| { |
| "paper_id": "C94-1001", |
| "header": { |
| "generated_with": "S2ORC 1.0.0", |
| "date_generated": "2023-01-19T12:50:02.543142Z" |
| }, |
| "title": "", |
| "authors": [ |
| { |
| "first": "Satoshi", |
| "middle": [], |
| "last": "Kinoshita", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "", |
| "institution": "Toshiba Corporation", |
| "location": {} |
| }, |
| "email": "" |
| }, |
| { |
| "first": "Akira", |
| "middle": [], |
| "last": "Kumano", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "", |
| "institution": "Toshiba Corporation", |
| "location": {} |
| }, |
| "email": "" |
| } |
| ], |
| "year": "", |
| "venue": null, |
| "identifiers": {}, |
| "abstract": "This paper outlines customization of a machine translation system using translation templates, which enable usm~ to represent the bilingual knowledge needed for complex translation. To evaluate their effectiveness, we analyzed a bilingual text to estimate tire improvement in eustomizability. The result shows that about 60% of mistranslated sentences can be translated as nrodel translations by combining the proposed fra,nework with the conventional customizing functions.", |
| "pdf_parse": { |
| "paper_id": "C94-1001", |
| "_pdf_hash": "", |
| "abstract": [ |
| { |
| "text": "This paper outlines customization of a machine translation system using translation templates, which enable usm~ to represent the bilingual knowledge needed for complex translation. To evaluate their effectiveness, we analyzed a bilingual text to estimate tire improvement in eustomizability. The result shows that about 60% of mistranslated sentences can be translated as nrodel translations by combining the proposed fra,nework with the conventional customizing functions.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Abstract", |
| "sec_num": null |
| } |
| ], |
| "body_text": [ |
| { |
| "text": "The ability of natural language processing (NI.P) systems is limited due to the knowledge they have, not their framework. This is reflected by recent intensive research on acquisition of linguistic knowledge from a corpus [2] [6] [91.", |
| "cite_spans": [ |
| { |
| "start": 222, |
| "end": 225, |
| "text": "[2]", |
| "ref_id": "BIBREF1" |
| }, |
| { |
| "start": 226, |
| "end": 229, |
| "text": "[6]", |
| "ref_id": "BIBREF6" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1." |
| }, |
| { |
| "text": "Machine translation (MT) systems are no exception.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1." |
| }, |
| { |
| "text": "Conrpared with monolingual knowledge, knowledge needed for translation is difficult to collect.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1." |
| }, |
| { |
| "text": "Knowledge acquisition from a bilingual or parallel corpus is considered to be a promising way to reduce tile painstaking task[l ] [10] .", |
| "cite_spans": [ |
| { |
| "start": 130, |
| "end": 134, |
| "text": "[10]", |
| "ref_id": "BIBREF10" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1." |
| }, |
| { |
| "text": "Without customization, no general-purpose MT system can output satisfactory translations; therefore it is essential to tune the system by developing a useroriented lexicon or by registering al)propriate target words.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1." |
| }, |
| { |
| "text": "The kind of customization needed depends on how the system is used. If a user translates a document to skim it, he can judge the ability of his MT system by semantic invariance: what percentage of the content of the source text is preserved in its t,anslation. If, on tile other hand, he requires translation o1\" publication qtmlity, semantically correct tnmslation is not sufficient; that is, translations should be wcllq'ormed so as to conform to a documcntational style\u2022 To minimize post-editing, more elaborate customizing functions than in the former ease are required.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1." |
| }, |
| { |
| "text": "In this paper, we will describe a customizing framework which uses 'translation templates.' This enables users to represent bilingual knowledge for complex translation where a drastic change in lingt, istic structures occu,s to generate natural translations. Then we will discuss the effectiveness of this framework by comparing it witl~ the practically used customizing functions based on the analysis of a bilingual text\u2022", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1." |
| }, |
| { |
| "text": "If a user wants publication-quality translations, stylistic wcll-fonnedness is as important as semantic invariance. Consider translating the Japanese sentence (1). Although its Iranslation (2), which is the result of our current MT system, is correct, (3) sounds more natural than (2); in (.3), the verb phrase \"using these detectors\" is nominalized to function as a subject to represent tile cause of the 'reduce' event. If tire user prefelx (3) to (2) as a translation of (1), (2) needs to be post-edited.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Machine Translation Using Translation Templates 2.1 Aim of Translation Templates", |
| "sec_num": "2." |
| }, |
| { |
