CN1100825A - Intelligent translating system - Google Patents
Intelligent translating system Download PDFInfo
- Publication number
- CN1100825A CN1100825A CN 93118098 CN93118098A CN1100825A CN 1100825 A CN1100825 A CN 1100825A CN 93118098 CN93118098 CN 93118098 CN 93118098 A CN93118098 A CN 93118098A CN 1100825 A CN1100825 A CN 1100825A
- Authority
- CN
- China
- Prior art keywords
- chinese
- dictionary
- machine translation
- translation
- knowledge base
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Machine Translation (AREA)
Abstract
There is provided an English to Chinese computer translation system, its hardware includes IBM series microcomputer and compatible machine with 1 M internal storage, 15 M free hard disc space, translation circuitboard and printer, it uses grammar with logic function to analyze sentence structure, grammar and generation of Chinese language, it can translate correctly and can translate complex sentence patterns, the present system can translate English received from fax, image scanner or Reyboard into Chinese.
Description
The present invention relates to a kind of machine translation system that English can be translated into automatically Chinese.
Twentieth century is the information age, and various information must be passed through language performance, and must carry out message exchange between the various countries, thereby the issues for translation of language becomes more and more important.
Because the development of electronic computer technology adds the development of speech recognition and voice compound technology, make the language translation machine device to realize in recent years.But the tempo in this field is slow, only make certain gains in the translation of the simple sentence pattern in indivedual limitation fields, and the translation of complicated sentence pattern is also very unsatisfactory.
The language translation is individual extremely complicated problems, and the polysemy of morpheme adds the ambiguousness of the syntactic structure of sentence, makes that the identification of language is difficult.Existing translation system does not all have comprehensive solution to these problems, has such or such shortcoming, thereby can not satisfy practical requirement, especially can not translate complicated sentence pattern.
Chinese patent CN 91108789.3 can only be translated into other target language with Chinese sentence, and should not estimate the translation of complicated sentence pattern.
Chinese patent CN 87106964A can only translate into another language from a kind of language with phrase.This translation system is not considered the grammatical relation between speech and the speech.
What Chinese patent CN 87104551 narrated is the paginal translation of Chinese and Japanese.Have than big difference to the identification of the Chinese and the syntactic structure of Japanese and to the identification of the syntactic structure of English.
Ying-Han the machine translation system that the purpose of this invention is to provide a kind of practicality.The English original text that it can accept to import with following four kinds of modes:
Directly accept the English original document image that fax is sent, convert thereof into the English text file of ASC II sign indicating number.
Accept the input of English original text from scanner, convert thereof into the English text file of ASC II sign indicating number.
English source text file from the keyboard typing.
Acceptance is from the english sentence of keyboard input.
This system can translate complicated sentence pattern.
The present invention finishes by following technical proposals:
After the English original text input, with text pretreatment system (PRELEX.EXE) English original text being resolved into sentence one by one earlier, is that unit translates sentence by sentence again with the sentence.Translation to sentence comprises that lexical analysis (LEX.EXE), grammatical analysis (DCG.EXE) and Chinese generate the process of (TRA.EXE) three complexity.After the translate end, can use document file management system provided by the invention (EDIT.EXE and DISP.EXE) that the result of translation is carried out post-editor, also can directly print or show.Can carry out Chinese and English simple sentence contrast, the contrast of entire article and the independent demonstration and the printout of translation.The present invention adopts the advanced functional unification grammar towards logic to carry out lexical analysis, grammatical analysis and Chinese first and generates, can handle very complicated sentence pattern, solved morphology, syntax and semantics ambiguity that many translation systems can't resolve, thereby the translation result accuracy is very high.
The block diagram of total system as shown in Figure 1.System gives system, lexical analysis subsystem, grammatical analysis subsystem, Chinese by the text pre-service and generates subsystem, document management subsystem, dictionary and linguistic context management tool, vocabulary function synthesized dictionary and linguistic context knowledge base number etc. and partly form.
The core of native system has adopted the functional unification method (Logic Oriented Functional Unification Grammar is called for short LOFUG) towards logic to carry out lexical analysis, grammatical analysis and Chinese and has generated.
Functional unification grammar (FUG) is the functional structure from morpheme, by functional unification calculation, determines the constituent structure of sentence.The functional structure of so-called morpheme is meant the restriction when various grammer composition done in part of speech, morpheme, the information such as constraint condition of getting justice when morpheme has polysemy.The constituent structure of sentence is the syntax tree of corresponding sentence then.The functional unification calculation is the functional structure according to morpheme, concerns the common factor of asking the corresponding function characteristic between the morpheme that satisfies the constituent structure constraint in part of speech, determines constituent structure according to the situation of occuring simultaneously.Give one example below the functional unification calculation is described.We know, in English, each both can do determiner (as each person), can do adverbial word (as The tickets are ￡ l each) again, but how to distinguish this two kinds of usages? in sentence Each PARTY must obey it, each is as adverbial word, be positioned at before the sentence, as if can be used as the adverbial modifier (as Now he might be in home.), but we give attribute of each, tell syntactic analyser, each can not be a preceding adverbial modifier, then grammatical analysis during the adverbial modifier, just can be done a preceding adverbial modifier possibility according to the negative each of this attribute, thereby obtain correct analysis result before getting sentence.
LOFUG then utilizes the item of logical type programming language PROLOG to define the various functional structures of morpheme and the various constituent structures of sentence, utilize " unification " and the set operation function of PROLOG, realization comprises the broad sense unification calculation of vocabulary functional unification calculation, thus with functional unification method (FUG) and logical type context-free grammar (DCG) in conjunction with as a whole.LOFUG is identical with the FUG principle, and what only use is the mechanism that program language PROLOG is provided.Can be referring to the example of explanation " linguistic context knowledge base " Shi Suoju.(be expressed as translate([" on " among the PROLOG in translation " on RMB basis ", " RMB ", " basis "]) time, with translate([" on " .X, " basis "]) unification, and remove to substitute X with " RMB ", in context bank, search for is_money(" RMB " then), search is_money(" RMB "), then unify successfully, the Chinese meaning of finding " RMB " again is " Renminbi ", therefore " on RMB basis " is translated as " by RMB ".
In illustrative view of functional configuration of the present invention shown in Figure 1, core " lexical analysis system ", " grammatical analysis system " have all adopted LOFUG.
Narrate the function and the structure of each ingredient below respectively.
1. text pretreatment system
The function of text pretreatment system is that english article is resolved into sentence, with the sentence be then unit submit to translation system (lexical analysis system, grammatical analysis system and Chinese generation system) go the translation.The foundation that article is resolved into sentence is that each may be as the punctuation mark (as fullstop, question mark, exclamation mark etc.) of sentence end mark, and the situation of several morphemes of the front and back of these symbols in article or punctuation mark.
For example, at following english sentence:
Among the He studied in U.S.A.in 1989.He came home in 1990, all there is fullstop " U ", " S ", " A " back, but is not the end mark of sentence.The fullstop of " 1989 " back is only the end mark of sentence.
2. vocabulary function synthesized dictionary
We have set up the corpus in an economy and trade field, utilize this corpus to set up the vocabulary function synthesized dictionary of economy and trade domain-specific then, and it contains 50,000 entries altogether.Each entry has not only comprised the meaning of a word, the part of speech of speech, has also comprised morphology, grammatical function characteristic, professional knowledge characteristic, various idiom usages of speech etc.So the data volume of each vocabulary on average is equivalent to more than 5 times of general machine translation system dictionary entry.
The file organization form of dictionary is B+ tree construction (briefly, the B+ tree is a kind of data structure that can be used to realize mass data is carried out the high-level efficiency ordering), can carry out sequential search and stochastic searching efficiently.Key length is 15 in the B+ tree, surpasses 15 prefix for length, confirms by finding out the overall length prefix that leaves in the dictionary.
3. linguistic context knowledge base
Provide knowledge, to guarantee high-quality morphology, grammatical analysis result at the mutual restriction relation of the relevant various features of semanteme in economy and trade field.Currency symbol RMB(Renminbi for example), the Y(dollar) etc., in specific word environment, can not only consider currency symbol is literal translated out, also to consider the influence of currency symbol to the translation of other word of front and back.
Represent that with PROLOG this linguistic context knowledge is as follows:
Translate([" on ", X, " basis "]), [" usefulness ", X_tran]):
is?Money（X），
dic（X，X_tran）.
is?Money（RMB）
Dic(RMB, " Renminbi ")
With X RMB substitution, the translation result that then draws on RMB basis is " by RMB ".
The file organization form of linguistic context knowledge base is with vocabulary function synthesized dictionary.
4. dictionary and context bank management tool
The file organization form of vocabulary function synthesized dictionary and linguistic context knowledge base is the same, and their unifications are managed by a management tool, are called dictionary and linguistic context management tool.Entry in the dictionary and a knowledge in the context bank all are one " items " from structure.Connect by doubly linked list between each.A B+ tree is formed in each key assignments and the address in doubly linked list thereof.
For fear of since the chain that insertion, modification and deletion caused reorganization and the huge system overhead of B+ tree reorganization, when carrying out aforesaid operations, do not change any other storage address.Carry out item when inserting, distribute a free space earlier, the item that insert is write this space, and this block space is linked to the last of chain, key assignments that then will this item and address are inserted into B+ and set in the index.During a deletion action of carrying out, former and later two of deleted directly are tied, thereby will remove among this Xiang Conglian, and in B+ tree index, corresponding node is deleted, do not carry out the reorganization of chain.The retouching operation of item is the insertion of item and the combination that work is rubbed in deletion.After through repeatedly modification and deletion action, understand more residual free space fragments in the storehouse like this.Can use the arrangement in storehouse to touch piece and remove these free space fragments.When the storehouse was very big, housekeeping operation was very time taking.
5. lexical analysis system
Morphology bulk-breaking system utilizes vocabulary function synthesized dictionary and linguistic context knowledge base, analyze meaning, lexical function characteristic and the linguistic context knowledge characteristics of each speech in the sentence, thereby summarize vocabulary functional characteristic (Funtional Sructure) expression formula of each english sentence.See embodiment for details.
Lexical analysis system first-selection resolves into one by one speech with a sentence, is that key assignments is searched the B+ tree with each speech again, therefrom finds the storage address of corresponding entry, reads corresponding entry according to this address from the chained list of storehouse again.What each entry in the dictionary was deposited is total functional structure of a morpheme, the lexical analysis system decomposes this total structure, filter mire worthless functional characteristic under concrete lexical characteristics and context environmental, and other functional characteristics are classified comprehensively, form the comprehensive function characteristic.
" hoped " among He hoped so for example, the hope in the dictionary has noun and two kinds of parts of speech of verb, but in this lexical characteristics, hope occurs with the past tense form, therefore can not make noun.The lexical analysis system can filter the noun part of speech of hope according to this environment, thereby has reduced the ambiguity of grammatical analysis process.
And for example in He is interested in something, intereted both can adjective, can be used as the past participle of verb interest again.But in this specific context environmental (there is auxiliary verb is the front), add that interested itself has the adjective part of speech, can assert that it should not get the verb part of speech.So just can filter the verb part of speech of interested, reduce the ambiguity of grammatical analysis.
Represent this section linguistic context knowledge following (part that finishes is represented to explain) with PROLOG from % to the delegation in PROLOG:
% restriction verb part of speech
Li?mit?Verb（Word1，Word2）：-
IsBe(Word1), %Word1 is " be "
Is Adj(Word2), %Word2 is an adjective
Get Rid OF Verb(Word2) .% removes the verb part of speech of Word2
Situation for polysemy, the lexical analysis system is according to functional structure total in the dictionary, decomposite the contextual information of each meaning of a word, thereby provide the functional characteristic expression formula of each meaning of a word and contextual information thereof, and each class contextual information of each meaning of a word asked union, form the comprehensive function characteristic expression formula of this morpheme.
6. grammatical analysis system
The grammatical analysis system utilizes the result of grammer, semantic analysis, by the reasoning from logic calculation of complexity, summarizes the constituent structure expression formula (referring to embodiment) of each english sentence.The reasoning from logic of the complexity here calculation comprises coupling, unification, recurrence, recall and their compound calculation.The content that these calculations are reflected also comprises some contents of linguistic psychology except the syntactic structure that comprises various complicated sentence patterns.
Psychological method is one of important method of artificial intelligence technology, also is the core methed of the FUG syntax.For example, the people when understanding language, can be constantly according to current acquired information to the grammer, the semantic phenomena that occur are made a prediction, and the notion of former formation is restricted.In our analytical algorithm, then utilize finite automaton theory with this prediction and the formalization in addition of restriction process, the former is in order to the scope of grammatical analysis after dwindling, and the latter is then in order to revise former analysis result, and Here it is has introduced the language analysis technology of psychological method.This technology is playing an important role aspect the solution polysemy.
In the language analysis process, except utilizing conventional deduction (Deduction), also made full use of the inverting reasoning.The so-called deduction is meant according to some known prerequisites, goes to prove some conclusion establishment.So-called inverting reasoning then refers to go counter those prerequisites that push away to set up according to some known conclusion (known phenomena), so it is the inverse operation of deducting.
The constituent structure expression formula of the sentence that the grammatical analysis system is produced is a syntactic structure tree of sentence, and it can be a multiple recursive structure.For example an adverbial clause can be with in a sentence, and can be with an attributive clause in the adverbial clause.The sentence structure of a multiple recurrence that Here it is.Owing to introduce the sentence structure of multiple recurrence, make the grammatical analysis system can analyze very complicated sentence.
7. Chinese generation system
The Chinese generation system utilizes the result of morphology, meaning of a word analytic system and grammatical analysis system, and appropriate Chinese meaning got in each morpheme, and according to the knowledge of English-Chinese sentence pattern contrast relationship, generates Chinese translation.Some English sentence when generating Chinese, need be readjusted its method for expressing, to meet the Chinese custom.See embodiment for details.
Below provide one embodiment of the present of invention:
The software inventory of system and hardware configuration
The hardware configuration of native system comprises:
IBM series microcomputer or its compatible
The above internal memory of IM (comprising the 640K conventional memory)
The above free hard drive space of 15M
The machine translation card that native system provides (comprising ROM and encipherment protection circuit), be equipped with three executable programs among the ROM:
The LEX.EXE-lexical analyzer
The DCG.EXE-syntactic analyser
TRA.EXE-Chinese generator program
Printer (do not print then and can not be equipped with this)
The software arrangements of native system is:
MS-DOS(PC-DOS) operating system
Hanzi system
6 executive routines that are placed on the hard disk that native system provides
The MAIN.EXE-primary control program
DISP.EXE-text display program
The EDIT.EXE-text editor
The HELP.EXE-program that shows help information
PRELEX.EXE-text preprocessor
DIC.EXE-dictionary and knowledge base management instrument
Dictionary and context bank DIC.EXT(leave on the hard disk)
When carrying out mechanical translation, primary control program reads in lexical analysis, grammatical analysis and Chinese generator program successively from ROM, operation after the deciphering.Hardware configuration as shown in Figure 2.
Operation primary control program MAIN.EXE can start C﹠amp on the PC that hardware and software possesses; The T machine translation system for example, supposes that the program that MAIN.EXE etc. is contained on the hard disk all is placed on C:/C﹠amp; Under the T TRAN catalogue, and the work at present catalogue is at C:/C﹠amp; Under the T TRAN, then can import TMAIN by keyboard
Thereby startup native system.Below can select required function, browse as full text translation, simple sentence translation, original text editor, result according to menu prompt, or the like.
Provided a translation example below from the former sentence of English to Chinese translation.
Original text
This?letter?will?be?handed?to?you?by?Mr.?Frank?Chadwick，our?senior?represen?tative，who?will?be?spending?much?of?his?time?during?April?and?May?in?the?Far?East，where?we?are?anxious?to?extend?our?interests.
The lexical analysis result
This part of speech: determiner
The meaning of a word: this
Letter part of speech: noun
The meaning of a word: letter
Will part of speech: auxiliary verb; Modal verb
The meaning of a word: will; Be willing to
The be meaning of a word: be
Handed part of speech: verb
Tense: past tense; Past participle
The meaning of a word: give; Pass
To part of speech: preposition
The meaning of a word: to; Arrive
(all the other omit)
The grammatical analysis result
Subject paragraph: This letter
Determiner This modification noun letter
Noun letter subject
The predicate paragraph:
will?be?handed?to?you?by?Mr.Frank?Chadwick，our?senior?representative，who?will?be?spending?much?of?his?time?during?April?and?May?in?the?Far?East，where?we?are?anxious?to?extend?our?interests
Predicate verb: will be handed tense; The future simple tense voice; Passive voice
Adverbial modifier's paragraph: to you modifies will be handed
Adverbial modifier's paragraph:
by?Mr.Frank?Chadwick，our?senior?representative，who?will?be?spenging?much?of?his?time?during?April?and?May?in?the?Far?East，where?we?are?anxious?to?extend?our?interests
Prepositional phrase is modified will be handed
Adverbial modifier's paragraph subject: Mr.Frank Chadiwick
Appositive: Our senior representative
Modify Mr.Franc Chadiwic
Attributive clause:
Who will be spending much of his time during Aprial and May in the Far East, where we are anxious to extend own interests modifies Mr.Franc Chadiwick
Subject who
Predicate verb will be spending
Object much of his time
Adverbial modifier during April and May modifies will be spending
Adverbial modifier in the Far East, where we are anxious to extend our interests
Prepositional phrase is modified will be spending
Object of preposition the Far East
Attributive clause where we are anxious to extend our interests
Modify Far East
Chinese generates
Who will be spending much of his time during April and May in the Far East, where we are anxious to extend our interests is an attributive clause, modifies representative.According to the Chinese custom, attribute should be placed on by before the modifier.But because this attributive clause is longer, and is in whole sentence final stage again, thus this attributive clause is placed directly in the modificand translation in subordinate clause, to guarantee semantic integrality.
