This is funding for a one-day workshop on scientific research applications of natural language technologies to be held at the annual meeting of the Association for Computational Linguistics (ACL), in Baltimore on June 26, 2014. A major growth area in applied computer science has been the application of automated techniques to massive datasets to answer scientific questions about people and society. Although much work in this area focuses on structured data or network data, linguistic data is also a key source of evidence for these phenomena. While some existing natural language processing (NLP) techniques have found use in this growing community, new techniques for discovering and analyzing social meanings and structures in text are in high demand.
Intellectual merit. Engagement between NLP researchers and domain scientists will introduce new problem formulations and new theoretical frameworks that will broaden and deepen applications of language technology to social science. Potential topics for presentations and discussion include (but are by no means limited to): inferring social relations from conversation and other linguistic behavior; automatic extraction of event data from text; inference of author and speaker properties from text and speech; relating text datasets to author social networks; tracking language change over space, time, and communities; measuring linguistic influence; computational analysis of literary and historical corpora; and tracking the flow of information, ideas, and sentiment through social networks. Much of the data to be studied comes from online communities, which are a focus of the Cyber-Human Systems program.
Broader Impact. This workshop will increase the visibility of the computational social science application area for ACL researchers and will help build connections between language technologists and social scientists. The workshop¹s format aims at fostering interactions among participants and invited speakers, contributing towards building a community interested in language technologies and domain scientists. This format is especially beneficial to student participants, who will leave with new ideas about guiding applications. As is typical for ACL workshops, an archival proceedings will be published openly through the ACL Anthology. The workshop organizers will report on the workshop, synthesizing the discussion at the workshop with an emphasis on research topics with greatest potential in the near future. The report will be published openly (e.g., posted to arXiv and possibly submitted for publication in a relevant journal) within a few months after the workshop.