| "text": "(1) korera-no kenshutsuki-wo tsukau kotoniyori these detectors-OP;J use by kakaku-ga teigen-shita price-SUBJ reduce-PAST", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Machine Translation Using Translation Templates 2.1 Aim of Translation Templates", |
| "sec_num": "2." |
| }, |
| { |
| "text": "(2) The price dropped by using these deteetm,'s.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Machine Translation Using Translation Templates 2.1 Aim of Translation Templates", |
| "sec_num": "2." |
| }, |
| { |
| "text": "(3) Use of these detectors reduced the price.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Machine Translation Using Translation Templates 2.1 Aim of Translation Templates", |
| "sec_num": "2." |
| }, |
| { |
| "text": "As the above example illustrates, when source and target languages have a significant difference in their linguistic features, linguistic structures of source sentences are drastically changed to generate natural translations.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Machine Translation Using Translation Templates 2.1 Aim of Translation Templates", |
| "sec_num": "2." |
| }, |
| { |
| "text": "In this paper, we will call translation which requires complex structural changes 'complex trallslalion.'", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Machine Translation Using Translation Templates 2.1 Aim of Translation Templates", |
| "sec_num": "2." |
| }, |
| { |
| "text": "This type of knowledge is stored in all MT systems, but hlsuJ'ficiently. Therefore, a framework for ct, slomizing complex translation should be incorporated into the system. For this purlmse, we have introduced a framework which uses 'translation templates' to represent such knowledge.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Machine Translation Using Translation Templates 2.1 Aim of Translation Templates", |
| "sec_num": "2." |
| }, |
| { |
| "text": "Using translation templates, a user can customize his MT system to deal with complex translation without any knowledge on the system's transkltion process because translation templates are created once the user specifics corresponding expressions in a source sentence and its expected trans]atlon.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Machine Translation Using Translation Templates 2.1 Aim of Translation Templates", |
| "sec_num": "2." |
| }, |
| { |
| "text": "A 'translation template' contains at least a pair of patterns, namely 'source' and 'target' patterns, each of which consists of 'constants' and 'variables.' A source pattern (SP) is a template to be compared with a source sentence, while a target pattern (TP) is used to generate a target sentence.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "2.2 Translation Templates", |
| "sec_num": "2.5" |
| }, |
| { |
| "text": "Several reports on machine translation using translation templates suggest that they are useful for translating fixed expressions [4] [7] [8] . Our translation template is more expressive in the following points:", |
| "cite_spans": [ |
| { |
| "start": 130, |
| "end": 133, |
| "text": "[4]", |
| "ref_id": "BIBREF4" |
| }, |
| { |
| "start": 138, |
| "end": 141, |
| "text": "[8]", |
| "ref_id": "BIBREF8" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "2.2 Translation Templates", |
| "sec_num": "2.5" |
| }, |
| { |
| "text": "\u2022 More parts of speech can be specified for variables.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "2.2 Translation Templates", |
| "sec_num": "2.5" |
| }, |
| { |
| "text": "\u2022 Conditions on translating expressions matched with variables can be specified.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "2.2 Translation Templates", |
| "sec_num": "2.5" |
| }, |
| { |
| "text": "These points will be explained below. Fig. 1 shows an example of a translatkm template.", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 38, |
| "end": 44, |
| "text": "Fig. 1", |
| "ref_id": null |
| } |
| ], |
| "eq_spans": [], |
| "section": "2.2 Translation Templates", |
| "sec_num": "2.5" |
| }, |
| { |
| "text": "'$1' and '$2', which appear in both the source and target patterns, are variables, and the remaining elements are constants.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "2.2 Translation Templates", |
| "sec_num": "2.5" |
| }, |
| { |
| "text": "All constants in the source pattern should appear in a source sentence in the same order. Strings which match with variables should satisfy parts of speech designated in the 'source condition.'", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "2.2 Translation Templates", |
| "sec_num": "2.5" |
| }, |
| { |
| "text": "In this example, the strings should be analyzed as 'rip' (noun phrases).", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "2.2 Translation Templates", |
| "sec_num": "2.5" |
| }, |
| { |
| "text": "The 'part of speech(POS)' of a template represents a syntactic category of a string matched with a source pattern. Currently, 'sentence' and 'sentence modifier' can be specified. Fig. 2 shows other examples of translation templates. Fig. 2 (a) shows a template which has a variable for a verb phrase. This template is created by referring to sentence (4) and its model translation (5) The target condition specifies that a verb phrase to be matched with the wniable '$2' is generated as a gerund. adjust by can be done (5) The frequency to be eliminated can be set by adjusting tile value of C by a trimmer capacitor.", |
