CN1254895A - Open full-information full-selection full-procedure template type man-machine complementary language translation method - Google Patents
Open full-information full-selection full-procedure template type man-machine complementary language translation method Download PDFInfo
- Publication number
- CN1254895A CN1254895A CN 98125015 CN98125015A CN1254895A CN 1254895 A CN1254895 A CN 1254895A CN 98125015 CN98125015 CN 98125015 CN 98125015 A CN98125015 A CN 98125015A CN 1254895 A CN1254895 A CN 1254895A
- Authority
- CN
- China
- Prior art keywords
- information
- full
- item
- man
- natural language
- 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
A man-machine complementary language translation method features that an open halographic, full-read, global template using sentences as objects for man-machine dialogue and universal for multi- natural languages' translation is created, said template contains all stand-by information terms needed by man-maching dialogue for multi-language translation, the all information terms of each natural language-sign is found out based on man-machine dialogue, and the system automatically generates natural language expression form of the language to be translated based on sentence unit while provides the explanation list of the original copy's sign information terms explained by the signs of translated language. Its advantages are man-machine complementing, multiple language universality, and correct meanings transfer.
Description
The present invention relates to a kind of machine translation method, relate to a kind of each network terminal that in computer network, is suitable for or rather and carry out the machine translation method that the information transmission exchanges with different natural languages.
The applicant proposes to Patent Office of the People's Republic of China on April 6th, 1998 that name is called " full-information full-selection full-procedure template type man-machine conversational language interpretation method ", application number is the patent of invention of 98101156.x, and the performing step of this method is: it is the natural language translation man-machine conversation template of comprising of object of various natural language necessary information key element with the sentence that a. sets up one; B. provide by dialog template different natural languages are unified commensurability all alternate information items after limited; C. earlier all the alternate information items after limited carry out automatic selection of optimal to unified commensurability by translation system, preferred result are manually adjusted on holographic dialog template by the user and are confirmed; D. by translation system according to determining that item of information translates the conversion of target language and generates, the option result of source language input side is offered user inquiring with translation.
A kind of general mode that full-information full-selection full-procedure template type man-machine conversational language interpretation method transmits as different natural language semantic informations, compare with other computing machine natural language translation technology, the semantic information transmission accurately, be suitable for multi-lingually having clear superiority aspect general.But, because the human-machine interactive information item is through the limited mistake of unified commensurability, it is various inadequately, flexible that the user carries out the method for man-machine interaction, and because the symbol that system is not taken in can not be handled, therefore can not satisfy the objective requirement of the continuous development and change of natural language.
The objective of the invention is to design a kind of open full-information full-selection full-procedure template type man-machine complementary language translation method, as to the replenishing of the 98101156.x of first to file " full-information full-selection full-procedure template type man-machine conversational language interpretation method " technical scheme, a personal-machine interactive mode is more versatile and flexible, expression content is unrestricted, generate the more natural and tripping man-machine conversation multilingual translation method of translation thereby full-information full-selection full-procedure template type man-machine conversational language interpretation method is expanded into.
The object of the present invention is achieved like this: a kind of open full-information full-selection full-procedure template type man-machine complementary language translation method is characterized in that comprising the steps:
What a. set up an opening is the multi-lingual general natural language translation man-machine conversation template of object with the sentence;
B. provide multi-lingual general whole alternate information items corresponding and the blank information item of expanding for the user by natural language translation man-machine conversation template with original text natural language symbol item;
C. by the original text user computerized optimization item of information in the natural language translation man-machine conversation template is carried out option, replenishes, confirms;
D. the item of information conversion of man-machine complementary being determined by translation system generates the translation symbol, and the translation user adopts the translation symbol that the solving result of the blank information item of whole alternate information items of original text natural language symbol item and expansion is inquired about.
Described open full-information full-selection full-procedure template type man-machine conversational language interpretation method also comprises: by program design the item of information by user's expansion is carried out the frequency of utilization statistics, and determine newly-increased general basic notion according to the frequency of utilization statistics, in the natural language translation man-machine conversation template of all languages of translation system, increase natural language symbol item and corresponding item of information synchronously.
Among the described step b, when having blank in the same languages alternate information item corresponding with original text natural language symbol item, it is described that the user can call the natural language symbol of being taken in by system.
Described whole alternate information item comprises the concept definition, temporal information, voice information, syntactic information, upper " lattice " information, part of speech information of natural language symbol item, single plural information harmonizing yinyang information by class, and the out of Memory except that above-mentioned information of the artificial design of confession that is provided by system.
Described original text natural language symbolic information item is-symbol code comprises the original text literal and with icon or with the fotmat code of numerical code expression; Original text natural language symbol comprises the sentence of individual character, speech, phrase, phrase and various regular collocations.
Conversion solving result process in the described steps d is to realize with artificial mark icon, digital mode.
Among the described step b, described natural language symbol and whole alternate information item thereof are the unions to various natural language conventional sign semantic information item basic concepts.
Multi-lingual general alternate information item among the described step b is according to languages difference and translation accuracy requirement difference, the user is carried out selectively open by system.
The technical characterstic of open full-information full-selection full-procedure template type man-machine conversational language interpretation method of the present invention is: because the basic goal of natural language translation is to transmit the actual semantic information of carrying of original text natural language symbol, the present invention program is by setting up open natural language holographic template, the channel that all semantic informations that may carry for each natural language symbol in each languages all provide man-machine interaction with flexible opening flag to find the solution is realized holographic translation; Setting up necessary natural language expressing standard is to finish the essential condition that ambiguity sign semantic item of information is found the solution task, guaranteed the automatic language translation quality, the present invention provides controllable alternate information item with holographic template to the user, thereby set up user oriented multilingual general natural language expressing standard, by in holographic template, carrying out the man-machine interaction option, find the solution the actual semantic information of carrying of natural language symbol string comprehensively, realize full choosing translation; Even the semantic information item to original text natural language symbol has carried out finding the solution comprehensively, after having converted the translation symbol string to, still may increase qi justice newly, Transfer Quality for the original text semantic information that guarantees translation system, the present invention can comprise intermediate code according to the final solving result of user's needs with any item of information man-machine interaction of the holographic inventory of original text, converts the translation sign format to, for the translation user inquiring, realize omnidistance translation.
Open full-information full-selection full-procedure template type man-machine conversational language interpretation method of the present invention has the value of widespread usage in network information interchange field, will have wide international market aspect network online service, the machine translation service opening.
Below in conjunction with one with english sentence " I saw a boy with a telescope near the bank ", translate into Chinese open full-information full-selection full-procedure template type man-machine conversational language translation process further specify technology of the present invention.
Fig. 1 is the basic process block diagram that full-information full-selection full-procedure template type man-machine conversational language translation human-machine interactive information is handled.
Fig. 2 is to be the natural language holographic template structural representation of object with the former sentence of English.
Fig. 3 is the Chinese contrast figure of formwork structure shown in Figure 2.
Fig. 4 is the actual process synoptic diagram that carries the syntactic information item of inquiry natural language symbol " with a telescope ".
Fig. 5 is the process synoptic diagram that carries out automatic or manual mark with icon, numeral or difformity, color lines, sash as code.
Fig. 6 carries out the process synoptic diagram of semantic description to it when the item of information symbol vacancy of certain language
Referring to Fig. 1, among the figure process description basic process handled of full-information full-selection full-procedure template type man-machine conversational language translation human-machine interactive information, wherein middle column frame 11 to 17 is main flows of translation system computing machine, left-hand column frame 21 to 16 illustrates user's participation process, right-hand column frame 31 to 35 illustrates the man-machine interaction process and reaches and internal database, relation between rule base, unidirectional arrow represents that man-machine interaction flows to, four-headed arrow is illustrated in the language translation process data and regular invoked procedure, the target N of institute represents that the system information processing needs man-machine interaction, the target Y of institute represents the next operation steps of automatic log-on flow process, and ##----## represents the information processing interface of this translation system and Internet system.Its top expression original text user side, its below expression translation user side.
Processing procedure begins, and execution in step 11 is imported natural language symbol to be translated in proper order by the original text user.
