CN103218353B

CN103218353B - Mother tongue personage learns the artificial intelligence implementation method with other Languages text

Info

Publication number: CN103218353B
Application number: CN201310086775.0A
Authority: CN
Inventors: 刘树根
Original assignee: Individual
Current assignee: Zero One Artificial Intelligence Technology Research Institute Nanjing Co Ltd
Priority date: 2013-03-05
Filing date: 2013-03-05
Publication date: 2018-12-11
Anticipated expiration: 2033-03-05
Also published as: WO2014134971A1; CN103218353A

Abstract

The present invention provides a kind of mother tongue personage and learns artificial intelligence implementation method with other Languages text.Mother tongue personage moves on on computer the reading understanding of bilingual text and carries out, it reads and understands and clicked simultaneously with mouse or software automatically identifies can make the word string in cabin, mark sentence cabin and cabin eye so as to retaining reading understands that achievement generation sentence understands template, containing component of expressing the meaning, the meaning of one's words is identical between template and component and mutual mapping；Also the gained of everybody is assembled as artificial intelligence its " " and understands template database in sentence.It is to express the meaning component that software, which splits source statement with the latter, and reconvert component of expressing the meaning is sentence of communicating with the eyes, and provides sentence to be corrected and carries out corrigendum running.Connecting without corrigendum handles next sentence；Need to correct has 11 kinds of simple and direct method for correcting and links with self-learning module；Corrigendum can make subsequent A more accurate to the conversion of B sentence.To provide, mother tongue reading foreign language, outer-female translation, mother-translation, sentence bone help the mother tongue personage for being writing for powerless reading foreign language to be read, with reference to the multinomial intelligent use of foreign language datum ability outside.

Description

Mother tongue personage learns the artificial intelligence implementation method with other Languages text

[technical field]

The present invention relates to a kind of implementation methods of artificial intelligence；In particular to a kind of tool mother tongue and professional knowledge and learn with outer The artificial intelligence implementation method that language person uses.

[background technique]

1) CN99804904.2 license, Liu Sha disclose " open holographic template MML man-machine language a translation Method and holographic semantic tagger system ".It establishes one and is limited dialog template by the natural language of object of sentence, include in template It is necessary to semantic information elements for the institute of various natural languages, by selecting formula human-computer dialogue to determine natural language symbol entirely in template The lexical information item and syntactic information item actually carried, completes the source language message and solves, and generates translation according to solving result, and will ask Solution result is converted into translation symbol, for translation semantic query.It attempts not depend on context of co-text and makees semantic analysis and sufficiently benefit With the effect of man-machine mutual supplement with each other's advantages, obtains and clear away semantic information conductive obstruction in network communication.It is established using " sentence is object " Template can use for it；It is too complicated, find it difficult to learn difficulty but comprising " various natural languages institute it is necessary to semantic information elements " in template With.And " the lexical information item " and " syntactic information item " actually carried is determined whereby；Expressing the meaning altogether between different language is not caught Property, it is impossible to obtain result.

2) CN208801286367 patent application discloses " a kind of production method of sentence component " in Liu Shugen.It utilizes Bilingual or multilingual words version the corpus of identical content is as training sample.A sentence pair is taken out, human-computer interaction side is utilized Formula carries out sentence pattern level and dissects comparison acquisition sentence pattern and small idiom；It carries out sentence cabin storey face and dissects comparison acquisition cabin mould and sense-group string. Then again with the next sentence pair of same methods anatomy.Simultaneously by the sentence pattern obtained in this process, cabin mould, sense-group string and small Idiom is referred to as sentence component, is built library preservation.This method obtains the process of sentence component and human brain reads understanding process (side Formula) it is close, it is significant.

3) CN200810086229.6 license, Liu Shugen disclose a kind of " mother tongue reading foreign language based on sentence component Method ".Sentence component is obtained using preceding same procedure；Mother tongue method for reading foreign language is generated based on sentence component.Generating people can With with the technical effect of respective mother tongue direct reading foreign language；Realize the long-awaited dream of people.

4) CN200910129096.0 patent application, Liu Shugen disclose a kind of " multilingual profession based on sentence component Interpretation method ".Based on sentence component；Make sentence component library that there are field structure mapping relations.The source of translating is treated using sentence component Sentence is split, and sentence is split into sentence component；Source language sentence component to be split is reflected by the mapping relations in corresponding library Sentence component of communicating with the eyes is penetrated and taken out, sentence of communicating with the eyes is assembled into the latter and is provided.Not only it can get high quality translation but also do not needed tool solely Vertical translation ability person operation.Whole process is very simple, and computer has only done simple and efficiently remembered, searches for and table look-up；It need not Directly processing is semantic, and it is unbridgeable " semantic obstacle " to have bypassed machine translation for a long time；The obvious progress of tool.

Between above-mentioned patent and application 2-4, there is a common feature, all seek one between human brain, computer, spoken and written languages A best joint, it is intended to obtain breakthrough technical progress whereby.Also there are a kind of progress and incidence relation.

The application further innovates on the basis of above-mentioned patent and application 2-4；Above-mentioned patent and application 2-4 are also this Shen Related patent and application that please be first, partial content need while implementing.

People often learn another spoken and written languages when having a good command of one's mother tongue and having certain professional knowledge background, they It is mother tongue personage (calling mother tongue personage in the following text)；In order to work and cause, need to learn again one or more text, so as to It can read with reference to foreign language datum；The study for especially carrying out model reading by bilingual text is read, enlarging one's vocabulary.Non- foreign language is special Each professional universities and colleges of industry, open up foreign language class bar none, and purpose is exactly to read and refer to foreign language datum.The purpose of " " is " use "；" " for opening up foreign language class is that foreign language datum is used in order to " read reference ".In fact, mother tongue personage's learns a foreign language, effect Fruit is generally undesirable, can make they learn a foreign language efficiency raising, that will obtain huge economical, societal benefits.

It is well known that human brain has unique understandability, computer does not have；Computer has outstanding memory search ability, human brain It is then much too late.If can be combined into one, breakthrough certainly will be broken out.

Orientation there are two types of being combined into one；First, interconnecting human brain computer by means of the instrument of certain brain wave, human brain can be with Using computer as the extender of memory；To which human brain can directly search computer memory body.Second, by human brain to limited mesh Target understandability tangibility becomes finite goal and understands product or template, this finite goal product or template are transferred to computer, Computer finite goal can be carried out based on the understanding that information processing.Although current someone, which explores, utilizes idea, pass through brain Electric wave robot command；But to realize the interconnection of the former this human brain computer, all too is remote.The latter, to the limited mesh of human brain Target understandability tangibility makes the way of computer simulation human brain close to understand that product or template are transferred to computer with tradition, It can be able to achieve.

Human brain is word for word scanned when reading understands by unit eyes of syntagma；They are divided into multiple semantic chunks by brain, Linear or multidimensional between block and block is managed out to contact.Block segmentation is over, and it is clear to contact between block；Exactly understand.If segmentation is not Agilely, it contacts unintelligible between block, does not understand；This process ... must be repeated by not understanding.It is contacted to reinforce dividing, manage out, Hand also helps --- and setting-out draws book, takes notes, and leaves the various markings read and understood.

When human brain reads this text (books) for leaving various markings again, connection moment shows between block segmentation, block, Understanding reading becomes simple and quick.

Although interpersonal, read understanding and leave the multifarious of the marking；It is also different to read the depth degree understood；But It is that this is rule, this is universal law.Otherwise all over the world just without drawing book, the thing that do not take notes；Naturally also just without drawing book Model, " Lenin's Note on Philosophy ".

In conclusion drawing book, taking notes is the marking of human brain unique understanding ability when reading understands.The prior art and existing The shortcomings that shape is: human brain, which is read, understands do not have any association with computer language Text extraction；Human brain reads the gained of understanding It cannot be borrowed by computer, memory search ability that cannot be outstanding with computer is combined into one, and the two disconnects completely.It can utilize Software approach understands reading and leaves more, unified and standard be imprinted as used in computer；It is the technology of a tool very meaning It attempts and challenges.

" female outer language philology use " refers to mother tongue personage, and foreign language studying uses foreign language --- reading, reference, translation foreign language Data --- overall process, reading is left by the means of human-computer interaction, software in this process and understands that more, specification is united " sentence understands template " of one marking, so as to realizing and reaching " artificial intelligence " effect and effect.

[summary of the invention]

The general plotting of the present patent application is:

Liu Shugen is moved in preceding patent and application 2-4, provide a kind of mother tongue personage learn with other Languages text it Artificial intelligence implementation method.There is 1. reading understanding to move on on computer to carry out；2. software marks sentence cabin and cabin eye；3. template and structure The meaning of one's words is identical between part and mutually maps；4. splitting source statement is component of expressing the meaning；5. conversion is expressed the meaning, component is sentence of communicating with the eyes；6. more Just with the six big steps such as machine learning.

Step 1.~be 2. intelligence " " embodiment manage the reading of mother tongue personage by the software approach of human-computer interaction Solution leaves more, the unified and standard marking, reads the product (achievement) understood to retain current sentence pair.To in " mother tongue Under the conditions of personage ", " sentence is unit " and " being to understand depth with A-B sentence transition translation " three finite constraints, limited mesh is generated Mark --- sentence understands template, saves the latter, is transferred to computer and follow-up work is enable to complete by software.Intelligence in this way A) appetite learned is big also feasible again, and the memory of computer magnanimity and outstanding search can play a role；B) intelligence learning it is easy to accomplish and Can be shared, mother tongue personage adds a computer, more meaning be everybody obtained by, can be added, can share；C) intelligence Follow development in science and technology, any pragmatic new results, new method, experiences in translation, it is easy to sum up in the point that sentence understands template closely.

3. the meaning of one's words is identical between template and component and mutually maps for step, step 1.~sentence that has 2. obtained is understood as Fruit, since the sentence component that can be expressed the meaning according to its frame-type is classified；Children learning is spoken, and people's sentence-making thoughts communication is all It is expressed the meaning the process of component using this frame-type.However, this understands that the sentence of example understands product from a sentence pair, in the known integrated circuit it is a fact that one Class sentence represents, and can extract " template " of a kind of sentence.

4. software is splitting source statement as component of expressing the meaning to step, and then 5. converting component of expressing the meaning is sentence of communicating with the eyes, and is 6. mentioned For corrigendum chance and carry out software self study, it is that these are completed in human brain, from A sentence to the conversion process of B sentence, by soft Part is realized in computer.4.~6. step constantly recycle be artificial intelligence " use " core；Four kinds can at least be generated for brain application: A) the vast mother tongue personage for powerless reading foreign language is read, with reference to the intelligent use of foreign language datum ability --- and mother tongue is read outer Text.B) the vast mother tongue personage for the translation of no masterpiece obtains the intelligent use of translation foreign language datum --- outer-female written translation.C) Intelligent use --- female-outer written translation of the larger translation ability of difficulty is obtained for the mother tongue personage that no masterpiece mother-translates outside.D) The mother tongue personage write with foreign language is needed to obtain with the intelligent use of foreign language writing ability for inability --- sentence bone helps writing.

The realization of female outer language philology manually intelligence obtains mother tongue personage and had both mitigated the time-consuming expense brain that learns a foreign language, read Read with reference to the ineffective difficulty of foreign language datum, at the same again for expand it is quite huge, spoken and written languages can be covered express the meaning all of component Component database of expressing the meaning is done one's bit；Later mother tongue personage is set to obtain more more preferably artificial intelligence benefits.

The technical problem to be solved by the present invention is

The present invention provides a kind of mother tongue personage and learns artificial intelligence implementation method with other Languages text.It to be solved Technical problem is the prior art (status), human brain read understandings obtained by, cannot by computer language Text extraction benefit With；The unique understandability of human brain and the outstanding memory search ability of computer cannot be combined into one；Computer cannot be sentence by sentence by language Text A is converted into the technical issues of meaning of one's words constant spoken and written languages B.

The technical solution adopted by the present invention to solve the technical problems is:

A kind of mother tongue personage is provided and learns artificial intelligence implementation method with other Languages text, feature includes following step Suddenly；

S1. reading understands to move on on computer and carry out, and mother tongue personage moves on to the reading understanding of identical content bilingual text It is carried out on computer, bilingual text A and B are respectively placed on the area source document A and B by software；

Software from the area source document A and B according to the cutting of end of the sentence symbol and take out mono- sentence pair of bilingual text A and B set reading understand Area A and B, the area source document A and B accordingly reduce one；

If mother tongue personage thinks to cut sentence improper, source document area A or B point are clicked, part before the point is moved into and read by software Understand that area follows closely after A or B taken sentence；

If clicking reading understands that area's A or B point, software will partially retract source document area A or B after the point；

S2. software mark sentence cabin and cabin eye, mother tongue personage, which reads understanding and clicked to read with mouse simultaneously, understands A and B, area It currently can be used as express the meaning identical word string or several continuous word strings in cabin, software is by them with capitalization according to successively for sequence Add brace at A and B while being designated as a sentence cabin in advance；

Or software automatically identifies A and B currently express the meaning identical word string or several continuous word strings for them with capitalization Letter for sequence according to successively adding brace at A and B while being designated as a sentence cabin in advance；

Mother tongue personage reads understanding or software automatic identification is successively searched for quantity string, proper name string, using names string, bilingual Mark a sentence cabin in advance simultaneously, if these three strings have all been searched for and bone example compares between 15-50%, pre- sentence cabin of marking terminates；

If these three strings all search for and bone example ratio > 50%, then with no using names string, the matching string continuation of other meaning of one's words according to The secondary pre- mark sentence cabin of search, until bone example compares between 15-50%, pre- sentence cabin of marking terminates；

When mark sentence cabin terminates in advance, software carries out sentence cabin, sentence bone format checking, with 16 binary digits replacement capitalization with A Sentence from left to right arranges for sequence and aforementioned pre- is designated as formally marking；

Bilingual sentence pair through formally marking further detects each cabin, if there is the big sentence cabin greater than original 5 word strings to need Further mark cabin eye or group compound word, mark cabin eye are followed successively by target word string with no using names string, other meaning of one's words matching string, and mould example, which compares, is Between 50-70%, ophthalmically acceptable 16 binary digits in cabin from left to right add square brackets to mark with A for sequence, be above the first round sentence cabin and The step of cabin eye understands mark using poster kind sentence as template software automatic identification, marks not since the second wheel understands mark Poster kind sentence accordingly express the meaning identical sentence cabin and cabin eye；

Every wheel mark selects the corpus of the identical AB bilingual text version of content, and the second wheel starts every wheel and understands that mark is new Increase a kind of spoken and written languages, another kind is the spoken and written languages marked, and A language distributes to the spoken and written languages marked, B language point The spoken and written languages that dispensing is newly added；

The formal mark of each sentence pair terminates, and carries out the mark of next sentence pair, connects S1；

S3. the meaning of one's words is identical between template and component and mutually maps, and software must be understood this by the reading that human brain is dominated Current reading is understood that the bilingual sentence pair in area marks upper sentence cabin, cabin eye in the process, to retain current sentence pair read understand at Fruit is transferred to computer in such a way that sentence understands template 100, and follow-up work is enable to generate Intelligential efficiency by software completion, Other sentences are marked in order to which sentence understands that template can be more particularly for understanding, they are stored in sentence and understand that template database replaces by software Human brain is remembered, and sentence understands template database at least decomposable sentence bone, cabin mould, four kinds of sense-group string, idiom components of expressing the meaning out, point Gou Cheng not express the meaning component base accordingly, they are point libraries that sentence understands template database；

S4. splitting source statement is component of expressing the meaning, and it is structure of expressing the meaning that software, which understands that template 100 splits source statement using sentence, Part understands instead of human brain and splits sentence, first carries out operating with sentence bone, presses the void nucleation sentence bone of sentence bone word string and sentence cabin in advance String list simultaneously indexes, when with sentence bone source statement from left to right word for word or by word string remove search sentence bone string list bone lead-in section simultaneously It is stored in interim table, then takes out interim table content by record, with sentence bone word section inquiry source statement, every section of sentence bone string can be It is found in source statement and order is mutually all matching, the source statement bone of current matching is taken out further according to sentence bone string list；

It is that component of expressing the meaning makes current source statement and understands mould that source statement sub-set, which is entered source statement bone to split source statement, Plate 101；

Sentence sentence bone of accordingly communicating with the eyes is taken out using current source statement sentence bone Mappings；

S5. converting component of expressing the meaning is sentence of communicating with the eyes, and software searches the sentence bone for sentence of communicating with the eyes using source statement sentence bone, will be current Source statement understands that the sentence cabin content in template is transferred to the corresponding sentence cabin of a bone of communicating with the eyes；

All cabins are detected, without skipping for cabin mould, if containing cabin mould, this cabin content is inserted in cabin mould, utilizes the cabin The Mappings in mould point library take out cabin mould of accordingly communicating with the eyes, and current source language cabin is intraocularly held to the corresponding cabin eye for moving on to cabin mould of communicating with the eyes；

Sentence cabin or the cabin eye for searching for mesh sentence comprehension template 102 one by one, search for sense-group string library with words and phrases string in source therein, will All explanation results of communicating with the eyes searched are saved in correction sheet, and using pre-selection module, one of pre-selection is as sense-group string of communicating with the eyes The source words and phrases string in current sentence cabin or cabin eye is replaced, until all cabins, cabin eye replacement terminate.

