| { | |
| "paper_id": "1987", | |
| "header": { | |
| "generated_with": "S2ORC 1.0.0", | |
| "date_generated": "2023-01-19T03:23:58.628768Z" | |
| }, | |
| "title": "", | |
| "authors": [], | |
| "year": "", | |
| "venue": null, | |
| "identifiers": {}, | |
| "abstract": "", | |
| "pdf_parse": { | |
| "paper_id": "1987", | |
| "_pdf_hash": "", | |
| "abstract": [], | |
| "body_text": [ | |
| { | |
| "text": "System Reinforced by Semantics), especially about its grammar system, are presented. A new grammar system developed for TAURAS enables the large-scale grammars be easily written for treating a large amount of documents which have various varieties of sentences. The goal of this system is multilingual translation for science and technical documents. The system is written in C for the sake of portability, efficiency, and readability, and implemented on Toshiba minicomputer AS3000 with UNIX*.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Design and philosophy of TAURAS(Toshiba Automatic Translation", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "The translation system has three main elements: 1) Translation unit 2) Bilingual editor 3) Software utilities, e.g. Japanese/English word-processors. The total software configuration is shown in figure 2-1. The bilingual editor is equipped with a man-machine-interface devices that makes possible the efficient processing of various problems related to translation, such as where and why errors occurred, for instance.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "SYSTEM CONFIGURATION", | |
| "sec_num": "2." | |
| }, | |
| { | |
| "text": "The morphological analyzer divides a word into morphemes and constructs a word structure as shown in figure 3-1. SW: source word (infinitive) POS: category NUM: number GEN: gender PSN: person TW: target words (translations) SM: semantic markers OTHERS: tense, aspect, modality and so on PLR: pointer to the lexical rules SW, POS, TW, SM, and PLR are necessarily provided by the dictionaries. NUM is also provided by the dictionaries only if the word has an irregular form such as \"feet.\" The input sentence (1) is transformed into a string of word structures (2), for instance .", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "TRANSLATION METHOD 3.1 MORPHOLOGICAL ANALYSIS", | |
| "sec_num": "3." | |
| }, | |
| { | |
| "text": "(1) He is a singer.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "TRANSLATION METHOD 3.1 MORPHOLOGICAL ANALYSIS", | |
| "sec_num": "3." | |
| }, | |
| { | |
| "text": "( Though the syntactic and semantic analyzer are separate in our system, they are not completely independent: rather than working sequentially, they proceed in an interactive manner. Figure 3 -3 shows the flow of the syntactic and the semantic processing and their relation. The features of the syntactic analysis of the system are as follows:", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 183, | |
| "end": 191, | |
| "text": "Figure 3", | |
| "ref_id": "FIGREF2" | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "TRANSLATION METHOD 3.1 MORPHOLOGICAL ANALYSIS", | |
| "sec_num": "3." | |
| }, | |
| { | |
| "text": "1) The syntactic analyzer always derives only one syntactic structure for a string of categories of a sentence. Structural ambiguities are implicitly represented in the syntactic structure. Semantic analyzer will construct a plausible conceptual structure, resolving such implicit ambiguities.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "TRANSLATION METHOD 3.1 MORPHOLOGICAL ANALYSIS", | |
| "sec_num": "3." | |
| }, | |
| { | |
| "text": "2) The syntactic analyzer is purely syntactic, and syntactic rules have no semantic conditions. A well-known example is the following: 1) He promised her to go. 2) He persuaded her to go.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "TRANSLATION METHOD 3.1 MORPHOLOGICAL ANALYSIS", | |
| "sec_num": "3." | |
| }, | |
| { | |
| "text": "These two sentences have the same surface syntactic structure, but they have different conceptual structures, corresponding to the different interpretations of the deep subject of \"to go.\"", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "TRANSLATION METHOD 3.1 MORPHOLOGICAL ANALYSIS", | |
| "sec_num": "3." | |
| }, | |
| { | |
| "text": "In our method, the syntactic analyzer makes a unique syntactic structure for the same sequence of categories with different conceptual interpretations. Thus, syntactic rules do not need to have semantic conditions. The syntactic rules are described by an Augmented Transition Network Grammar(ATNG) formalism.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "TRANSLATION METHOD 3.1 MORPHOLOGICAL ANALYSIS", | |
| "sec_num": "3." | |
| }, | |
| { | |
| "text": "Our method of semantic analysis is based on the following hypothesis.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "SEMANTIC ANALYSIS 3.3.1 METHOD OF SEMANTIC ANALYSIS", | |
| "sec_num": "3.3" | |
| }, | |
