation and selectively present it to optimize uptake. From a theoretical view, it will develop techniques to (1) model topically-focused temporal textual corpora in the context of knowledge graphs for measuring semantic association and causal relations, and (2) effectively retrieve content by considering aspects of coverage, trust, persuasion, and discord.