| { | |
| "paper_id": "W97-0208", | |
| "header": { | |
| "generated_with": "S2ORC 1.0.0", | |
| "date_generated": "2023-01-19T04:36:38.110542Z" | |
| }, | |
| "title": "Sense Tagging: Semantic Tagging with a Lexicon", | |
| "authors": [ | |
| { | |
| "first": "Yorick", | |
| "middle": [], | |
| "last": "Wilks", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "University of Sheffield", | |
| "location": { | |
| "addrLine": "Regent Court, 211 Portobello Street", | |
| "postCode": "S1 4DP", | |
| "settlement": "Sheffield", | |
| "country": "UK" | |
| } | |
| }, | |
| "email": "[email protected]" | |
| }, | |
| { | |
| "first": "Mark", | |
| "middle": [], | |
| "last": "Stevenson", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "University of Sheffield", | |
| "location": { | |
| "addrLine": "Regent Court, 211 Portobello Street", | |
| "postCode": "S1 4DP", | |
| "settlement": "Sheffield", | |
| "country": "UK" | |
| } | |
| }, | |
| "email": "marks@[email protected]" | |
| } | |
| ], | |
| "year": "", | |
| "venue": null, | |
| "identifiers": {}, | |
| "abstract": "Sense tagging, the automatic assignment of the appropriate sense from some lexicon to each of the words in a text, is a specialised instance of the general problem of semantic tagging by category or type. We discuss which recent word sense disambignation algorithms are appropriate for sense tagging. It is our belief that sense tagging can be carried out effectively by combining several simple, independent, methods and we include the design of such a tagger. A prototype of this system has been implemented, correctly tagging 86% of polysemous word tokens in a small test set, providing evidence that our hypothesis is correct.", | |
| "pdf_parse": { | |
| "paper_id": "W97-0208", | |
| "_pdf_hash": "", | |
| "abstract": [ | |
| { | |
| "text": "Sense tagging, the automatic assignment of the appropriate sense from some lexicon to each of the words in a text, is a specialised instance of the general problem of semantic tagging by category or type. We discuss which recent word sense disambignation algorithms are appropriate for sense tagging. It is our belief that sense tagging can be carried out effectively by combining several simple, independent, methods and we include the design of such a tagger. A prototype of this system has been implemented, correctly tagging 86% of polysemous word tokens in a small test set, providing evidence that our hypothesis is correct.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Abstract", | |
| "sec_num": null | |
| } | |
| ], | |
| "body_text": [ | |
| { | |
| "text": "This workshop is about semantic tagging: marking each word token 1 in a text with some marker identifying its semantic category, similar to the way a partof-speech tagger assigns a grammatical category to each token in a text. Our recent work has been concerued with sense tagging, a particular instance of this problem. Sense tagging is the process of assigning, to each content word in a text, its particular sense from some lexicon. This differs from the more general case of semantic tagging, where the tags for each word (type) are not be specific to that type and do not correspond to word senses in a lexicon. For example the tags may be broad semantic categories such as HUMAN or ANI14ATE or WordNet synsets.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Sense tagging", | |
| "sec_num": "1" | |
| }, | |
| { | |
| "text": "Another, broader, class of algorithms are word sense disambiguation (WSD) algorithms. By WSD algorithm we mean any procedure which carries out semantic disambignation on words, these may not necessarily be tagging algorithms, in that they do IOften loosened to each content word in a text.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Sense tagging", | |
| "sec_num": "1" | |
| }, | |
| { | |
| "text": "not attempt to mark every token in a text but may be restricted to disambiguating small sets of word types.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "47", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Sense tagging is a difficult problem: each word (type) has its own set of tags which may be quite large. This rules out approaches which rely on a discriminator being created for each semantic tag which is then applied to text, although this is a valuable technique when there are a small number of tags which are broad semantic categories.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "47", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "However, sense tagging is an extremely useful procedure to carry out since the tags which are associated during sense tagging are rich in knowledge and therefore likely to be extremely useful for further processing. Indeed, the lack of reliable, large-scale, sense taggers has often been blamed for the failure of machine translation for the last 30 years.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "47", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "In this paper we shall discuss some recent approaches to the WSD problem and examine their usefulness for the more specialised task of sense tagging. We then propose an approach which makes use of several different types of information and report a partial implementation of this system which produces very encouraging results.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "47", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Recent word sense disambignation (WSD) algorithms can be categorised into two broad types:", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Recent Word Sense Disambiguation algorithms", | |