| { | |
| "text": "HYPl: Meaning is lexical The typical example is shown in 3.2. That is, 1) He promised her to go. 2) He persuaded her to go.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "SEMANTIC ANALYSIS 3.3.1 METHOD OF SEMANTIC ANALYSIS", | |
| "sec_num": "3.3" | |
| }, | |
| { | |
| "text": "Though both sentences have the same sequence of categories, they have different deep subjects for the infinitive \"to go.\" This means these two sentences have different conceptual structures and this difference results from the difference of the meanings of \"promise\" and \"persuade\". HYPl implies more than this fact. From the computational linguistic point of view, it insists that semantic rules should not be mixed with syntactic rules. According to this schema, semantic rules are attached to words in the dictionary as lexical rules. If semantic rules are incorporated into syntactic rules in the form of conditions or the like, syntactic rules will be overly intricate. Moreover, syntactic rules, which originally only define the order of words in a sentence, must be written considering the meaning of a resultant sentence. For these reasons, we adopt lexical rules for semantic processing. The conceptual diagram of our semantic processing system is provided in figure 3-4.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "SEMANTIC ANALYSIS 3.3.1 METHOD OF SEMANTIC ANALYSIS", | |
| "sec_num": "3.3" | |
| }, | |
| { | |
| "text": "Semantic rules consist of tree-to-tree conversion together with conditions and actions. The form of a semantic rule is as follows: MP = TP; COND; ACT; CTRL Here, MP is a matching pattern which must match against with a subtree of syntactic structure. If conditions, which are represented by COND, also hold true, then the subtree is converted into the target pattern represented by TP, and actions represented by ACT are executed. CTRL is the control of flow of lexical rules attached to a word. More detailed explanations are given below, using a simple example.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "NOTATION OF SEMANTICS RULES", | |
| "sec_num": "3.3.2" | |
| }, | |
| { | |
| "text": "Here, the notation using square brackets is an abbreviation of the word structure and the relevant features. [SW:I] This condition denotes condition that the translation of \"take\" into Japanese depends on the nature of the object.", | |
| "cite_spans": [ | |
| { | |
| "start": 109, | |
| "end": 115, | |
| "text": "[SW:I]", | |
| "ref_id": null | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "NOTATION OF SEMANTICS RULES", | |
| "sec_num": "3.3.2" | |
| }, | |
| { | |
| "text": "ACT: ACT treats features in a word, adding and deleting some features, and so on. E.g.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "NOTATION OF SEMANTICS RULES", | |
| "sec_num": "3.3.2" | |
| }, | |
| { | |
| "text": "ACT: set-feature(nodel;TW;no-ru)", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "NOTATION OF SEMANTICS RULES", | |
| "sec_num": "3.3.2" | |
| }, | |
| { | |
| "text": "CTRL: This decides the type of rules: accept-type or reject-type", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "NOTATION OF SEMANTICS RULES", | |
| "sec_num": "3.3.2" | |
| }, | |
| { | |
| "text": "In the above example, the lexical rules for \"take\" can select a proper translation by referring to the SW slot or SM slot.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "ROLES OF SEMANTIC RULES 1) Selection of translation", | |
| "sec_num": "3.3.3" | |
| }, | |
| { | |
| "text": "2) Processing of idioms For our purpose, an \"idiom\" is any sequence of more than one word which must be treated as a single unit for translation purposes. Idioms can be non-contiguous or have variables, such as \"put on\", \"abandon oneself to\", and so on. Such idioms can not be registered in the dictionary as a single word, because \"put\" and \"on\" may sometimes be separated by intervening words, and \"abandon\" and \"oneself\" may both undergo grammatical variability, e.g. inflection, person concord. Syntactic analysis does not treat these as idioms, but simply as individual words and makes a syntactic structure in the usual way. Semantic analysis interprets these words in the syntactic structure as an idiom using lexical rules which are attached to the head word of the idiom. Hence, the idioms are represented as subtrees in the rules, not as a string of words.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "ROLES OF SEMANTIC RULES 1) Selection of translation", | |
| "sec_num": "3.3.3" | |
| }, | |
| { | |
| "text": "3) Lexically structural transfer Lexical rules for idioms are an example of lexical rules with structural transfer. More generally lexical rules other than idioms have structural transfer. For example, English has a negative determiner \"no\" while Japanese does not have such a determiner and such negation must be expressed by negation of the predicate. Hence, if this negative determiner appears in a sentence, a lexical rule which transfers the noun negation to the predicate is needed. One of lexical rules of \"no\" is as follows; Here, \"*\" in \"v*\" is a wild-card character and \"v*\" means verb group. The arc depicted by a dotted line means an ancestor-descendent relation while a rigid line indicates a parent-child relation.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "ROLES OF SEMANTIC RULES 1) Selection of translation", | |
| "sec_num": "3.3.3" | |
| }, | |
| { | |
| "text": "So far we have computed a conceptual structure for the target language. The next stage is the generation of the target sentence. The roles of syntactic generation are as follows:", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "SYNTACTIC GENERATION", | |
| "sec_num": "3.4" | |
| }, | |
| { | |
| "text": "1) Determine word order in a conceptual tree 2) Attach postpositions (Joshi in Japanese)", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "SYNTACTIC GENERATION", | |
| "sec_num": "3.4" | |
| }, | |
| { | |
| "text": "Word order is represented by an augmented context free grammar. Postpositions which represent cases are usually given as literals in rules, since there are typical postpositions representing each specific case. For example, \"ga\" , \"wo\", and \"e\" represent subject, object, and goal respectively. But some verbals do not take these typical postpositions; for example, \"suki-da\" (\"like\") takes \"ga\" as an object case marker instead of \"wo\". For such special cases, the variable postposition mechanism is used. In the case of \"ga\" for object, a postposition is not given by a rule but from the case slot in the word structure of \"like.\"", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "SYNTACTIC GENERATION", | |
| "sec_num": "3.4" | |
| } | |
| ], | |
| "back_matter": [ | |
| { | |
| "text": "Most information for morphological generation is included in the dictionaries. For example, one of the translations for \"write\" is \"yo-mu\". This is an infinitive form and the following information is needed to get the conjugated form: 1) Stem: \"yo\" 2) Conjugation type: \"5-dan\" 3) Kind of conjugational part: \"ma\" 4) Other information (such as an irregular form in special use): \"onbin-kei\"Information which is got during syntactic analysis, such as tense, aspect, modality and so on, is attached to the head verb as morphological information. Figure 3-7 gives a rough sketch of the above process. 1) Semantics with low computation cost and high performance is introduced, which is a new semantico-syntactic approach.2) TAURAS has been developed as a total translation system on a engineering workstation.", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 544, | |
| "end": 554, | |
| "text": "Figure 3-7", | |
| "ref_id": null | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "MORPHOLOGICAL GENERATION", | |
| "sec_num": "3.5" | |
| } | |
| ], | |
| "bib_entries": { | |
| "BIBREF0": { | |
| "ref_id": "b0", | |
| "title": "Transition Network Grammar for Natural Language Analysis", | |
| "authors": [ | |
| { | |
| "first": "Alain", | |
| "middle": [], | |
| "last": "Colmerauer", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "; Les Systemes-Q Ou Un Formalisme Pour Analyser Et Synthetiser Des Phrases", | |
| "middle": [], | |
| "last": "Sur Ordinateur'", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Taum", | |
| "middle": [], | |
| "last": "Universite De Montreal", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "W", | |
| "middle": [ | |
| "A" | |
| ], | |
| "last": "Woods", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1970, | |
| "venue": "Communication of the ACM", | |
| "volume": "13", | |
| "issue": "10", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Alain Colmerauer(1970).'LES SYSTEMES-Q OU UN FORMALISME POUR ANALYSER ET SYNTHETISER DES PHRASES SUR ORDINATEUR', TAUM, Universite de Montreal W.A.Woods(1970).'Transition Network Grammar for Natural Language Analysis',Communication of the ACM Vol.13, No.10", | |
| "links": null | |
| }, | |
| "BIBREF1": { | |
| "ref_id": "b1", | |
| "title": "ARBORESCENCES ET TRANSFORMATIONS', GETA", | |
| "authors": [ | |
| { | |
| "first": "J", | |
| "middle": [], | |
| "last": "Chauche", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1972, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "J.Chauche(1972).'ARBORESCENCES ET TRANSFORMATIONS', GETA, Universite Scientifique et Medicale de Grenoble", | |
| "links": null | |
| }, | |
| "BIBREF2": { | |
| "ref_id": "b2", | |
| "title": "Le language de programmation REZO', Department d'Informatique Faculte des Arts et des Sciences", | |
| "authors": [ | |
| { | |
| "first": "Gilles", | |
| "middle": [], | |
| "last": "Stewart", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1975, | |
| "venue": "META", | |
| "volume": "21", | |
| "issue": "2", | |