| "cite_spans": [ |
| { |
| "start": 381, |
| "end": 384, |
| "text": "(5)", |
| "ref_id": "BIBREF5" |
| } |
| ], |
| "ref_spans": [ |
| { |
| "start": 179, |
| "end": 185, |
| "text": "Fig. 2", |
| "ref_id": null |
| }, |
| { |
| "start": 233, |
| "end": 239, |
| "text": "Fig. 2", |
| "ref_id": null |
| } |
| ], |
| "eq_spans": [], |
| "section": "2.2 Translation Templates", |
| "sec_num": "2.5" |
| }, |
| { |
| "text": "The introduction of variables which match with verb phrases improves the flexibility of translation templates. Without these variables, we must create restricted source patterns, in which the word order of postpositional phrases like \"-de\" and \"-wo\" is fixed. Fig. 2 (b) shows a template which has a variable appearing only in the target pattern. Tiffs template is created by referring to sentences (6) and (7) below. The target word (tw) of variable '$3' is specified as 'be' and its surface form is determined according to the 'numbe,\" feature of the exp~ession of variable '$1 ' T-type low-pass filter-by be done 7The carrier component is eliminated by T-type low-pass filters. Fig. 3 shows a conceptual flow of translation process using translation templates. (The actual implementation is different from the flow.)", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 260, |
| "end": 266, |
| "text": "Fig. 2", |
| "ref_id": null |
| }, |
| { |
| "start": 681, |
| "end": 687, |
| "text": "Fig. 3", |
| "ref_id": null |
| } |
| ], |
| "eq_spans": [], |
| "section": "2.2 Translation Templates", |
| "sec_num": "2.5" |
| }, |
| { |
| "text": "Fil.'st, the 'translation template dictionary' is searched for applicable templates.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Translation Process", |
| "sec_num": "2.3" |
| }, |
| { |
| "text": "If no applicable template is found, the source sentence is translated using the conventional translation module; if found, strings matched with variables are parsed and translated. Finally, translations of variables are embedded into the target pattern.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Translation Process", |
| "sec_num": "2.3" |
| }, |
| { |
| "text": "This process is implemented in the conventiom, I translation module of our transfer-based MT system [3] .", |
| "cite_spans": [ |
| { |
| "start": 100, |
| "end": 103, |
| "text": "[3]", |
| "ref_id": "BIBREF2" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Translation Process", |
| "sec_num": "2.3" |
| }, |
| { |
| "text": "Tile morphological analyzer first constructs a word lattice for an input sentence by referring to the word dictionaries and the Japanese morphological grammar, and then produces a sequence of words from the lattice until the syntactic analyzer parses it snccessfully.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "(a) Morphological Analysis", |
| "sec_num": null |
| }, |
| { |
| "text": "Constants in tile source pattern of translation templates are stored in the 'template constant dictionary' used in the first phase of morphological analysis to create the word lattice. Fig. 4 shows a simplified example of a word lattice for sentence (1) .", |
| "cite_spans": [ |
| { |
| "start": 250, |
| "end": 253, |
| "text": "(1)", |
| "ref_id": "BIBREF0" |
| } |
| ], |
| "ref_spans": [ |
| { |
| "start": 185, |
| "end": 191, |
| "text": "Fig. 4", |
| "ref_id": "FIGREF2" |
| } |
| ], |
| "eq_spans": [], |
| "section": "(a) Morphological Analysis", |
| "sec_num": null |
| }, |
| { |
| "text": "Constants of transhltion templates in a word lattice should be selected if and only i[\" all the constants of a particuk~r template are selected simultaneously to form a valid sequence of words. In Fig. 4 , we can obtain two valid word sequences froln the word lattice.", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 197, |
| "end": 203, |
| "text": "Fig. 4", |
| "ref_id": "FIGREF2" |
| } |
| ], |
| "eq_spans": [], |
| "section": "(a) Morphological Analysis", |
| "sec_num": null |
| }, |
| { |
| "text": "The present implementation permits one al3plicable template for each source sentence. If more than one templates are applicable, the priority for each template is calculated based o,i the total length of constants and tile scope of the source sentence covered by the template, and a word seqt,ence is produced in the order of their priorities. (c) TransFer and (;eneration In the transfer phase, a translation template is transfornmd into a lexical transfer rule in the conventional form, so that the new matching pattern matches with the struelure produced by tile syntactic analyzer. The result of applying this rule is a target structure;* its direct constitt, ents are given tile word order and ready to output as a target sentence.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "(a) Morphological Analysis", |
| "sec_num": null |
| }, |
| { |
| "text": "In principle, all translation can be described by translation templates.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Criteria for Using Translation Templates", |
| "sec_num": "3." |
| }, |
| { |