| "sec_num": "2" | |
| }, | |
| { | |
| "text": "1. WSD using information in an explicit lexicon. This is usually a Machine Readable Dictionary (MRD) such as the Longman Dictionary o] Contemporary English (LDOCE) (Procter, 1978) , WordNet (Miller (Ed.), 1990) or handcrafted. Recent examples of this work include (Bruce and Guthrie, 1992) , (Bruce and Wiebe, 1994) , (McRoy, 1992) .", | |
| "cite_spans": [ | |
| { | |
| "start": 164, | |
| "end": 179, | |
| "text": "(Procter, 1978)", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 190, | |
| "end": 210, | |
| "text": "(Miller (Ed.), 1990)", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 264, | |
| "end": 289, | |
| "text": "(Bruce and Guthrie, 1992)", | |
| "ref_id": "BIBREF1" | |
| }, | |
| { | |
| "start": 292, | |
| "end": 315, | |
| "text": "(Bruce and Wiebe, 1994)", | |
| "ref_id": "BIBREF3" | |
| }, | |
| { | |
| "start": 318, | |
| "end": 331, | |
| "text": "(McRoy, 1992)", | |
| "ref_id": "BIBREF11" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Recent Word Sense Disambiguation algorithms", | |
| "sec_num": "2" | |
| }, | |
| { | |
| "text": "2. WSD using information gained from training on some corpus. This approach can be further sub-m m classified:", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Recent Word Sense Disambiguation algorithms", | |
| "sec_num": "2" | |
| }, | |
| { | |
| "text": "(a) Supervised training, where information is gathered from corpora which have already been semantically disambiguated. As such corpora are hard to obtain, usually requiring expensive hand-tagging, research in this area has concentrated on other forms of lexical ambiguities, eg. (Gale, Church, and Yarowsky, 1992) . (b) Unsupervised training, where information is gathered from raw corpora which has not been semantically disambiguated. The best examples of this approach has been the resent work of Yarowsky - (Yarowsky, 1992) , (Yarowsky, 1993) , (Yarowsky, 1995) .", | |
| "cite_spans": [ | |
| { | |
| "start": 280, | |
| "end": 314, | |
| "text": "(Gale, Church, and Yarowsky, 1992)", | |
| "ref_id": "BIBREF6" | |
| }, | |
| { | |
| "start": 512, | |
| "end": 528, | |
| "text": "(Yarowsky, 1992)", | |
| "ref_id": "BIBREF21" | |
| }, | |
| { | |
| "start": 531, | |
| "end": 547, | |
| "text": "(Yarowsky, 1993)", | |
| "ref_id": "BIBREF22" | |
| }, | |
| { | |
| "start": 550, | |
| "end": 566, | |
| "text": "(Yarowsky, 1995)", | |
| "ref_id": "BIBREF23" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Recent Word Sense Disambiguation algorithms", | |
| "sec_num": "2" | |
| }, | |
| { | |
| "text": "These approaches are not mutually exclusive and there are, of course, some hybrid cases, for example Luk (Luk, 1995) uses information in MRD definitions (approach 1) and statistical information from untagged corpora (approach 2b).", | |
| "cite_spans": [ | |
| { | |
| "start": 105, | |
| "end": 116, | |
| "text": "(Luk, 1995)", | |
| "ref_id": "BIBREF8" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Recent Word Sense Disambiguation algorithms", | |
| "sec_num": "2" | |
| }, | |
| { | |
| "text": "Approach 2a is the least promising since text tagged with word senses is practically non-existent and is both time consuming and difficnlt to produce manually. Much of the research in this area has been compromised by the fact that researchers have focussed on lexical ambiguities that are not true word sense distinctions, such as words translated differently across two languages (Gale, Church, and Yarowsky, 1992) or homophones ~ (Yarowsky, 1993) .", | |
| "cite_spans": [ | |
| { | |
| "start": 382, | |
| "end": 416, | |
| "text": "(Gale, Church, and Yarowsky, 1992)", | |
| "ref_id": "BIBREF6" | |
| }, | |
| { | |
| "start": 433, | |
| "end": 449, | |
| "text": "(Yarowsky, 1993)", | |
| "ref_id": "BIBREF22" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Comparing Different Approaches", | |
| "sec_num": "3" | |
| }, | |
| { | |
| "text": "Even in the cases where data with the appropriate sense distinctions is available, the text is unliicely to be from the desired domain: a word sense discriminator trained on company news text will be much less effective on text about electronics products. A discriminator trained on many types of text so as to be generic will not be particularly successful in any specific domain.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Comparing Different Approaches", | |
| "sec_num": "3" | |
| }, | |
| { | |
| "text": "Approach 2b has received much attention recently. Its disadvantage is that sense disambiguation is not carried out relative to any well defined set of senses, but rather an ad hoc set. Although this research has been the most successful of all approa~es, it is difficult to see what use could be made of the word sense distinctions produced.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Comparing Different Approaches", | |
| "sec_num": "3" | |
| }, | |
| { | |