S6. corrigendum and self study, software take out conduct sentence to be corrected of communicating with the eyes from mesh sentence comprehension template 102 and show and provide, Receive operator's confirmation or corrigendum, while carrying out self study；

Or software takes out from template of communicating with the eyes and show and provide as sentence to be corrected of communicating with the eyes, and takes out from the language template of source as source language It shows and provides with reference to sentence, receive operator's confirmation or corrigendum, software carry out self study simultaneously；

When human brain is read, the marking moment for reading understanding shows, and reading understanding becomes simple and quick, software compounding practice Into correction running 201；

Software is the component S4 that expresses the meaning splitting source statement, and then conversion expresses the meaning component as the sentence S5 that communicates with the eyes, software cooperation Make corrigendum and self study S6, these are turned what human brain was completed from the sentence of spoken and written languages A103 to the sentence of spoken and written languages B104 It changes, software is given to realize to be recycled between S4-S6, can generate multinomial intelligent use.

In the above method, sentence described in step S3 understands that template database further feature is: sentence understands that template includes sentence Example, sentence cabin, sentence bone, cabin eye, cabin mould, sense-group string, idiom, they are the member that sentence understands template altogether, directly to express the meaning at Member is component of expressing the meaning, and is connected with them and constitutes the long-chain that a multilingual sentence understands template；

Software save with database and be remembered instead of human brain while acquisition sentence understands template, first round reason Solution mark sentence understands that template database contains A moulds, B mould fields, and the second wheel understands that mark starts every wheel and increases a kind of language newly Text, sentence understand that template database first adds the sentence mould field of a corresponding language, and every wheel understands that mark must take same content, difference Text version corpus ensures that the meaning of one's words expressed by each languages sentence mould of identical recordings is identical, and mutual phase mapping；

Subordinate clause understands that template database decomposites several species and expresses the meaning component, respectively constitutes component base of expressing the meaning, they are sentences Point library for understanding template database divides library to have inheritance, inherits the characteristic that sentence understands template database, divide same record sheet in library The component meaning of one's words of anticipating is also identical, their also mutual phase mappings.

In the above method, using pre-selection module described in step S5, one of pre-selection is worked as sense-group string replacement of communicating with the eyes The source words and phrases string of preceding sentence cabin or cabin eye, further feature is:

Correction sheet at least word string section, search string, explanation, the first field of going here and there, preselect module and spoken and written languages type are close It is related；

The pre-selection 312 of english translation is that 1. with current sentence cabin or cabin, intraocularly English word string inquires " experience selects word " library, has looked into and has taken It explains that the first item Chinese word string of field contents provides as the string of communicating with the eyes of 313 maximum adaptations out, looks into and be difficult this word string of correction sheet section head The string of communicating with the eyes for being recorded as maximum adaptation provides；2. inquiring " middle word with the Chinese content 314 that current sentence cabin or cabin eye have obtained The former sequence field in sequence " library is looked into and is provided with the string 315 of communicating with the eyes that positive sequence field contents replace with maximum adaptation, looked into without constant；

In translate the pre-selection of English with sentence cabin in source, transition type of communicating with the eyes or cabin eye source language content search correction sheet search string word Section, have congruence explains that field contents are provided as the string of communicating with the eyes of maximum adaptation using same record, and 317 inquire " cabin again when no congruence Experience " library middle deck field has 318 to provide by the string of communicating with the eyes of maximum adaptation of English cabin field contents, looks into " cabin experience " library and is difficult to work as Preceding sentence cabin or cabin eye Chinese content are that the string of communicating with the eyes of maximum adaptation provides；

All pre-selection modules selected correction sheet check in item change it is former provide content, while it being changed in correction sheet First of position to this word string section.

In the above method, software described in step S6 understands that template is taken out as communicate with the eyes sentence to be corrected and/or conduct from mesh, original sentence Source language refers to sentence, receives operator's confirmation or corrigendum, while carrying out self study, further feature is:

When user reads communicate with the eyes sentence to be corrected and corresponding source language with reference to sentence, i.e. correction starts, and the meaning of one's words, which corrects, has sentence to be corrected complete Total correctness is without corrigendum and needs to correct two kinds of situations, receives user click the Back button when without correcting and connects S4, needs Entering after correction operates 201 for corrigendum connects S4 again；

Software enters correction running 201, and links with self-learning module 301, to the human-computer interaction of correction course It practises, remember and provide data for pre-selection module.

The correction that enters for needing to correct is stated in the above method and operates 201, further comprises following steps:

A) the replacement 202 of improper word string --- when user thinks that sentence to be corrected word string is improper, click its software search more Positive table, pop-up drop-down list provide whole correlations and check in that item is to be selected, and then to be clicked the item replacement that checks in chosen, to translate sentence improper Word string, while being replaced into first of word string section；

B) the cutting 203 in sequence sentence cabin --- when encounter without reference to the sequence sentence cabin that can not carry out cutting automatically provide to It cuts, to be separated two sentence cabin contents as cut-off by click-point after being clicked；

C) intervene optionally sentence bone 204 --- when encountering a cabin and cabin eye unshapeliness, there is incomplete or stacking mistake to select sentence bone table Now or user does not approve and clicks the button " after rear " when currently translating, and software provides multipair matched sentences bone to be selected, is selected with clicking Sentence bone after taking is inserted in connecting S6 again to be continued to operate；

D sense-group string 205) is added --- when improper word string is optionally replaced and lacked, receives user and use the extension meaning of a word, glue Band leading or following word word method add sense-group string, are replaced to add string and add languages component word corresponding in sense-group string Section；

E) group choosing goes here and there 206 again --- and when improper word string is optionally replaced and lacked, receives user using group compound word or reelect The method of compound word adds compound word string, is replaced to add string and adds languages component field corresponding in the compound word library of sense-group string；

F) sentence cabin word order 207 --- when translating sentence sentence cabin has word order mistake, receives user click and wherein go here and there, then click " ← Move " or " shifting → " order button, string Forward will be clicked or move back a displacement；

G) compile word replacement 208 --- when much more continuous string is improper, cannot with click choose when, choose string with drag-and-drop operation When, when then filling out other word strings, editor at word replacement or be empty compiling, then clicking " ∧ " button, before software is replaced with the latter Person；

H) rhetoric 209 --- when the rhetoric phenomenon for occurring to be solved with above-mentioned steps, receive user click " rhetoric " Then current sentence to be corrected is copied to edit box by button, receive user using the method for editor and carry out rhetoric operation；

I) cabin conversion 210 --- translate that inch receives user click cabin or cabin intraocularly holds in the middle, search correction sheet is all What current sentence cabin or cabin eye contained check in item, list, which provides, waits user is repeatedly orderly to click selections, by their displacements to school The corresponding sentence cabin of sentence or cabin intraocularly hold；

J) mend quantifier 211 --- when english translation due to English quantifier-free and Chinese have quantifier thus need sentence to be corrected need Place is wanted to fill quantifier, the need for receiving user click sentence to be corrected add place, Shuo Chuan rear center of the software search closest to click-point Noun when " ▲ " button while being clicked, with head noun volumes of searches dictionary, takes out corresponding quantifier and add sentence to be corrected；If Nothing is looked into quantifier library, is provided dialogue and is received to input to add and typing quantifier library；

K) mend hat it 212 --- translate in the middle inch due to Chinese without and English has article, verb to/be, to/have Deng, these needs are added, software provide retain sentence cabin mark label communicate with the eyes sentence to be corrected when, it has been shown that " a/an ", " the ", " to/be ", " to/have ", when one of they are clicked, and sentence to be corrected is also clicked, software by one of they to School sentence is added the word string of appropriate form by the place of clicking；

The above A-H state step is unrelated with the type of spoken and written languages, has common feature, can appear in A between multilingual During translating the two-way translation that B or B translates A；I state step is related to the type of spoken and written languages, expresses the meaning for interval between no word Among the translation of text；J-K state step and spoken and written languages type are closely related, they have characteristics of personality and are set according to specific languages It sets.

It is described to link with self-learning module in the above method, the human-computer interaction of correction course is learnt, is remembered and Data are provided for pre-selection module, further feature is:

A, experience selects word, 303 after in english translation meaning of one's words correction operation when the improper word string of user click sentence to be corrected, Software search correction sheet, list provide whole correlations check in item it is to be selected while, all related constraint terms are separated by life with space At through channeling test of cement sheath, end is attached corresponding English word string；

While user chooses one of list displacement, with choosing entry and English string to inquire " experience selects word " library, frequency has been looked into Degree=frequency+1, looks into nothing, and entry will be chosen to move on to through first of channeling test of cement sheath and pop down formula is stored in the middle explanation field in the library 305, English string Deposit English string field is removed from through channeling test of cement sheath, current sentence bone number is stored in sentence bone field；

B, Chinese word order, when the Chinese sentence to be corrected of generation being provided in the english translation meaning of one's words correction operation backup in " providing sentence " is write down when measuring user and using " ← shifting " or the mobile sentence cabin of " shifting → " button or cabin eye word order 306 by shifting sentence cabin Cabin label or cabin eye number, when meaning of one's words correction terminates, search adjusts word order by displacement when " returns " command button is clicked Sentence cabin or cabin eye, are that will shift forward and backward result as unit of sentence cabin and be stored in " the former sequence " in middle word order library 308, " positive sequence " word respectively Section, and remove current sentence cabin or cabin eye and select the relative recording in dictionary 305 in experience；

C, cabin experience, when translating " cabin conversion " running of English meaning correction in 309 it is standby deposit current sentence cabin, cabin label, when " return " order button is clicked, by current sentence it is all by cabin conversion turn forward after content look into cabin experience library 311, looked into frequency Degree+1；It looks into no pop down formula and is stored in the library.

In the above method, after step S1 is completed, before S2 step starts, there are one steps:

First operate with sentence bone with current A language example sentence, search is when the whether existing matched sentences bone of precedent, if searching A language Sentence has matched sentences bone and obtains skipping for operator's approval after being inserted in, it is not necessary to understand mark, read in next sentence pair；

If not searching A sentence has matched sentences bone or has searched but looked into when being denied after being inserted in by operator with mark fortune After work, into the step S2 for understanding mark sentence cabin.

In the above method, circulation generates multinomial intelligent use between S4-S6 described in step S6, and further feature is It is read, for the vast mother tongue personage of powerless reading foreign language with reference to the intelligent use of foreign language datum ability:

Mother tongue reads foreign language --- and allow people to utilize the mother tongue direct reading foreign language datum of oneself.By the circulation of step 4-6, Source language is converted into communicating with the eyes and be provided by one ground, for people one one is read, if having, word string is improper, is carried out with mouse-click Corrigendum；It can only take two fields of A, B language component of multilingual component base of expressing the meaning to generate dedicated mother tongue and read foreign language special edition, it will Regular History Frequency component of expressing the meaning is stored in user terminal, remaining better adapts to user individual by network support.

Outside-mother's translation --- the vast mother tongue personage for powerless translation foreign language datum is referred to, translates foreign language datum Intelligent use, software cutting foreign language datum sentence by sentence carry out the conversion cycle between S4-S6, mother tongue sentence are converted out, with mother tongue sentence Understand that the mode of template is shown and provide sentence to be corrected, while showing source language with reference to sentence；

When user, which thinks that somewhere is improper, uses mouse-click, software starts correction running and Machine self-learning immediately；

User terminal reserved high-frequency is expressed the meaning component, when to reserved high-frequency express the meaning component base close to personal-machine at full capacity when, from It is dynamic to be cleared up, 20% vacating space of low frequency end is deleted, again meter frequency, after rest part is carried out by network by corporate server Platform real-time support.

In the above method, circulation generates multinomial intelligent use between S4-S6 described in step S6, and further feature is The intelligent use of female-outer translation ability is obtained for the mother tongue personage that no masterpiece mother-translates outside:

Female-outer written translation --- mother tongue information translation at foreign language text, software cutting mother tongue data sentence by sentence, progress Conversion cycle between S4-S6, conversion are gone out sentence, and sentence understands that the mode of template is shown and provides sentence to be corrected in addition, simultaneously Show that female sentence comprehension understands that template is prepared for the examination；

When user, which thinks that somewhere is improper, uses mouse-click, software starts correction running immediately, makees corresponding necessity when necessary Interaction, the translation brief meaning of one's words is accurate, provides the constant foreign language translation of the original text meaning of one's words；

Corrigendum operation some is directly related with concrete syntax text type, and in addition to publicly-owned more orthofunction, software is according to specific language Kind gives suitable personalized corrigendum operation.

In the above method, circulation generates multinomial intelligent use between S4-S6 described in step S6, and further feature is The mother tongue personage write with foreign language is needed to obtain the intelligent use with foreign language writing ability for inability, sentence bone helps writing, software It receives mother tongue personage to illustrate in a manner of three kinds, is judged to 1. illustrating with mother tongue sentence if according with no lower horizontal line containing end of the sentence in signal, If also thering is lower horizontal line to be determined as 2. with fuzzy sentence bone signal, if containing space in the Chinese word strings of signal containing end of the sentence symbol in signal Or the above two are after the failure of sentence bone with being judged to 3. continuing the running of following one respectively with the signal of several word strings:

1. writing out complete mother tongue sentence expression with the signal of mother tongue sentence in " mother tongue meaning area " and wanting the meaning expressed with foreign language Think.One can be only write, more can also be write, it is necessary to accord with using end of the sentence.Software accords with cutting, the work to get off with cutting according to end of the sentence For example sentence, operated with sentence bone, such as do not search matched sentences bone, jump and 3. handle；It searches, provides search result progress The circulation of S4-S6；

After being wholly converted into English word string, software push buttons " tail adds " receive user click, and current sentence tail is added to English Area is write, operator can also be received in english writing area and carry out edit-modify；

2. software reception is clicked " with sentence bone " after order button running with fuzzy sentence bone signal, a list is popped up, is listed Sentence bone containing " fuzzy sentence bone " composition, and cursor is fallen on desirable possible maximum record, while its China and British sentence bone is It is shown in the respective block of top.User click " ↓ tail adds " order button is received, English sentence bone tail is added on English composition window Mouthful, it waits user to fill in a cabin content, completes this English composition；

, can be alternative in list if first record does not simultaneously meet the meaning to be expressed of user, it clicks single again after the record Hit " ← optionally " order button, which will display；

3. with the signal of several word strings, software will illustrate cutting to be word string, after being searched for according to word string as a result, arriving bilingual text Or sentence pair library, data base etc. using sentence as searched in material existing for unit containing sentence, and with the number containing word string length Sequence takes high-end 10 record lists to be selected to provide with reference to sentence.

Compared with prior art, the beneficial effects of the present invention are:

1) mother tongue personage, which obtains, had both mitigated the time-consuming expense brain that learns a foreign language, has read the difficulty ineffective with reference to foreign language datum；It is again simultaneously Expand it is quite huge, express the meaning all component databases of expressing the meaning of component of spoken and written languages can be covered do one's bit, make later Mother tongue personage obtains the benefit of more artificial intelligence.

2) so that human brain is read understanding by software and leave more unified markings, read the sentence reason for understanding product to retain Template is solved, the latter, which is transferred to computer, enables follow-up work to complete by software；Such " ", a) efficiency is very high；B) this Kind learn easy to accomplish and can share, mother tongue personage adds an a computer, more meaning be everybody obtained by, can be added, It can share；C) this follows development in science and technology, any pragmatic new results, experiences in translation closely, it is easy to which end, which is formed a complete sentence, understands template.

3) spoken and written languages A can be converted into the constant spoken and written languages B of the meaning of one's words sentence by sentence by computer, carry out high translation quality whereby Machine translation；Mother tongue personage can directly be read with mother tongue and refer to foreign language datum whereby.

4) Current software can " intelligence " people also when imagination and arguement, take the lead in realizing artificial under finite goal Intelligence, develops artificial intelligence product, has positive learning value and realistic meaning.

5) present invention tool carries forward Chinese Chinese, promotes the integrated beneficial effect of universal language text.Human Natural Language Very long " textual " period is had gone through, " textual " does not complete, i.e., there are no the language of text, will completely disappear. Natural language after textual can be rated as " spoken and written languages ", will enter this " closing to express the meaning using modern science and technology as background Second period that connection " technology is come together.It is universal language text integration after this period；This period is just in other words It is the spoken and written languages integrated primary stage.Spoken and written languages after " being associated with expressing the meaning " are no longer " in respective independent system Interior slow development ", but " being associated with expressing the meaning " it is common afterwards, rapidly develop.

, using modern science and technology as background, being entirely possible to for the technology of " being associated with expressing the meaning " is exactly that institute of the invention is public for this The technology opened.It is using Chinese-English as first spoken and written languages pair and kernel language text pair.Other Languages text will be become It expresses the meaning associated template, the association expressed the meaning like that other Languages text as DNA replication dna.This point is for carrying forward Chinese Chinese is significant.

Chinese Chinese is once worried due to the character set too bitter brain greatly in face of computer and is difficult to input.Natural language is a large amount of Annexed, many compatriots honest patriots worry for it.Sigh with feeling and be spaced between no word, ideograph lags behind alphabetic writing etc.；Even Appeal to be changed to alphabetic writing before being annexed danger association and being come.Although these are anxious reasonable, it I had never expected and sent out in this series Bright appealing becomes extra, the anxious change advantage of institute here, is spaced between no word and is easier association of being expressed the meaning instead；Although alphabetic writing elder generation Into, the world's spoken and written languages integration in face of it is not necessarily more superior.So supporting, implementing the invention of this series (including subsequent hair It is bright) there is the beneficial effect for carrying forward Chinese Chinese.

[Detailed description of the invention]

Fig. 1 sentence understands prototype pattern figure；

Fig. 2 corrects operation workflow schematic diagram；

Fig. 3 self-learning module and pre-selection block process schematic diagram；

[specific embodiment]

A kind of mother tongue personage learns the artificial intelligence approach of another spoken and written languages, below to grasp the Chinese people of Chinese Chinese Scholar, each enterprises and institutions professional technician and student enrollment.They have grasped Chinese Chinese, there is corresponding professional knowledge With certain english foundation, they need to learn or pursue the study of English, and reference read English data is Chinese personage.Below Referring to attached drawing, using them as mother tongue personage, it is embodiment that study read with reference to English data using English.Divide and 1. reads reason Solution is moved on on computer and is carried out；2. software marks sentence cabin and cabin eye；3. the meaning of one's words is identical between template and component and mutually maps；4. tearing open Dividing source statement is component of expressing the meaning；5. conversion is expressed the meaning, component is sentence of communicating with the eyes；6. corrigendum and machine learning；7. multinomial intelligent use Software realization etc. the contents of the present invention are further described below:

One, reading understands to move on on computer and carry out

When human brain reads this text (books) for leaving various markings again, connection moment shows between block segmentation, block, Reviewing reading understanding becomes simple and quick.

The reading understanding of human brain is moved on on computer and is carried out, many people have been to do so；In computer or net Read electronic documents on network.Key problem in technology is how to retain reading to understand " marking "；Firstly the need of the depth for determining a understanding Degree and purpose.Our purpose is to be associated different language text with " expressing the meaning ", and bilingual sentence pair is marked and is expressed the meaning accordingly Component.The reading of mother tongue personage understands that weight marks bilingual sentence pair with hook and expresses the meaning component as degree, can mark out a cabin, cabin eye i.e. It can.However, it is very high for understanding depth requirements not to human brain, as long as finding out the identical word string expressed the meaning in two sentences.By soft The part mother tongue personage not high using electronic dictionary progress automatic identification or L proficiency, can accomplish.

The document of the ready-made identical content different language text version of computer and networks is good sample portion, Secondary, the document of written identical content different language text version enters computer by scanning recognition or keyboard input.With Human brain afterwards, which is read, to be understood and can carry out on computers.

In the software interface of this method, four regions are set.Bilingual text A and B are read in software by software respectively, are set respectively In the area source document A and B.Then, it according to the cutting of end of the sentence symbol and takes out mono- sentence pair of bilingual text A and B from the area source document A and B and sets and read Reading understands area A and B, and the area source document A and B accordingly reduces one.Certain special circumstances are clicked if mother tongue personage thinks to cut sentence improper Source document area A or B point, software, which moves into part before the point to read, understands that area follows closely after A or B taken sentence.If clicking reading Understand that area's A or B point, software will partially retract source document area A or B after the point.

After getting ready, it is necessary first to search for matched sentences bone for sentence pair, have matched sentences bone and be inserted in rear each sentence cabin and cabin Eye is well-balanced, without it is incomplete, without stacking skipping for phenomenon；No matched sentences bone has incompleteness, has the sentence pair for stacking phenomenon just to be understood Mark.It carries out operating with sentence bone first.With sentence bone in advance using languages as unit, by the void nucleation sentence bone of sentence bone word string and sentence cabin String list simultaneously indexes, such as " I know { 1 } got crush on { 2 }, { 3 } you could { 4 } " etc. make a bone to sentence bone Word string " I know...got crush on... ... you could.... "；Sentence bone string list at least " sentence bone word string, sentence bone Code (sentence bone library recording mechanism), bone are first (such as example=" I ") " field；It indexes.Sentence bone word string forms a complete sentence bone word section such as by the segmentation of sentence cabin " I know ", " got crush on ", ", ", " you could ", " "；Notice that ", ", " " are equal with " got crush on " It is a sentence bone word section.This just constitutes English sentence bone string list, can be used for the search using English as original sentence.For example " I knows sentence bone again Road morning dew { 1 } likes upper { 2 }, and { 3 } your { 4 } are seen." etc. make a bone word string " I, which knows ... likes upper ... ... You ... have."；It tabulates jointly index with sentence bone code, bone lead-in section.Sentence bone word string forms a complete sentence bone word section as " I knows by the segmentation of sentence cabin Road ", ", ", " you ", " is seen " on liking "."；This just constitute in sentence bone string list, can be used for the search using Chinese as original sentence.

Operated with sentence bone --- in advance by the void nucleation sentence bone string list in sentence bone word string and sentence cabin and after indexing, match sentence bone When A sentence example from left to right English by word for word removed in word search sentence bone string list bone lead-in section (English bone containing sentence word string first A word or symbol；In containing first character or punctuation mark) and be stored in interim table, then interim table content is taken out by record, with sentence Bone word section inquires sentence pair example, and every section of sentence bone string can be found in sentence pair example and order is mutually all matched sentences bone, further according to The sentence bone code field of sentence bone string list takes out corresponding languages sentence bone.With A sentence data search, A, B sentence bone are taken out；Current sentence pair example is right Ground of number taking a seat all is inserted in.

Matched sentences bone is searched with the running of sentence bone, current sentence pair example is inserted in and is shown provides " cabin detection " and " cut distribution mark " Button, think after operator reads each sentence cabin and cabin eye it is well-balanced, without it is incomplete, without clicking " cabin when stacking phenomenon and the correct meaning of one's words Detection " button, expresses one's approval, and software connects cabin detecting step., if it is not, not approving, " cut distribution mark " button is clicked, it is soft Part enters the running of cut distribution mark, sees whether B sentence contains with A words and phrases string inquiry sense-group string library (or with conditional electronic dictionary), generates Current sentence pair meaning of one's words matching list；It gets ready for mark sentence cabin, cabin eye.

If searching A sentence has matched sentences bone and obtains skipping for operator's approval after being inserted in, it is not necessary to understand mark, read in Next sentence pair.If not searching A sentence has matched sentences bone or has searched but cut when being denied after being inserted in by operator After mark running, into the step of understanding mark sentence cabin.

Cut distribution mark running --- cut distribution mark operate machine utilize containing part of speech, A words and phrases, B words and phrases field meaning of one's words matching list, First using A sentence sub-word string as unit cutting, successively fill matching list A words and phrases field, then by record take out inquire sense-group string Library or conditional electronic dictionary explain whether search B sentence contains with the corresponding B language checked in, contain and are longest string filling B words and phrases Its A words and phrases is inserted part of speech field simultaneously by field, is sky without B words and phrases field is contained.One A word string has a plurality of B language to explain, Their A words and phrases word field is identical.This just performs the preparation in identification mark sentence cabin, and sentence cabin counter (N=0) is reset.It connects The step of continuous mark sentence cabin.

Two, software mark sentence cabin and cabin eye

The essence of spoken and written languages is to express the meaning, meaning mankind's intercommunication of institute's table；Sentence is can express the complete meaning of one's words substantially single Position.The sentence of different language text can express the identical meaning of one's words.Different language text is expressed the meaning by several common structures of expressing the meaning Part is realized.However, multilingual text is associated with expressing the meaning, so that it may mark off their corresponding components of expressing the meaning.Sentence Subordinate sentence bone and sentence cabin two parts, sentence bone are the bone and frame of sentence, sentence cabin be on a bone frame flexibly, the cabin that is often replaced. They are stems as gap-filling questions, sentence bone；Sentence cabin is the sky on stem., as mathematical formulae, sentence bone is formula for they, and sentence cabin is to become Amount.However take different word string replacement sentence cabin contents that can generate many new sentences；In other words, sentence bone is a kind of sentence Bone and frame.It is to imitate, attempting to look for a bone, with sentence bone that child, which babbles,；Adult wording and phrasing rap off, and are Use the process of sentence bone；Conventional machines translation is formed a complete sentence with word, word combination, and shortcoming is exactly sentence bone frame.

Sentence cabin is marked, identical decide whether as a sentence of whether expressing the meaning according to current sentence pair, using A, B sentence word string Cabin.Wherein part of speech is a reference, is most successively frequently as sentence cabin or most suitable as the word string in sentence cabin: quantity string indicates The word string or several continuous word strings of number, amount；Proper name string, proprietary name word string；Using names string, the noun word string carried by article are a variety of Spoken and written languages have article (can make distinguishing mark)；Without using names string, the noun word string carried without article；Other meaning of one's words matching strings, Other word strings in addition to preceding 4 kinds of word strings, if express the meaning in sentence pair it is identical all can serve as a cabin, in bone example than suitable model In enclosing.Mother tongue personage reads understanding or software automatic identification is successively searched for quantity string, proper name string, using names string, bilingual same When mark a sentence cabin in advance.It is preferred between sentence cabin and sentence cabin with there is word string to be separated by, no word string is mutually divided into sequence sentence cabin；Only allow Two cabins are connected.Once A, three cabins of one of B language appearance are connected, software prompts " doing over again " immediately, otherwise cannot continue to run.A, B language There are two sentence cabins sequentially simultaneously in sentence, such as A language has 2,3 cabins sequentially；B language also has 2,3 or 3,2 sequentially, and Ying Gonghan Word string (including interval word string) <=5, software prompt operator consider to merge；It is only " consideration ".

Bone example is than the mark sentence cabin obtained for prolonged and repeated many experiments, the index of cabin eye；It is that a bone and sentence example contain word Accord with the ratio of number.The first round marks A language choosing English, and B language chooses text.It is calculated with English character number, if a Sino-British sentence pair, English Chinese character sum (including punctuation mark) is 100；Starting its bone example ratio when mark is 100%；After marking a sentence cabin, in cabin Containing 20 characters, bone example ratio=20/100 at this moment is equal to 20%；A sentence cabin is marked again, and total number of characters in sentence cabin increases To 45, bone example compares equal to 45%；And so on.Second wheel starts to know as DNA replication dna to have marked sentence as template Sentence cabin and the cabin eye of non-poster sentence are not marked, and bone example is than still with the calculating of English sentence.

Sentence pair is marked out a cabin, and there are two types of methods:

One of method is read by operator, mother tongue personage and understands sentence pair, the consistent word string that will express the meaning in sentence pair or continuous Several word strings be set to a sentence cabin, with its head and the tail of mouse-click, A, B sentence carry out simultaneously；Software will be clicked two points it Between content, be labeled as a sentence cabin；One distich cabin of mark marks second pair ... again.It is disclosed in preceding application 2 " a kind of production method of sentence component ", component of expressing the meaning is the progress of sentence component, and labeling operation method and step is consistent, operator Can refer to above-mentioned most suitable sentence cabin word string of making successively to consider to search for, using them as sentence cabin, when bone example than it is suitable when terminate mark Sentence cabin.Which is not described herein again.

The two of method are pressed the above-mentioned word string most suitable as sentence cabin by software automatically, and successively search is marked as sentence cabin Note.It with English sentence is from left to right that sequence changes and pre- is designated as formally marking after up to standard first to be successively that sequence is marked in advance.Second wheel starts with Sentence through marking is that template carries out identification mark to non-poster sentence, and sentence cabin label also replicates simultaneously；This is also different language text One of identical technical characteristic guarantee of the similar component meaning of one's words of expressing the meaning of word.

Using one of above method, mother tongue personage, which reads, understands that clicking reading with mouse simultaneously understands the current table in A and B, area They are added brace same at A and B by identical word string of anticipating or several continuous word strings, software according to successive with capitalization for sequence When be designated as a sentence cabin in advance.

Using the two of the above method, software automatically identifies A and B currently express the meaning identical word string or several continuous words String is by them with capitalization according to successively adding brace at A and B for sequence while being designated as a sentence cabin in advance.

Automatic identification marks the running in sentence cabin, such as sentence pair example in advance:

" 800 jin in 1970 of In Hengtung County, its per-mu_grain_yield surpassed, double that before 1965.

Hengdong County is more than 800 jin in every per mu yield in 1970, is pervious twice of nineteen sixty-five."

The current sentence pair meaning of one's words matching list of machine queries, first search identification quantity string, have, by it not in A, B sentence In simultaneously according to N=N+1 with " A, B, C ... " in advance mark；This example has 3 pairs of quantity strings, is designated as the 3 sentence cabins " A, B, C " in advance, such as:

" In Hengtung County, its per-mu_grain_yield surpassed A { 800 jin } in B { 1970 }, double that before C { 1965 }

Hengdong County is more than A { 800 jin } in B { 1970 } every per mu yield, is C { nineteen sixty-five } pervious twice."

Without or identification search complete sentence quantity string after identification search proper name string, have, equally by it not in A, B sentence It is marked in advance according to N=N+1 with " A, B, C ... " simultaneously；This example has 1 pair of proper name string, and connecting is designated as a cabin " D " for example in advance:

" In D { Hengtung County }, its per-mu_grain_yield surpassed A { 800 jin } in B { 1970 }, double that before C { 1965 }

D { Hengdong County } is more than A { 800 jin } in B { 1970 } every per mu yield, is C { nineteen sixty-five } pervious twice."

Without or identification search complete sentence proper name string after identification search using names string, have, equally by it not in A, B sentence It is marked in advance according to N=N+1 with " A, B, C ... " simultaneously.This example without；Without or identification search complete sentence using names string after show symbol " | " and " ← " is shown respectively in two sides, " → " order button, bone example ratio and " √ " order button is calculated and be shown, and receives operator Corrigendum is approved；Show that bone example compares for " 64% " simultaneously；Index of this example because not reaching 15-50% needs to continue identification mark Infuse sentence cabin；Receive operator and click " √ " button, continues identification mark.

It is bilingual successively while to identify quantity string, proper name string, using names string above；After the identification of these three word strings, bone is calculated Example ratio, together with pre- mark as a result, number order button is shown together provides.If these three strings have all been searched for and bone example compares in 15- Between 50%, pre- sentence cabin of marking terminates.If these three strings have all been searched for and bone example ratio > 50%, then with no using names string, Qi Tayu Meaning matching string continues the pre- mark sentence cabin of successively search, until bone example compares between 15-50%, pre- sentence cabin of marking terminates.Mother tongue personage's core Confirm in fact or give directions and corrigendum is executed by software.Without corrigendum, bone example is than exceeded, such as this example, then continues identification mark.Such as Fruit bone example compares higher than 15-50%, is that part of speech field is noun and A words and phrases B words and phrases field is not empty for language according to searching for matching list It anticipates matched no using names string, or is not noun and not empty match for other meaning of one's words of A words and phrases B words and phrases field is gone here and there.Ibid and add Show current sentence pair meaning of one's words matching list, " | hand | ", " ∨ ", " < ⌒ ", " format checking " button.

This example continues identification without using names string and other meaning of one's words matching string, it is noted that now begins to be every identification one sentence of mark in advance Cabin is suspended, and interaction approval or corrigendum are waited.If this example searches for no using names string from left to right, there is " per-mu_grain_ Yield ", " every per mu yield ", and there is word string to be separated by with the forward and backward sentence of mark cabin, thus further pre- mark sentence cabin " E ", such as:

" In D { Hengtung County }, its E { per-mu_grain_yield } surpassed A { 800 jin } In B { 1970 }, double that before C { 1965 }

D { Hengdong County } more than A { 800 jin }, is C { nineteen sixty-five } pervious twice in B { 1970 } E { every per mu yield }."

(continued)

At this moment bone example is more up to standard than for " 42% ", but do not have=< 15% can terminate to know by operator according to meaning of one's words situation Not, " format checking " button is clicked；It can also identify that pre- mark clicks " √ " button again；Also other order buttons can be used to carry out more Just.The order button of corrigendum operation and above-mentioned display is associated, and it is as follows to patrol column:

" ← | → " button: it is clicked when A or B sentence and is inserted into display " | " then differentiation " ← ", " → " button in click-point Whether clicked；When " ← " is clicked, the word string on " | " left side is moved to left, removes its left side word string if " | " is in sentence cabin Sentence cabin；If " | " sentence out of my cabin if by its left side word string move into sentence cabin.When " → " is clicked, the word string on the right of " | " is moved to right, If " | " is in sentence cabin by it on the right of word string remove sentence cabin；If " | " sentence out of my cabin if by it on the right of word string move into sentence cabin. The number of sentence cabin content is corrected whereby.

" √ " button: approve that sentence cabin is marked in current identification in advance, continue to identify and mark new sentence cabin in advance.

" | hand | " button: manual identification sentence cabin, while the head and the tail for being intended to mark a cabin are respectively clicked in A, B language sentence to be corrected, it is then single Hit " | hand | " button, they are marked into a distich cabin in advance again, it is automatic to modify bone example ratio.

" < ⌒ " button: clicking to go back on one's word finally identifying pre- target sentence cabin every time, and reduction is pre- to the end to mark preceding state and bone Example ratio；It repeats until leaving out completely.

Sentence cabin is marked in the identification of above-mentioned sentence pair example in advance has part content to be also not directed to, then is illustrated below:

Such as read in sentence pair " for three years, there is been a running fight between the Tory and Labour members of the Housing Committee about raising council ho House rents. ", " over 3 years, it is mutual that the Conservative Party member and Labour Party member of the housing committee just improve corporations' rent matters The warfare of pursuit and attack is mutually carried out."

Quantity string, proper name string, using names string are successively identified simultaneously automatically as above-mentioned；Wherein there is prompt: " using names string ' running ' is mismatched！"；After these three word strings identify pre- mark sentence cabin, it is 59% that display bone example, which compares, marks result in advance:

" for A { three years }, there is been a running fight between B { the Tory } and C{Labour members}of D{the Housing Committee}about raising council house rents."；

" come within A { 3 years }, the B { the Conservative Party member } and C { Labour Party member } of D { the housing committee } just improves corporations' rent Matters have carried out mutually the warfare of pursuit and attack."

There is the content that upper example be not related in this example: ibid and add show current sentence pair meaning of one's words matching list, " | hand | ", " ∨ ", " < ⌒ " button；Receive operator and uses the sense-groups alignment schemes such as group compound word, the extension meaning of a word, adhesive tape leading or following word word method Matching list is modified, the supplement meaning of one's words matches word string, continues pre- mark sentence cabin.

There is prompt " using names string ' running ' mismatch in number, special, three kinds of pre- marks of word strings identification of hat！"；See matching list, language " running " corresponding languages B field in kind A is sky so not matching also；Look into the explanation of conditional electronic dictionary " running " Have: " n is run；Race；Operating；Rotation；Outflow；Adj is run；Constantly；Connection；Flowing；Race ".They are all Without containing without matching in search B sentence.According to the meaning of one's words of current sentence pair example, it expresses the meaning of " pursuing and attacking "；It is " to run quickly Run ", the meaning of a word of " race " extension.The operation for meeting " sense-group alignment " requires (after in detail).So augmenting " pursuing and attacking " in languages B field (part of speech field filling " t " show it is other add part of speech, similarly hereinafter)；" running ", " pursuing and attacking " is set to be matched as sense-group string.However it is preced with Name string " a running fight " is matched with " warfare of pursuit and attack "；It is identified to be designated as a cabin " E " in advance；It is 48% that bone example, which compares,.Such as:

" for A { three years }, there is been E { a running fight } between B { the Tory}and C{Labour members}of D{the Housing Committee}about raising council house rents."；

" come within A { 3 years }, the B { the Conservative Party member } and C { Labour Party member } of D { the housing committee } just improves corporations' rent Matters have carried out mutually E { warfare of pursuit and attack }."

See the big section in rear portion one of sentence pair, without sentence cabin, while bone example is than 48%；It can continue to identify pre- mark sentence cabin.Thus See matching list, after " raising council house rents " and " improve corporations' rent matters " wherein due to " council " fails to match.Conventional dictionaries it have " n, senate；Council；Council；Council；Council；Committee Member's meeting；Council；The meaning of a word such as Congressman "；The meaning for expressing " corporations " herein is that the meaning of a word of original word string extends.Meet " sense-group The operation requirement of alignment ".So being explained in languages B field supplement " corporations "；When mother tongue personage clicks " √ " button, continuation is automatic Identification is designated as in advance:

For A { three years }, there is been E { a running fight } between B { the Tory}and C{Labour members}of D{the Housing Committee}about F{raising council house rents}.

Come within A { 3 years }, the B { the Conservative Party member } and C { Labour Party member } of D { the housing committee } { improves corporations room with regard to F Rent matters } carry out E { warfare of pursuit and attack } mutually.

At this moment it is 26% that bone example, which compares,；Sentence cabin is reasonably distributed (interval is generally 1-5 string between cabin).The pre- mark of identification terminates；It searches Rope, which receives " format checking " button, to be clicked.Carry out format checking running.Then formal mark are as follows:

3 { the of for 1 { three years }, there is been 2 { a running fight } between Tory}and 4{Labour members}of 5{the Housing Committee}about 6{raising council house rents}.

1 comes for { 3 years }, and 3 { the Conservative Party members } and 4 { Labour Party members } of 5 { the housing committees } { improve corporations room with regard to 6 Rent matters } 2 { warfares of pursuit and attack } have been carried out mutually.

" format checking " button: terminate the corrigendum operation of current sentence pair, sentence cabin, sentence bone format checking are carried out, with 16 systems Number replacement capitalization from left to right pre- is designated as formally marking with A for sequence arrangement is aforementioned.Bilingual sentence pair through formally marking Each cabin is further detected, if there is the big sentence cabin greater than original 5 word strings that need to further mark cabin eye or group compound word, marks cabin Eye also with quantity string, proper name string, using names string, without using names string, other meaning of one's words matching string be followed successively by target word string, mould example is than (imitating bone Example ratio) between 50-70%, ophthalmically acceptable 16 binary digits in cabin from left to right add square brackets to mark with A for sequence.

Sentence cabin is detected one by one, and 5 original word strings of=> will enter group compound word running.Group compound word --- according to current sentence pair, Sentence cabin meaning of one's words judgement, as one needs original, there are two entries or more to merge, and when the meaning of one's words is just equal to another, the former is with "-" They are connected and merges into an entry, referred to as compound word.In other words, when two or more original word strings it is total it is semantic not Acquisition can be added by word string justice is connected to compound word (multiple string) with " _ ", such as: " works little " " work is slack ", in sentence pair " the unsuccessful person, on the other hand, works little and just waits to see pass by."；" and loser works slack, opportunity quietly mistake soon." in express the meaning it is identical.Although " works " has " work " Justice, but " little " does not have the meaning of " sluggishness "；Their total semanteme cannot be added by word string justice and be obtained, therefore is connected with " _ ", group For compound word.For another example: knew_nothing_about_it knows nothing；Compelled_to_go simply must go；Late_at_ The night late into the night；Works_little work is slack；Etc..

If original 8 word strings of=> with current sentence cabin make " cut distribution mark " running generate current sentence cabin meaning of one's words matching list with into Onestep extraction cabin mould is to have mould sentence cabin, in the following example:

1 { the American } 2 { economic } system is, organized around, 3 { a basically Private-enterprise }, 4 { market-oriented economy } in which 5 { consumers } largely determine 6{what shall be produced}by 7{spending their money in the marketplace for those goods and services that they want most}.

1 { U.S. } 2 { economic } is with 3 { basic private firms } and 4 { market oriented economies } for framework, this In 2 { economic }, 5 { consumers } are to great extent by means of 7 { on the market for those their most desired cargos and services pay } To determine 6 { what should be manufactured }.

Successively sentence cabin is detected one by one, and 1-6 sentence cabin is all not above 8 original strings；Sentence { the spending their of cabin 7 Money in the marketplace for those goods and services that they want most } it is super 8 original word strings are crossed, makees the running of cut distribution mark and generates current sentence cabin meaning of one's words matching list in case further extracting cabin mould is to have mould sentence cabin.

Extracting cabin mould, also with quantity string, proper name string, using names string it is bilingual and meanwhile successively identification, be designated as cabin eye in advance, such as currently Cabin is noted as:

spending their money in A[the marketplace]for those goods and services that they want most

It is those they most desired cargo and services pay in A [in the market]

At this moment mould example is than 81%；Greater than 50-70%, then with no using names string or other meaning of one's words matching string in spaced premise Under be designated as cabin eye one by one, mother tongue personage except approving that corrigendum is outer, can also use a group compound word, the extension meaning of a word, adhesive tape leading or after with The method of the sense-groups such as words alignment modifies matching list, and the supplement meaning of one's words matches word string, continues pre- mark cabin eye B are as follows:

spending their money in A[the marketplace]for those B[goods and services]that they want most

It is that those their most desired B [cargo and service] pay in A [in the market]

At this moment mould example is than 65%；Mould example compares between 50-70%；It sees meaning of one's words needs, can terminate or mark in advance again sentence Cabin C；Again in advance mark cabin eye C must mould example than also between 50-70%；Otherwise go back on one's word reduction, mark eye terminates.

spending their money in A[the marketplace]for those B[goods and services]that C[they want most]

It pays in the B [cargo and service] that A [in the market] is those C [they are most desired]

At this moment mould example is than 53%；Also between 50-70%, allow.Finally, being from left to right sequence modification cabin with English sentence cabin Eye is designated as formally marking in advance, such as current sentence cabin formally marks are as follows:

spending their money in 1[the marketplace]for those 2[goods and services]that 3[they want most]

It pays in 2 [cargo and the services] that 1 [in the market] was those 3 [they are most desired]

Current sentence pair is included in current sentence cabin, is become:

1 { the American } 2 { economic } system is, organized around, 3 { a basically Private-enterprise }, 4 { market-oriented economy } in which 5 { consumers } largely determine 6{what shall be produced}by 7{(623786)spending their money in 1[the marketplace]for those 2[goods and services]that 3[they want most]}.

1 { U.S. } 2 { economic } is with 3 { basic private firms } and 4 { market oriented economies } for framework, this In 2 { economic }, { (623786) are the 2 of those 3 [they are most desired] in 1 [in the market] to 5 { consumers } to great extent by means of 7 [cargo and service] payment } determine 6 { what should be manufactured }.

Be above the first round sentence cabin and cabin eye mark the step of, second wheel sentence cabin and cabin eye mark start, do not have to according to word string But using poster kind sentence as template software automatic identification, mark non-poster kind sentence and accordingly currently express the meaning identical sentence cabin and cabin Eye；

Every wheel mark selects the corpus of the identical AB bilingual text version of content, and the first polling is most representative English-Chinese Two kinds of spoken and written languages, the second wheel start every wheel mark and increase a kind of spoken and written languages newly, and another kind is the spoken and written languages marked, A language distributes to the spoken and written languages marked, and B language distributes to the spoken and written languages being newly added.

Three, the meaning of one's words is identical between template and component and mutually maps

Software understands current reading in this reading understanding process that must be dominated by human brain the bilingual sentence pair mark in area Sentence cabin, cabin eye on note are that sentence understands template to retain current sentence pair to read the achievement understood.

Sentence understands that template includes sentence example；The sentence cabin included by brace；Sentence bone outside brace.And by side in big sentence cabin The cabin eye that bracket is included；Cabin mould outside square brackets.Original several word strings or connected multiple by group compound word are intraocularly contained in sentence cabin and cabin Word string, they are the bilingual sense-group strings down to the matching alignment of the multilingual meaning of one's words.Sense-group string is the component of expressing the meaning of a kind of fillibility, is filled out It fills among sentence cabin or cabin eye.Though cabin mould has the case where frame, be filled in big sentence cabin, itself also have fillibility, fill In sentence cabin.Sentence bone is the component of expressing the meaning of frame-type, and sentence bone is that sentence understands that the unique fixed part of template, including cabin label and flower include Number, the part or example that other all on-fixeds can be replaced.There are some special sentences, they do not include the meaning of one's words as described above The sentence cabin of alignment is matched, the idiom used that express the meaning directly is matched with sentence.They become the member that sentence understands template jointly.With For upper example, such as: "

1 { the American } 2 { economic } system is, organized around, 3 { a basically Private-enterprise }, 4 { market-oriented economy } in which 5 { consumers } large] y determine 6{what shall be produced}by 7{spending their money in 1[the marketplace]for those 2[goods and services]that 3[they want most]}.

1 { U.S. } 2 { economic } is with 3 { basic private firms } and 4 { market oriented economies } for framework, this In 2 { economic }, 5 { consumers } are to great extent by means of 7 { in 2 [the cargo kimonos that 1 [in the market] was those 3 [they are most desired] Business] payment } determine 6 { what should be manufactured }."

This is that current sentence pair reads the achievement understood.It is carried out since above-mentioned reading understanding has been moved on computer, It is labeled by software, is realized so " being transferred to computer " is naturally implicit on computers.This software by human-computer interaction Means make the reading of mother tongue personage understand the achievement for leaving more, the unified and standard marking；Be " mother tongue personage ", " sentence is single Under the conditions of member " and " being to understand depth with A-B sentence transition translation " three finite constraints, finite goal is generated --- sentence understands Template 100.Sentence understands template as shown in Figure 1, there is A sentence comprehension template 101；B sentence comprehension template 102；This is first round reason Contained spoken and written languages type when solution mark, and a certain segment in long company later；Every wheel understands the newly-increased one kind of mark later Spoken and written languages, it will gradually increase, and multilingual sentence understands that template can form long-chain.The sentence of spoken and written languages A 103, passes through A Sentence comprehension template 101, by their member, express the meaning component and another spoken and written languages, the structure of expressing the meaning of B sentence comprehension template 102 The sentence 104 of part generation spoken and written languages B.In short, the sentence of garden spoken and written languages A103, can convert out the constant language of the meaning of one's words The sentence of text B；It is reverse same, by the sentence of spoken and written languages B104, the constant spoken and written languages A103 of the meaning of one's words can be converted out Sentence.Sentence understands that template includes sentence example, such as: "

The American economic system is, organized around a basically private- Enterprise, market-oriented economy in which consumers largely determine what shall be produced by spending their money in the marketplace for those goods and services that they want most.

The economy in the U.S. in this economy, is consumed using basic private firm and market oriented economy as framework Person is to great extent by means of determining that should be manufactured on the market for those their most desired cargo and services pay Out."

Including the sentence cabin included by brace, as: " the U.S. the American "；" economic is economical "；"a Basically private-enterprise basic private firm "；" market-oriented economy market orientation It is economical "；" consumers consumer "；" what should be manufactured what shall be produced "；"spending their money in the marketplace for those goods and services that they want Most is those they most desired cargo and services pay on the market "；

Sentence bone outside including brace, such as "

{ 1 } { 2 } system is, organized around { 3 }, { 4 } in which { 5 } largely determine {6}by{7}.

{ 1 } { 2 } are with { 3 } and { 4 } for framework, and in this { 2 }, { 5 } are determined to great extent by means of { 7 } {6}。”

And the eye of the cabin including being included in big sentence cabin by square brackets, " the marketplace is in the market "；"goods and Services cargo and service "；" they want most they most desired "

With the cabin mould outside square brackets, "

spending their money in[1]for those[2]that[3]

It is [2] payment of those [3] in [1] ".

Sentence cabin and cabin intraocularly contain original one or several word strings and by the connected compound word strings of group compound word, they be it is bilingual with To the sense-group string of multilingual meaning of one's words matching alignment, " U.S. the American " as listed above；" economic is economical "； " a basically private-enterprise basic private firm "；" etc..Separately there are some special sentences, they Sentence cabin not comprising the meaning of one's words as described above matching alignment directly matches the idiom used that express the meaning with sentence, such as: "

" Buddhist monk carries water and drinks, and two Buddhist monks lift water and drink, and three Buddhist monks do not have water to drink."；" one boy is a boy, two Boys half a boy, three boys no boy. "；" there can never be too much deception in war "；"There can never be too much Deception in war. " etc.；It does not express the meaning the equal word string for dividing a cabin into, a bone can not be marked out, sentence cabin belongs to Idiom.Chinese idiom, common saying, proverb, the two-part allegorical saying etc. of Chinese are nearly all idioms.

As above-mentioned, sentence understands that template includes sentence example, sentence cabin, sentence bone, cabin eye, cabin mould, sense-group string and idiom, they are total The member of template is understood for sentence.As shown in Figure 1, A sentence comprehension template 101, B sentence comprehension template 102 respectively containing this 7 kinds at Member.To the member that expresses the meaning it is directly wherein component of expressing the meaning, there is a bone, cabin mould, sense-group string and four kinds of idiom；A, phase in B template The component of expressing the meaning answered has horizontal line to be connected, and indicates that their meaning of one's words are identical and can mutually map.Multilingual sentence understands that template passes through table Meaning component is connected to form the long-chain that a sentence understands template.

Wherein the horizontal line of connection " sentence bone " is extraordinarily thick indicates that it is the unique fixed member of template.There are four to be in 7 members It expresses the meaning component, that is to say, that subordinate clause, which understands, can at least decomposite a bone, four type table of cabin mould, sense-group string and idiom in template Meaning component.The essence of spoken and written languages is to express the meaning.Sentence is the basic unit that can express the complete meaning of one's words；Different spoken and written languages Sentence can express the identical meaning of one's words；Expressing the meaning for different language text is realized with identical component of expressing the meaning.

Sense-group string is the component of expressing the meaning of a kind of fillibility, be filled in a cabin or cabin eye among, including original vocabulary, term, Phrase and the word string added by the whole word of sense-group.Though cabin mould has the case where frame, be filled in big sentence cabin, also tool is filled out for itself Filling property is filled in a cabin.Sentence bone is the component of expressing the meaning of frame-type, including cabin label and brace.Sentence bone is that sentence understands in template Unique fixed member；Other members all on-fixeds are all the part that can be replaced or example.Idiom is to match carry out table with sentence The special sentence of one kind that meaning uses, it is one of component of expressing the meaning.Sentence example is not component of expressing the meaning, and sentence example is can to express the complete meaning of one's words Basic unit, be the sentence example of template.

Sentence understands template objective reality, regardless of whether people recognize and evaluate it.Child babbles, people's choice of word Sentence-making thoughts communication is all conscious or is unconsciously to understand template using sentence, with the process for component of expressing the meaning.This is that sentence understands The case where template is applied alone.In addition, one-to-one can use over the ground；It is one-to-one over the ground using can produce sentence between A, B spoken and written languages Conversion.Since the similar component meaning of one's words of expressing the meaning is identical and mutual phase mapping, so the sentence meaning of one's words of A, B spoken and written languages converted It is constant.

One distich understands that template includes many sentence examples.A kind of sentence with identical sentence bone can understand mould in this distich It is split on plate；It is assembled to come out.One distich understands template can cover the number of sentence example, is the representative that sentence understands template and sentence bone Property, represent the embodiment of ability.Obviously, sentence can classify according to its bone；Sentence understands that template can come from some sentence To example, and its unique fixed member, sentence bone can represent and cover a kind of sentence, can be applied by a kind of sentence.

Such as this distich example " The fisherman consents to return the feather suit, on Condition that fairy dance and play heavenly music for him. dances and drills for him in fairy maiden Under conditions of playing melody in the sky, fisherman promises to give back the garment or robe made of feathers.", referred to as " original example " is retained in sentence and understands template.Using it as sentence A pair of of template that example is generated: "

1 { The fisherman } consents to return 2 { the feather suit }, on condition that 3{(00205)1[fairy]2[dance]and 3[play heavenly music]for him}.

3 under conditions of { 1 [fairy maiden] is his 2 [dancings] and 3 [playing melodies in the sky] }, 1 { fisherman } promises to give back 2 { plumages Clothing }." and their sentence bone: "

{ 1 } { 2 } consents to return, on condition that { 3 }

Under conditions of { 3 }, { 1 } is promised to give back { 2 }.

"；Their sentence bone can represent and cover a kind of sentence, can be applied by a kind of sentence, can such as apply out: "

1 { The child } consents to return 2 { Ipad }, on condition that 3 { (00205) 1 [father]2[buy the toys]and 3[play games together]for him}.

3 under conditions of { (00205) 1 [father] is his 2 [buying toy] and 3 [playing game together] }, 1 { child } promises to return Also 2 { Ipad computers }.

1 { The boy } consents to return 2 { her bag }, on condi tion that 3 { (00205) 1 [his female classmate]2[prepare the homework]and 3[not inform others]for him}.

3 under conditions of { (00205) 1 [schoolgirl of the same class] is his 2 [doing the homework] and 3 [not informing other people] }, 1 { boy } is answered 2 { her school bags } should be given back.

1 { The employer } consents to return 2 { the delinquent wages }, on condition that 3{the workers not reveal}.

Under conditions of 3 { rural migrant worker does not accuse }, 1 { employer } promises to give back 2 { wages being in arrears with }.

3 { the of 1 { Tom } consents to return 2 { calculator }, on condition that homework is finished}.

Under conditions of 3 { completing homework }, 1 { Tom } promises to give back 2 { calculators }.

1 { Mother } consents to return 2 { his favorite doll }, on condition that 3 {the child get A greed in the next exam}.

Under conditions of 3 { child tests A grades of acquisition } next time, 1 { mother } promises to give back 2 { his favorite toys }.

"；Etc. many sentence pairs；For generate it is many between in English-or in-English sentence conversion.

Why claim sentence to understand template in view of its characteristic: 1) a kind of product is read for one by human brain and is understood by software mark More the sentence pair of the unified marking understands product on note.2) a kind of place, a kind of place that can be split sentence and assemble sentence.3) There are fixation means, a kind of special fixation means --- sentence bone, sentence bone is the bone and frame of sentence；Sentence bone is different language text Word is expressed the meaning indispensable composition of expressing the meaning.4) there are operating mode, a kind of understanding mode；It is a kind of based on sentence understand by A sentence The mode that son is converted to B sentence, or the operating mode for splitting A sentence, assembling B sentence.5) can be tired out with prolonged application Product can be shared.

It can be more particularly for understanding the other sentences of mark in order to which sentence understands template, understand that template carries out A, B sentence by A, B Conversion, they are stored in sentence and understand that template database replaces human brain memory by software, and effect is naturally higher.Moreover, it protects It is stored in sentence and understands that template can guarantee " sentence understands between template, component of the same race of expressing the meaning that the meaning of one's words is identical and mutual phase mapping ".Current data It is not so difficult that library technology will reach this purpose.It is to establish " sentence understands template database " first round understanding mark to choose most first The Chinese-English bilingual text of meaning is represented, should mutually there is two middle template, heroes and models' plate fields on the database.Example English-Chinese two as above A understands that template is stored in respectively under the corresponding language template field of current record.Its Chinese and English sentence understand template "

1 { The fisherman } consents to return 2 { the feather suit }, on condition That 3 { (00205) 1 [fairy] 2 [dance] and 3 [play heavenly music] for him } " is stored in currently Record " heroes and models' plate " field；In corresponding sentence understand template "

3 under conditions of { (00205) 1 [fairy maiden] is his 2 [dancings] and 3 [playing melodies in the sky] }, 1 { fisherman } promises to return Also 2 { garments or robes made of feathers }." it is stored in current record " middle template " field.This storage is executed by software, can be errorless with entirely accurate.And And as long as passing through database search, so that it may find the record, and can mutually be mapped between record field.Such as search " English Under template " field when precedent heroes and models, by mapping take out with record " middle template " field contents, certainly be work as precedent in Mould.Their meaning of one's words is centainly identical；This " certain identical " there are also correlation steps to guarantee, such as " every wheel mark takes same content, not identical text Letter this corpus " carries out；After obtaining template, step store method and step etc. described in this section.

Since the second wheel understands mark, every wheel increases a kind of spoken and written languages newly, and sentence understands that template database accordingly adds one Certain a language template field.Such as second take turns newly-increased Russian, adds " Russia's template " field, using Sino-Russian bilingual sample sentence pair, Sino-Russian two sentences, which are obtained, after understanding mark understands template: "

Does 1 { John } { make great efforts } 2 { work } 3 as 4 { Henry }?

1 { Р а б o т а e т } 2 { Д ж o н } т а к 3 { у с e р н o } к а к 4 { Γ e н л и }? "；With the Chinese marked Sentence understand template search to current record, it is seen that under " middle template " field " 1 { John } 3 as 4 { Henry } make great efforts } 2 { work }? "；Can also see that has " 3 { hard } as 4 of Does 1 { John } 2 { work } as under " heroes and models' plate " field { Henry }? "；Then current Russia's sentence is understood into template " 1 { Р а б o т а e т } 2 { Д ж o н } т а к 3 { у с e р н o } к а к, 4 { Γ e н л и }? " it is stored in " Russia's template " field.Hereafter, not only the meaning of one's words is identical by their threes, and can mutually hint obliquely at.

Sentence understands that template includes at least sentence bone, cabin mould, four kinds of sense-group string, idiom components of expressing the meaning.Here they are decomposited Come, be stored in corresponding component database respectively, corresponding component database equally has languages component field, is equally same record The meaning of one's words is identical and mutual mapping.Such as above example decomposites China and British, three kinds of sentence bones of Russia are stored in the corresponding China and British in bone library, Russia's sentence bone Field, such as: " Does { 1 } { 2 } as { 3 } as { 4 }? " it is stored in English sentence bone field；" { 1 } { 3 } { 2 } as { 4 }? " it is stored in Middle sentence bone field；" { 1 } { 2 } т а к { 3 } к а к { 4 }? " it is stored in Russia's sentence bone field.Other components of expressing the meaning, cabin mould, are practised sense-group string Language and so on.

These component databases of expressing the meaning are point libraries that sentence understands template database.Divide library to inherit total library, inherits sentence understanding The characteristic of template database, also identical with the record component meaning of one's words of expressing the meaning in point library, their also mutual phase mappings.Fen Ku and divide between library Independently of each other.Zong Ku is upper and lower concept and has the division of labor with base relation is divided；Total library storage sentence understands template, them are stored in a point library Component of expressing the meaning；Such arrangement, the applications such as search, access are convenient, flexible.

Four, splitting source statement is component of expressing the meaning

It is component of expressing the meaning that software, which understands that template 100 splits source statement using sentence, sentence is understood and split instead of human brain, first Operated with sentence bone.With the running of sentence bone in advance by sentence bone word string and sentence cabin void nucleation sentence bone string list and index, with sentence bone When source statement from left to right word for word or is removed the bone lead-in section of search sentence bone string list by word string and is stored in interim table, then by note Enroll out interim table content, with sentence bone word section inquiry source statement, every section of sentence bone string can be found in source statement and Order is mutually all matching, and the source statement bone of current matching is taken out further according to sentence bone string list.Such as:

Sentence bone is become a bone word string by languages in advance, as English sentence bone " draw { 1 } to divide { 2 } into { 3 } " becomes The bone word string that forms a complete sentence " draw to divide into. " is stored in English sentence bone string list (at least string of bone containing sentence, bone head, three word of sentence bone code Section) distich bone string field index (bone lead-in section " draw ").Source statement is from left to right taken word for word or by word string when with sentence bone The bone lead-in section of lower search sentence bone string list is simultaneously stored in interim table, then takes out interim table content by record, with sentence bone word section query source Sentence.When precedent takes out every section of " draw to divide into. " sentence bone string from interim table sentence bone string field can be It is found in source statement and order is mutually all matching.When 4 sections of precedent (wherein " " is also a section), they can be in source language It is found in sentence " Draw a diagonal line to divide the square into two triangles. " and secondary Sequence is identical, so matching.Further according to sentence bone string list current record sentence bone code field to sentence bone library take out current matching source language Sentence bone: " draw { 1 } to divide { 2 } into { 3 } ".

Then, by source statement " Draw a diagonal line to divide the square into two Triangles. it " is inserted in source statement bone " draw { 1 } to divide { 2 } into { 3 } " and splits source statement are as follows:

draw 1{a diagonal line}to divide 2{the square}into 3{two triangles}.

Their component of expressing the meaning is high-visible, it is easy to by software decomposite Lai.Such as example obtains source statement bone " draw {1}to divide{2}into{3}."；Sense-group string " diagonal ", " line ", " square ", " two ", " triangles " Deng.

This just obtains current source statement and understands template；Source language example is split into component of expressing the meaning.

Five, conversion expresses the meaning component as sentence of communicating with the eyes

After previous step, which splits into source language example, expresses the meaning component.Software understands that template 100 converts structure of expressing the meaning using sentence Part is sentence of communicating with the eyes, as shown in Figure 1, the component of expressing the meaning of A, B languages is corresponding one by one, has lines to be connected.Software is managed using source statement The sentence bone that sentence bone " draw { 1 } to divide { 2 } into { 3 } " in solution template searches sentence of accordingly communicating with the eyes " is drawn { 1 }, { 2 } It is divided into { 3 }.", current source statement is understood that the sentence cabin content in template is transferred to the corresponding sentence cabin of a bone of communicating with the eyes.Such as:

" 1 { a diagonal line } is drawn, 2 { the square } are divided into 3 { two triangles }."

All cabins are detected, without skipping for cabin mould, if containing cabin mould, this cabin content is inserted in cabin mould, utilizes the cabin The Mappings in mould point library take out cabin mould of accordingly communicating with the eyes, and current source language cabin is intraocularly held and moves on to the corresponding cabin of cabin mould of communicating with the eyes in the eyes. When precedent is free of cabin mould, continuation next step.

Sentence cabin or the cabin eye for searching for mesh sentence comprehension template one by one, search for sense-group string library with words and phrases string in source therein, will search Rope to all explanation results of communicating with the eyes be saved in correction sheet.Such as " a " Chinese explain have "an", "one", " one "； The Chinese of " diagonal " explains there is " diagonal line ", " twills "；" square " Chinese explain have " square ", " square ", " square ", " square ", " just ",；The Chinese of " triangles " explains there is " triangulable ", " triangle " etc. Correction sheet is enumerated Deng all searched.Correction sheet at least word string section, search string (example is English string), explanation, the first word of going here and there Section.Wherein word string section, that is, source document word string and its all explanations occupy continuous recording section of correction sheet.

Using pre-selection module, one of them replaces the source words and phrases string in current sentence cabin or cabin eye as sense-group string of communicating with the eyes for pre-selection, Until all cabins, cabin eye conversion end.

It preselects module and spoken and written languages type is closely related.The pre-selection 312 of english translation is 1. intraocular with current sentence cabin or cabin English word string inquires " experience selects word " library, looks into have and takes out the first item Chinese word string for explaining field contents as the mesh of 313 maximum adaptations Language string provides, look into be difficult correction sheet this word string section owner record be maximum adaptation string of communicating with the eyes provide；2. with current sentence cabin or cabin eye The Chinese content 314 obtained inquires the former sequence field in " middle word order " library, has looked into and has replaced with maximum fit with positive sequence field contents The string 315 of communicating with the eyes matched provides, and looks into without constant.In translate the pre-selection of English with sentence cabin in source, transition type of communicating with the eyes or cabin eye source language content It inquires correction sheet and searches for string field, have congruence explains that field contents are provided as the string of communicating with the eyes of maximum adaptation using same record, without complete 317 inquire " cabin experience " library middle deck field again whens equal, have 318 to provide using English cabin field contents as the string of communicating with the eyes of maximum adaptation, Look into " cabin experience " library be difficult current sentence cabin or cabin eye Chinese content be maximum adaptation string of communicating with the eyes provide；

All pre-selection modules selected correction sheet check in item change it is former provide content, while it being changed in correction sheet First of position to this word string section.It is described on such as to obtain sentence of communicating with the eyes:

" 1 { diagonal line } is drawn, 2 { squares } are divided into 3 { two triangles }."；It shows and provides as sentence to be corrected.

Six, corrigendum and self study

Start to correct when mother tongue personage reads communicate with the eyes sentence to be corrected and corresponding source language with reference to sentence, while the school of software Positive module (attached drawing 2) borrows correction sheet to carry out meaning of one's words correction running, and self-learning module 301 links, the human-computer interaction to correction course Learnt, remembered and provides data for pre-selection module；

Meaning of one's words correction has sentence to be corrected correctly without correcting and needing to correct two kinds of situations, to receive user when without correcting completely (referring to mother tongue personage together, similarly hereinafter) clicks the Back button and connects S4, and entering after correction operates for needing to correct connects S4 again.

The correction running 201 that enters for needing to correct is exemplified below:

A) the replacement 202 of improper word string: for example, the sentence to be corrected read as user and corresponding source language are with reference to such as:

If " your 1 { buying that house }, you will 3 where } tided over for 2 { your old ages }? "

" if you 1 { buy that home }, 2 { the_rest_of_your_life } 3 of will you spend { there }? "

When user thinks that sentence to be corrected word string " that " is improper, its software search correction sheet is clicked, pops up drop-down list Providing whole correlations, to check in item to be selected.It is first according to the character position number query search string, the string that are hit word string and its place sentence to be corrected Bit field is met record and is provided with drop-down list；Such as " that, due to, that, so ".

When list is a certain by user click, such as " that " is clicked, with current improper word string " that of " that " displacement sentence to be corrected It is a "；And " that " and " that " 2 is recorded into content transposition in correction sheet；Due to " that ", " that " Length discrepancy, also to modify with The value of record " string is the first " field afterwards.

When the Back button is clicked, above-mentioned corrigendum relevant information is recorded in experience, the standby pre-selection block search of dictionary is selected to make With；It finally returns to, carries out next translation running.

B) the cutting 203 in sequence sentence cabin --- when encounter provided without reference to the sequence sentence cabin that can not carry out cutting automatically, Operator is waited to intervene.If sequence sentence cabin contains 2 word strings, two sentence cabins are respectively taken first, automatic segmentation.In another example two sentence cabins It is connected, one of them is reiterant sentences cabin again, and with reference to another of reiterant sentences cabin, sequence sentence cabin content is marked and reiterant sentences cabin phase Same part, remaining part belong to another among sequence sentence cabin, automatic segmentation, it is not necessary to interaction.In addition to this case where, regards For " without reference to that can not carry out cutting automatically ".

Such as sentence " When will he go there, tomorrow or some_other_day? " to be translated half when running Way is stopped, and sequence sentence cabin " 1 { 2 { he go there } " occurs, without reference to can not automatic segmentation；So display " sequence sentence cabin, It please click cut-off: ":

" 12 { he go there } " and following information:

When will 1 { 2 { he go there }, 3 { tomorrow } or 4 { some_other_day }?

[1]+when+[2]+,+[3]+still+[4]+?

It obviously, should be " he ", " go there " by " he go there " cutting；Be placed to respectively [1]+when+ [2] in.

To be separated two sentence cabin contents as cut-off by click-point after " he go there " is clicked；Become:

" when will 1 { he } 2 { go there }, 3 { tomorrow } or 4 { some_other_day }? "

Continue to operate.

C) intervene optionally sentence bone 204 --- when encountering a cabin and cabin eye unshapeliness, there is incomplete or stacking mistake to select sentence bone table Now or user does not approve and clicks the button " after rear " when currently translating, and software provides multipair matched sentences bone to be selected, is selected with clicking Sentence bone after taking is inserted in connecting S6 again to be continued to operate；Such as sentence to be translated:

I went to see_my_doctor for a check-up yesterday.

Automatically sentence to be corrected is provided are as follows:

I go to see 1 { I has gone ■ to ask ■ physical examination yesterday of doctor }.

1{I went to see_my_doctor for a check-up yesterday}.

A sentence cabin is stacked to reference to sentence full content；Minimum " sentence has a bone, two, sentence cabin part " is not supported； Select sentence bone wrong, " after rear " button receives user click automatically；Multipair matched sentences bone is provided to be selected；

After user selects a sentence bone else, continues running and provides sentence to be corrected and respective sources language with reference to sentence are as follows:

3 { yesterdays }, I went 1 { asking doctor } to make 2 { ■ physical examinations }.

I went to 1{see_my_doctor}for 2{a check-up}3{yesterday}.

Corresponding subsequent action is connected with subsequent 213 by differentiating by user click according to mentioned order button." ■ " above It indicates the word string do not expressed the meaning, the exhausted choosing (similarly hereinafter) of list can be popped up after user feels improper and clicks.

D sense-group string 205) is added --- when improper word string is optionally replaced and lacked, receives user and use the extension meaning of a word, glue The band sense-groups alignment schemes such as leading or following word word add sense-group string, are replaced to add string and add language corresponding in sense-group string Kind component field.Such as sentence to be translated:

Now doctors have discovered a cause of ulcers.

Automatically the sentence to be corrected provided are as follows:

1 { present doctors } has found the reason of one kind leads to 2 { ulcer }.

It can not find constraint term in correction sheet as user click " doctors "；It is provided when user, which chooses, to consult the dictionary " doctors is that name answers or move 3 " prompt.Our law regulation, it is all to there are paradigmatic conduct neologisms to log in.However it is defeated to receive user Entering " doctors " is " doctors " and " doctors " add as sense-group string, is stored in sense-group string library.Sentence to be translated is replaced with simultaneously:

1 { present doctors } have found the reason of one kind leads to 2 { ulcer }.

Corresponding subsequent action is connected with subsequent 213 by differentiating by user click according to mentioned order button.

Such as sentence to be translated:

It usually happens before eating or during the night.

Automatically the sentence to be corrected and corresponding source statement provided is referenced as:

It typically occur in 1 had a meal in the past or 2 ... during ■ night.

it usually happens 1{before eating}or 2{during the night}.

When one of the two cabins of user click word string, the alternate item user that correction sheet is listed is not chosen.When with dragging Put and choose " before eating " to receive user they are formed multiple string " before meals " using the method for group compound word, replacement here and It adds in compound word library.Then user chooses " during the night " to inquire compound word automatically with drag and drop method again " during The night is at night " it provides；Receive user click selection, sentence to be corrected corrigendum at this moment are as follows:

It typically occurs in 1 { before meals } or 2 { at night }.

Such as sentence to be translated:

Doctors have been able to help lessen the pain of ulcers.

1 { doctors } can help 2 { ulcer for mitigating ■ pain } already.

1{Doctors}have been able to help 2{lessen the pain of ulcers}.

If user click " pain ", then click " shifting → " button for continuous 2 times, it moves to right 2 times, sentence to be corrected becomes:

1 { doctors } can help 2 { canker sores for mitigating ■ } already.

Then user click " " after click " move → " button, sentence to be corrected corrigendum are as follows:

1 { doctors } can help 2 { pain for mitigating ■ ulcer } already.

When user click the Back button, first differentiates that user uses " ← shifting " or " shifting → " button to move word order, open Dynamic self-learning module, the sentence cabin that will be currently moved through, content is saved in " middle word order " library before shifting, after moving.Then it returns.

G) compile word replacement 208 --- when much more continuous string is improper, it is inconvenient with click choose when, choose string with drag-and-drop operation Afterwards, when then filling out other word strings, editor at word replacement or be empty compiling, then clicking " ∧ " button, before software is replaced with the latter Person.

Such as sentence to be translated:

1911 Nobel Peace Prize for his work of Alfred Herman, he shared the toward world peace.

1 { Alfred Herman } obtains 3 { 1911 Nobels because of the contribution that 2 { they } are done by 4 { world peaces } Peace Prize}。

1 { Alfred Herman }, 2 { he } shared 3 { 1911 Nobel Peace Prize of the } for his work toward 4{world peace}.

When user thinks " Nobel Peace Prize "；And chosen with drag-and-drop operation, they are shown in and " choose by software Text box under string " and " compiling word replacement " printed words；Receiving is compiled as " year Nobel Peace Prize ".It mends and clicks when " ∧ " button, with The latter replaces the former, corrects sentence to be corrected are as follows:

1 { Alfred Herman } obtains 3 { promises in 1911 because of the contribution that 2 { they } are done by 4 { world peaces } Bell's peace prize }.

H) rhetoric 209 --- when there is the rhetoric phenomenon of inconvenient listed other step process, receives user click and " repair Diction " button, then copies to edit box for current sentence to be corrected, receives user using the method for editor and carries out rhetoric operation.

Such as when translating sentence or rhetoric phenomenon after other listed step process, such as:

Dickens ' language, at once rich colourful and varied, is like fine and sensitive musical instrument.

1 { Dickensian language } not only 2 { varied and colorful } but also 3 { changeful }, much like 4 { exquisite and moving sounds Happy tool }.

1 { Dickens ' language }, atonce 2 { rich colourful } and 3 { varied }, is like 4 {fine and sensitive musical instrument}.

When user click " rhetoric " button, current sentence to be corrected is copied into edit box, receives user and uses the side of editor Method carries out rhetoric operation.Sentence to be corrected of communicating with the eyes after operation are as follows:

1 { Dickensian language } not only 2 { varied and colorful } but also 3 { changeful }, much like 4 { exquisite and moving pleasures Device }.

In addition to function as described above, it can also scan aim at editor set by rhetoric before returning " return " order button Frame；When the edit box is not empty, the frame content is taken to return；Otherwise sentence to be corrected is taken to return.

J) mend quantifier 211 --- when english translation due to English quantifier-free and Chinese have quantifier thus need sentence to be corrected need Place is wanted to fill quantifier, the need for receiving user click sentence to be corrected add place, Shuo Chuan rear center of the software search closest to click-point Noun when " ▲ " button while being clicked, with head noun volumes of searches dictionary, takes out corresponding quantifier and add sentence to be corrected；If Nothing is looked into quantifier library, is provided dialogue and is received to input to add simultaneously typing quantifier library.Such as sentence to be translated:

24,1970, China successfully launched its first man-made earth of On April satellite.

1 { in April 24 }, 2 { 1970, China } have succeeded in sending up 3 { the first artificial earths satellite }.

3 { first man- of 1 { on April 24 }, 2 { 1970, China } successfully launched its made earth satellite}.

Head noun after user goes here and there numeric string or number is chosen using drag-and-drop operation.Such as user choose it is " artificial Earth satellite "；They are shown under " choosing string " and " compiling word replacement " by software；

When user click order button " ▲ ", with head noun (being " satellite " here) volumes of searches dictionary, take out corresponding Quantifier, " " add sentence to be corrected；At this moment sentence to be corrected becomes:

2 { 1970 } years 1 { in April 24 }, 3 { China } have succeeded in sending up 4 { first man made earth satellites }.

If nothing is looked into quantifier library, dialog prompt is provided: " looking into nothing；The addition in the form of ' amount/name ' can be displaced compiling word." such as Example input "/satellite ".

The above A-H state step is unrelated with the type of spoken and written languages, has common feature, can be applied to multilanguage two-way and turn over It translates；As english translation or in translate the two-way translation of English during.J state step and spoken and written languages type are closely related, are used for English In translating or use when other languages are translated into Chinese.Hereafter there are also I, K state step are related to the type of spoken and written languages.I state step It is rapid between no word among the translation of interval ideograph, such as in translate English；K state step translates English in being used for.I-K has individual character Feature is arranged according to specific languages.Illustrate by taking english translation as an example above；Hereafter I, K appear in translate English during.So hereafter To translate English example explanation in corresponding:

Such as sentence to be translated:

" theory is inherently to a kind of abstract and symbolism expression of the reality recognized."

1 { essentially }, a 2 { theory } is an 3 { abstract }, 4 { expression of symbolism } of what is conceived to be 5{reality}.

2 { theoretical } are in one kind 3 { abstract } and 4 { expression of symbolism } that 1 { substantially } is 5 { reality } to having recognized.

Sentence to be corrected has 4 sentence cabins；Wherein 3 sentence cabins have all turned into English；Only sentence cabin 4 also maintains source statement cabin content " 4 { expression of symbolism } "

When user click cabin content " expression of symbolism ", during generating correction sheet due to inquiry sense-group library, From left to right possible arrangement is used as word string to inquire sense-group library to the ideograph being spaced between no word, and all items that check in are stored in corrigendum Table." expression of symbolism " be divided into " symbol ", " symbolism ", " symbolism ", " ", the Chinese word string such as " expression ", check in Item is explained comprising they and they correlation English.It is successively chosen for user according to English habit.

Here user should first choose " symbolic symbolism "；" representation expression " is chosen again；However work as Preceding sentence cabin is disposed, current sentence to be corrected are as follows:

1 { essentially }, a 2 { theory } is an 3 { abstract }, 4 { symbolic representation}of what is conceived to be 5{reality}.

Regardless of whether there is or not other corrigendums；Software all detect mentioned order button and provide associated frame, table whether by user list It hits to differentiate subsequent 213, connects corresponding subsequent action.

K) mend hat it 212 --- translate in the middle inch due to Chinese without article, verb to/be, to/have distinguish etc., In translate inch, these needs are added；Software provide retain sentence cabin mark label communicate with the eyes sentence to be corrected when, it has been shown that " a/ An ", " the ", " to/be ", " to/have ", " ### ", when (being end mark except " ### ") is clicked in one of they, to school When sentence is also clicked, one of they are added the word string of appropriate form by software in sentence to be corrected by the place of clicking；

Such as sentence to be translated:

We are tired, this is obvious to all.

1 { we are tired }, as anyone can see.

1 { we are tired }, this is obvious to all.

Software may arrange inquiry sense-group string library with " we are tired ", and all items that check in are stored in the choosing of correction sheet standby user It takes.Here user can choose " we we " and " tired is tired "；At this moment sentence to be corrected becomes:

1 { we tired }, as anyone can see.

Such requirement for obviously not meeting English.User can click " mend and be preced with it " lower section and list " a/an the to/ Be to/have### " " to/be " therein；And the intermediate space " we tired " is clicked again.Software judgement needs exist for " mend and be preced with it " operation, and according to the requirement of " to/be ", it is automatically inserted into verb " are " here；However current sentence to be corrected becomes Are as follows:

1 { we are tired }, as anyone can see.

The state step for arranging and illustrating to need to correct is patrolled above.There are also a kind of situations, and the sentence to be corrected provided is completely correct, Operator does not click any of above button and the related page control for providing display, just clicks the Back button；Provide Sentence to be corrected it is completely correct, the case where without correcting；Such case the more is used the more more with the operation accounting regular meeting of self study.

It should be added that sentence to be corrected referred to above is all the mesh to remain a cabin, cabin eye mark, label Sentence comprehension template form provides.When " return " order button is clicked return, these cabins, cabin eye mark, label and Unwanted space will all discard, as traditional approach: such as: " 1 { we are tired }, as anyone can see. "； " 2 { 1970 } years 1 { in April 24 }, 3 { China } have succeeded in sending up 4 { first man made earth satellites }." etc. become: " we are tired, as anyone can see.

On April 24th, 1970, China succeeded in sending up first man made earth satellite." etc. return；It is given to and translates Literary frame, so that ultimately produce the output of target translation.Such as by providing the constant language of the meaning of one's words after the sentence running of spoken and written languages A103 The sentence of text B104.

Self-learning module 301 learns the human-computer interaction of correction course, is remembered and provides data for pre-selection module, Specific running includes following content.Such as:

A, experience selects word, 303 after in english translation meaning of one's words correction operation when the improper word string of user click sentence to be corrected, Software search correction sheet, list provide whole correlations check in item it is to be selected while, all related constraint terms are separated by life with space At through channeling test of cement sheath, end is attached corresponding English word string.While user chooses one of list displacement, with choosing entry and English string " experience selects word " library is inquired, frequency=frequency+1 has been looked into, has looked into nothing, entry will be chosen to move on to through first of channeling test of cement sheath and the deposit of pop down formula should The middle explanation field in library 305, English string remove deposit English string field from through channeling test of cement sheath, and current sentence bone number is stored in sentence bone field.

B, Chinese word order, when the Chinese sentence to be corrected of generation being provided in the english translation meaning of one's words correction operation backup in " providing sentence " is write down when measuring user and using " ← shifting " or the mobile sentence cabin of " shifting → " button or cabin eye word order 306 by shifting sentence cabin Cabin label or cabin eye number, when meaning of one's words correction terminates, search adjusts word order by displacement when " returns " command button is clicked Sentence cabin or cabin eye, are that will shift forward and backward result as unit of sentence cabin and be stored in " the former sequence " in middle word order library 308, " positive sequence " word respectively Section, and remove current sentence cabin or cabin eye and select the relative recording in dictionary 305 in experience.

Seven, the software realization of multinomial intelligent use

The application embodiment further illustrates the contents of the present invention referring to seven themes of attached drawing point.This last When a theme starts, 6 themes further illustrated above are looked back:

It is carried out 1. reading and understanding to move on on computer；2. software marks sentence cabin and cabin eye；3. meaning of one's words phase between template and component It maps together and mutually.Mother tongue personage is read the achievement understood and is transferred to computer by the means of software by these three themes of this 1-3, It is the implementation of artificial intelligence its " ".It is same that assembly obtained by everybody " " is understood in sentence by template database as software；It protects Card, which is studied in order to practise, to be achieved.

4. splitting source statement is component of expressing the meaning；5. conversion is expressed the meaning, component is sentence of communicating with the eyes；6. corrigendum and machine learning.4- 6 these three themes are the technological cores of artificial intelligence its " use ", are studied in order to practise by the implementation of this core, " " is artificial intelligence ?；" with " is also that artificial intelligence is used.This 4-6 circulation is the technological core of artificial intelligence its " use ".

Implementation and embodiment are the software realization of 7. multinomial intelligent use；By assembling the gained of everybody's " " in sentence Understand template database and 4-6 is the technological core of artificial intelligence its " use "；The setoff that also add the prior art, generates more Kind intelligent use true, hereafter chooses wherein representative be further described below:

(1) mother tongue reads foreign language

This is to be read, for the vast mother tongue personage of powerless reading foreign language with reference to the intelligent use of foreign language datum ability. Mother tongue reads foreign language, and people is allowed to utilize the mother tongue direct reading foreign language datum of oneself.Pass through the circulation of step 4-6, a ground Source language is converted into communicating with the eyes and is provided, for people one one is read.Wherein computer is during conversion, a source words and phrases string There may be a plurality of corresponding explanation of communicating with the eyes, pre-selection module pre-selection one provides, remaining whole remaines in buffer area (correction sheet).People When reading, when thinking that certain word string is improper, inappropriate, with mouse-click, it enters correction running, and software is at once below buffer area It draws the form of list to provide to be selected, as long as reader clicks certain entry to be selected, is replaced at once；And make that subsequent to translate sentence more accurate.

Such software understands template database by the resulting sentence for assembling everybody's " "；The artificial intelligence of 4-6 circulation uses it Technological core；Prior art composition relevant with part.Three parts it is specific selection have stress it is different.

The prior art is needed using combination mainly in relation to being suitble to the interface read and corresponding function button, Here it will not go into details.