Full sentence translation result is as follows:
This envelope letter will be by Frank Mr. Chadwick, and our senior delegate gives you, and with expensive his time, we thirst for expanding our interests to this Frank Mr. Chadwick April And May in the Far East in the Far East.
Advantage of the present invention
(1) translation accuracy height
We are from three aspect definition accuracy.These three aspects are:
Letter: translation is verily expressed the original text semanteme
Reach: be ignorant of the semanteme that English people also can understand original text from translation fully
Refined: translation is natural and tripping, conforms with Chinese habit expression mode
According to above definition, native system is rate of accuracy reached 95% aspect believing, reaching, the rate of accuracy reached 60-70% aspect fidelity, fluency, elegance.
(2) translation speed is fast
Reach 18,000 words/hour
Below provide the comparative example of some other Ying-Han machine translation system on a native system translation result and the market.
Original text: Girls who like music like boys who like football
The native system translation result: the girl of delight in music likes the boy of football
High upright intelligent machine translation system (Gao Li scientific ﹠ technical corporation) translation result: the girl of delight in music likes boy, and who resembles football
863AE intelligence machine translation system (instant translator) translation result: who resembles music and resembles these boys who resembles football these girls
As can be seen, translation result of the present invention is the most accurate.
Description of drawings:
Fig. 1 is this machine translation system FB(flow block).
Fig. 2 is system hardware structure figure.
Claims (7)
1, a kind of can exactly English be translated into Chinese intelligent machine translation system, it is characterized in that total system as follows step carry out work:
1) English input
2) text gives processing
3) lexical analysis
4) grammatical analysis
5) Chinese generates
6) document management (printing and post-editor)
When carrying out lexical analysis, by the relevant knowledge content of vocabulary function synthesized dictionary and linguistic context knowledge base, input morphology and grammer, and this dictionary and linguistic context knowledge base are managed by a dictionary and linguistic context knowledge base management instrument.
2, intelligent machine translation system as claimed in claim 1 is characterized in that lexical analysis, grammatical analysis and Chinese generate functional unification grammar (LOFUG) technology that all adopts towards logic, use logical type programming language (PROLOG).
3, intelligent machine translation system as claimed in claim 1 is characterized in that system software comprises: MS-DOS(PC-DOS) operating system, hanzi system, vocabulary function synthesized dictionary, linguistic context knowledge base, deposit in six executable programs on the hard disk and deposit in three executable programs among the ROM.
4, intelligent machine translation system as claimed in claim 3 is characterized in that described six executable programs of depositing in hard disk are:
Primary control program, text display program, text editor, the program that shows help information, text preprocessor, dictionary and knowledge base management instrument
5, intelligent machine translation system as claimed in claim 3 is characterized in that three executable programs among the described ROM of depositing in are: lexical analyzer, syntactic analyser and Chinese generator program
6, intelligent machine translation system as claimed in claim 1 is characterized in that dictionary and linguistic context knowledge base management instrument comprise the creation module in storehouse, the insert module of chain, the retrieval module of item, the removing module of item, modified module, the administration module in storehouse and the removing module in storehouse of item.
