Jason Ginsburg

Text analysis and entity extraction in asymmetric threat response and prediction

The Asymmetric Threat Response and Analysis Program (ATRAP) is a software system for intelligence fusion, visualization, reasoning, and prediction. ATRAP consists of a set of tools for annotating and automatically extracting entities and relationships from documents, visualizing this information in relational, geographic, and temporal dimensions, and determining future courses of action of adversaries by creating situational threat templates and applying customized prediction algorithms. In this paper, we first describe the task of analyzing data in intelligence reports, and then provide an overview of major system components: the Text Highlighter tool, the ThoughtSpace™ visualization environment, and the Template Builder and prediction tool. Subsequently, we describe linguistic characteristics of intelligence reports, and describe ATRAPs named entity recognition system.

Automatic generation of English lesson materials for native speakers of Japanese

This paper describes a computer program, a Lesson Material Builder, that is designed to automatically generate editable lesson materials for teaching English to speakers of Japanese. This program has two functions, a) an Exercise Creator, and b) a Vocabulary Quiz Creator. The Exercise Creator automatically creates an editable reading exercise document, from an input text, in which target vocabulary words are highlighted and definitions in Japanese are provided. Taking a reading exercise, of the sort produced by the Exercise Creator, the Vocabulary Quiz Creator can create vocabulary quizzes that require students to match vocabulary words with their definitions.

Locality in minimalist syntax

Chomsky, Noam. 2000. Minimalist inquiries: The framework. In Step by step: Essays on minimalist syntax in honor of Howard Lasnik, ed. Roger Martin, David Michaels, and Juan Uriagereka, 89–155. Cambridge, MA: MIT Press. Chomsky, Noam. 2008. On phases. In Foundational Issues in linguistic theory: Essays in honor of Jean-Roger Vergnaud, ed. Robert Freidin, Carlos P. Otero, and Maria Luisa Zubizarreta, 133–167. Cambridge, MA: MIT Press. Cinque, Guglielmo. 1999. Adverbs and functional heads: A cross-linguistic perspective. Oxford: Oxford University Press. Raposo, Eduardo and Juan Uriagereka. 1996. Indefinite se. Natural Language and Linguistic Theory 14:749–810. van Riemsdijk, Henk. 2006. Grafts follow from Merge. In Phases of interpretation, ed. Mara Frascarelli, 17–44. Berlin: Mouton de Gruyter. Rizzi, Luigi. 1997. The fine structure of the left periphery. In Elements of grammar, ed. Liliane Haegeman, 281–337. Dordrecht: Kluwer.

A Wikipedia-based corpus reference tool

This paper describes a dictionary-like reference tool that is designed to help users find information that is similar to what one would find in a dictionary when looking up a word, except that this information is extracted automatically from large corpora. For a particular vocabulary item, a user can view frequency information, part-of-speech distribution, word-forms, definitions, example paragraphs and collocations. All of this information is extracted automatically from corpora and most of this information is extracted from Wikipedia. Since Wikipedia is a massive corpus covering a diverse range of general topics, this information is probably very representative of how target words are used in general. This project has applications for English language teachers and learners, as well as for language researchers.

A computational model of language generation applied to English wh-questions

This paper presents a computational model of language generation, based on Phase Theory, that automatically constructs sentences from underlying numerations. This model incorporates explicit algorithms that determine selection and merger of Lexical Items from a subnumeration, determine the labels of Merged syntactic elements, account for movement of elements within a derivation, and account for when phrases are sent to Spell-Out. This paper shows how this model automatically produces the derivation of an English wh-question.