In conjunction with referring to Fig. 2 and contrast Fig. 3, in the tagmeme 1 to 10 of template, insert ten natural language symbol I saw a boy with a telescope near the bank in this example successively; The step 12 of system's master routine is in the semantic information item symbolic library 31 of extendible multi-lingual correspondence, each natural linguistic notation is carried out the search of standby (semanteme) item of information, when search then can be described semanteme with the semantic symbol that system has taken on template by the original text user by step 21, said process finally generate in the template by the concept definition item, the specific properties item, the tense item, the semantic standby message item that voice item etc. constitute, if concept definition item of information blank under certain natural language symbol, occurs, occur as the bank place? then the system user vocabulary symbol that provided item of information in can the employing system carries out semantic description to it, i.e. concept definition item " institution for keeping or lending money " in the template; The step 13 of system's master routine, according to the rule in the semantic information item probabilistic preference rule storehouse 32, as using boldface type specified message item in the template, by computing machine a plurality of semantic standby message item of listing each the natural linguistic notation in the template in is carried out automatic selection of optimal, can select to confirm to not obtaining definite preferred semantic information item by the original text user by step 22; The step 14 of system's master routine, by calling the automatic marking model of sentence structure composition item of information storehouse 33, the syntactic information item of listing each the natural linguistic notation in the template in is marked automatically, and said process finally generates sentence structure composition item, part of speech item, upper " lattice " item, the scene item in the template; The step 15 of system's master routine, call sentence structure composition item of information automatic selection of optimal rule base 34, sentence structure composition item of information to each natural linguistic notation carries out automatic selection of optimal, can call syntactic information item 3 d structure model storehouse 23 by step 24 therebetween, the syntactic information item that does not obtain unique preferred result is selected to confirm on template by the original text user, as using boldface type specified message item in the template; System's master routine so far just can transmit determined above-mentioned item of information with the coding form of making by oneself on network.
Full detail item in the template has expanded to all items of information that the natural language symbol can carry, and its whole standby message items not only comprise concept definition, temporal information, voice information, syntactic information, upper " lattice " information, part of speech information, single plural information, the yin and yang attribute information of natural language symbol but also the out of Memory that can manually design and mark in open template bottom expansion.
In conjunction with referring to Fig. 4, the step 16 of translation user terminal system master routine, call translation and change create-rule storehouse 35 automatically, according to multi-lingual symbol and tagmeme transformation rule, to convert the natural language translation of translation customer requirements through the item of information solving result that the original text user confirms automatically to, Chinese conversion as shown in Figure 4 generates near result's " I see that a boy is with telescope bank "; System's master routine will inquire in step 17 whether user's translation has not had qi justice, if qi justice is arranged, the translation user can determine the query context of related information items in step 26 by the man-machine interaction process, can call multi-lingual corresponding informance item symbolic library 25 between base, is to modify subject or modification object as the translation user on earth in order to find the solution " band telescope ", in Fig. 4? shown in, just can directly inquire about the actual syntactic information item that carries of this symbol, thereby determine it is to modify object.So far translation process finishes.
All alternate information Xiang Junke in the template utilize the lines of icon or numeral or different colours and sash to carry out artificial or automatic mark as code, as shown in Figure 5 use numerical reference, with 1 mark " I ", with near～1 mark " bank " (syntactic information), with 2 mark " seeing ", " seeing one " (syntactic informations), with 3 marks " boy ", mark " being with telescopical " (syntactic information) etc. with～3.
In conjunction with referring to Fig. 6, " laughing at " with the natural language symbol shown in the figure is example, and the various natural language conventional sign of union semantic information item forms the process of basic concept.Only list the basic concept union result of its English, Chinese among the figure, Russian, Japanese, French, German, Spanish and other languages can be carried out semantic description by same procedure.Three main processes shown in the figure are: when vacancy occurring as certain language basis item of information symbol, carry out semantic description with other semantic item of information symbol commonly used of this kind language, form the Back ground Information item, as " laughing at " and " laugh "; When the non-Back ground Information item of certain language vacancy, then set up fuzzy corresponding relation with Back ground Information item symbol, form non-Back ground Information item, as " laugh, smile, covered with smiles, laughingly " and " burst, smile "; When other nearly adopted item of information vacancy of certain language, then carry out semantic description with this language semantic item of information symbol commonly used, form other nearly senses of a dictionary entry, as " laugh foolishly, grin and laugh at " and " snicker, grin ".Unified coding in the table is exactly the coding form of making by oneself among Fig. 1, is used for transmission information on network.
When the original text user utilizes semantic description method to find the solution the original text symbol in Fig. 1 step 21, system program also will carry out the statistics of its frequency of utilization automatically, when frequency of utilization reaches certain level, promptly in the natural language symbolic library of all languages that translation system is taken in, increase newly-increased natural language symbol or newly-added information item synchronously.When the frequency of utilization of finding the solution bank as artificial description reaches certain level, system promptly adds newly-increased symbol " banque " in the natural language symbolic library of French and the corresponding French symbol that utilizes system to take in carries out semantic description, and provides other relevant alternate information item.The extended method of other Languages is identical with it.