| "text": "That is, users can make a lranslalion template by substituting corresponding expressions in source and target sentences wilh variables.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Criteria for Using Translation Templates", |
| "sec_num": "3." |
| }, |
| { |
| "text": "The question is the appropriateness of te m plates.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Criteria for Using Translation Templates", |
| "sec_num": "3." |
| }, |
| { |
| "text": "The fi~.'st criterion is its 'applicability.' ill the following cases, translation templates are inal)propriate because the sonrce pattern is too specific to be applied to other sentences. (C1) A source sentence is translated into two target sentences or a compound sentence. (C2) Two source sentences are translated into one t:,rget sentence. (C3) A source sentence contains a parenthesis or a gapping.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Criteria for Using Translation Templates", |
| "sec_num": "3." |
| }, |
| { |
| "text": "In slJcb cases, the source pattern may contain more conslants than that of the ordinary translation templates. Sentence (8) and its model translation (9) show an example of (C3), where the source sentence contains a gapping. The source pattern created from this sentence will be of low applicability. Another criterion is the 'contextual independence.' It is often the case in Japanese-to-English translation that a zero-pronoun in a source sentence is resolved from the context and its translation equivalent appears in the target sentence. A translation template created from such translation may generate a contextually inappropriate translation.", |
| "cite_spans": [ |
| { |
| "start": 120, |
| "end": 123, |
| "text": "(8)", |
| "ref_id": "BIBREF8" |
| }, |
| { |
| "start": 150, |
| "end": 153, |
| "text": "(9)", |
| "ref_id": "BIBREF9" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Criteria for Using Translation Templates", |
| "sec_num": "3." |
| }, |
| { |
| "text": "Note that these criteria are not absolute; templates which do not meet these criteria should be used if they lead to correct translation of other sentences.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Criteria for Using Translation Templates", |
| "sec_num": "3." |
| }, |
| { |
| "text": "A statistical method could be introduced to objectively determine the appropriateness.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Criteria for Using Translation Templates", |
| "sec_num": "3." |
| }, |
| { |
| "text": "This section briefly describes customizing functions which have been adopted ill our MT system [3] [5].", |
| "cite_spans": [ |
| { |
| "start": 95, |
| "end": 98, |
| "text": "[3]", |
| "ref_id": "BIBREF2" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "4, Conventional Customizing Functions", |
| "sec_num": null |
| }, |
| { |
| "text": "\u2022 User-defined word dictionary A user-defined word dictionary (or simply a user dictionary) is the basis for improving the quality of MT output.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "4, Conventional Customizing Functions", |
| "sec_num": null |
| }, |
| { |
| "text": "Translation parameters are introduced to give preference or default interpretation in the translation process. In general, all of the processing are based on the system's linguistic knowledge, which is not open to users.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "\u2022 Translation parameters", |
| "sec_num": null |
| }, |
| { |
| "text": "For example, users cannot change the application order of syntactic rules used by the parser. Therefore the system derives the same syntactic tree for a given sentence to generate one particular translation.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "\u2022 Translation parameters", |
| "sec_num": null |
| }, |
| { |
| "text": "Translation parameters enable usm~ to partially control the t,'anslation process.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "\u2022 Translation parameters", |
| "sec_num": null |
| }, |
| { |
| "text": "One of the parameters used in Japanese-to-F.nglish translation treats subjectless sentences, which are common linguistic phenomena ill Japanese. With this parameter, users can specify the sentence type of a target sentence (imperative or declarative) and, if necessary, the voice and translation equivalents for the omitted subject (personal pronouns, \"it\" or a userdefined string).", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "\u2022 Translation parameters", |
| "sec_num": null |
| }, |
| { |
| "text": "For example, sentence (10) is translated into sentences (11) to (15) according to the specified parameter vah,es. \u2022 User-defined rules User-defined rules are used for representing knowledge to determine an appropriate translation equivalent for a source word (or an expression) by referring to its related words. There are three types of user-defined rt, les available: (R1) Rules for verbs (R2) P, ules for functional phrases (R3) Rules for conjunctional phrases Rule (R1) determines a translation equivalent of a verb based on its case fillers. A translation for a functional phrase is determined based on its preceding noun vnd tile verb phrase it modifies, whereas a translation for a conjunctional phrase is based on its preceding verb phrase and the verb phrase it modifies. Additionally, rules (R2) and (R3) can specify where translation eqvivzdents for functional and con junctional phr.'lses are generated.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "\u2022 Translation parameters", |