| "pages": "127--133", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Gilles Stewart(1975).'Le language de programmation REZO', Department d'Informatique Faculte des Arts et des Sciences, Universite de Montreal John Chandioux(1976).'METEO: un systeme operationnel pour la traduction automatisee des bulletins meteorologiques', META, Vol.21, No.2 pp.127-133", | |
| "links": null | |
| }, | |
| "BIBREF3": { | |
| "ref_id": "b3", | |
| "title": "Description d'un systeme de traduction automatisee des manuels d'entretien en aeronautique", | |
| "authors": [ | |
| { | |
| "first": "Bernard", | |
| "middle": [], | |
| "last": "Vauquois", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "; L'evolution Des Logiciels Et Modeles Linguistiques Pour La Traduction Automatisee'", | |
| "middle": [], | |
| "last": "", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Geta", | |
| "middle": [], | |
| "last": "", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Universite", | |
| "middle": [], | |
| "last": "Scientifique", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": ";", | |
| "middle": [ | |
| "P" | |
| ], | |
| "last": "Medicale De Grenoble John Lehrberger", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "L", | |
| "middle": [], | |
| "last": "Bourbeau", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Chevalier", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "S", | |
| "middle": [], | |
| "last": "Lepage", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1977, | |
| "venue": "American Journal of Computational Linguistics", | |
| "volume": "6", | |
| "issue": "1", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Bernard Vauquois(1977).'L'EVOLUTION DES LOGICIELS ET MODELES LINGUISTIQUES POUR LA TRADUCTION AUTOMATISEE', GETA, Universite Scientifique et Medicale de Grenoble John Lehrberger(1978).'Automatic Translation and the Concept of Sublanguage', TAUM, Universite de Montreal Isabelle,P., Bourbeau,L., Chevalier,M., Lepage,S.(1978). 'Description d'un systeme de traduction automatisee des manuels d'entretien en aeronautique, TAUM, Universite de Montreal William Woods(1980).'CASCADED ATN GRAMMARS', American Journal of Computational Linguistics, Vol.6,No.1", | |
| "links": null | |
| }, | |
| "BIBREF4": { | |
| "ref_id": "b4", | |
| "title": "Analysis Grammar of Japanese in the Mu-Project -A Procedural Approach to Analysis Grammar", | |
| "authors": [ | |
| { | |
| "first": "Jonathan", | |
| "middle": [], | |
| "last": "Slocum ; Makato", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Nagao", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1981, | |
| "venue": "THE METAL PARSING SYSTEM',Linguistic Research Center", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Jonathan Slocum (1981).'THE METAL PARSING SYSTEM',Linguistic Research Center, The University of Texas Jun-ichi Tsujii, Jun-ichi Nakamura, Makato Nagao(1984).'Analysis Grammar of Japanese in the Mu-Project -A Procedural Approach to Analysis Grammar',Proceedings of COLING84", | |
| "links": null | |
| }, | |
| "BIBREF5": { | |
| "ref_id": "b5", | |
| "title": "Dealing with Incompleteness of Linguistic Knowledge on Language Translation -Transfer and Generation Stage of a Mu Machine Translation Project", | |
| "authors": [ | |
| { | |
| "first": "Makoto", | |
| "middle": [], | |
| "last": "Nagao", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Toyoaki", | |
| "middle": [], | |
| "last": "Nishida", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1984, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Makoto Nagao, Toyoaki Nishida, Jun-ichi Tsujii(1984).Dealing with Incompleteness of Linguistic Knowledge on Language Translation -Transfer and Generation Stage of a Mu Machine Translation Project',Proceedings of COLING84", | |
| "links": null | |
| } | |
| }, | |
| "ref_entries": { | |
| "FIGREF0": { | |
| "type_str": "figure", | |
| "uris": null, | |
| "text": "Figure 3-1 WORD STRUCTURE * UNIX is a Trademark of Bell Laboratories", | |
| "num": null | |
| }, | |
| "FIGREF1": { | |
| "type_str": "figure", | |
| "uris": null, | |
| "text": "Figure 3-2 SENTENCE STRUCTURE", | |
| "num": null | |
| }, | |
| "FIGREF2": { | |
| "type_str": "figure", | |
| "uris": null, | |
| "text": "-6 TYPICAL TRANSFER PROBLEM CAUSED BY LANGUAGE PAIR DIFFERENCE", | |
| "num": null | |
| }, | |
| "TABREF1": { | |
| "type_str": "table", | |
| "content": "<table><tr><td>MP: nodel [SW:take]</td><td>TP:</td><td>[TW:no-ru]</td></tr><tr><td>OBJ</td><td/><td/></tr><tr><td>node2 [SM:vehicle]</td><td/><td/></tr><tr><td colspan=\"3\">COND: While MP represents conditions about syntactic relations</td></tr><tr><td colspan=\"3\">in a syntactic structure, COND represents conditions about</td></tr><tr><td colspan=\"2\">features in a word structure.</td><td/></tr><tr><td>E.g.</td><td/><td/></tr><tr><td colspan=\"3\">COND: semantic marker of node no.2 = \"vehicle\"</td></tr></table>", | |
| "text": "means that the element under question is only SW and other features are irrelevant.", | |
| "num": null, | |
| "html": null | |
| } | |
| } | |
| } | |
| } |