CN101855630A - Formalization of a natural language - Google Patents
Formalization of a natural language Download PDFInfo
- Publication number
- CN101855630A CN101855630A CN200880115885A CN200880115885A CN101855630A CN 101855630 A CN101855630 A CN 101855630A CN 200880115885 A CN200880115885 A CN 200880115885A CN 200880115885 A CN200880115885 A CN 200880115885A CN 101855630 A CN101855630 A CN 101855630A
- Authority
- CN
- China
- Prior art keywords
- natural language
- text
- key concept
- word
- clear
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/40—Processing or translation of natural language
- G06F40/55—Rule-based translation
- G06F40/56—Natural language generation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Machine Translation (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
It is disclosed a method for formalization of a natural language allowing creation of an unambiguous model of a natural language text. It is determined the basic notions for entities that are named by a natural language and for each basic notion it is attached an unique number or name and a description, in addition it is attached a list of words which can name the basic notion for each used natural language. The unambiguous model uses only basic notions. In this way it is possible a machine to interpret the unambiguous model and to input knowledge and data in a base or to make a text generation in another natural language using the unambiguous model. Also it can be generated a text in artificial language such as a program language.
Description
Technical field
The present invention relates in the machine that uses natural language, import knowledge.It can be as the machine translator of natural language.
Background technology
The most frequently used mechanism is that the set of words-all artificial languages that define with the machine interpretation natural language all are the type.The trial of the grammer connotation of definition word has been arranged.Developed for given text and provided subject fields, can also define the preferred connotation of word in this way, therefore for example can in mechanical translation, realize better result.Also attempt according to other words in the text and this word the use in other words and added up the connotation that defines a word.Attempted that also therefore the numerical value of the setting of the word in given natural language and other natural languages from identity set had similar connotation from macaronic word with same appropriate value.
Summary of the invention
Technical matters
Come the problem of clear and definite explanation natural language also not solve by machine, this is obstacles for import knowledge and data to the machine that uses natural language.Machine can not be used for the regular translation of file, because this is not the reliable fashion of translation.It can not generate the natural language text that has clear and definite explanation for different people, but this is extremely important again when writing textbook or patented claim.Computing machine can not be programmed and use natural language, because a sentence of natural language has a lot of possible connotations in view of form, so correct sentence can be explained by different way on the grammer.Existing human knowledge can not optimally be used, because there is not formal mode to make machine directly explain the knowledge of writing with natural language.
Technical scheme
Explain that natural language generally comprises the machine mould of the knowledge that structure explained.Therefore explain natural language text by different modes, can define the connotation of word in grammer part, sentence and the sentence of speech.Problem is not recall relation, and the people can not exert one's influence to the model that forms.This be because, without comparison basis between this model and natural language text.Therefore this model still is a kind of structure that can not only be explained by a kind of mode.The core of technology provides a kind of method that is used to generate clear and definite model.The model of Xing Chenging can only be explained with unique mode in this way.
This method comprises five steps.
In first step, a large amount of language is studied, its objective is the basis of the human employed notion of definition.To consider also whether the word in the natural language is key concept.Key concept is represented some entities or action.Usually several different key concepts represented in the same word in the natural language, so word has different connotations.Expression " c л ъ H ц e=1 " (" c л ъ H ц e " looks like in English is the sun) and " sun=1 " can help to carry out mechanical translation on the technological layer, but it can not help the clear and definite translation of connotation.In this type systematic, the possibility of result of translation is such: " User rights=п paBaTa Ha Hap к oMaHa " (" п paBaTa Ha Hap к oMaHa " look like in English is drug-addict's right), but in fact in given context " user rights " to look like be consumer's right.This class word of enumerating has produced the intermediate language with indeterminate connotation.What enumerate here is entity rather than word.Entity according to this method has unique title.Title can be a numeral, but also can be the word from the natural language of wide-scale distribution.It shall yet further be noted that the given word in the natural language can only represent an entity with a kind of mode.Under this mode, " c л ъ H ц e " (" c л ъ H ц e " looks like in English is the sun) can only have a connotation-star, for the every other connotation of word " c л ъ H ц e ", must select other words.Should be understood that this class names not influence of natural language connotation.For entity, carry out characterization with their description according to this method.Provide the description of entity in an identical manner with natural language, this dictionary by natural language is finished.Each entity has a word list, can come by word list that named entity-some is as dictionary with natural language, but it is at entity rather than word.
About the structure of entity, this entity has unique tags-title or numeral, description and represents the word list of described entity with natural language, and this structure further is called as key concept.
Second step of described method is only to utilize key concept to generate natural language text.In this step of this method, it has utilized the method applicatory of all in the background technology, and this syntax and semantics connotation that can define the word in the text also can generation model.During generation model, can utilize the overall situation of the different connotations of word to use statistics or at each user's of this method partial statistics.Can use the similar text of clear and definite word connotation.The human translation that given text is translated another kind of language from a kind of language also can be used for defining the key concept that natural language text uses, because the word that uses in the translation is studied, and in view of they connotation and with compare from the word in the urtext.
The 3rd step of this method is to recall relation.In this step, the model that is generated in second step is used to produce the text with the used identical natural language of urtext.The operator can utilize computer program to change the model that is produced, so that the model that is produced meets the expectation that he understands the text.Can directly change model,, for example use the relational tree between the entity because it directly acts on represented entity.This working method needs strict training.In another is realized, can change model by attempting to be changed to which entity of computer interpretation.Urtext can be compared with the text of generation, and the difference between the text of mark urtext and generation.For from each the mark word in the dictionary, it exports Alphabetical List, can filter the synonym that those are abandoned because having unsuitable connotation.The operator selects and repeats in real time this process-new product is therefore arranged and have new correction from Alphabetical List.Yet, select synonym always not to be enough to define a given entity.Therefore can consider that some modes change the explanation relation between two key concepts in the given text.In this way, can utilize the visual manner of mark and discriminating to carry out.For example, what the subject that can specify sentence is, or what the meaning is and explains what is.Can generate the mode of the temporal relationship that is used for indicating text.Can generate the mode of the surface that is used to change text, explain and product so that can easily manage.For example, such situation is arranged, the wherein real standard that is different from explained, as play with words and satirize-in this way,, must provide two explanations according to surface: standard with improved, and they become the part of clear and definite model.Can generate a lot of modes of the type, the personage that receives an education shows in his brain to computing machine and thinks.Purpose is the clear and definite model that obtains to represent with accurate way the text connotation.
The clear and definite model of the natural language text of the 4th step-generation of this method is affixed on the file that comprises natural language text.This has produced the clearly explanation of natural language text, and this is very useful for patented claim and mechanical translation.When utilization has the method generation textbook text of additional clear and definite model, owing to utilized the definition of the entity that uses in the text, and the recurrence of the definition of employed entity with more high-grade definition entity the time uses, and computer program can produce the explanation of any complexity level.
The 5th step of this method is to utilize the clear and definite model of natural language text to be used for machine learning and to generate notion and theory, and machine uses the center of the formalization knowledge that obtains from the clear and definite model of natural language text.
Beneficial effect
The present patent application can be used for mechanical translation, retrieval knowledge, and the retrieval word that is not based on text and comprised wherein, but under current technical merit, retrieval is the clear and definite model similar with the text of being retrieved.Can utilize to the analysis of the clear and definite model of text retrieve-so the fieldworker can answer a question, for example retrieve information according to the transfer property of Bulgarian law foreign nationals.
Embodiment
The exemplary realization of first step of this method
Utilize computer program to determine the key concept of language, the Alphabetical List of each word of the natural language that examination is examined.Its synon definition that gives in the definition of each word of the language that provides in the dictionary and the dictionary is compared.Utilize the simple comparison that defines with retrieval in similar text of comparing.Purpose is the different connotations that define given word according to the synonym of each connotation.In this way, utilize the definition of each word that provides in the dictionary and the comparison between its synon definition, come the relevant similar text of definition from two definition (they have formed different connotations), " entity " of name this method.The definition of entity is formed by the similar text in two synon definition usually.When finding such entity, check whether data center has registered similar entity, and compare the description of registered entities and the description of novel entities.If novel entities is not registered in the data center, then register this entity.
After the description with entity forms the center of entity automatically, come named entity also to determine their description by the expert.For entity, providing can be in definite condition their word list of giving a definition, and this determines that condition depends on the text that comprises this word and the surface of the text, for example text whether be science or text whether be to play with words etc.But, can utilize the clear and definite model of natural language description to describe each entity when the center of all entities time spent all.This can utilize the key concept of language by the philologist of the clear and definite model of the description that generates entity, utilizes the description of the natural language that forms automatically to finish.After finding the key concept of natural language, ensuing natural language utilizes the center of formed key concept.How the linguist can easily define the entity of name registration in the language of determining, and in the heart final entity sets in must additionally being increased to.When entity is increased to the center, should notify the conforming linguist who seeks natural language, so that they can provide the suitable name of novel entities, they are responsible for this.Can the name of novel entities be described.
Can explain second kind of more natural language such as grade automatically.First kind of process that language is identical with research is set.Entity according to registration is provided with new center.The title that the entity at center of making a fresh start has is the word from second kind of language.From the dictionary of second kind of language to the first kind of language, can find the possible translation of each title of the entity at second center.For each translation-from the word of first kind of language at first center, the entity that taking-up can be named with this word.Owing to produced all combinations that the institute with first kind of language might translate that substitute of each word of describing, therefore vacation is carried out in the description of the entity of second kind of language and translated.Vacation translation from the description of the entity of second kind of language is compared with the description of the entity at first center of taking-up.Find out and the mark consistance best.Each consistance that finds is in this way ratified by the linguist.After the consistance approval, from second, remove this entity in the heart.Mark is in second language to show, and this entity is increased to the entity at first center for the name tabulation of this entity of second language.After handling all consistance, still be in second these entities in the heart or in first, be registered as novel entities in the heart, perhaps by manually finding out their in the heart consistance in first.
In official document, must obtain the consistance of the natural language text of generation from clear and definite model.This can be that cost realizes with the simplicity of the text of paying generation, although can produce a plurality of natural language texts with identical connotation from the language viewpoint, and by safeguarding that clear and definite model realizes that unique generation text represents identical knowledge.Linguist's task is that the feature that will obtain the required text of unique generation text joins in the clear and definite model.
Such method is for translating another kind of language with official document from a kind of language, and particularly patented claim is even more important.
On the other hand, in the translation of literary works, preferably produce a plurality of natural language texts, and be used to select best one to be used to make up concrete language from the statistics of the literary works of special language from clear and definite model.
The exemplary realization of second step of this method
Text can be expressed as tree list, and each tree is a sentence of text.Relation between can independently being set.Each element of tree is the object with extra feature, and this feature extracts from text automatically and the person of being operated adds by hand.The part of these features is each element of tree and the relation between other elements of tree.Some elements of the tree of the sentence in the expression text, for example noun can have relation with the element that belongs to other trees.The order of tree in the tabulation is extremely important.It has represented the order of the sentence in the urtext, and the order in the text that finally produces according to clear and definite model.
The exemplary realization of the 3rd step of this method
Generate the superstructure of text editor, it has other functions, helps to make that the variation in the clear and definite model of the text that forms automatically is more prone to.For example, screen is divided into three zones.The first area is used for whole urtext-plain text editing machine.Second area is used for the relation of recalling when clear and definite model has generated.In this zone, be that the machine of the sentence after the processing of text produces text.In the time of on mouse being covered definite word that machine produces text, the description that shows the key concept of naming with this word is as prompting.Identical sentence in urtext by mark suitably.The 3rd zone is a toolbar, is used to change the clear and definite model that is applied on the second area.These instruments comprise the entity that changes explanation when providing the synonym of word, and this synonym is the synonym with another entity of the word name in handling.It can be used as the description that prompting provides the key concept of being named by synonym.It comprises the mode of the feature of selecting text, for example plays with words, satire, poem or scientific and technological text.The concrete connotation that comprises the replacer who defines employed noun for example is actually he, she or it.This concrete connotation is defined within the whole range of text, because it has set the relation at preceding sentence that gives text with clear and definite noun.Text is examined continuously from the beginning to the end, because it has provided all required feature and relations, has therefore formed clear and definite model.Machine has produced the text that has synonymous with urtext at least when handling sentence.This processing comprises the setting that changes and produce thing.
The exemplary realization of the 4th step of this method
The clear and definite model of the given text that produces is affixed on the source document.Can realize described additional with multiple mode.Can in source document, be increased to the link of the clear and definite model of the text.The file of file in the urtext and clear and definite model can be written in the file package.Must remember that the general text of natural language can have the clear and definite model of a plurality of formation.This is that what the operator used is his understanding because come numerous explanations of the given text of natural language are filtered by people-operator, so that he is a natural language with text translation in clear and definite machine mould.Therefore prediction can append to natural language text on a lot of clear and definite models.When being patented claim, the object of being protected is unique clear and definite model of the text of this application naturally, identical with its application.
The exemplary realization of the 5th step of this method
The clear and definite model of natural language text goes for formal processing.Can generate different types of expression of clear and definite model, it is applicable to different types of machine processing.Clear and definite model can be defined as the new kind of computer software, because they can be the objects of form of explanation.In this way, can realize machine learning, because it obtains truth and relation from the clear and definite model of natural language text.Can all with formal application clearly mechanism of in artificial intelligence, being studied.In this way, can substitute traditional software with dedicated system, this dedicated system interrelates with the natural language and the domestic consumer that increase clear and definite model easily, and provides service to be used for producing application software according to user's request.
Industrial applicibility
Disclosed method is by special software performing. Computer software can be used by the professional person, generates and the supported data center with the basic conception of human usefulness. Another computer software is used by all users, generates and use the clear and definite model of natural language text. Last computer software can enough basic conception be connected to data center.
The method can be used for machine translation, from a kind of natural language to another kind of natural language, perhaps arrives for example artificial language of computer language. The method can be used for retrieval and process natural language.
Especially, the method be applied in particular importance in the patent system, not only be used for clearly defining object of protection and automatically retrieving and investigate, also be used in up-to-date and valuable human knowledge machine processing, this is the motivation that automatically produces new knowledge for the mankind.
Claims (10)
1. the formalization method of natural language makes it possible to carry out machine interpretation and produces natural language text by the machine mould that generates text, it is characterized in that, generates the clear and definite model of natural language text, and it can only be explained by the unique method that comprises the steps:
Utilize the previous human employed key concept of determining, key concept comprises all key concepts of unique expression of all entities or action, and described key concept is unique label one numeral or word, and described key concept has the description of natural language, and for being used each natural language that this method is handled, described key concept has additional word list, and its title is given natural language;
Use the Computer Analysis natural language text, utilize key concept, particularly, find employed key concept with the word list of the definite key concept of given natural language name, and utilize grammatical analysis and semantic analysis, produce the first clear and definite model of natural language text;
Use the first clear and definite model to produce the text of identical natural language once more with computing machine;
The natural language text and the urtext that relatively produce with computing machine, and mark difference from the first clear and definite model;
The operator utilizes him can check the computer program of key concept, select key concept and change by computing machine, he also determines the relation and the feature of the indiscoverable text of computing machine, speech part for example, action in the compound sentence is attitude regularly really, perhaps time of the action in two continuous sentences, the correct substitute of noun, associated speech part, and how related;
Computing machine utilizes operator's the remarks and the first clear and definite model, and produces the second clear and definite model;
Computing machine utilizes the second clear and definite model to produce the text of identical natural language once more;
Natural language text and urtext that computing machine relatively produces from the second clear and definite model, and mark difference;
The operator revises, and repetition of explanation-generation-correction step, admits the connotation that can represent natural language text well enough that produces recently from the clear and definite model of computing machine up to the operator.
2. the formalization method of natural language according to claim 1, it is characterized in that, also comprise step: with the clear and definite model of the natural language text that forms, by link or by the file of natural language text is put into a file package together with the file that comprises its clear and definite model, append on the identical text.
3. the formalization method of natural language according to claim 1 is characterized in that, also comprises step: the clear and definite model of natural language text is used for machine processing, for example retrieves, extract truth and relation, determine the legal connotation of text.
4. the formalization method of natural language according to claim 1, it is characterized in that, also comprise step: the human translation that compares the urtext of one or more language, exactly and automatically determining employed key concept, the tense of speech part and the relation between them, property, number, action and with the temporal relationship of other actions.
5. the formalization method of natural language according to claim 1 is characterized in that, also comprises step: produce the artificial language text from the clear and definite model of natural language text.
6. a method that is used for determining human employed key concept is used for enforcement of rights and requires 1 described method, it is characterized in that, comprises step:
For each word of natural language, its synonym in the computing machine synonymicon is found out and extracted to computing machine;
For each to word-synonym, computing machine comparison dictionary provide for this word and this synon description;
For identical word or word-synon per two similar texts of the given text that has comprised given number percent, assert that they have described a key concept;
The key concept tabulation that computing machine output is assert, and the description of making this decision;
Key concept to each identification is checked data center, with the similar text found in the step formerly with in the heart the description of key concept compare, determine whether this key concept is registered, if the word or the word-synonym of given number percent are arranged, think that then this key concept is registered, the description of the key concept that finds and other two the similar descriptions that cause retrieval are exported by computing machine;
The operator checks whether the text of exporting by word consistance mode has semantic consistency, if find such consistance, he judges that the key concept provide registers, that he only will register or with two words-synonym adding of certain natural language name key concept;
If in data center, do not find given key concept, from two similar texts, select an adding, perhaps determine to describe by the operator.
7. method that is used for increasing new natural language to the key concept center that forms, it is used to satisfy the described needs that utilize the method for human used key concept of claim 1, it is characterized in that, comprises step:
Utilize the method that is used for newspeak according to claim 6, form the second key concept center;
Formation is from the dictionary of second kind of language to the first kind of language (in the heart), and each title of the key concept at new center is found out possible translation;
For each translation-word of first kind of language, extraction can be named the key concept of this word;
Vacation-translation is carried out in description to the key concept of second kind of language, produces all combinations with all possible translation of first kind of language of substituting of each word in describing;
Vacation-the translation of the description of the key concept of second kind of language of comparison and extraction same word or word-synon number percent in the description of the key concept at first center;
Find out best consistance and mark;
Improve each consistance of being found out in this way by the operator, whether the similar description that operator's decision is found by similar word has semantic consistency;
After the consistance of improving in second in the heart, key concept is removed, and the name tabulation of the key concept of second kind of language of mark shows it is in second kind of language, and is increased in the key concept at first center;
After handling all consistance, still in the heart those key concepts register to first center as new key concept in second, perhaps find out their in the heart consistance in first by the operator.
8. be used to realize the special software of the described method of claim 1, it can Edit Text, it is characterized in that having following function:
Open a link, write in the data center before according to claim 6 or according to the set of claim 6 and 7 key concepts of preparing to data center;
Utilization produces the clear and definite model of natural language text for the previous key concept of preparing of given natural language;
Produce natural language text from clear and definite model;
Setting produces natural language from clear and definite model;
Relevant sentence in the text of mark urtext and generation;
Difference between the relevant sentence in the text of mark urtext and generation;
For definite word of natural language, the description of the selected key concept of expression computing machine is with specifying word in the urtext or specifying word to finish this in the text that produces according to clear and definite model and represent;
Make the operator can directly change or point out the key concept of synonym, computing machine appends to this key concept on the word of natural language text;
Make the operator can point out speech part and relation from the part of speech to another part;
Make the operator can point out to meet the temporal relationship between the action in the sentence or the temporal relationship of the action in two adjacent sentences;
Make the operator can point out the content that substitutes with special title;
Make the operator can point out the surface of text, for example the subject fields of text, whether be irony, satirize or play with words.
9. special software according to claim 8, it is characterized in that, also has function: along with the description that utilizes the key concept of using in the text, produce the description of any complexity level, and the baseline concept description that recursively is used for determining more high-grade key concept, and describe with this and to substitute this key concept.
10. special software according to claim 8, it is characterized in that, also has function: retrieve or handle clear and definite model, rather than retrieval or processing natural language text, can also represent retrieval or result by producing natural language or artificial language text in addition, perhaps the result is expressed as the consistance of natural language text.
Applications Claiming Priority (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| BG10109996A BG66255B1 (en) | 2007-11-14 | 2007-11-14 | Natural language formalization |
| BG109996 | 2007-11-14 | ||
| PCT/BG2008/000022 WO2009062271A1 (en) | 2007-11-14 | 2008-11-12 | Formalization of a natural language |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101855630A true CN101855630A (en) | 2010-10-06 |
Family
ID=40638266
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN200880115885A Pending CN101855630A (en) | 2007-11-14 | 2008-11-12 | Formalization of a natural language |
Country Status (8)
| Country | Link |
|---|---|
| EP (1) | EP2220572A4 (en) |
| JP (2) | JP2011503730A (en) |
| KR (1) | KR101506757B1 (en) |
| CN (1) | CN101855630A (en) |
| BG (1) | BG66255B1 (en) |
| CA (1) | CA2705345A1 (en) |
| EA (1) | EA201070614A1 (en) |
| WO (1) | WO2009062271A1 (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104011711A (en) * | 2011-12-27 | 2014-08-27 | 皇家飞利浦有限公司 | Text analysis system |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10303769B2 (en) * | 2014-01-28 | 2019-05-28 | Somol Zorzin Gmbh | Method for automatically detecting meaning and measuring the univocality of text |
| CN111401033B (en) * | 2020-03-19 | 2023-07-25 | 北京百度网讯科技有限公司 | Event extraction method, event extraction device and electronic equipment |
| CN112861548B (en) * | 2021-02-10 | 2023-06-23 | 百度在线网络技术（北京）有限公司 | Training method, device, equipment and storage medium for natural language generation and model |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6014680A (en) * | 1995-08-31 | 2000-01-11 | Hitachi, Ltd. | Method and apparatus for generating structured document |
| US20020128816A1 (en) * | 1997-09-30 | 2002-09-12 | Haug Peter J. | Probabilistic system for natural language processing |
| CN1484217A (en) * | 2003-07-11 | 2004-03-24 | 中国科学院声学研究所 | Hierarchical classifying and logic combined natural colloquial conversation dessription method |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1325430A2 (en) * | 2000-09-29 | 2003-07-09 | Axonwave Software Inc. | A method and system for adapting synonym resources to specific domains |