| "sec_num": null |
| }, |
| { |
| "text": "Sentences (16) to (18) below show a customization example using a user-defined rule for a functional phrase. In sentence (17), which is the initial ot, tput by our system, the functional phrase \"hi doukishite\" is translated into a verb phrase. Contrast this with the customized sentence (18), in which tim phrase is translated into the prepositional phrase \"in synchronisnt with.\" User-defined rules have limitations in that they cannot represent complex structural changes. However, this is intentionally designed to prevent mistranslation possibly caused by adding these structural rules into tile system's knowledge. Alternatively, the proposed framework has been introduced to represent knowledge for more complex trartslation.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "\u2022 Translation parameters", |
| "sec_num": null |
| }, |
| { |
| "text": "To confirm tile effectiveness of translation templates, we analyzed a parallel text, namely a service mant, al on an electronic eqvipment written in Japanese and its English translation, and estimated the improvement in cvstomizability.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Outline of Analysis", |
| "sec_num": "5.1" |
| }, |
| { |
| "text": "The analysis was done as follows:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Outline of Analysis", |
| "sec_num": "5.1" |
| }, |
| { |
| "text": "(i) Translate the sot,rce sentences using the MT system, which is in the default state except that undefined words are registered in tile user dictionary.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Outline of Analysis", |
| "sec_num": "5.1" |
| }, |
| { |
| "text": "(ii) Compare the 'sentence structure' of the MT output in (i) aqd its cmresponding sentence in tile English manual, and find out sentences for customization.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Outline of Analysis", |
| "sec_num": "5.1" |
| }, |
| { |
| "text": "(iii) Categorize the above sentences according to the type of customization needed to translate them into sentences Itaving tile same sentence structures as the model translations.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Outline of Analysis", |
| "sec_num": "5.1" |
| }, |
| { |
| "text": "The 'sentence structure' used for judging tile necessity of customization includes tile following linguistic features: (21) Tim FM unit has two frame memories tlmt can store two different i,nages.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Outline of Analysis", |
| "sec_num": "5.1" |
| }, |
| { |
| "text": "\u2022", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Outline of Analysis", |
| "sec_num": "5.1" |
| }, |
| { |
| "text": "We have analyzed 492 sentences excluding titles and figure captions. The average sentence length was 52 Kanji characters. Table I shows the overall result. Out of 492 sentences, 42% have tile same sentence structures as the model translations, while the remaining 58% ilave different sentence stnmtures and require customization of the system. The latter is further divided into fonr categories according to the type of customization needed to improve the MT output, as shown in Table 2 .", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 122, |
| "end": 129, |
| "text": "Table I", |
| "ref_id": null |
| }, |
| { |
| "start": 479, |
| "end": 487, |
| "text": "Table 2", |
| "ref_id": null |
| } |
| ], |
| "eq_spans": [], |
| "section": "A nalysis Result", |
| "sec_num": "5.2" |
| }, |
| { |
| "text": "By the conventional customizing functions, namely, translation parametens and user-defined rules, 14% are customizable.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "A nalysis Result", |
| "sec_num": "5.2" |
| }, |
| { |
| "text": "In addition, translation telnplates can improve 45%, which suggests that 59% will improve in tolzd. This also means that, t,sing all customizing functions, 76% of the given sentences can be translated as in the Fnglish ,nanual, while only 51%, can be done so t,sing the conventional ft, nctions. Tlmse figures suggest that a translation template is useful to deal with complex translation.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "A nalysis Result", |
| "sec_num": "5.2" |
| }, |
| { |
| "text": "Sentences which cannot be ct, stomized are divided into four categories:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "A nalysis Result", |
| "sec_num": "5.2" |
| }, |
| { |
| "text": "\u2022 Failed application of parameters (20%)", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "A nalysis Result", |
| "sec_num": "5.2" |
| }, |
| { |
| "text": "\u2022 Inadequate syntax for templates (9%)", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "A nalysis Result", |
| "sec_num": "5.2" |
| }, |
| { |
| "text": "\u2022 Inappropriate templates (65%)", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "A nalysis Result", |
| "sec_num": "5.2" |
| }, |
| { |