| "text": "Using approach 1 with hand crafted lexicons has the disadvantage of being expensive to create: indeed Small and Rieger (Small and Rieger, 1982) attempted WSD using \"word experts\", which were 2Words pronounced identically but spelled di~erently.", | |
| "cite_spans": [ | |
| { | |
| "start": 119, | |
| "end": 143, | |
| "text": "(Small and Rieger, 1982)", | |
| "ref_id": "BIBREF16" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Comparing Different Approaches", | |
| "sec_num": "3" | |
| }, | |
| { | |
| "text": "essentially hand crafted disambiguators. They reported that the word expert for \"throw\" is \"currently six pages long, but should be ten times that size\", making this approach impractical for any system aiming for broad coverage.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "48", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Word senses are not absolute or Platonic but defined by a given lexicon, as has been known for many years from early work on WSD, even though the contrary seems widely believed: \".. it is very difficult to assign word occurrences to sense classes in any manner that is both general and determinate. In the sentences \"I have a stake in this country.\" and \"My stake in the last race was a pound\" is \"stake\" being used in the same sense or not? If \"stake\" can be interpreted to mean something as vague as 'Stake as any kind of investment in any enterprise' then the answer is yes. So, if a semantic dictionary contained only two senses for \"stake\": that vague sense together with 'Stake as a post', then one would expect to assign the vague sense for both the sentences above. But if, on the other hand, the dictionary distinguished 'Stake as an investment' and 'Stake as an initial payment in a game or race' then the answer would be expected to be different. So, then, word sense disambiguation is relative to the dictionary of sense choices available and can have no absolute quality about it.\" (Wilks, 1972) There is no general agreement over the number of senses appropriate for lexical entries: at one end of the spectrum Wierzbicka (Wierzbicka, 1989) claims words have essentially one sense while Pustejovsky believes that \"... words can assume a potentially ini~nite number of senses in context.\" (Pustejovsky, 1995) How, then, are we to get an initial lexicon of word senses? We believe the best resource is still a Machine Readable Dictionary: they have a relatively well-defined set of sense tags for each word and lexical coverage is high.", | |
| "cite_spans": [ | |
| { | |
| "start": 1095, | |
| "end": 1108, | |
| "text": "(Wilks, 1972)", | |
| "ref_id": "BIBREF18" | |
| }, | |
| { | |
| "start": 1236, | |
| "end": 1254, | |
| "text": "(Wierzbicka, 1989)", | |
| "ref_id": "BIBREF17" | |
| }, | |
| { | |
| "start": 1402, | |
| "end": 1421, | |
| "text": "(Pustejovsky, 1995)", | |
| "ref_id": "BIBREF14" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Proposed Approach", | |
| "sec_num": "4" | |
| }, | |
| { | |
| "text": "MRDs are, of course, normally generic, and much practical WSD work is for sub-domains. We are adhering to the view that it is better to start with such a generic lexicon and adapt it automatically with specialist words and senses. The work described here is part of ECRAN (Wilks, 1995) , a European LRE project on tuning lexicons to domains, with a general sense tagging module used as a first stage.", | |
| "cite_spans": [ | |
| { | |
| "start": 272, | |
| "end": 285, | |
| "text": "(Wilks, 1995)", | |
| "ref_id": "BIBREF19" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Proposed Approach", | |
| "sec_num": "4" | |
| }, | |
| { | |
| "text": "An interesting fact about recent word sense disambiguation algorithms is that they have made use of different, orthogonal, sources of information: the in-formation provided by each source seems independent of and has no bearing on any of the others. We propose a tagger that makes use of several types of information (dictionary definitions, parts-of-speech, domain codes, selectional preferences and collocates) in the tradition of McRoy (McRoy, 1992) although, the information sources we use are orthogonal, unlike the sources she used, making it easier to evaluate the performance of the various modules.", | |
| "cite_spans": [ | |
| { | |
| "start": 433, | |
| "end": 452, | |
| "text": "McRoy (McRoy, 1992)", | |
| "ref_id": "BIBREF11" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Knowledge Sources", | |
| "sec_num": "5" | |
| }, | |
| { | |
| "text": "It has already been shown that part-of-speech tags are a useful discriminator for semantic disambiguation (Wilks and Stevenson, 1996) , although they are not, normally, enough to fully disambiguate a text. For example knowing \"bank\" in \"My bank is on the corner.\" is being used as a noun will tell us that the word is not being used in the 'plane turning corner' sense but not whether it is being used in the 'financial institution' or 'edge of river' senses. Partof-speech tags can provide a valuable step towards the solution to sense tagging: fully disambiguating about 87% of ambiguous word tokens and reducing the ambiguity for some of the rest.", | |
| "cite_spans": [ | |
| { | |
| "start": 106, | |
| "end": 133, | |
| "text": "(Wilks and Stevenson, 1996)", | |
| "ref_id": "BIBREF20" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Part-of-speech", | |
| "sec_num": "5.1" | |
| }, | |
| { | |
| "text": "Pragmatic domain codes can be used to disambiguate (usually nominal) senses, as was shown by (Bruce and Guthrie, 1992) and (Yarowsky, 1992) . Our intuition here is that disambiguation evidence can be gained by choosing senses which are closest in a thesanral hierarchy. Closeness in such a hierarchy can be effectively expressed as the number of nodes between concepts. We are implementing a simple algorithm which prefers close senses in our domain hierarchy which was derived from LDOCE (Bruce and Guthrie, 1992) .", | |
| "cite_spans": [ | |
| { | |
| "start": 93, | |
| "end": 118, | |
| "text": "(Bruce and Guthrie, 1992)", | |
| "ref_id": "BIBREF1" | |
| }, | |
| { | |
| "start": 123, | |
| "end": 139, | |
| "text": "(Yarowsky, 1992)", | |
| "ref_id": "BIBREF21" | |
| }, | |
| { | |
| "start": 489, | |
| "end": 514, | |
| "text": "(Bruce and Guthrie, 1992)", | |
| "ref_id": "BIBREF1" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Domain codes (Thesaural categories)", | |
| "sec_num": "5.2" | |
| }, | |
| { | |
| "text": "Recent work has been done using collocations as semantic disambiguators, (Yarowsky, 1993) , (Dorr, 1996) , particularly for verbs. We are attempting to derive disambiguation information by examining the prepositions as given in the subcategorization frames of verbs, and in the example sentences in LDOCE.", | |
| "cite_spans": [ | |
| { | |
| "start": 73, | |
| "end": 89, | |
| "text": "(Yarowsky, 1993)", | |
| "ref_id": "BIBREF22" | |
| }, | |
| { | |
| "start": 92, | |
| "end": 104, | |
| "text": "(Dorr, 1996)", | |
| "ref_id": "BIBREF5" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Collocates", | |
| "sec_num": "5.3" | |
| }, | |
| { | |
| "text": "There has been a long tradition in NLP of using selectional preferences for WSD (Wilks, 1972) . This approach has been recently used by (McRoy, 1992) and (Mahesh and Beale, 1996) . At its best it disambiguates both verbs, adjectives and the nouns they modify at the same time, but we shall use this information late in the disambiguation process when 49 we hope to be reasonably confident of the senses of nouns in the text from processes such as 5.2 and 5.5.", | |
| "cite_spans": [ | |
| { | |
| "start": 80, | |
| "end": 93, | |
| "text": "(Wilks, 1972)", | |
| "ref_id": "BIBREF18" | |
| }, | |
| { | |
| "start": 136, | |
| "end": 149, | |
| "text": "(McRoy, 1992)", | |
| "ref_id": "BIBREF11" | |
| }, | |
| { | |
| "start": 154, | |
| "end": 178, | |
| "text": "(Mahesh and Beale, 1996)", | |
| "ref_id": "BIBREF9" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Selectionai Preferences", | |
| "sec_num": "5.4" | |
| }, | |
| { | |
| "text": "Lesk (Lesk, 1986) proposed a method for semantic disambiguation using the dictionary definitions of words as a measure of their semantic closeness and proposed the disambiguation of sentences by computing the overlap of definitions for a sentence. Simmulated annealing, a numerical optimisation algorithm, was used to make this process practical (Cowie, Guthrie, and Guthri~, 1992) , choosing an assignment of senses from as many as 10 l\u00b0 choices.", | |
| "cite_spans": [ | |
| { | |
| "start": 5, | |
| "end": 17, | |
| "text": "(Lesk, 1986)", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 346, | |
| "end": 381, | |
| "text": "(Cowie, Guthrie, and Guthri~, 1992)", | |
| "ref_id": null | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Dictionary definitions", | |
| "sec_num": "5.5" | |
| }, | |
| { | |
| "text": "The optimisation is carried out by minimising an evaluation function, computed from the overlap of a given configuration of senses. The overlap is the total number of times each word appears more than once in the dictionary definitions of all the senses in the configuration. So that if the word \"bank\" appeared three times in a given configuration we would add two to the overlap total. This function has the disadvantage that longer definitions are prefered over short ones, since these simply have more words which can contribute to the overlap. Thus short definitions or definitions by synonym are penalised. We attempted to solve this problem by making a slight change to the method for calculating the overlap. Instead of each word contributing one we normalise it's contribution by the number of words in the definition it came from, so if a word came from a definition with three words it would add one third to the overlap total. In this way long definitions have to have many words contributing to the total to be influential and short definitions are not penalised.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Dictionary definitions", | |
| "sec_num": "5.5" | |
| }, | |
| { | |
| "text": "We found that this new function lead to a small improvement in the results of the disambiguation, however we do not believe this to be statistically significant.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Dictionary definitions", | |
| "sec_num": "5.5" | |
| }, | |
| { | |
| "text": "We have recently implemented a basic version of this tagger, initially incorporating only the part-ofspeech (5.1) and dictionary definition (5.5) stages in the process, with further stages to be added later. Our tagger currently consists of three modules:", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "A Basic Tagger", | |
| "sec_num": "6" | |
| }, | |
| { | |
| "text": "\u2022 Dictionary look-up module", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "A Basic Tagger", | |
| "sec_num": "6" | |
| }, | |
| { | |
| "text": "\u2022 Part-of-speech filter", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "A Basic Tagger", | |
| "sec_num": "6" | |
| }, | |
| { | |
| "text": "\u2022 Simulated annealing 1. We have chosen to use the machine readable version of LDOCE as our lexicon. This has been used extensively in NLP research and provides a broad set of senses for sense tagging.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "A Basic Tagger", | |
| "sec_num": "6" | |
| }, | |
| { | |
| "text": "The text is initially stemmed, leaving only morphological roots, and split into sentences. Then words belonging to a list of stop words (prepositions, pronouns etc.) are removed. For each of the remaining words, e~ of its senses are extracted from LDOCE and stored with that word. The textual definitions in each sense is processed to remove stop words and stem remaining words.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "A Basic Tagger", | |
| "sec_num": "6" | |
| }, | |
| { | |
| "text": "2. The text is tagged using the Brill tagger (Brill, 1992 ) and a translation is carried out using a manually defined mapping from the syntactic tags assigned by Briil (Penn Tree Bank tags (Marcus, Santorini, and Marcinkiewicz, 1993) ) onto the simpler part-of-speech categories associated with LDOCE senses. We then remove all senses whose part-of-speech is not consistent with the one assigned by the tagger, if none of the senses are consistent with the part-of-speech we assume the tagger has made an error and do not remove any senses.", | |
| "cite_spans": [ | |
| { | |
| "start": 45, | |
| "end": 57, | |
| "text": "(Brill, 1992", | |
| "ref_id": "BIBREF0" | |
| }, | |
| { | |
| "start": 189, | |
| "end": 233, | |
| "text": "(Marcus, Santorini, and Marcinkiewicz, 1993)", | |
| "ref_id": "BIBREF10" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "A Basic Tagger", | |
| "sec_num": "6" | |
| }, | |
| { | |
| "text": "3. The final stage is to use the simulated annealing algorithm to optimise the dictionary deftnition overlap for the remaining senses. This algorithm assigns a single sense to each token which is the tag assodated with that token.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "A Basic Tagger", | |
| "sec_num": "6" | |
| }, | |
| { | |
| "text": "Below is an example of the senses assigned by the system for the sentence \"A rapid rise in prices soon eventuated unemployment.\" We show the homograph and sense numbers from LDOCE with the stemmed content words from the dictionary definitions which are used to calculate the overlap following the dash.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Example Output", | |
| "sec_num": "7" | |
| }, | |
| { | |
| "text": "\u2022 rapid homograph 1 sense 2 -done short time", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Example Output", | |
| "sec_num": "7" | |
| }, | |
| { | |
| "text": "\u2022 rise homograph 2 sense 1 -act grow greater powerful", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Example Output", | |
| "sec_num": "7" | |
| }, | |
| { | |
| "text": "\u2022 soon homograph 0 sense 1 -long short time", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Example Output", | |
| "sec_num": "7" | |
| }, | |
| { | |
| "text": "\u2022 prices homograph 1 sense 1 -amount money which thing be offer sell buy", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Example Output", | |
| "sec_num": "7" | |
| }, | |
| { | |
| "text": "\u2022 unemployment homograph 0 sense 1 -condition lack job", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Example Output", | |
| "sec_num": "7" | |
| }, | |
| { | |
| "text": "The senses have additional information associated which we do not show here: domain codes, part of speech and grammatical information as well as semantic information.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Example Output", | |
| "sec_num": "7" | |
| }, | |
| { | |
| "text": "The senses for a word in LDOCE are grouped into homographs, sets of senses realeated by meaning. For example, one of the homographs of \"bank\" means roughly 'things piled up', the different senses distinguishing exactly what is piled up.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Example Output", | |
| "sec_num": "7" | |
| }, | |
| { | |
| "text": "We have conducted some preliminary testing of this approach: our tests were run on 10 handdisambiguated sentences from the Wall Street Journal amounting to a 209 word corpus. We found that of, the word tokens which had more than 1 homograph, 86% were assigned the correct homo. graph and 57% of tokens were assigned the correct sense using our simple tagger. These figures should be compared to 72% correct homograph assignment and 47% correct sense assignment using simulated annealing alone on the same test set (see (Cowie, Guthrie, and Guthrie, 1992) ). It should he noted that the granularity of sense distinctions at the LDOCE homograph level (eg. \"bank\" as 'edge of river' or 'financial institution') is the same as the distinctions made by current small-scale WSD algo-ritbm~ (eg. (Gale, Church, and Yarowsky, 1992) , (Yarowsky, 1993) , (Schfitze, 1992)) and our system is a true tagging algorithm, operating on free text.", | |