| WO2006128238A1 (en) | 2005-06-02 | 2006-12-07 | Newsouth Innovations Pty Limited | A method for summarising knowledge from a text |
-
2007
- 2007-11-14 BG BG10109996A patent/BG66255B1/en unknown
-
2008
- 2008-11-12 EP EP08850309A patent/EP2220572A4/en not_active Withdrawn
- 2008-11-12 JP JP2010533390A patent/JP2011503730A/en active Pending
- 2008-11-12 KR KR1020107013115A patent/KR101506757B1/en active IP Right Grant
- 2008-11-12 EA EA201070614A patent/EA201070614A1/en unknown
- 2008-11-12 CN CN200880115885A patent/CN101855630A/en active Pending
- 2008-11-12 WO PCT/BG2008/000022 patent/WO2009062271A1/en active Application Filing
- 2008-11-12 CA CA2705345A patent/CA2705345A1/en not_active Abandoned
-
2014
- 2014-02-21 JP JP2014031296A patent/JP2014139799A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6014680A (en) * | 1995-08-31 | 2000-01-11 | Hitachi, Ltd. | Method and apparatus for generating structured document |
| US20020128816A1 (en) * | 1997-09-30 | 2002-09-12 | Haug Peter J. | Probabilistic system for natural language processing |
| CN1484217A (en) * | 2003-07-11 | 2004-03-24 | 中国科学院声学研究所 | Hierarchical classifying and logic combined natural colloquial conversation dessription method |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104011711A (en) * | 2011-12-27 | 2014-08-27 | 皇家飞利浦有限公司 | Text analysis system |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2009062271A1 (en) | 2009-05-22 |
| BG66255B1 (en) | 2012-09-28 |
| EP2220572A1 (en) | 2010-08-25 |
| BG109996A (en) | 2009-05-29 |
| KR20100108338A (en) | 2010-10-06 |
| JP2011503730A (en) | 2011-01-27 |
| EA201070614A1 (en) | 2010-10-29 |
| KR101506757B1 (en) | 2015-03-27 |
| EP2220572A4 (en) | 2011-03-09 |
| CA2705345A1 (en) | 2009-05-22 |
| JP2014139799A (en) | 2014-07-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR100533810B1 (en) | Semi-Automatic Construction Method for Knowledge of Encyclopedia Question Answering System | |
| CN109344236A (en) | One kind being based on the problem of various features similarity calculating method | |
| CN106919689A (en) | Professional domain knowledge mapping dynamic fixing method based on definitions blocks of knowledge | |
| US7853595B2 (en) | Method and apparatus for creating a tool for generating an index for a document | |
| CN101855630A (en) | Formalization of a natural language | |
| JP6409071B2 (en) | Sentence sorting method and calculator | |
| Westphal et al. | Countering language attrition with PanLex and the Web of Data | |
| JP2011503730A5 (en) | ||
| Zinn et al. | The isocat registry reloaded | |
| Anstein | Computational approaches to the comparison of regional variety corpora: prototyping a semi-automatic system for German | |
| Soni et al. | Toward a neural semantic parsing system for EHR question answering | |
| Dimitriadis et al. | How to integrate databases without starting a typology war: The Typological Database System | |
| Cui et al. | Combine unsupervised learning and heuristic rules to annotate organism morphological descriptions | |
| KR20170115406A (en) | Apparatus and method for processing biological system information | |
| Parvez | Named entity recognition from bengali newspaper data | |
| Rechenmacher et al. | The Contibution of Wolfgang Richter to Current Developments in the Study of Biblical Hebrew | |
| Baud et al. | Knowledge sources for Natural Language Processing. | |
| Ahmad et al. | Terminology management: a corpus-based approach | |
| Abdelkader et al. | How Existing NLP Tools of Arabic Language Can Serve Hadith Processing | |
| Hudon | An assessment of the usefulness of standardized definitions in a thesaurus through interindexer terminological consistency measurements | |
| Schloen et al. | An Item-Based Approach: Propertize | |
| Bi et al. | Construction of a Descriptive Framework for Grammatical Attributes of Relational Mark of Korean Compound Sentence Based on Knowledge Representation | |
| Zhao et al. | Automatic Recognition and Extraction of English Verb Types Based on Index Line Clustering | |
| Alade et al. | ISSUES IN KNOWLEDGE REPRESENTATION IN YORÙBÁ NARRATIVES | |
| Biffi | Accademia della Crusca’s Online Dictionaries |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| AD01 | Patent right deemed abandoned |
Effective date of abandoning: 20171219 |
|
| AD01 | Patent right deemed abandoned |