| "text": "\u2022 Others (6%)", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "A nalysis Result", |
| "sec_num": "5.2" |
| }, |
| { |
| "text": "First, a translation parameter does not work when the condition on its application is not customizable. One example is a translation parameter of sentence types for enumerated items.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "A nalysis Result", |
| "sec_num": "5.2" |
| }, |
| { |
| "text": "If the system can recognize such a specific form, its translation can be customized. Otherwise tile specified parameter is not used.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "A nalysis Result", |
| "sec_num": "5.2" |
| }, |
| { |
| "text": "Second, an extended syntax for translation templates is needed to represent more complex translation. An example is to extend the syntax so that conversion of grammatical categories, such as nominalization of verb phrases, can be specified.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "A nalysis Result", |
| "sec_num": "5.2" |
| }, |
| { |
| "text": "Third, translation templates are not utilized in light of the criteria explained in 3. The statistics of the rejected sentences is as follows.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "A nalysis Result", |
| "sec_num": "5.2" |
| }, |
| { |
| "text": "\u2022 Division or concatenation of sentences (57%)", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "A nalysis Result", |
| "sec_num": "5.2" |
| }, |
| { |
| "text": "\u2022 Resolution of zero-pronouns (24%)", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "A nalysis Result", |
| "sec_num": "5.2" |
| }, |
| { |
| "text": "\u2022 Parenthesis / gapping (11%)", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "A nalysis Result", |
| "sec_num": "5.2" |
| }, |
| { |
| "text": "\u2022 Others (8%)", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "A nalysis Result", |
| "sec_num": "5.2" |
| }, |
| { |
| "text": "A translation template proposed in this paper is more flexible than others due to variables to match with 'verb phrases' and 'clauses.' Basically, a pattern matching approach like the template-based translation has a disadvantage on word order when it is applied to a language that has relatively free word order like Japanese. This problem is partially solved by using these variables because the word order of the constituents of verb phrases and clauses is not fixed.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Discussion \u2022 Flexibility of translation templates", |
| "sec_num": "5.3" |
| }, |
| { |
| "text": "The question about the appropriateness of a translation template is also raised in case of a translation example in Example-based Machine Translation (EBMT). It is easy to measure the system performance, but is difficult to evaluate tile appropriateness of examples based on their amount and the performance. Tbis issue Ires been ignored so far.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Appropriateness of translation templates", |
| "sec_num": null |
| }, |
| { |
| "text": "Our criteria will be the first approacb to this issue.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Appropriateness of translation templates", |
| "sec_num": null |
| }, |
| { |
| "text": "Although every translation can be described using translation templates, some criteria to determine its appropriateness should be provided because without them automatic template learning will soon lead to tile explosion of the template database.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Appropriateness of translation templates", |
| "sec_num": null |
| }, |
| { |
| "text": "In this paper, we have presented a framework for customizing a machine translation system using userdefined translation templates. This enables users to represent bilingual knowledge for complex translation. We have conducted a preliminary analysis to evaluate tile effectiveness of the proposed framework based on a bilingual text.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Conclusion", |
| "sec_num": "6." |
| }, |
| { |
| "text": "Tile result shows that about 60% of mistranslated sentences can be properly translated by combining the proposed framework with the convention:d ct, stomizing functions, while only 14% can be achieved using tbe conventional customizing functions.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Conclusion", |
| "sec_num": "6." |
| }, |
| { |
| "text": "One of our current concerns is to extend translation templates and make them more expressive to deal with more complex translation. The proposed framework does not permit variables in a template to be changed into other grammatical categories. Another concern is to improve the user interface for registering translation templates.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Conclusion", |
| "sec_num": "6." |
| }, |
| { |
| "text": "Through tile analysis of source and target sentences, initial values in the interface will be more acct,rate and need less correction.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Conclusion", |
| "sec_num": "6." |
| } |
| ], |
| "back_matter": [], |
| "bib_entries": { |
| "BIBREF0": { |
| "ref_id": "b0", |
| "title": "Two l,anguages Are More Informative Than One", |
| "authors": [ |
| { |
| "first": "I", |