| "cite_spans": [ | |
| { | |
| "start": 519, | |
| "end": 554, | |
| "text": "(Cowie, Guthrie, and Guthrie, 1992)", | |
| "ref_id": "BIBREF4" | |
| }, | |
| { | |
| "start": 789, | |
| "end": 823, | |
| "text": "(Gale, Church, and Yarowsky, 1992)", | |
| "ref_id": "BIBREF6" | |
| }, | |
| { | |
| "start": 826, | |
| "end": 842, | |
| "text": "(Yarowsky, 1993)", | |
| "ref_id": "BIBREF22" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Results", | |
| "sec_num": "8" | |
| }, | |
| { | |
| "text": "Our evaluation is unsatisfactory due to the small test set, but does demonstrate that the use of independent knowledge sources leads to an improvement in the quality of disambignation. We fully expect our results to improve with the addition of further, independent, modules.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Results", | |
| "sec_num": "8" | |
| }, | |
| { | |
| "text": "In this paper we have argued that semantic tagging can be carried out only relative to the senses in some lexicon and that a machine readable dictionary provides an appropriate set of senses.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Conclusion", | |
| "sec_num": "9" | |
| }, | |
| { | |
| "text": "We reported a simple semantic tagger which achieves 86% correct disambignation using two independent sources of information: part-of-speech tags and dictionary definition overlap. A proposal to extend this tagger is developed, based on other, mutually independent, sources of lexical information.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Conclusion", | |
| "sec_num": "9" | |
| } | |
| ], | |
| "back_matter": [ | |
| { | |
| "text": "This research was supported by the European Union Language Engineering project ECRAN, number LE2110. We are grateful to Jim Cowie at CRL for providing the simulated annealing code and to Kevin Humphries and Hamish Cnnningham of the Sheffield I NLP group for advice with the implementation of the tagger.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Acknowledgements", | |
| "sec_num": null | |
| } | |
| ], | |
| "bib_entries": { | |
| "BIBREF0": { | |
| "ref_id": "b0", | |
| "title": "A simple rule-based pad-of-speech tagger", | |
| "authors": [ | |
| { | |
| "first": "E", | |
| "middle": [], | |
| "last": "Brill", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1992, | |
| "venue": "Proceeding of the Third Conference on Applied Natural Language Processing", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Brill, E. 1992. A simple rule-based pad-of-speech tagger. In Proceeding of the Third Conference on Applied Natural Language Processing, Trento, Italy.", | |
| "links": null | |
| }, | |
| "BIBREF1": { | |
| "ref_id": "b1", | |
| "title": "Genus disambiguation: A study in weighted preference", | |
| "authors": [ | |
| { | |
| "first": "R", | |
| "middle": [], | |
| "last": "Bruce", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "L", | |
| "middle": [], | |
| "last": "Guthrie", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1992, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Bruce, R. and L. Guthrie. 1992. Genus dis- ambiguation: A study in weighted preference.", | |
| "links": null | |
| }, | |
| "BIBREF2": { | |
| "ref_id": "b2", | |
| "title": "Proceesings of COLING-9~", | |
| "authors": [], | |
| "year": null, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "1187--1191", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "In Proceesings of COLING-9~, pages 1187-1191, Nantes, France.", | |
| "links": null | |
| }, | |
| "BIBREF3": { | |
| "ref_id": "b3", | |
| "title": "Word-sense disambiguation using decomposable models", | |
| "authors": [ | |
| { | |
| "first": "R", | |
| "middle": [], | |
| "last": "Bruce", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "J", | |
| "middle": [], | |
| "last": "Wiebe", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1994, | |
| "venue": "Proceedings of the 3~nd Annual Meeting of the Association for Computational Linguistics", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "139--145", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Bruce, R. and J. Wiebe. 1994. Word-sense disam- biguation using decomposable models. In Proceed- ings of the 3~nd Annual Meeting of the Associ- ation for Computational Linguistics, pages 139- 145, Los Cruces, New Mexico.", | |
| "links": null | |
| }, | |
| "BIBREF4": { | |
| "ref_id": "b4", | |
| "title": "Lexical disambiguation using simulated annealing", | |
| "authors": [ | |
| { | |
| "first": "J", | |
| "middle": [], | |
| "last": "Cowie", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "L", | |
| "middle": [], | |
| "last": "Guthrie", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "J", | |
| "middle": [], | |
| "last": "Guthrie", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1992, | |
| "venue": "Proceedings of COLING-9", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "359--365", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Cowie, J., L. Guthrie, and J. Guthrie. 1992. Lex- ical disambiguation using simulated annealing. In Proceedings of COLING-9,~, pages 359-365, Nantes, France.", | |
| "links": null | |
| }, | |
| "BIBREF5": { | |
| "ref_id": "b5", | |
