Beyond Translation: Design and Evaluation of an Emotional and Contextual Knowledge Interface for Foreign Language Social Media Posts


Although many social media sites now provide machine translation (MT) for foreign language posts, translation of a post may not suffice to support understanding of, and engagement with, that post. We present SenseTrans, a tool that provides emotional and contextual annotations generated by natural language analysis in addition to machine translation. We evaluated SenseTrans in a laboratory experiment in which native English speakers browsed five Facebook profiles in foreign languages. One group used the SenseTrans interface while the other group used MT alone. Participants using SenseTrans reported significantly greater understanding of the posts, and greater willingness to engage with the posts. However, no additional cognitive load was associated with using an interface that provided more information. These results provide promising support for the idea of using computational tools to annotate communication to support multilingual sense making and interaction on social media.

Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems