| { | |
| "paper_id": "1987", | |
| "header": { | |
| "generated_with": "S2ORC 1.0.0", | |
| "date_generated": "2023-01-19T03:20:54.701507Z" | |
| }, | |
| "title": "The Current Stage of The Mu-Project", | |
| "authors": [ | |
| { | |
| "first": "Jun-Ichi", | |
| "middle": [], | |
| "last": "Tsujii", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "Kyoto Univ", | |
| "location": { | |
| "settlement": "Kyoto", | |
| "country": "Japan" | |
| } | |
| }, | |
| "email": "" | |
| } | |
| ], | |
| "year": "", | |
| "venue": null, | |
| "identifiers": {}, | |
| "abstract": "", | |
| "pdf_parse": { | |
| "paper_id": "1987", | |
| "_pdf_hash": "", | |
| "abstract": [], | |
| "body_text": [ | |
| { | |
| "text": "The Japanese government MT project (the MU project), which was a four year project supported by the special coordination fund of STA (Science and Technology Agency), was completed at the end of March, 1986. The main objective of the project was to develop a proto-type system which would demonstrate the feasibility of MT systems for certain restricted subject domains and document types. We considered that this initial objective of the project was well fulfilled so that we started another four year project (the MU-II project) from April, 1986. The new project aims to develop & system which will be used for daily translation services in JICST (Japan Information Center for Science and Technology) from 1990.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Introduction", | |
| "sec_num": "1" | |
| }, | |
| { | |
| "text": "Because the technical details of the MU-project have been given in the other papers (1) (2) (3) (4) (5), we will report overall results of the Mu Project and brief introduction of the outline of the new project.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Introduction", | |
| "sec_num": "1" | |
| }, | |
| { | |
| "text": "The following gives the rough sketch of the current MU systems.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Outline of the Mu Systems", | |
| "sec_num": "2" | |
| }, | |
| { | |
| "text": "Neutral Dictionary Heuristically Guided Processing 3. Language Pairs: Japanese to English, English to Japanese Two systems have been developed, both of which uses same software systems", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Design Principle: lexicon Driven Processing", | |
| "sec_num": "2." | |
| }, | |
| { | |
| "text": "The subject fields will be extended in the other scientific and technological fields in the new project.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Subject Fields: Electrical Engineering", | |
| "sec_num": "4." | |
| }, | |
| { | |
| "text": "5. Document Type: Abstracts of Scientific and Technological Papers 6. Implementation Language:", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Subject Fields: Electrical Engineering", | |
| "sec_num": "4." | |
| }, | |
| { | |
| "text": "\u2022 Core translation system:", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Subject Fields: Electrical Engineering", | |
| "sec_num": "4." | |
| }, | |
| { | |
| "text": "-Uti-LISP (University of Tokyo Interactive Lisp) with some augmentations for treating Chinese characters, etc.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Subject Fields: Electrical Engineering", | |
| "sec_num": "4." | |
| }, | |
| { | |
| "text": "-Zeta-LISP (Symbolics 3600 series) The main systems were developed first by Uti-LISP on FACOM 382, and then, they were converted into the Zeta-LISP version.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Subject Fields: Electrical Engineering", | |
| "sec_num": "4." | |
| }, | |
| { | |
| "text": "\u2022 Support Systems including interactive editing systems for texts and dictionaries, management of the dictionary data bases, etc: PLI 7. Grammar Writing Language: A special language called GRADE was designed and implemented by LISP for writing grammars in the project. GRADE facilitates flexible tree structure matching, backtracking, etc.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Subject Fields: Electrical Engineering", | |
| "sec_num": "4." | |
| }, | |
| { | |
| "text": "8. Size of Dictionaries: Verbs, Adjective, Adverbs, etc. = 5,000 Nouns (incl. terminology) = 70,000", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Subject Fields: Electrical Engineering", | |
| "sec_num": "4." | |
| }, | |
| { | |
| "text": "We just show the size of the English-Japanese system. The system of the other direction is of almost same size.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "The Transition Process and the Size of the Grammars:", | |
| "sec_num": "3" | |
| }, | |
| { | |
| "text": "\u2022 Total Number of SG = 500", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "English Analysis Grammar:", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "\u2022 Total Number of RR = 2298", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "English Analysis Grammar:", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "The following are the main steps of analysis which are performed in the order shown.