Balancing human and system visualization during document triage

Abstract

People must frequently sort through and identify relevant materials from a large set of documents. Document triage is a specific form of information collecting where people quickly evaluate a large set of documents from the Internet by reading (or skimming) documents and organizing them into a personal information collection. During triage people can re-read documents, progressively refine their organization, and share results with others. People usually perform triage using multiple applications in concert: a search engine interface presents lists of potentially relevant documents; a document reader displays their content; and a text editor or a more specialized application records notes and assessments. However, people often become disoriented while switching between these subtasks in document triage. This can hinder the interaction between the subtasks and can distract people from focusing on documents of interest. To support document triage, we have developed an environment that infers users’ interests based on their interactions with multiple applications and on an analysis of the characteristics and content of the documents they are interacting with. The inferred user interest is used to relieve disorientation by generating visualizations in multiple applications that help people find documents of interest as well as interesting sections within documents.