CN107340999A - Software automation method and system and the method in structure natural language understanding storehouse - Google Patents
Software automation method and system and the method in structure natural language understanding storehouse Download PDFInfo
- Publication number
- CN107340999A CN107340999A CN201710013071.9A CN201710013071A CN107340999A CN 107340999 A CN107340999 A CN 107340999A CN 201710013071 A CN201710013071 A CN 201710013071A CN 107340999 A CN107340999 A CN 107340999A
- Authority
- CN
- China
- Prior art keywords
- semantic
- knowledge
- vocabulary
- understanding
- lexical
- 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
- G06F8/00—Arrangements for software engineering
- G06F8/20—Software design
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/205—Parsing
- G06F40/211—Syntactic parsing, e.g. based on context-free grammar [CFG] or unification grammars
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/253—Grammatical analysis; Style critique
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F8/00—Arrangements for software engineering
- G06F8/40—Transformation of program code
- G06F8/41—Compilation
- G06F8/44—Encoding
- G06F8/447—Target code generation
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Machine Translation (AREA)
Abstract
The invention discloses a kind of method and system of the automatic software production based on natural language understanding, content includes：First, the structure model in a natural language understanding storehouse is pre-established, builds lexical semantic knowledge base；2nd, the system for establishing a natural language understanding, to software process descriptive information, machine first carries out understanding parsing, builds syntactic structure corresponding to sentence semantics；3rd, establish and understand deduction model, on the basis of grammatical and semantic structure, machine is deduced, intractability element, search rule or semantic values are constantly extracted, insertion rewrites intractability element, the process repeats, and untill no intractability element, is formed and solves string；4th, formalized model is established, the solution string that the model is formed by arranging these solution processes, after being formalized to solution string extraction operation semanteme, software code row set corresponding to output.Using the present invention, the development rate of software and the reliability of software can be effectively improved.
Description
【Technical field】The present invention relates to a kind of automatic software production method, more particularly to natural language description
Demand analysis directly understands, deduced and form turns to the software automation method of software, software automation system, and to nature
Language carries out substantive understanding and the method for building natural language understanding storehouse.
【Background technology】Programming is a cumbersome mental labour, and software automation is then that one kind allows machine to enter instead of the mankind
The method of row automated programming, efficiency can be improved, liberate mental labour.But current software automation degree is still within semi-automatic add
The situation of manual intervention, the degree of automatic software production are still relatively low.It is embodied in, first, software automation realizes way
Footpath is converted to functional specification firstly the need of software requirement description, because the process of non-formalization to formalization is very tired
Difficulty, the specification that the software requirement of non-formalization is converted into formalization still need manual mode, and this is that software production is automatic
One of difficult point of change.Most system effectivenesies are relatively low at present from requirements specification to functional specification, also easily make mistakes.The
Two, currently, the either software factory of Component- Based Development or model-driven framework MDA, it is all based on the specification of formalization
On, and the Deductive Synthesis on the basis of formalization, the scale also very little of the software automation of this respect, from the perspective of from functional specification
It is bright to be still difficult to automate to generation design specification；
【The content of the invention】
The present invention provides a kind of automatic software production method and system so that software project or application are achieved.Root
According to one aspect of the present invention, there is provided one kind directly exports soft from natural language description information by understanding, deduction, formalization
The software automation method of part, it is different from the describing mode of natural language.Natural language is limited natural language, is belonged to
Formalization method, this method are directly based upon natural language description demand analysis, as long as what program done, how to operate retouches by hand
State out, the software automation method can automatically generates required software.Wherein, description information include operating process,
Working specification and knowledge description, these knowledge descriptions in descriptive text can also be individually placed.This method can reduce labor
Fatigue resistance, improve programming efficiency.
According to an aspect of the present invention, there is provided a kind of software automation system, held first, can increase according to description information
Capable accuracy, improve the correctness to demand processing；Second, it is wrong to reduce the subtle machine of the mankind caused by human factor
By mistake, the accuracy of programming is improved；Third, improving programming efficiency, the labor intensity manually programmed is reduced.
