Towards a Computational History of the ACL: 1980-2008

Proceedings of the ACL-2012 Special Workshop on Rediscovering 50 Years of Discoveries, pages 13–21, Jeju, Republic of Korea, 10 July 2012. c⃝2012 Association for Computational Linguistics Towards a Computational History of the ACL: 1980–2008 Ashton Anderson Stanford University ashtona@stanford.edu Dan McFarland Stanford University dmcfarla@stanford.edu Dan Jurafsky Stanford University jurafsky@stanford.edu Abstract We develop a people-centered computational history of science that tracks authors over top- ics and apply it to the history of computa- tional linguistics. We present four findings in this paper. First, we identify the topical subfields authors work on by assigning auto- matically generated topics to each paper in the ACL Anthology from 1980 to 2008. Next, we identify four distinct research epochs where the pattern of topical overlaps are stable and different from other eras: an early NLP pe- riod from 1980 to 1988, the period of US government-sponsored MUC and ATIS eval- uations from 1989 to 1994, a transitory period until 2001, and a modern integration period from 2002 onwards. Third, we analyze the flow of authors across topics to discern how some subfields flow into the next, forming dif- ferent stages of ACL research. We find that the government-sponsored bakeoffs brought new researchers to the field, and bridged early top- ics to modern probabilistic approaches. Last, we identify steep increases in author retention during the bakeoff era and the modern era, suggesting two points at which the field be- came more integrated. 1 Introduction The rise of vast on-line collections of scholarly pa- pers has made it possible to develop a computational history of science. Methods from natural language processing and other areas of computer science can be naturally applied to study the ways a field and its ideas develop and expand (Au Yeung and Jatowt, 2011; Gerrish and Blei, 2010; Tu et al., 2010; Aris et al., 2009). One particular direction in computational history has been the use of topic models (Blei et al., 2003) to analyze the rise and fall of research top- ics to study the progress of science, both in general (Griffiths and Steyvers, 2004) and more specifically in the ACL Anthology (Hall et al., 2008). We extend this work with a more people-centered view of computational history. In this framework, we examine the trajectories of individual authors across research topics in the field of computational linguistics. By examining a single author’s paper topics over time, we can trace the evolution of her academic efforts; by superimposing these individual traces over each other, we can learn how the entire field progressed over time. One goal is to investi- gate the use of these techniques for computational history in general. A second goal is to use the ACL Anthology Network Corpus (Radev et al., 2009) and the incorporated ACL Anthology Reference Corpus (Bird et al., 2008) to answer specific questions about the history of computational linguistics. What is the path that the ACL has taken throughout its 50-year history? What roles did various research topics play in the ACL’s development? What have been …

Content truncated to 500 words