During enforcement, the natural linguistic notation of in the template each top can be an individual character, as " I ", also can be speech, as " bank " (or phrase " state bank "), phrase, as " being with telescopical ", with various regular collocations, as " ... near ", can certainly be sentence (as " a blind person lights a candle ").
In addition, semantic substantially if the user only requires translation, then the option class for man-machine interaction of open system then can correspondingly reduce; Otherwise, when the fluency of customer requirements translation is high, then want corresponding increasing for the class of man-machine interaction option.
Method of the present invention is a characteristic accurately with man-machine complementary, multi-lingual general, semantic information transmission, is particularly useful for only understanding that the user of mother tongue carries out multi-lingual widely interchange on computer network.
Claims (8)
1. an open full-information full-selection full-procedure template type man-machine complementary language translation method is characterized in that comprising the steps:
What a. set up an opening is the multi-lingual general natural language translation man-machine conversation template of object with the sentence;
B. provide multi-lingual general whole alternate information items corresponding and the blank information item of expanding for the user by natural language translation man-machine conversation template with original text natural language symbol item;
C. by the original text user computerized optimization item of information in the natural language translation man-machine conversation template is carried out option, replenishes, confirms;
D. the item of information conversion of man-machine complementary being determined by translation system generates the translation symbol, and the translation user adopts the translation symbol that the solving result of the blank information item of whole alternate information items of original text natural language symbol item and expansion is inquired about.
2. open full-information full-selection full-procedure template type man-machine complementary language translation method according to claim 1, it is characterized in that: described open full-information full-selection full-procedure template type man-machine conversational language interpretation method also comprises: by program design the item of information by user's expansion is carried out the frequency of utilization statistics, and determine newly-increased general basic notion according to the frequency of utilization statistics, in the natural language translation man-machine conversation template of all languages of translation system, increase natural language symbol item and corresponding item of information synchronously.
3. open full-information full-selection full-procedure template type man-machine complementary language translation method according to claim 1, it is characterized in that: among the described step b, when having blank in the same languages alternate information item corresponding with original text natural language symbol item, it is described that the user can call the natural language symbol of being taken in by system.
4. open full-information full-selection full-procedure template type man-machine complementary language translation method according to claim 1, it is characterized in that: described whole alternate information items comprise the concept definition, temporal information, voice information, syntactic information, upper " lattice " information, part of speech information of natural language symbol item, single plural information harmonizing yinyang information by class, and the out of Memory except that above-mentioned information of the artificial design of confession that is provided by system.
5. open full-information full-selection full-procedure template type man-machine complementary language translation method according to claim 1, it is characterized in that: described original text natural language symbolic information item is-symbol code comprises the original text literal and with icon or with the fotmat code of numerical code expression; Original text natural language symbol comprises the sentence of individual character, speech, phrase, phrase and various regular collocations.
6 open full-information full-selection full-procedure template type man-machine complementary language translation methods according to claim 1 is characterized in that: the conversion solving result process in the described steps d is to realize with artificial mark icon, digital mode.
7. open full-information full-selection full-procedure template type man-machine complementary language translation method according to claim 1, it is characterized in that: among the described step b, described natural language symbol and whole alternate information item thereof are the unions to various natural language conventional sign semantic information item basic concepts.
8. open full-information full-selection full-procedure template type man-machine complementary language translation method according to claim 1, it is characterized in that: the multi-lingual general alternate information item among the described step b, be according to languages difference and translation accuracy requirement difference, the user carried out selectively open by system.