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "English Analysis Grammar:", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "Step 1 Post-Morphological Analysis including the processing of number expressions, parenthetical expressions, compound words, etc. (SG = 52, RR = 138)", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "English Analysis Grammar:", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "Step 2 Treatment of Question and Imperative sentences (SG = 8, RR = 20)", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "English Analysis Grammar:", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "Step 3 Disambiguation of parts of speech based on local environment (SG = 68, RR = 374)", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "English Analysis Grammar:", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "The decision of parts of speech at this step is tentative. The rules in this step are heuristic ones only based on local environments around the ambiguous words, which are expected to give correct interpretations for most cases.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "English Analysis Grammar:", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "However, because the decisions given at this stage are sometimes wrong, the following stages revise the decisions (see below).", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "English Analysis Grammar:", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "Step 4 Adverbial Phrases (SG = 4, RR = 43)", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "English Analysis Grammar:", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "Step 5 Simple Noun Phrases (Phrases whose heads are nouns and which consist of prenominal modifiers and the heads) (SG = 11, RR = 27)", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "English Analysis Grammar:", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "Note that we can find in abstracts many examples of long sequences of words including nouns, ed/ing form verbs, adjectives, adverbs, etc. which form simple noun phrases.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "English Analysis Grammar:", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "Step 6 Verb Group (main verbs and auxiliaries) (SG = 3, RR = 27)", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "English Analysis Grammar:", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "Step 7 Determination of the Scopes of Coordinated Verb Groups and Clauses (SG = 56, RR = 145) It is often quite difficult to determine whether the coordination is a clausal one or phrasal one. So, this SG calls recursively the SG for phrasal coordination to check the possibility The two SGs call each other in coordinated ways.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "English Analysis Grammar:", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "Step 8 Determination of the Scopes of Coordinated Phrases (SG = 29, RR = 199)", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "English Analysis Grammar:", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "Step 9 Appositional and Inserted Phrases/Clauses (SG = 46, RR = 1159)", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "English Analysis Grammar:", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "Step This", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "English Analysis Grammar:", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "Step first identifies several cue words (or expressions) such as WH words, subordinate conjunctions, 'that', -ing/-ed verbs, to-infinitives, etc. each of which indicates existence of clausal constituents of certain types.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "English Analysis Grammar:", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "Wrong decision makings in the preceding steps are to be detected and revised by rules referring to these global structures.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "English Analysis Grammar:", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "Step 11 Disambiguation of Parts of Speech based on Local Environment-2 (SG = 5, RR = 13)", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "English Analysis Grammar:", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "Step 12 Case Pattern Matching (SG = 96, RR = 240)", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "English Analysis Grammar:", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "The scopes of clausal constituents given in the global structures are tentative. The exact scopes are to be determined at the same time when the internal structures of individual clauses are determined.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "English Analysis Grammar:", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "Step 13 Treatment of Circumstantial Elements (SG = 44, RR = 517)", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "English Analysis Grammar:", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "Semantic interpretation of circumstantial elements is crucial in getting natural translation results.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "English Analysis Grammar:", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "The interpretations are performed by invoking lexical rules defined for individual prepositions and group prepositions.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "English Analysis Grammar:", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "Step 14 Conversion from phrase Structures to Dependency Structures (SG = 18, RR = 78)", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "English Analysis Grammar:", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "Step 15 Semantic Interpretation including deep tense, aspect interpretation (SG = 14, RR = 38)", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "English Analysis Grammar:", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "Step 16 Checking Rules of final analysis results (SG = 5, RR = 34)", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "English Analysis Grammar:", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "Special care is taken to adjust structural differences of the two languages, English and Japanese, which belong to quite different language families. In order to support global structural changes during the transfer, two separate phases, Pre-Transfer Loop and Post-Transfer Loop, are provided, which are executed before and after the main transfer phase respectively. The detailed construction of the transfer phase and some discussions are given in (5) . ", | |