| "middle": [], |
| "last": "Agan", |
| "suffix": "" |
| }, |
| { |
| "first": "A", |
| "middle": [], |
| "last": "Itai", |
| "suffix": "" |
| }, |
| { |
| "first": "U", |
| "middle": [], |
| "last": "Scbwall", |
| "suffix": "" |
| } |
| ], |
| "year": 1991, |
| "venue": "Proc. cf ACI,-91", |
| "volume": "", |
| "issue": "", |
| "pages": "130--137", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "1)agan, I., Itai, A. and Scbwall, U. : Two l,anguages Are More Informative Than One, Proc. cf ACI,-91, pp. 130-137, 1991.", |
| "links": null |
| }, |
| "BIBREF1": { |
| "ref_id": "b1", |
| "title": "Acquisition of Selectional Patterns", |
| "authors": [ |
| { |
| "first": "R", |
| "middle": [], |
| "last": "Grishman", |
| "suffix": "" |
| }, |
| { |
| "first": "J", |
| "middle": [], |
| "last": "Sterling", |
| "suffix": "" |
| } |
| ], |
| "year": 1992, |
| "venue": "Proc. rf COLING-92", |
| "volume": "", |
| "issue": "", |
| "pages": "658--664", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Grishman, R. and Sterling, J. : Acquisition of Selectional Patterns, Proc. rf COLING-92, pp. 658- 664, 1992.", |
| "links": null |
| }, |
| "BIBREF2": { |
| "ref_id": "b2", |
| "title": "EJ/JE Machine Translation System ASTRANSAC", |
| "authors": [ |
| { |
| "first": "I", |
| "middle": [ |
| "I" |
| ], |
| "last": "Hirakawa", |
| "suffix": "" |
| }, |
| { |
| "first": "I", |
| "middle": [ |
| "I" |
| ], |
| "last": "Nogami", |
| "suffix": "" |
| }, |
| { |
| "first": "S", |
| "middle": [], |
| "last": "Amano", |
| "suffix": "" |
| } |
| ], |
| "year": null, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Hirakawa, II., Nogami, II. and Amano, S. : EJ/JE Machine Translation System ASTRANSAC--", |
| "links": null |
| }, |
| "BIBREF3": { |
| "ref_id": "b3", |
| "title": "Extensions toward Personalization", |
| "authors": [], |
| "year": 1991, |
| "venue": "Proc. of MT SUMMITIII", |
| "volume": "", |
| "issue": "", |
| "pages": "73--80", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Extensions toward Personalization, Proc. of MT SUMMITIII, pp. 73-80, 1991.", |
| "links": null |
| }, |
| "BIBREF4": { |
| "ref_id": "b4", |
| "title": "Learning Tr:inslation Templates from Bilingual Text", |
| "authors": [ |
| { |
| "first": "It", |
| "middle": [], |
| "last": "Kaji", |
| "suffix": "" |
| }, |
| { |
| "first": "Y", |
| "middle": [], |
| "last": "Kida", |
| "suffix": "" |
| }, |
| { |
| "first": "Y", |
| "middle": [], |
| "last": "Morimoto", |
| "suffix": "" |
| } |
| ], |
| "year": 1992, |
| "venue": "Proc. rf COLING-92", |
| "volume": "", |
| "issue": "", |
| "pages": "672--678", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Kaji, It., Kida, Y. and Morimoto, Y. : Learning Tr:inslation Templates from Bilingual Text, Proc. rf COLING-92, pp. 672-678, 1992.", |
| "links": null |
| }, |
| "BIBREF5": { |
| "ref_id": "b5", |
| "title": "Customiz:ltion of Machine Translation System with User-defined Rules", |
| "authors": [ |
| { |
| "first": "A", |
| "middle": [], |
| "last": "Kumano", |
| "suffix": "" |
| }, |
| { |
| "first": "S", |
| "middle": [], |
| "last": "Kinoshila", |
| "suffix": "" |
| }, |
| { |
| "first": "I", |
| "middle": [ |
| "I" |
| ], |
| "last": "Llirakawa", |
| "suffix": "" |
| } |
| ], |
| "year": 1993, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Kumano, A., Kinoshila, S. and llirakawa, II. : Customiz:ltion of Machine Translation System with User-defined Rules, Japan Soc. Artif Intell. Technical Report, SIG-SI,UD-9301-6, 1993 (in Japanese).", |
| "links": null |
| }, |
| "BIBREF6": { |
| "ref_id": "b6", |
| "title": "Automatic Acquisition of a l.arge Snbeategorization Dictionary from Corpora", |
| "authors": [ |
| { |
| "first": "C", |
| "middle": [ |
| "D" |
| ], |
| "last": "Manning", |
| "suffix": "" |
| } |
| ], |
| "year": 1993, |
| "venue": "Proc. rfACL-93", |
| "volume": "", |
| "issue": "", |
| "pages": "235--242", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Manning, C. D. : Automatic Acquisition of a l.arge Snbeategorization Dictionary from Corpora, Proc. rfACL-93, pp. 235-242, 1993.", |
| "links": null |
| }, |
| "BIBREF7": { |
| "ref_id": "b7", |
| "title": "Pattern-Based Translation: Context-Free Transducer and Its Applications to Practical NLP", |
| "authors": [ |
| { |
| "first": "H", |
| "middle": [], |
| "last": "Maruyama", |
| "suffix": "" |
| } |
| ], |
| "year": 1993, |
| "venue": "Proc. of Natural Language Processing Pacific Rim Symposium", |
| "volume": "", |
| "issue": "", |
| "pages": "232--237", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Maruyama, H.: Pattern-Based Translation: Context-Free Transducer and Its Applications to Practical NLP, Proc. of Natural Language Processing Pacific Rim Symposium, pp. 232-237, 1993.", |
| "links": null |
| }, |
| "BIBREF8": { |
| "ref_id": "b8", |
| "title": "Extraction of Fixed Patterns from AP Economy News", |
| "authors": [ |
| { |
| "first": "N", |
| "middle": [], |
| "last": "Ljratani", |
| "suffix": "" |
| }, |
| { |
| "first": "N", |
| "middle": [], |
| "last": "Katoh", |
| "suffix": "" |
| }, |
| { |
| "first": "T", |
| "middle": [], |
| "last": "Aizawa", |
| "suffix": "" |
| } |
| ], |
| "year": 1991, |
| "venue": "Proc. of the 42th t4nmml Convention IPS Japan", |
| "volume": "", |
