|Title||On automated semantic and syntactic annotation of texts for lexicographic purposes|
|Abstract||The main idea of this paper is that automatic annotation is the only means to secure an efficient access to the whole set of linguistic productions rather than merely a small subset of such productions annotated manually. |
Why is it necessary for a lexicographer to turn to open unannotated corpora? There are two valid concurrent reasons for that: the ever-growing rate of linguistic changes, on the one hand, and, on the other, the regional, social and professional ‘segmentation’ of the language, requiring a differential approach to the language phenomena under analysis.
For the past 10 years or so, the line of research based on the ‘Internet as a Corpus’ approach has seen booming growth. As far as technologies are concerned, the means of access available to the researcher are much more modest in this case. The methods currently used for indexing the World Wide Web by search engines are based on principles that are far from being linguistic. In spite of the fact that there are projects like Semantic Web, the Internet remains so far a raw text corpus with rather unreliable data about the frequency of occurrence.
We are presenting ongoing project ABBYY Syntactic and Semantic Parser that offers technologies for the automated linguistic annotation of text corpora. These technologies make a seamless addition to the technologies for the production of representative sub-corpora relating to the major Internet segments. ABBYY Syntactic and Semantic Parser (SSP) is built on linguistic technologies developed within the scope of the ABBYY Compreno project. It is planned to be part of LingvoPro portal (http://lingvopro.abbyyonline.com/en). Compreno is a multi-language (at the moment English, Russian, German, Spain, French, Chinese) ongoing NLP project based on the combination of sophisticated linguistic modeling and modern methods of language structure analysis (recognition). It is a scalable linguistic technology to use at a basic level for a range of NLP applications. As far as lexicography is concerned, the most important feature of this system is that automatic linguistic annotation is derived from a thorough syntactic and semantic analysis of a sentence.
|Session||Lexicography and language technology|
|Keywords||automated linguistic annotation, syntax and semantic analysis, corpus-based lexicography, Internet as a corpus.|