Analyzing the Dynamics of Research by Extracting Key Aspects of Scientific Papers

Proceedings of the 5th International Joint Conference on Natural Language Processing, pages 1–9, Chiang Mai, Thailand, November 8 – 13, 2011. c⃝2011 AFNLP Analyzing the Dynamics of Research by Extracting Key Aspects of Scientific Papers Sonal Gupta Department of Computer Science Stanford University sonal@cs.stanford.edu Christopher D. Manning Department of Computer Science Stanford University manning@cs.stanford.edu Abstract We present a method for characterizing a research work in terms of its focus, do- main of application, and techniques used. We show how tracing these aspects over time provides a novel measure of the in- fluence of research communities on each other. We extract these characteristics by matching semantic extraction patterns, learned using bootstrapping, to the depen- dency trees of sentences in an article’s abstract. We combine this information with pre-calculated article-to-community assignments to study the influence of a community on others in terms of tech- niques borrowed and the ‘maturing’ of some communities to solve other prob- lems. As a case study, we show how the computational linguistics community and its sub-fields have changed over the years with respect to their foci, methods used, and domain problems. For instance, we show that part-of-speech tagging and parsing have increasingly been adopted as tools for solving problems in other do- mains. We also observe that speech recog- nition and probability theory have had the most seminal influence. 1 Introduction The evolution of ideas and the dynamics of a re- search community can be studied using the sci- entific articles published by the community. For instance, we may be interested in how methods spread from one community to another, or the evo- lution of a topic from a focus of research to a problem-solving tool. We might want to find the balance between technique-driven and domain- driven research within a field. Establishing such a rich insight of the development and progress of scientific research requires an understanding of more than just the “topics” of discussion or cita- tion links between articles, which have been used in the previous work to study trend and impact of articles. As an example, to determine whether technique-driven researchers have greater or lesser impact, we need to be able to identify styles of work. To achieve this level of detail and to be able to connect together how methods and ideas are be- ing pursued, it is essential to move beyond bag- of-words topical models. This requires an under- standing of sentence and argument structure, and is therefore a form of information extraction. To study the application domains, the tech- niques used to approach the domain problems, and the focus of scientific articles in a community, we propose to extract the following concepts from the articles FOCUS: an article’s main contribution TECHNIQUE: a method or a tool used in an article, for example, expectation maximiza- tion and conditional random fields DOMAIN: an article’s application domain, such as speech recognition and classification of documents. For example, if an article concentrates on reg- ularization in support vector machines and shows improvement in parsing …

Content truncated to 500 words