| "cite_spans": [ | |
| { | |
| "start": 450, | |
| "end": 453, | |
| "text": "(5)", | |
| "ref_id": "BIBREF4" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Transfer Grammar:", | |
| "sec_num": "3.2" | |
| }, | |
| { | |
| "text": "Detailed evaluation criteria of translation results and results of evaluation are given in (1) . About 80 are evaluated as \"understandable\" by native speakers, which fulfils the initial objectives of the project.", | |
| "cite_spans": [ | |
| { | |
| "start": 91, | |
| "end": 94, | |
| "text": "(1)", | |
| "ref_id": "BIBREF0" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Evaluation Results:", | |
| "sec_num": "3.4" | |
| }, | |
| { | |
| "text": "The results of the MU project show economical and technical feasibility of MT in the field of 'translation of abstracts'. The translation results are not so natural as you can expect of those produced by professional (highly qualified) human translators, but they are good enough to get rough ideas about the contents of the papers. The MU systems as they are, however, have several defects as systems for practical services, because they were initially designed as research and development prototypes. In order to remove certain defects, a new project, the MU-II project, was started. The main objectives of the MU-II are as follows;", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Brief Outline of the Mu-II Project", | |
| "sec_num": "4" | |
| }, | |
| { | |
| "text": "The new version of GRADE is to be implemented by C in order to improve the processing speed and memory requirements. Some facilities the current version of GRADE provides, which are too flexible and too powerful and which have been rarely used by grammar writers, will be removed. This will also improve the efficiency of processing.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Improvement of Processing Speed and Reduction of Memory Requirement", | |
| "sec_num": "4.1" | |
| }, | |
| { | |
| "text": "The current dictionaries contain about 80,000 lexical items, but they cover only the ordinary words of Japanese, and the vocabularies of certain restricted areas of electrical engineering fields. The dictionaries of the new project are planned to gather 300,000 lexical items including additional ordinary words, terminological expressions in several technological fields.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Augmentation of Dictionaries", | |
| "sec_num": "4.2" | |
| }, | |
| { | |
| "text": "The current version of grammars in the MU systems is designed to translate input sentences whatever strange the inputs are. In fact, we can find many sentences which are difficult even for ordinary native speakers (not the specialists of the fields the input abstracts deal with) to understand. Not a few SGs and RRs are prepared simply for treating very idiosyncratic constructions which even native speakers feel difficulties in understanding. Existence of such SGs and RRs make the systems clumsy and difficult to maintain. In the new systems, such (almost ungrammatical) sentences are supposed not to be in the text. We will give some trainings to the people who prepare abstracts. However, we keep the principle that MT systems such as systems for translation of abstracts should not rely heavily on pre-editing.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Restricted Language", | |
| "sec_num": "4.3" | |
| }, | |
| { | |
| "text": "Various human interaction facilities are to be integrated around the core systems, including translation oriented text editors, effective utilization of domain specific dictionaries and user oriented dictionaries, maintenance tools for dictionary data bases, etc.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Other Facilities in the New Systems", | |
| "sec_num": "4.4" | |
| }, | |
| { | |
| "text": "The evaluation of the MU project is currently quite positive for further research and development of MT systems. The experiments made with the MU systems show that the basic design principles, especially lexicon driven processings, neutral dictionaries, and heuristically guided processings are extremely effective in the actual development of large scale MT systems. The MU-II project will concentrate on making the current systems practically usable by refining the present softwares and grammars.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Conclusion", | |
| "sec_num": "5" | |
| }, | |
| { | |
| "text": "However, we also noticed through experiences that there are still many problems, theoretically or practically, which are obviously beyond our current framework. Among others, appropriate target lexical selections, processing of elliptical expressions frequent in Japanese, etc. require a lot further investigations on possible frameworks of semantic and contextual processings in MT. The discussions about integrations of such components is given in (6) .", | |