Priority Applications (4)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 98125015 CN1254895A (en) | 1998-11-20 | 1998-11-20 | Open full-information full-selection full-procedure template type man-machine complementary language translation method |
| PCT/CN1999/000046 WO1999052041A1 (en) | 1998-04-06 | 1999-04-06 | Opening and holographic template type of language translation method having man-machine dialogue function and holographic semanteme marking system |
| AU33249/99A AU3324999A (en) | 1998-04-06 | 1999-04-06 | Opening and holographic template type of language translation method having man-machine dialogue function and holographic semanteme marking system |
| CN99804904A CN1111814C (en) | 1998-04-06 | 1999-04-06 | Opening and alli-information template type of language translation method having man-machine dialogue function and all-information semanteme marking system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 98125015 CN1254895A (en) | 1998-11-20 | 1998-11-20 | Open full-information full-selection full-procedure template type man-machine complementary language translation method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN1254895A true CN1254895A (en) | 2000-05-31 |
Family
ID=5228968
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 98125015 Pending CN1254895A (en) | 1998-04-06 | 1998-11-20 | Open full-information full-selection full-procedure template type man-machine complementary language translation method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN1254895A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100346337C (en) * | 2002-12-27 | 2007-10-31 | 联想(北京)有限公司 | Dynamic forming system of open type natural language |
| CN1457041B (en) * | 2002-05-10 | 2010-05-26 | 微软公司 | System for automatically annotating training data for natural language understanding system |
| CN102236645A (en) * | 2010-05-06 | 2011-11-09 | 上海五和际软件信息有限公司 | Semantic logic-based pseudo-natural language human-computer dialogue device |
| CN107436889A (en) * | 2016-05-29 | 2017-12-05 | 陈勇 | A kind of human-computer dialogue platform exchanged with word |
-
1998
- 1998-11-20 CN CN 98125015 patent/CN1254895A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1457041B (en) * | 2002-05-10 | 2010-05-26 | 微软公司 | System for automatically annotating training data for natural language understanding system |
| CN100346337C (en) * | 2002-12-27 | 2007-10-31 | 联想(北京)有限公司 | Dynamic forming system of open type natural language |
| CN102236645A (en) * | 2010-05-06 | 2011-11-09 | 上海五和际软件信息有限公司 | Semantic logic-based pseudo-natural language human-computer dialogue device |
| CN107436889A (en) * | 2016-05-29 | 2017-12-05 | 陈勇 | A kind of human-computer dialogue platform exchanged with word |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104679850B (en) | Address structure method and device | |
| CN102193643B (en) | Word input method and input method system having translation function | |
| CN101520786B (en) | Method for realizing input method dictionary and input method system | |
| CN101694603B (en) | Cross-platform Mongolian display and intelligent input method based on Unicode | |
| JP2732563B2 (en) | Machine translation method and apparatus | |
| CN109145276A (en) | A kind of text correction method after speech-to-text based on phonetic | |
| CN101008864A (en) | Multifunctional and multilingual input system for numeric keyboard and method thereof | |
| CN102165437A (en) | Information processing device and information processing method | |
| CN102165438A (en) | Information processing device and information processing method | |
| JPH07114558A (en) | Chinese character conversion correcting process system | |
| Aswani et al. | A hybrid approach to align sentences and words in English-Hindi parallel corpora | |
| JP2011181109A (en) | Information retrieval support program, computer having information retrieval support function, server computer and program storage medium | |
| CN103164398A (en) | Chinese-Uygur language electronic dictionary and automatic translating Chinese-Uygur language method thereof | |
| CN103164397A (en) | Chinese-Kazakh electronic dictionary and automatic translating Chinese- Kazakh method thereof | |
| CN1254895A (en) | Open full-information full-selection full-procedure template type man-machine complementary language translation method | |
| AU2023206192A1 (en) | Generative prompt expansion for image generation | |
| CN103164396A (en) | Chinese-Uygur language-Kazakh-Kirgiz language electronic dictionary and automatic translating Chinese-Uygur language-Kazakh-Kirgiz language method thereof | |
| CN103164395A (en) | Chinese-Kirgiz language electronic dictionary and automatic translating Chinese-Kirgiz language method thereof | |
| CN110362691A (en) | A kind of tree bank building system | |
| CN1111814C (en) | Opening and alli-information template type of language translation method having man-machine dialogue function and all-information semanteme marking system | |
| KR100519748B1 (en) | Method and apparatus for internet navigation through continuous voice command | |
| Barberis et al. | Improving accessibility for deaf people: an editor for computer assisted translation through virtual avatars. | |
| CN1466039A (en) | Electronic remote controller capable of inputting Chinese and various characters | |
| CN109284012A (en) | A kind of Gu Yi nationality's text language in-put control system and method, information data processing terminal | |
| CN110874527A (en) | Cloud-based intelligent paraphrasing and phonetic notation system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C01 | Deemed withdrawal of patent application (patent law 1993) | ||
| WD01 | Invention patent application deemed withdrawn after publication |