LingBrowser: An active, Interactive and Intelligent Linguistic Content Browser Employing Natural Language Processing Technology

点击翻译按钮获取中文摘要

Intelligent, interactive and internet-accessible tools can be helpful in learning (about) other languages. With support from both the Linguistics Program and the Office of International Science and Engineering, a research team led by Dr. Kemal Oflazer, a Turkish computational linguist, and Dr. Sharon Inkelas, an American linguist specializing in phonology and morphology, will develop a new, multidimensional tool, called LingBrowser, whose aim is to make available to linguists and language students instantaneous information about Turkish words and texts. LingBrowser can analyze anything from a single word to an entire text. Once given a text by the user, LingBrowser will provide, for any words that the user highlights, instantaneous information about pronunciation, morphological analysis (showing the parts of complex words and their meanings), semantics (the meanings of words and some phrases) and, to some degree, syntax (determining the part of speech of highlighted words and their grammatical role in the sentence containing them). LingBrowser can show the user other words in the same text which share the same root or even the same endings as the highlighted word, and offer assistance in cases of ambiguity. Many words in Turkish are ambiguous, at least when viewed out of the context of the sentence containing them. LingBrowser will list all of the possible meanings of a given word and rank, for the user, the meanings in order of plausibility in context. LingBrowser can also instantly list all of the sentences in the text containing a highlighted word or words, for users seeking more information about how exactly a particular word or phrase is used in Turkish. LingBrowser will be a highly useful research tool both to professional linguists seeking to learn more about the grammar (phonology, morphology, syntax) of Turkish and to advanced students of the language as well as translators. It embodies all of the functionality of dictionaries but goes far beyond what a dictionary can offer. LingBrowser, which will be developed during a series of collaborative visits both in Turkey and in the United States, is computationally innovative; it will be the first advanced natural language processing (NLP)-based text browsing and exploration tool for use by linguistics students. Its platform independence makes it highly adaptable, usable on different types of computers, and its modular design makes it easily adaptable not only to languages other than Turkish but also, down the road, to additional types of uses, including language instruction drills for the more elementary language learner. The broader impacts of LingBrowser are twofold: (a) an increase in activity in research into Turkish grammar that LingBrowser will make possible, and (b) an increase in the development of related computational tools for other languages, based on the computational innovations embodied in LingBrowser.This award is co-funded by NSF's Office of International Science and Engineering.