| "cite_spans": [ | |
| { | |
| "start": 450, | |
| "end": 453, | |
| "text": "(6)", | |
| "ref_id": "BIBREF5" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Conclusion", | |
| "sec_num": "5" | |
| } | |
| ], | |
| "back_matter": [], | |
| "bib_entries": { | |
| "BIBREF0": { | |
| "ref_id": "b0", | |
| "title": "The Japanese Government Project of Machine Translation", | |
| "authors": [ | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Nagao", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "J", | |
| "middle": [], | |
| "last": "Tsujii", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "J", | |
| "middle": [], | |
| "last": "Nakamura", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1985, | |
| "venue": "Journal of Computational Linguistics", | |
| "volume": "11", | |
| "issue": "2-3", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Nagao, M., Tsujii, J. and Nakamura, J.: The Japanese Government Project of Machine Translation, Journal of Computational Linguistics, Vol-11, No.2-3, 1985.", | |
| "links": null | |
| }, | |
| "BIBREF1": { | |
| "ref_id": "b1", | |
| "title": "Dealing with Incompleteness of Linguistic Knowledge in Language Translation", | |
| "authors": [ | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Nagao", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1984, | |
| "venue": "Proc. of Coling84", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Nagao, M., et al.: Dealing with Incompleteness of Linguistic Knowledge in Language Trans- lation, in Proc. of Coling84, 1984.", | |
| "links": null | |
| }, | |
| "BIBREF2": { | |
| "ref_id": "b2", | |
| "title": "Analysis Grammar of Japanese in the MU Machine Translation Systems", | |
| "authors": [ | |
| { | |
| "first": "J", | |
| "middle": [], | |
| "last": "Tsujii", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1984, | |
| "venue": "Proc. of Coling", | |
| "volume": "84", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Tsujii, J., et al.: Analysis Grammar of Japanese in the MU Machine Translation Systems, in Proc. of Coling 84, 1984.", | |
| "links": null | |
| }, | |
| "BIBREF3": { | |
| "ref_id": "b3", | |
| "title": "Grammar Writing System (GRADE) of MU-Machine Translation Project and its Characteristics", | |
| "authors": [ | |
| { | |
| "first": "J", | |
| "middle": [], | |
| "last": "Nakamura", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1984, | |
| "venue": "Proc. of Coling", | |
| "volume": "84", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Nakamura, J., et al.: Grammar Writing System (GRADE) of MU-Machine Translation Project and its Characteristics, in Proc. of Coling 84, 1984.", | |
| "links": null | |
| }, | |
| "BIBREF4": { | |
| "ref_id": "b4", | |
| "title": "Transfer Phase of a Machine Translation system", | |
| "authors": [ | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Nagao", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1986, | |
| "venue": "Proc. of Coling", | |
| "volume": "86", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Nagao, M., et al.: Transfer Phase of a Machine Translation system, in Proc. of Coling 86, 1986.", | |
| "links": null | |
| }, | |
| "BIBREF5": { | |
| "ref_id": "b5", | |
| "title": "Future Direction of Machine Translation", | |
| "authors": [ | |
| { | |
| "first": "J", | |
| "middle": [], | |
| "last": "Tsujii", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1986, | |
| "venue": "Proc. of Coling", | |
| "volume": "86", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Tsujii, J.: Future Direction of Machine Translation, in Proc. of Coling 86, 1986.", | |
| "links": null | |
| } | |
| }, | |
| "ref_entries": { | |
| "FIGREF0": { | |
| "type_str": "figure", | |
| "text": "Pre-Transfer Loop: Japanese oriented structures are transformed into more neutral ones (SG = 9, RR = 23) \u2022 Processing of structures governed by predicates (SG = 33, RR = 577) \u2022 Processing of structures governed by nominal concepts (SG = 10, RR = 192) \u2022 Processing of adverbial phrases (SG = 2, RR = 121) \u2022 Processing of Prenominal Modifiers (SG = 21, RR = 180) \u2022 Post-Transfer Loop : The neutral structures are Transformed into English-oriented structures (SG = 11, RR = 46) Conversion of parts of speech (SG = 22, RR = 47) \u2022 Selection of Appropriate Case Particles (SG = 16, RR = 103) \u2022 Realization of Voice Information (SG = 28, RR = 34) \u2022 Determination of Phrase Ordering (SG = 33, RR = 44) \u2022 Insertion of Punctuation Marks, etc. (SG = 1, RR = 6) \u2022 Building of Phrase Structure Trees (SG = 79, RR = 160) \u2022 Calculation of inflectional forms for each word and adjustment of structures for morphological generation (SG = 15, RR = 189)", | |
| "num": null, | |
| "uris": null | |
| }, | |
| "TABREF0": { | |
| "content": "<table/>", | |
| "type_str": "table", | |
| "text": "Determination of Global Structures of Sentences and Revision of Tentative Decisions in Preceding Steps (SG = 43, RR = 153) All the processing up to this step analyze sentences from the bottom to the top. On the other hand, this stage tries to hypothesize the global structures of input sentences in a topdown manner. That is, this step constructs tree structures which show the relationships of clausal constituents in input sentences, while the internal structures of individual clausal constituents have not been determined yet.", | |
| "num": null, | |
| "html": null | |
| } | |
| } | |
| } | |
| } |