| "title": "Role of word sense disambiguation in lexical aquisition: Predicting semantics from syntactic cues", | |
| "authors": [ | |
| { | |
| "first": "B", | |
| "middle": [], | |
| "last": "Dorr", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1996, | |
| "venue": "Proceedings ol COLING-96", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "322--327", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Dorr, B. 1996. Role of word sense disambiguation in lexical aquisition: Predicting semantics from syn- tactic cues. In Proceedings ol COLING-96, pages 322-327, Copenhagen, Denmark.", | |
| "links": null | |
| }, | |
| "BIBREF6": { | |
| "ref_id": "b6", | |
| "title": "One sense per discourse", | |
| "authors": [ | |
| { | |
| "first": "W", | |
| "middle": [], | |
| "last": "Gale", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "K", | |
| "middle": [], | |
| "last": "Church", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "D", | |
| "middle": [], | |
| "last": "Yarowsky", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1992, | |
| "venue": "Proceedings of the DARPA Speech and Natural Language Workshop", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "233--237", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Gale, W., K. Church, and D. Yarowsky. 1992. One sense per discourse. In Proceedings of the DARPA Speech and Natural Language Workshop, pages 233-237, Harriman, New York, February.", | |
| "links": null | |
| }, | |
| "BIBREF7": { | |
| "ref_id": "b7", | |
| "title": "Automatic sense disambiguation using machine readable dictionaries: how to tell a pine cone from an ice cream cone", | |
| "authors": [ | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Lesk", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "I986", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": null, | |
| "venue": "Proceedings ol ACM SIGDOC Conference", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "24--26", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Lesk, M. I986. Automatic sense disambiguation us- ing machine readable dictionaries: how to tell a pine cone from an ice cream cone. In Proceed- ings ol ACM SIGDOC Conference, pages 24-26, Toronto, Ontario.", | |
| "links": null | |
| }, | |
| "BIBREF8": { | |
| "ref_id": "b8", | |
| "title": "Statstical sense disambiguation with relatively small corpora using dictionary definitions", | |
| "authors": [ | |
| { | |
| "first": "A", | |
| "middle": [], | |
| "last": "Luk", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1995, | |
| "venue": "Proceedings of ACL 95", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "181--188", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Luk, A. 1995. Statstical sense disambiguation with relatively small corpora using dictionary defini- tions. In Proceedings of ACL 95, pages 181-188.", | |
| "links": null | |
| }, | |
| "BIBREF9": { | |
| "ref_id": "b9", | |
| "title": "Evaluation of the Mikrokosmos system as of 1996", | |
| "authors": [ | |
| { | |
| "first": "K", | |
| "middle": [], | |
| "last": "Mahesh", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "S", | |
| "middle": [], | |
| "last": "Beale", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1996, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Mahesh, K. and S. Beale. 1996. Evaluation of the Mikrokosmos system as of 1996. Technical Report MCCS-96-300, Computing Research Laboratory.", | |
| "links": null | |
| }, | |
| "BIBREF10": { | |
| "ref_id": "b10", | |
| "title": "Building a large annotated corpus of English: The Penn Tree Bank", | |
| "authors": [ | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Marcus", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "R", | |
| "middle": [], | |
| "last": "Santorini", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "M", | |
| "middle": [ | |
| "A" | |
| ], | |
| "last": "Marcinkiewicz", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1993, | |
| "venue": "Computational Linguistics", | |
| "volume": "19", | |
| "issue": "2", | |
| "pages": "313--330", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Marcus, M., R. Santorini, and M. A. Marcinkiewicz. 1993. Building a large annotated corpus of En- glish: The Penn Tree Bank. Computational Lin- guistics, 19(2):313-330.", | |
| "links": null | |
| }, | |
| "BIBREF11": { | |
| "ref_id": "b11", | |
| "title": "Using multiple knowledge sources for word sense disambiguation", | |
| "authors": [ | |
| { | |
| "first": "S", | |
| "middle": [], | |
| "last": "Mcroy", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1992, | |
| "venue": "Computational Linguistics", | |
| "volume": "", | |
| "issue": "1", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "McRoy, S. 1992. Using multiple knowledge sources for word sense disambiguation. Computational Linguistics, 18(1).", | |
| "links": null | |
| }, | |
| "BIBREF12": { | |
| "ref_id": "b12", | |
| "title": "WordNet: An on-line lexical database", | |
| "authors": [ | |
| { | |
| "first": ";", | |
| "middle": [ | |
| "G A" | |
| ], | |
| "last": "Miller", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1990, | |
| "venue": "International Journal of Lexicography", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Miller (Ed.), G. A. 1990. WordNet: An on-line lexical database. International Journal of Lexi- cography.", | |