7, a kind of intelligent machine translation system is characterized in that system hardware comprises IBM series microcomputer or its compatible, the above internal memory of 1M (comprising the 640K conventional memory), the above free hard drive space of 15M, native system machine translation card and printer (do not print and can not be equipped with).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 93118098 CN1100825A (en) | 1993-09-25 | 1993-09-25 | Intelligent translating system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 93118098 CN1100825A (en) | 1993-09-25 | 1993-09-25 | Intelligent translating system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN1100825A true CN1100825A (en) | 1995-03-29 |
Family
ID=4992337
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 93118098 Pending CN1100825A (en) | 1993-09-25 | 1993-09-25 | Intelligent translating system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN1100825A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102609410A (en) * | 2012-04-12 | 2012-07-25 | 传神联合（北京）信息技术有限公司 | Authority file auxiliary writing system and authority file generating method |
| CN104516939A (en) * | 2013-10-04 | 2015-04-15 | 高霆科技股份有限公司 | Parallel hardware search system for constructing artificial intelligent computer |
| CN104679735A (en) * | 2013-11-30 | 2015-06-03 | 赵会军 | Pragmatic machine translation method |
| CN104873413A (en) * | 2015-05-20 | 2015-09-02 | 周淑华 | Easy-to-dry nail polish |
| CN105018555A (en) * | 2015-08-06 | 2015-11-04 | 重庆馗丰食品有限公司 | Preparation method of giant salamander skin collagen peptide |
-
1993
- 1993-09-25 CN CN 93118098 patent/CN1100825A/en active Pending
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102609410A (en) * | 2012-04-12 | 2012-07-25 | 传神联合（北京）信息技术有限公司 | Authority file auxiliary writing system and authority file generating method |
| CN102609410B (en) * | 2012-04-12 | 2014-12-17 | 传神联合（北京）信息技术有限公司 | Authority file auxiliary writing system and authority file generating method |
| CN104516939A (en) * | 2013-10-04 | 2015-04-15 | 高霆科技股份有限公司 | Parallel hardware search system for constructing artificial intelligent computer |
| CN104679735A (en) * | 2013-11-30 | 2015-06-03 | 赵会军 | Pragmatic machine translation method |
| CN104873413A (en) * | 2015-05-20 | 2015-09-02 | 周淑华 | Easy-to-dry nail polish |
| CN105018555A (en) * | 2015-08-06 | 2015-11-04 | 重庆馗丰食品有限公司 | Preparation method of giant salamander skin collagen peptide |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US6278967B1 (en) | Automated system for generating natural language translations that are domain-specific, grammar rule-based, and/or based on part-of-speech analysis | |
| US6760695B1 (en) | Automated natural language processing | |
| US9672206B2 (en) | Apparatus, system and method for application-specific and customizable semantic similarity measurement | |
| US6782384B2 (en) | Method of and system for splitting and/or merging content to facilitate content processing | |
| US7526424B2 (en) | Sentence realization model for a natural language generation system | |
| KR101130444B1 (en) | System for identifying paraphrases using machine translation techniques | |
| US9262409B2 (en) | Translation of a selected text fragment of a screen | |
| US20070011132A1 (en) | Named entity translation | |
| JP2003050797A (en) | Scaleable machine translation system | |
| WO2002039318A1 (en) | User alterable weighting of translations | |
| Onyenwe et al. | A Basic Language Resource Kit Implementation for the Igbo NLP Project | |
| Starko et al. | VESUM: A Large Morphological Dictionary of Ukrainian As a Dynamic Tool. | |
| WO1997040452A1 (en) | Automated natural language translation | |
| CN1100825A (en) | Intelligent translating system | |
| Bangalore | Complexity of lexical descriptions and its relevance to partial parsing | |
| Ismail et al. | ALIF editor for generating Arabic normalized lexicons | |
| Yeshambel et al. | Evaluation of corpora, resources and tools for Amharic information retrieval | |
| JP2008077512A (en) | Document analysis device, document analysis method and computer program | |
| WO1997048058A9 (en) | Automated translation of annotated text | |
| WO1997048058A1 (en) | Automated translation of annotated text | |
| JP2005025555A (en) | Thesaurus construction system, thesaurus construction method, program for executing the method, and storage medium with the program stored thereon | |
| Fransen | Past, present and future: Computational approaches to mapping historical Irish cognate verb forms | |
| JP4635585B2 (en) | Question answering system, question answering method, and question answering program | |
| JP2005092615A (en) | Natural language processing system, natural language processing method, and computer program | |
| JP3892227B2 (en) | Machine translation system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C01 | Deemed withdrawal of patent application (patent law 1993) | ||
| WD01 | Invention patent application deemed withdrawn after publication |