| "issue": "", |
| "pages": "6--10", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "lJratani, N., Katoh, N. and Aizawa, T. : Extraction of Fixed Patterns from AP Economy News, Proc. of the 42th t4nmml Convention IPS Japan, 6E-4, 1991.", |
| "links": null |
| }, |
| "BIBREF9": { |
| "ref_id": "b9", |
| "title": "Lexical Knowledge Acquisition from I~ilingual Corpora", |
| "authors": [ |
| { |
| "first": "T", |
| "middle": [], |
| "last": "Utsuro", |
| "suffix": "" |
| }, |
| { |
| "first": "Y", |
| "middle": [], |
| "last": "Matsumoto", |
| "suffix": "" |
| }, |
| { |
| "first": "M", |
| "middle": [], |
| "last": "Nagao", |
| "suffix": "" |
| } |
| ], |
| "year": 1992, |
| "venue": "Proc. of COLING-92", |
| "volume": "", |
| "issue": "", |
| "pages": "581--587", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Utsuro, T., Matsumoto, Y. and Nagao, M. : Lexical Knowledge Acquisition from I~ilingual Corpora, Proc. of COLING-92, pp. 581-587, 1992.", |
| "links": null |
| }, |
| "BIBREF10": { |
| "ref_id": "b10", |
| "title": "A Method for Extracting Translation Pattern from Translation Examples, l'roc. of TMI-93", |
| "authors": [ |
| { |
| "first": "", |
| "middle": [], |
| "last": "Watanabe", |
| "suffix": "" |
| } |
| ], |
| "year": 1993, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "292--301", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Watanabe, 1t. : A Method for Extracting Translation Pattern from Translation Examples, l'roc. of TMI-93, pp. 292-301,1993.", |
| "links": null |
| } |
| }, |
| "ref_entries": { |
| "FIGREF1": { |
| "type_str": "figure", |
| "uris": null, |
| "text": "kyariaseibun-no jokyo-wa, carrier component-of elimination-TOP T-gata roopasufiruta-niyori okonawareru.", |
| "num": null |
| }, |
| "FIGREF2": { |
| "type_str": "figure", |
| "uris": null, |
| "text": "c_;rt.o ..... --~.,--~,..q,,,i~,,;,.,'~...~'m'\"'\"\" / (wo). (tsukau) (kotoniyori)~ {i'li,l'~. (ga) (teigenshita) Example of a Word Lattice", |
| "num": null |
| }, |
| "FIGREF3": { |
| "type_str": "figure", |
| "uris": null, |
| "text": "The preamplifiers are decentralized for 8 dements in P1 and for 24 elements in P2.", |
| "num": null |
| }, |
| "FIGREF4": { |
| "type_str": "figure", |
| "uris": null, |
| "text": "I0) sono botan-wo oshimasu the button-OllJ press (11) Press tim button. (I 2) Tile button is pressed. (13) I press the button. (14) It presses the bt, tton. (15) It presses the button. (imperative) (passive) (personal pronouns) (\"it\") (\"#\" as user-defined string)", |
| "num": null |
| }, |
| "FIGREF5": { |
| "type_str": "figure", |
| "uris": null, |
| "text": "I6) kono kairo-wa shingmL-ni doukishite this circuit-TOP signal-with synchronize parttxll-wo hass#i-sltrlt pulse-OBJ genehate (17) This cirenit genm,ates a pulse synclu-onizing with a signal. (18) This circuit genenltes a pulse in synchronism with a signal.", |
| "num": null |
| }, |
| "TABREF0": { |
| "type_str": "table", |
| "html": null, |
| "content": "<table><tr><td>The conditions on variables in the 'source pattern.' The 'source condition(SCND)' represents grammatical categories of variables currently in use are noun, noun phrase, number, clause and verb phrase. A string matched with a variable should be parsed as the specified category. The 'target condition(TCND)' represents conditions on variables in tile 'target pattern.' Two types are available: 'attribute' and 'relation.' Attributes specify information on one variable. For example, variables for nouns can be specified as having a 'default article' and a 'default number' to be used if there are no explicit clues to determine the article and the number. Similarly, the form of verb phrases in generation can be specified as 'to-infinitive' or 'gerund.' Relations represent the number agreements between a subject and a verb in the target pattern, for example. Variables may appear only in tile source or target pattern. POS TP SCND : $1 .pos=np/$2.pos=np : s (wo tsukau kotoniyori) : use of $I reduced $2 Fig. 1 Template Example (ga teigen shita) POS SP TP SCND TCND POS SP TP SCND TC N D s (no settei wa) (kotoniyori okonaeru) $1 can beset by $2 $1 .pos =np / $2.pos=vp $2.vpgcnd=ING (a) Template with a variable for verb phrase :s (no jokyo wa) (niyori okonawareru )</td></tr></table>", |
| "num": null, |
| "text": "Variables which appear only in the source pattern are used to represent expressions which have relations with another variable but disappear in the target sentence. Variables which appear only in the target pattern are used to represent a target word which is inflected by tile number agreement with tile" |
| }, |
| "TABREF2": { |
| "type_str": "table", |
| "html": null, |
| "content": "<table><tr><td>(b) Syntactic Analysis When a translation template is applicable, the syntactic analyzer plays two roles. First is to analyze part of Source. selltellce { I Tomplate Sea ch I 1 No ~ Yes Translation of [Embedding into TP 1 Output Fig. 3 Translation Process Variables Conventional MT</td></tr></table>", |
| "num": null, |
| "text": "the word sequeuce which should be matched with variables of the template. Words in the word sequence, except ['or template constants, should be parsed as syntactic categories specified in each" |
| } |
| } |
| } |
| } |
