Use of Machine Readable Dictionaries for Word-Sense Disambiguation in SENSEVAL-2

Use of Machine Readable Dictionaries for Word-Sense Disambiguation in SENSEV AL-2 Kenneth C. Litkowski CL Research 9208 Gue Road · Damascus, MD 20872 ken@clres. com Abstract CL Research's word-sense disambiguation (WSD) system is part of the DIMAP dictionary software, designed to use any full dictionary as the basis for unsupervised disambiguation. Official SENSEV AL-2 results were generated using WordNet, and separately using the New Oxford Dictionary of English (NODE). The disambiguation functionality exploits whatever information is made available by the lexical database. Special routines examined multiword units and contextual clues (both collocations, definition and example content words, and subject matter analyses); syntactic constraints have not yet been employed. The official coarse- grained precision was 0.367 for the lexical sample task and 0.460 for the all-words task (these are actually recall, with actual precision of 0.390 and 0.506 for the two tasks). NODE definitions were automatically mapped into WordNet, with precision of0.405 and 0.418 on 75% and 70% mapping for the lexical sample and all-words tasks, respectively, comparable to WordNet. Bug fixes and implementation of incomplete routines have increased the precision for the lexical sample to 0.429 (with many improvements still likely). Introduction CL Research's participation in SENSEV AL-2 was designed to ( 1) extend WSD techniques from SENSEV AL-l (Litkowski, 2000 ), (2) generalize WSD mechanisms to rely on a full dictionary rather than a small set of entries where individual crafting might intrude, and (3) investigate WSD using one dictionary mapped into another (WordNet). Results indicate positive achievements for each of these goals. Time constraints precluded a complete assessment of the upper limits that can be achieved. In particular, although the general architecture from SENSEV AL-l was retained, several specific WSD routines were notreimplemented. Incomplete testing, debugging, and implementation of new routines significantly affected the official results. Several of these problems are investigated more fully below. CL Research's WSD functionality is implemented in DIMAP1, designed primarily for creation and maintenance of lexicons for natural language processing. In particular, DIMAP is designed to make machine-readable dictionaries (MRDs) tractable and to create semantic networks (similar to WordNet (Fellbaum, 1998) and MindNet (Richardson, 1997)) automatically by analyzing and parsing definitions. Section 1 describes the dictionary preparation techniques for WordNet and NODE (The New Oxford Dictionary of English, 1998), as well as the mapping from NODE to WordNet. Section 2 describes the WSD techniques used in SENSEV AL-2. Section 3 describes the SENSEV AL-2 results and section 4 discusses these results .. 1 Dictionary Preparation DIMAP can disambiguate any text against WordNet or any other dictionary converted to DIMAP, with a special emphasis on corpus instances for specillc lemmas. The dictionaries used for disambiguation operate in the background (as distinguished from the foreground development and maintenance of a dictionary), with rapid btree lookup to access and examine the characteristics of multiple senses of a word after a sentence has been parsed. DIMAP allows multiple senses for each entry, with fields for the definitions, usage notes, hypemyms, hyponyms, 1Dictionary MAintenance :frograms, available from CL …

Content truncated to 500 words