| "links": null | |
| }, | |
| "BIBREF14": { | |
| "ref_id": "b14", | |
| "title": "The Generative Lexicon", | |
| "authors": [ | |
| { | |
| "first": "J", | |
| "middle": [], | |
| "last": "Pustejovsky", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1995, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Pustejovsky, J. 1995. The Generative Lexicon. MIT Press, Cambridge, MA.", | |
| "links": null | |
| }, | |
| "BIBREF15": { | |
| "ref_id": "b15", | |
| "title": "Dimensions of meaning", | |
| "authors": [ | |
| { | |
| "first": "H", | |
| "middle": [], | |
| "last": "Schiitze", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1992, | |
| "venue": "Proceedings of Supercomputing 'g2", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Schiitze, H. 1992. Dimensions of meaning. In Pro- ceedings of Supercomputing 'g2.", | |
| "links": null | |
| }, | |
| "BIBREF16": { | |
| "ref_id": "b16", | |
| "title": "Parsing and comprehending with word experts (a theory and its realisation)", | |
| "authors": [ | |
| { | |
| "first": "S", | |
| "middle": [], | |
| "last": "Small", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "C", | |
| "middle": [], | |
| "last": "Rieger", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1982, | |
| "venue": "W. Lehnert and M. Ringle, editors, Strategies for Natural Language Processing", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Small, S. and C. Rieger. 1982. Parsing and com- prehending with word experts (a theory and its realisation). In W. Lehnert and M. Ringle, edi- tors, Strategies for Natural Language Processing.", | |
| "links": null | |
| }, | |
| "BIBREF17": { | |
| "ref_id": "b17", | |
| "title": "Semantics, Culture and Cognition", | |
| "authors": [ | |
| { | |
| "first": "A", | |
| "middle": [], | |
| "last": "Wierzbicka", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1989, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Wierzbicka, A. 1989. Semantics, Culture and Cog- nition. Oxford University Press, Oxford.", | |
| "links": null | |
| }, | |
| "BIBREF18": { | |
| "ref_id": "b18", | |
| "title": "Grammar, Meaning and the Machine Analysis of Language", | |
| "authors": [ | |
| { | |
| "first": "Y", | |
| "middle": [], | |
| "last": "Wilks", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1972, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Wilks, Y. 1972. Grammar, Meaning and the Ma- chine Analysis of Language. Rontledge, London.", | |
| "links": null | |
| }, | |
| "BIBREF19": { | |
| "ref_id": "b19", | |
| "title": "ECRAN: Extraction of Content: Research at Near-market. EU funded LRE Research Proposal", | |
| "authors": [ | |
| { | |
| "first": "Y", | |
| "middle": [], | |
| "last": "Wilks", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1995, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Wilks, Y. 1995. ECRAN: Extraction of Content: Research at Near-market. EU funded LRE Re- search Proposal.", | |
| "links": null | |
| }, | |
| "BIBREF20": { | |
| "ref_id": "b20", | |
| "title": "The grammar of sense: Is word-sense tagging much more than part-of-speech tagging?", | |
| "authors": [ | |
| { | |
| "first": "Y", | |
| "middle": [], | |
| "last": "Wilks", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Stevenson", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1996, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Wilks, Y. and M. Stevenson. 1996. The grammar of sense: Is word-sense tagging much more than part-of-speech tagging? Technical Report CS-96- 05, University of Sheffield.", | |
| "links": null | |
| }, | |
| "BIBREF21": { | |
| "ref_id": "b21", | |
| "title": "Word-sense disambiguation ushag statistical models of Roget's categories trained on large corpora", | |
| "authors": [ | |
| { | |
| "first": "D", | |
| "middle": [], | |
| "last": "Yarowsky", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1992, | |
| "venue": "COLING-92", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Yarowsky, D. 1992. Word-sense disambiguation us- hag statistical models of Roget's categories trained on large corpora. In COLING-92.", | |
| "links": null | |
| }, | |
| "BIBREF22": { | |
| "ref_id": "b22", | |
| "title": "One sense per collocation", | |
| "authors": [ | |
| { | |
| "first": "D", | |
| "middle": [], | |
| "last": "Yarowsky", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1993, | |
| "venue": "Proceedings ARPA Human Language Technology Workshop", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "266--271", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Yarowsky, D. 1993. One sense per collocation. In Proceedings ARPA Human Language Technology Workshop, pages 266--271.", | |
| "links": null | |
| }, | |
| "BIBREF23": { | |
| "ref_id": "b23", | |
| "title": "Unsupervised word-sense disambiguation rivaling supervised methods", | |
| "authors": [ | |
| { | |
| "first": "D", | |
| "middle": [], | |
| "last": "Yarowsky", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1995, | |
| "venue": "Proceedings of ACL95", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "189--196", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Yarowsky, D. 1995. Unsupervised word-sense dis- ambiguation rivaling supervised methods. In Pro- ceedings of ACL95, pages 189-196.", | |
| "links": null | |
| } | |
| }, | |
| "ref_entries": {} | |
| } | |
| } |