Template database, which is mainly used for the optimization to component of expressing the meaning, to be understood for sentence, mother tongue is read foreign language and do not related to generally here And.It needs software to reduce weight because mother tongue reads foreign language, is especially suitable for small memory space and the limited applications of speed penalty such as mobile phone. Point library --- the component base of expressing the meaning of template database is mainly understood using sentence.Even only take multilingual component base of expressing the meaning A, Two fields of B language component generate dedicated mother tongue and read foreign language version.Such as " Chinese reads English ", " English reading Chinese " etc. are specially Version is read English for Chinese personage and is used.So subdivision specific versions not enough, take two component fields of A, B language also to dislike greatly, will They are divided into user terminal and server two parts, and user's Regular History Frequency part is stored in user terminal, and rest part is by server reality When support.It is to better conform to user individual that this user's Regular History Frequency part, which is stored in user terminal, and software is special with operator Industry, knowledge, hobby correlation reach maximization.

Mother tongue reads foreign language class software, need not all include to be accepted or rejected for correction operational part.Such as the replacement of volume word, Repair word, mend quantifier, mend hat it four corrigendum can be incorporated in compile word replacement.Obviously more troublesome in poeration, but function simplifies succinct boundary Face mother tongue read foreign language with greater need for.

For the circulation core of step 4-6, sentence to be corrected of communicating with the eyes is provided after circulation, it is not necessary to provide source language again with reference to sentence.Certainly, It can also be made setting option, make user free.

When user only browses, regardless of needing not need to correct, all click the Back button and handle next sentence, fair current and Under；Mouse-click, software starting corrigendum running, since self study links can be used when user thinks that somewhere is improper or does not understand also Subsequent conversion can be made more accurate, to be suitable for intensive reading, model reading, browse different demands.

Mother tongue reads foreign language in addition to realizing on PC, can also all be placed on website and run, and mobile phone is only to send out It goes wrong, obtains the terminal of answer.

(2) outside-mother's translation

This is to obtain the intelligence with reference to outer-female translation of foreign language datum, work for the vast mother tongue personage of powerless translation foreign language datum It can apply.Mother tongue reads foreign language emphasis point and is " to read ", and the emphasis of outer-female translation is as " translating ", compared with outside mother tongue reading, Corrigendum and human-computer interaction are more focused on, and translation high quality is pursued.

Outside-female translation flow is: software cutting foreign language datum sentence by sentence carries out the conversion cycle between S4-S6, converts out mother Sentence, is shown in a manner of female sentence comprehension template and provides sentence to be corrected, while showing source language with reference to sentence.When user feels somewhere It is improper with mouse-click when, software start corrigendum running immediately, makees corresponding necessary interaction, Machine self-learning, generation when necessary It is more more clever with software, translate more efficient application effect.

The support of template database, which also uses two parts to carry out, to be understood for sentence, user terminal divides library to retain high in use process Frequency is expressed the meaning component, when to reserved high-frequency original idea component base close to personal-machine at full capacity when (turned over unified example subscriber computer completion Translate time measurement), it is cleared up, is sorted with usage frequency automatically, delete 20% vacating space of low frequency end, again meter frequency.Remaining Part carries out backstage real-time support by corporate server by network.

Outside-female translate is mainly used for PC, can also be developed according to user demand just with A, B language component field The special editions such as multiple mobile phone, personal digital assistant meet the needs of personalized.Two-way intertranslation between multilingual version can be by user It chooses and carries out.

Full mobilization is operated for corrigendum, it is complete.And the prior art for preferentially combining related translation, editing, interface with And corresponding function, also it will not go into details here.

(3) female-outer translation

This is the intelligent use that the mother tongue personage translated outside for no masterpiece mother-obtains female-outer translation ability.It is " female-to turn up Translate " it is essentially the same with " outer-female to translate " principle, software realization, the two difference is mother tongue personage, makees " outer-female translation " purpose Language is mother tongue, is the spoken and written languages of operator's acquistion since the childhood, and mother tongue personage is not only to have a good command of one's mother tongue, and also has corresponding profession Knowledge, even if L proficiency is very poor, it is easy to play the advantage of mother tongue and professional knowledge, it is easy to be obtained by this software high-quality The translation of amount.Work " female-outer translation " is exactly the opposite, although this method software is intelligently done a lot of work, due to target language Speech text is the spoken and written languages that user does not grasp, if foreign language basis is too poor, mistake selects the corrigendum of foreign language word string that may increase, and is encountered The different sentence cabin of female outer word order, corrigendum foreign language word order may be incorrect.That is, making to operator's foreign language fundamental importance " female-outer translation " is more some higher than " outer-female translation ".

Mother tongue information translation at foreign language text, cutting mother tongue data, the conversion carried out between S4-S6 follow software sentence by sentence Ring, conversion are gone out sentence, and sentence understands that the mode of template is shown and provides sentence to be corrected in addition, while showing that female sentence comprehension understands Template appendix.When user, which thinks that somewhere is improper, uses mouse-click, software starts corrigendum running immediately, makees corresponding necessity when necessary Interaction, the translation brief meaning of one's words is accurate, provides the foreign language translation for meeting the original text meaning of one's words.Corrigendum operation some and concrete syntax text Type is directly related, and in addition to publicly-owned more orthofunction, software gives suitable personalized corrigendum content according to specific languages.

Spoken and written languages personalized question, although " female-outer translation " is also needed with " outer-female to translate ".But in contrast, Mother-outer translation brief is thinner, such as the corrigendum partial content directly related with spoken and written languages described above not only all includes, and And it should be extended as far as possible in conjunction with the prior art.

(4) sentence bone helps writing

The mother tongue personage write with foreign language is needed to obtain the intelligent use with foreign language writing ability for inability.Directly with outer Text writing, expressing thoughts often has words not know how to say, writes shortage sentence bone and is difficult to start writing.Because patrolling list word does not become sentence Son；The exactly sentence bone that traditional machine translation many decades lack almost without progress.This method has a bone to express the meaning component, Ke Yijin Row sentence bone helps the intelligent use of writing.There is sentence bone, inserts corresponding word string toward sentence cabin and be obtained with genuine foreign language sentence.It is past Sentence cabin filling word string is readily appreciated that also easy to accomplish.Need exist for explanation key be software how to receive operator signal, Search out desirable sentence bone.Software searches out desirable sentence bone after receiving operator's signal.Below by taking " Chinese-English to help writing " as an example It further illustrates.

Mother tongue personage sentences in " mother tongue meaning area " input mother tongue word string signal if according with no lower horizontal line containing end of the sentence in signal It is set to 1. with the signal of mother tongue sentence, if also having lower horizontal line to be determined as 2. with fuzzy sentence bone signal, if signal containing end of the sentence symbol in signal Chinese word strings in be judged to 3. continuing following one respectively with the signal of several word strings and transporting containing space or after with the failure of sentence bone Make:

1. with the signal of mother tongue sentence

Writing out complete mother tongue sentence expression in " mother tongue meaning area ", you want the meaning expressed with foreign language.One can only be write Sentence, can also write more, it is necessary to accord with using end of the sentence.Software according to end of the sentence accord with cutting, using cutting get off as example sentence, matched Sentence bone running, does not such as search matched sentences bone, jumps and 3. handle；It searches, provides search result.Such as:

It " studies medicine at Darwin 16 years old to Edinburg, Cambridge is arrived after 3 years." when the sentence be input to mother tongue the meaning area.It clicks Shown after " match sentence bone " button mother tongue sentence bone " [1]+year when to+[2]+[3]+,+[4]+arrive afterwards+[5]+." English is provided simultaneously Sentence bone: " at the age of+ [1]+went to+ [2]+to+ [3]+and+ [4]+later to+ [5]+".

Then middle sentence cabin content is transferred to English sentence bone.Sentence bone has sequence sentence cabin [2] [3] to need without reference because in Intervene cutting, show " 23 { studying medicine Edinburg } ".When operator clicks left button between " fort " two words, software cuts them It is divided into " 2 { Edinburg } 3 { studying medicine } "；Their content is filled into the corresponding sentence cabin of English sentence bone by software respectively.Subsequent intermediate state Display box shows the hybrid-type foreign language intermediate sleeve sentence of bone example: "

16 went to Edinburg to of at the age of Darwin studies medicine 3 years later to Cambridge " of and

When operator successively click one by one or drag and drop choose certain Chinese word strings " Darwin ", " 16 ", " Edinburg ", etc., Software provides corresponding English entry in the list of lower part, after user click list items are chosen, software by it replace just now by The Chinese word strings clicked or chosen only have the corresponding not list software of a word string directly to replace it.Such as drag and drop choose " ten Six ", after " Darwin ", by software prosthesis are as follows:

At the age of sixteen Darwin went to Edinburg to studies medicine 3 years later to Cambridge of and

Certainly clicking should be arranged successively with English word order.Then continue to " Edinburg ", " studying medicine ", " three ", " year ", " Cambridge " Running.

When wherein the existing word of certain word string also has compound word that can indicate the selected word string meaning of one's words, such as " studying medicine " word list has " learn ", " studying ", " medicine " etc.；Compound word list has " studying medicine ", " study medicine ", should click " ↑ choosing It is multiple " button, choose compound word displacement " studying medicine ".When this operation terminates.English sentence are as follows:

“At the age of sixteen Darwin went to Edinburgh to study_medicine and three years later to Cambridge.”

Software push buttons " tail adds " receive user click, this tail is added to english writing area.Here it supplements, in centre When state display box chooses word string, it can repeat to select word operation until sentence all over Britain is completed.When needs inquire English word string If, the method for equally dragging choosing can be used and carry out.A word can be chosen, several can also be once chosen.

2. with fuzzy sentence bone signal

Some complete sentence bone need not exactly be remembered and be write out to so-called fuzzy sentence bone, write out a part of certain bone.

Such as to express that " life, which exists, needs the atmosphere of correct amount type." this meaning.To it write one it is fuzzy, Possible sentence bone.Can mother tongue look like area write " _ need it is appropriate _."；" _ exist _ it needs."；" _ presence _ need it is appropriate _ _." etc..The sentence bone word for after order button, taking out " with sentence bone " and being separated by lower horizontal line, end of the sentence symbol is clicked in software reception, with Sentence bone word search " sentence bone string list ".It is 3. handled as do not searched to jump；It searches, the record searched is kept in into interim table, Again with interim table sentence bone word length statistics, sequence, long preceding, list provides high-end (most 10), successively takes out corresponding sentence Bone, list provide, and cursor falls in most upper record, while its China and British sentence bone has been shown in the respective block of top.If with Family, which is approved, clicks " ↓ tail adds " order button, and English sentence bone tail is added on English composition window by software, and user is waited to fill in a cabin Hold, completes this English composition.

If first record does not simultaneously meet the meaning to be expressed of user, can be alternative in list.It clicks single again after the record Hit " ← optionally " order button, which will display.

3. with the signal of several word strings

This method is used as sample using " identical content bilingual text ", directly reads bilingual text and is placed in A, the area B Yuan Yu.This Outside, also identical content bilingual text is read in, it is temporary with " sentence pair library " or " data base " that sentence pair can also be cut into；And it is other with Sentence is that these materials existing for unit can be directly utilized.

If the signal of user is the Chinese Chinese word strings comprising space, they are cut into word using these spaces as separator String；Or following horizontal line, end of the sentence symbol are separation cutting word string.Then, it is scanned for these word strings；If both of which does not have, Three words are from left to right cut, is searched for, is taken, discards and cut again；It is cut without changing with two words, searches for, take, discard and cut again；Nothing Change and cut with four words, searches for, take, discard and cut again；It is cut without changing with a word, searches for, take, discard and cut again.Until showing Mean zero.

It is after being searched for according to word string as a result, to bilingual text or sentence pair library, data base etc. using sentence as material existing for unit It is middle to search for the sentence that contains, and with how much sequences containing word string length, take high-end 10 record lists with reference to sentence provide to Choosing.

Such as think that " independent studies find that compared with parent has heart disease medical history, siblings suffer from a heart complaint can for expression It can be the bigger predictive factors of people's risk." this meaning, it can only write that " research finds that parent has heart disease brother Younger brother sister suffers from a heart complaint."；" research finds that parent has heart disease siblings to suffer from a heart complaint."；" research finds that parent has heart Sick siblings suffer from a heart complaint." etc..Display, which provides, as above to be stated, and it will not go into details.

Listed above is representative embodiment in artificial intelligence.Technical characteristic is opposed according to the present invention, can be developed very much More intelligence software products out.

Refer to that sense-group is aligned, and makees a supplement here at number above.

Sense-group alignment --- sense-group alignment schemes are across the languages features using sense-group, by multilingual word, word, phrase or Phrase etc. carries out semantic alignment under the support of current sentence pair example, they become as sense-group string after alignment, qualified to be stored in Sense-group string library.Mainly there is the following method:

1. organizing compound word --- when two or more original word strings total semanteme cannot by word string justice be added obtain with " _ " is connected to compound word, referred to as multiple string, such as:

Knew_nothing_about_it knows nothing

Compelled_to_go simply must go

The Late_at_night late into the night

Works_little work is slack

Wherein " works little " " work is slack ", although " works " has the justice of " work ", " little " does not have The meaning of " sluggishness "；Their total semanteme cannot be added by word string justice and be obtained, therefore is connected with " _ ", be organized as compound word.

2. according to current sentence pair example, extension or the supplement meaning of a word

Nothing can be 1{wholly beautiful}that is not 2{useful}.

It is all to be impossible to 1 { perfect } without 2 { applications }.

Wherein " useful " meaning of a word only has " useful, useful, beneficial "；But it expresses and " is answered really in the sentence pair " semantic；" application " and " useful " meaning of a word are close simultaneously, give to augment or extend meaning of a word item " application ".

3 { small } of 1 { She } was 2 { strong }, { she } was of for all 1 so

1 { she } is although 3 { slight of statures }, very 2 { solid }.

Wherein " small is small " augments " slight of stature " meaning of a word item according to the sentence pair example.

3. string length is added and subtracted under the premise of not changing original word, word, convenient for splicing

I ask you to teach me every other day. I ask you every other day to teach me.

Wherein " teach " v has " lecturing, professor " meaning of a word；It is a length of " religion " to subtract word string, increases " religion " meaning of a word item.

4. adhesive tape attaches word string

Such as " good is good " adhesive tape is at " benefit, good thing, good intention ".Such as " word word " adhesive tape is at " word ".

5. morphological change separately makees entry typing, in library, (participle of English, the meaning of one's words of the expression such as adjective adverbial word comparative degree increase Add entry and the corresponding meaning of a word).

Been supplement " still, how " meaning of a word；Punished augments " punishment " meaning of a word；

Etc. and so on.

Claims

1. a kind of mother tongue personage learns the artificial intelligence implementation method with other Languages text, feature includes the following steps；

S1. reading understands to move on on computer and carry out, and mother tongue personage moves on to calculating to the reading understanding of identical content bilingual text It is carried out on machine, bilingual text A and B are respectively placed on the area source document A and B by software；

Software from the area source document A and B according to the cutting of end of the sentence symbol and take out mono- sentence pair of bilingual text A and B set reading understand area A and Accordingly reduce one in the area B, source document A and B；

If mother tongue personage thinks to cut sentence improper, source document area A or B point are clicked, part before the point is moved into read by software to be understood Area follows closely after A or B taken sentence；

S2. software mark sentence cabin and cabin eye, mother tongue personage, which reads understanding and clicked to read simultaneously with mouse, understands that A and B, area is current It can be used as express the meaning identical word string or several continuous word strings in cabin, they are added flower according to successive with capitalization for sequence by software Bracket is designated as a sentence cabin at A and B in advance simultaneously；

Or software automatically identifies A and B currently express the meaning identical word string or several continuous word strings for them with capitalization According to successively adding brace at A and B for sequence while being designated as a sentence cabin in advance；

Mother tongue personage read understand or software automatic identification be all successively to be searched for quantity string, proper name string, using names string, it is bilingual simultaneously One sentence cabin of mark in advance, if these three strings have all been searched for and bone example compares between 15-50%, pre- sentence cabin of marking terminates；

If these three strings have all been searched for and bone example ratio > 50%, then continue successively to search with no using names string, other meaning of one's words matching string Rope marks sentence cabin in advance, until bone example compares between 15-50%, pre- sentence cabin of marking terminates；

When in advance mark sentence cabin terminate, software carry out sentence cabin, sentence bone format checking, with 16 binary digits replace capitalization with A from Left-to-right arrange for sequence aforementioned pre- is designated as formally marking；

Bilingual sentence pair through formally marking further detects each cabin, if there is the big sentence cabin greater than original 5 word strings need to be into one Step mark cabin eye or group compound word, mark cabin eye are followed successively by target word string with no using names string, other meaning of one's words matching string, and mould example compares for 50- Between 70%, ophthalmically acceptable 16 binary digits in cabin from left to right add square brackets to mark with A for sequence, are first round sentence cabin and cabin eye above The step of understanding mark using poster kind sentence as template software automatic identification, marks non-poster since the second wheel understands mark Kind of sentence is accordingly expressed the meaning identical sentence cabin and cabin eye；

Every wheel mark selects the corpus of the identical AB bilingual text version of content, and the second wheel starts every wheel and understands mark newly-increased one Kind spoken and written languages, another kind are the spoken and written languages marked, and A language distributes to the spoken and written languages marked, and B language is distributed to The spoken and written languages being newly added；

S3. the meaning of one's words is identical between template and component and mutually maps, software it is this must be by reading understanding process that human brain is dominated It is middle that current reading is understood that the bilingual sentence pair in area marks upper sentence cabin, cabin eye, the achievement understood is read to retain current sentence pair, with Sentence understands that the mode of template 100 is transferred to computer, so that follow-up work is generated Intelligential efficiency by software completion, for sentence Understand that template can be more particularly for understanding and mark other sentences, software by they be stored in sentence understand template database replace human brain into Row memory, sentence understand template database at least decomposable sentence bone, cabin mould, four kinds of sense-group string, idiom components of expressing the meaning out, respectively constitute Corresponding component base of expressing the meaning, they are point libraries that sentence understands template database；

S4. splitting source statement is component of expressing the meaning, and it is component of expressing the meaning, generation that software, which understands that template 100 splits source statement using sentence, Sentence is understood and split for human brain, is first carried out operating with sentence bone, is pressed the void nucleation sentence bone string list of sentence bone word string and sentence cabin in advance And index, when with sentence bone source statement from left to right word for word or by word string remove search sentence bone string list bone lead-in section and be stored in Interim table, then interim table content is taken out by record, with sentence bone word section inquiry source statement, every section of sentence bone string can be in source language It is found in sentence and order is mutually all matching, the source statement bone of current matching is taken out further according to sentence bone string list；

It is that component of expressing the meaning makes current source statement and understands template that source statement sub-set, which is entered source statement bone to split source statement, 101；

S5. converting component of expressing the meaning is sentence of communicating with the eyes, and software searches the sentence bone for sentence of communicating with the eyes using source statement sentence bone, by current source language Sentence understands that the sentence cabin content in template is transferred to the corresponding sentence cabin of a bone of communicating with the eyes；

All cabins are detected, without skipping for cabin mould, if containing cabin mould, this cabin content is inserted in cabin mould, utilizes the cabin mould point The Mappings in library take out cabin mould of accordingly communicating with the eyes, and current source language cabin is intraocularly held to the corresponding cabin eye for moving on to cabin mould of communicating with the eyes；

Sentence cabin or the cabin eye for searching for mesh sentence comprehension template 102 one by one, search for sense-group string library with words and phrases string in source therein, will search for To all explanation results of communicating with the eyes be saved in correction sheet, using pre-selection module, pre-selection one of them as communicate with the eyes sense-group string replacement The source words and phrases string of current sentence cabin or cabin eye, until all cabins, cabin eye replacement terminate；

S6. corrigendum and self study, software takes out from mesh sentence comprehension template 102 to be shown and provides as sentence to be corrected of communicating with the eyes, receiving Operator's confirmation or corrigendum, while carrying out self study；

Or software takes out from template of communicating with the eyes and show and provide as sentence to be corrected of communicating with the eyes, and is taken out from the language template of source as the reference of source language Sentence display provides, and receives operator's confirmation or corrigendum, software carry out self study simultaneously；

When human brain is read, the marking moment for reading understanding shows, and reading understanding becomes simple and quick, and software compounding practice enters Correction running 201；

Software is the component S4 that expresses the meaning splitting source statement, and then converting component of expressing the meaning is the sentence S5 that communicates with the eyes, and software matches cooperation more Just simultaneously self study S6, these are converted what human brain was completed from the sentence of spoken and written languages A103 to the sentence of spoken and written languages B104, Software is given to realize to be recycled between S4-S6, can generate multinomial intelligent use.

2. according to the method described in claim 1, it is characterized in that: sentence understands that template includes sentence example, sentence cabin, sentence bone, cabin eye, cabin Mould, sense-group string, idiom, they are the member that sentence understands template altogether, and directly the member to express the meaning is component of expressing the meaning, with their phases Even constitute the long-chain that a multilingual sentence understands template；

Software save with database and be remembered instead of human brain, the first round understands mark while acquisition sentence understands template Note sentence understands that template database contains A moulds, B mould fields, and the second wheel understands that mark starts every wheel and increases a kind of spoken and written languages newly, Sentence understands that template database first adds the sentence mould field of a corresponding language, and every wheel understands that mark must take same content, different literals Version corpus ensures that the meaning of one's words expressed by each languages sentence mould of identical recordings is identical, and mutual phase mapping；

Subordinate clause understands that template database decomposites several species and expresses the meaning component, respectively constitutes component base of expressing the meaning, they are that sentence understands Template database divides library, divides library to have inheritance, inherits the characteristic that sentence understands template database, same in point library to record structure of expressing the meaning The part meaning of one's words is also identical, their also mutual phase mappings.

3. according to the method described in claim 1, it is characterized in that: utilizing pre-selection module, one of pre-selection in the step S5 The source words and phrases string that current sentence cabin or cabin eye are replaced as sense-group string of communicating with the eyes includes: correction sheet at least word string section, search string, solution It releases, go here and there the first field, preselect module and spoken and written languages type is closely related；

The pre-selection 312 of english translation is that 1. with current sentence cabin or cabin, intraocularly English word string inquires " experience selects word " library, has looked into taking-up solution It releases the first item Chinese word string of field contents to provide for the string of communicating with the eyes of 313 maximum adaptations, looks into and be difficult correction sheet this word string section owner record It is provided for the string of communicating with the eyes of maximum adaptation；2. inquiring " middle word order " library with the Chinese content 314 that current sentence cabin or cabin eye have obtained Former sequence field, look into and 315 provided with the string of communicating with the eyes that positive sequence field contents replace with maximum adaptation, looked into without constant；

In translate the pre-selection of English with sentence cabin in source, transition type of communicating with the eyes or cabin eye source language content search correction sheet search for string field, have Congruence explains that field contents are provided as the string of communicating with the eyes of maximum adaptation using same record, and 317 inquire " cabin experience " again when no congruence Library middle deck field has 318 to provide by the string of communicating with the eyes of maximum adaptation of English cabin field contents, looks into " cabin experience " library and is difficult current sentence cabin Or cabin eye Chinese content provides for the string of communicating with the eyes of maximum adaptation；

All pre-selection modules selected correction sheet check in item change it is former provide content, while it being shifted in correction sheet First of this word string section.

4. according to the method described in claim 1, it is characterized in that: in the step S6 software from mesh original sentence understand template take out make For sentence to be corrected of communicating with the eyes, and/or as source language sentence is referred to, receives operator's confirmation or corrigendum, while carrying out self study and including:

When user reads communicate with the eyes sentence to be corrected and corresponding source language with reference to sentence, i.e. correction starts, and the meaning of one's words, which corrects, is having sentence to be corrected completely just Really without correcting and needing to correct two kinds of situations, receives user click the Back button when without correcting and connect S4, need to correct Enter correction running 201 after connect S4 again；

Software enters correction running 201, and links with self-learning module 301, is learnt to the human-computer interaction of correction course, is remembered Recall and provides data for pre-selection module.

5. according to the method described in claim 4, it is characterized in that: correction running includes the following steps:

A) the replacement 202 of improper word string --- when user thinks that sentence to be corrected word string is improper, its systematic search correction sheet is clicked, Pop-up drop-down list provides whole correlations and checks in that item is to be selected, then translates the improper word string of sentence to be clicked the item replacement that checks in chosen, It is replaced into first of word string section simultaneously；

B) the cutting 203 in sequence sentence cabin --- when encounter provided without reference to the sequence sentence cabin that can not carry out cutting automatically it is to be cut, To be separated two sentence cabin contents as cut-off by click-point after being clicked；

C) intervene optionally sentence bone 204 --- when encountering a cabin and cabin eye unshapeliness, there is incomplete or stackings mistake to select the performance of sentence bone When or user do not approve and click the button " after rear " when currently translating, system multipair matched sentences bone is provided it is to be selected, to click selection Sentence bone afterwards is inserted in connecting S6 again to be continued to operate；

D sense-group string 205) is added --- when improper word string is optionally replaced and lacked, before receiving user using the meaning of a word, adhesive tape is extended Draw or following word word method adds sense-group string, is replaced to add string and adds languages component field corresponding in sense-group string；

E) group choosing goes here and there 206 again --- and when improper word string is optionally replaced and lacked, receives user using group compound word or reelect compound word Method add compound word string, be replaced to add string and add languages component field corresponding in the compound word library of sense-group string；

F) sentence cabin word order 207 --- when translating sentence sentence cabin has word order mistake, receives user click and wherein go here and there, then click " ← move " or " shifting → " order button will click string Forward or move back a displacement；

G) compile word replacement 208 --- when much more continuous string is improper, cannot with click choose when, when choosing string with drag-and-drop operation, When then filling out other word strings, editor at word replacement or be empty compiling, then clicking " ∧ " button, system replaces the former with the latter；

H) rhetoric 209 --- when the rhetoric phenomenon for occurring to be solved with above-mentioned steps, receive user click " rhetoric " button, Then current sentence to be corrected is copied into edit box, receives user using the method for editor and carries out rhetoric operation；

I) cabin conversion 210 --- translate that inch receives user click cabin or cabin intraocularly holds in the middle, search correction sheet is all current What sentence cabin or cabin eye contained checks in item, and list, which provides, waits user is repeatedly orderly to click selections, by their displacement sentence to be corrected phases A cabin or cabin is answered intraocularly to hold；

J) mend quantifier 211 --- when english translation due to English quantifier-free and Chinese have quantifier thus need in sentence to be corrected to need it Place fills quantifier, and the need for receiving user click sentence to be corrected add place, systematic search closest to click-point Shuo Chuan rear center noun, It when " ▲ " button while being clicked, with head noun volumes of searches dictionary, takes out corresponding quantifier and add sentence to be corrected；If quantifier library Nothing is looked into, dialogue is provided and receives to input to add and typing quantifier library；

K) mend hat it 212 --- translate in the middle inch due to Chinese without and English has article, verb to/be, to/have, these Need to add, system provide retain sentence cabin mark label communicate with the eyes sentence to be corrected when, it has been shown that " a/an ", " the ", " to/ Be ", " to/have ", when one of they are clicked, and sentence to be corrected is also clicked, system is clicked one of they in sentence to be corrected Place adds the word string of appropriate form；

The above A-H state step is unrelated with the type of spoken and written languages, have common feature, can appear in A between multilingual translate B or During B translates the two-way translation of A；I state step is related to the type of spoken and written languages, for being spaced ideograph between no word Among translation；J-K state step and spoken and written languages type are closely related, they have characteristics of personality and are arranged according to specific languages.

6. according to the method described in claim 4, it is characterized in that: the human-computer interaction to correction course is learnt, is remembered And data are provided for pre-selection module and include:

A, experience selects word, 303 after in english translation meaning of one's words correction operation when the improper word string of user click sentence to be corrected, system Search for correction sheet, list provide whole correlations check in item it is to be selected while, all related constraint terms are separated by generation warp with space Channeling test of cement sheath, end are attached corresponding English word string；

While user chooses one of list displacement, with choosing entry and English string to inquire " experience selects word " library, looked into frequency= Frequency+1 looks into nothing, and entry will be chosen to move on to through first of channeling test of cement sheath and pop down formula is stored in the middle explanation field in the library 305, and English string is from warp Deposit English string field is removed in channeling test of cement sheath, and current sentence bone number is stored in sentence bone field；

B, Chinese word order, backup is in " providing when providing the Chinese sentence to be corrected of generation in english translation meaning of one's words correction operation Sentence " is write down when measuring user and using " ← shifting " or the mobile sentence cabin of " shifting → " button or cabin eye word order 306 by the cabin in shifting sentence cabin Label or cabin eye number, when meaning of one's words correction terminates, the sentence cabin of displacement adjustment word order is passed through in search when " return " command button is clicked Or cabin eye, it is that will shift forward and backward result as unit of sentence cabin and be stored in " the former sequence " in middle word order library 308, " positive sequence " field respectively, And it removes current sentence cabin or cabin eye and selects the relative recording in dictionary 305 in experience；

C, cabin experience, when translating " cabin conversion " running of English meaning correction in 309 it is standby deposit current sentence cabin, cabin label, when " returning Return " order button clicked, by current sentence it is all by cabin conversion turn forward after content look into cabin experience library 311, looked into frequency+ 1；It looks into no pop down formula and is stored in the library.

7. according to the method described in claim 1, it is characterized in that S2. software marks sentence cabin and cabin eye after the completion of S1 step Before beginning, there are one steps:

First operate with sentence bone with current A language example sentence, search is when the whether existing matched sentences bone of precedent, if searching A sentence has Matched sentences bone and skipping for operator's approval is obtained after being inserted in, it is not necessary to understand mark, read in next sentence pair；

If not searching A sentence has matched sentences bone or has searched but looked into when being denied after being inserted in by operator with mark running Afterwards, into the step S2 for understanding mark sentence cabin.

8. according to the method described in claim 1, it is characterized in that: circulation generates multinomial intelligence between S4-S6 in the step S6 Can include: to obtain the intelligent use read with reference to foreign language datum ability for the vast mother tongue personage of powerless reading foreign language:

Mother tongue reads foreign language --- and allow people to utilize the mother tongue direct reading foreign language datum of oneself；By the circulation of step 4-6, one Source language is converted into communicating with the eyes by one ground to be provided, and for people one one is read, if having, word string is improper, is carried out more with mouse-click Just；It can only take two fields of A, B language component of multilingual component base of expressing the meaning to generate dedicated mother tongue and read foreign language special edition, it will be normal It is stored in user terminal with high frequency component of expressing the meaning, remaining better adapts to user individual by network support；

Outside-mother's translation --- the intelligence of foreign language datum is referred to, translated for the powerless vast mother tongue personage for translating foreign language datum Using, software cutting foreign language datum sentence by sentence, the conversion cycle between S4-S6 is carried out, mother tongue sentence is converted out, with female sentence comprehension The mode of template, which is shown, provides sentence to be corrected, while showing source language with reference to sentence；

User terminal reserved high-frequency is expressed the meaning component, when to reserved high-frequency express the meaning component base close to personal-machine at full capacity when, automatically into Row cleaning, deletes 20% vacating space of low frequency end, counts frequency again, and rest part carries out backstage in fact by corporate server by network When support.

9. according to the method described in claim 1, it is characterized in that: circulation generates multinomial intelligence between S4-S6 in the step S6 It can be using the intelligent use for including: female-outer translation ability of mother tongue personage acquisition translated outside for no masterpiece mother-:

Female-outer written translation --- mother tongue information translation at foreign language text, software cutting mother tongue data sentence by sentence, progress S4-S6 Between conversion cycle, conversion goes out sentence, and sentence understands that the mode of template is shown and provides sentence to be corrected in addition, while showing mother Sentence comprehension understands that template is prepared for the examination；

When user, which thinks that somewhere is improper, uses mouse-click, software starts correction running immediately, makees corresponding interactive, translation brief The meaning of one's words is accurate, provides the constant foreign language translation of the original text meaning of one's words；

Corrigendum operation some is directly related with concrete syntax text type, and in addition to publicly-owned more orthofunction, software is given according to specific languages With the personalized corrigendum operation for being suitable for.

10. according to the method described in claim 1, it is characterized in that: circulation generates multinomial intelligence between S4-S6 in the step S6 It can apply and include:

The mother tongue personage write with foreign language is needed to obtain the intelligent use with foreign language writing ability for inability, sentence bone helps writing, Software is received mother tongue personage and is illustrated in a manner of three kinds, is judged to 1. showing with mother tongue sentence if according with no lower horizontal line containing end of the sentence in signal Meaning, if also thering is lower horizontal line to be determined as 2. with fuzzy sentence bone signal, if containing in the Chinese word strings of signal containing end of the sentence symbol in signal Space or the above two with being judged to 3. continuing the running of following one respectively with the signal of several word strings after the failure of sentence bone:

1. writing out complete mother tongue sentence expression with the signal of mother tongue sentence in " mother tongue meaning area " and wanting the meaning expressed with foreign language；It can Only to write one, more can also be write, it is necessary to accord with using end of the sentence；Software accords with cutting according to end of the sentence, using cutting get off as example Sentence operate with sentence bone, does not such as search matched sentences bone, jump and 3. handle；It searches, provides search result and carry out S4- The circulation of S6；

After being wholly converted into English word string, software push buttons " tail adds " receive user click, and current sentence tail is added to english writing Area can also receive operator in english writing area and carry out edit-modify；

2. with fuzzy sentence bone signal, software reception is clicked " with sentence bone " after order button running, and a list is popped up, list containing The sentence bone of " fuzzy sentence bone " composition, and cursor is fallen on desirable possible maximum record, while its China and British sentence bone has been shown Show in the respective block of top；User click " ↓ tail adds " order button is received, English sentence bone tail is added on English composition window, etc. A cabin content is filled in user, completes this English composition；

, can be alternative in list if first record does not simultaneously meet the meaning to be expressed of user, click and clicked after the record again " ← Optionally " order button, the record detailed content will be shown；

3. software will illustrate cutting to be word string, after being searched for according to word string as a result, depositing to by unit of sentence with the signal of several word strings Material in search for containing sentence take high-end 10 record lists to refer to and with how much sequences containing word string length Sentence provides to be selected.