According to an aspect of the present invention, there is provided one kind structure natural language understanding storehouse is semantic knowledge-base (general knowledge storehouse)
Method.The purpose of the present invention is to be achieved through the following technical solutions a kind of software automation method, and this method includes A, understands (machine
Device).Description information is understood, abbreviation script understands；Script understand include Word Understanding, sentence comprehension, sentence group understanding with
And semantic expressiveness；Morphological analysis is carried out on the basis of understanding, can reach the reasonability of word analysis；Implement syntax point by understanding
Analysis, to reach the reasonability of syntactic analysis；By morphological analysis and syntactic analysis, the final text for building semantic understanding target is known
Feel semantic structure, prepared for subsequent process；B, (machine) is deduced.Intractability element, i.e., still unsolved element, it is known that condition
In the element that can not search out be then intractability element.Transaction is extracted on the basis of the semantic expressiveness of step A description information, is solved
Transaction is read, judgement rewriting is carried out to intractability element, so as to constantly expansion transaction, problem is obtained deducing and solved.Detailed process is：
Transaction (program header) is gone out by semantics recognition first, transaction is problem string；Machine is analyzed problem string, finds to catch hardly possible
Solution property element.If capturing, if can match, it is known that if be set to solvability element；Otherwise, to intractability element in this text
Rule or what values are scanned for, i.e. search semanteme or cause and effect string, rewrite intractability element.Reanalyse problem string, see whether
There is intractability element.This catches rewrite process and is carried out continuously, untill without intractability element；So as to obtain the problem of final
String, also known as reasoning string or solution string.Replacement policy, to reduce the approach of solution, avoids solving way also including the use of heuristic knowledge
The multiple shot array in footpath.C, (machine) is formalized：Formalization is the arrangement and semiosis to reasoning string, whole from reasoning string
Manage out program circuit.Reasoning string, which arranges, includes conversion translation of identification variables, operator and function etc. etc..Identification variables pass through handle
The preceding variable words recognition that during what values are semantic there is machine value to point to feature is variable, so as to realize the semanteme according to concept
The variable format that feature is carried out.The conversion translation of operator is then to carry out operator according to semanteme directly to understand conversion, is
The one-to-one corresponding process of natural language and program language is carried out using consciousness semantic feature as intermediary.Pass through identification variables AND operator
Deng conversion translation, so as to extract the operational semantics in reasoning string, form program code row set.
It is preferred that further comprise A ', collection formatting or unformatted lexical information, generation before the step A
Lexical semantic knowledge base.In the above method, the step A ' includes A ' 1, judges whether the vocabulary definitions method is to format
Method, if it is, performing step A ' 2, the lexical information that collection formats, generate lexical semantic knowledge base；Otherwise, step is performed
Rapid A ' 3, the lexical information for gathering unformatted, generate lexical semantic knowledge base.In the above method, the collection lattice of step A ' 2
The lexical information of formula, generation lexical semantic knowledge base includes A ' 21, determines to need the vocabulary point of automatic data collection, needs to gather to be described
Vocabulary point distribution vocabulary ID；A ' 22, according to the vocabulary point for needing automatic data collection, lexical information record is acquired, root
According to mark acquisition and the corresponding lexical semantic information of vocabulary point for needing automatic data collection and relevant knowledge information is formatted, for institute
The lexical semantic point distribution lexical semantic point ID that predicate remittance packet contains；A ' 23, by the lexical information in the form of five-tuple
Stored；The five-tuple of any vocabulary point includes semantic, the semantic point ID of word ID (vocabulary), consciousness, example and relevant
Knowledge.In the above method, step A ' 3 is described to be acquired to lexical information record, is obtained and the vocabulary for needing automatic data collection
The lexical information that lexical semantic information corresponding to point includes A ' 31, pair determination is acquired carries out morphological analysis and syntactic analysis；
The vocabulary that the lexical information point after A ' 32, identification formatting includes, obtains vocabulary point；A ' 33, from the meaning description record in
Semantic point information of the selection comprising the vocabulary；A ' 34, the semantic point letter for obtaining and becoming more meticulous is understood by example (relevant knowledge)
Breath, and extract relevant knowledge semanteme.A ' 35, the lexical information stored in the form of five-tuple；Any vocabulary
The five-tuple of point includes word ID (vocabulary), consciousness semantic, semantic point ID, example and knowledge.Lexical information is with the shape of five-tuple
Formula is stored, and it will carry out vocabulary playback, semantic playback before storing；Vocabulary playback includes inquiring about general knowledge storehouse, sees that the vocabulary is
It is no existing, the semantic similarities and differences should be differentiated if existing, then consistently increase definition；Semanteme playback includes inquiring about general knowledge storehouse
Semantic component, see whether the semanteme existing, and consistently increase semantic interlink.Knowledge playback includes inquiring about general knowledge storehouse,
See whether the knowledge is existing, and consistently increase the knowledge semantic.Semantic and knowledge, its storage form are all structurings
's；Semanteme therein is consciousness semanteme.
【Brief description of the drawings】
Fig. 1 is the structure block schematic diagram of existing software automation system, automatic to existing software in conjunction with Fig. 1
The structure of change system illustrates, specific as follows：Existing software automation system includes Understanding Module 101, deduces module 102
With formalization module 103.Understanding Module 101 receives description information, and according to what-why values (the consciousness semanteme of description information
Collection), be word syntactic structure description information cutting, and its it is corresponding understand result, the understanding result be corresponding to semantic structure
Collection.Deduce module 102 and receive the understanding result that natural language understanding system (Understanding Module 101) is sent, and to the understanding result
The problem of being deducted, forming expansion is gone here and there, or is process of problem solving string.Module 103 is formalized to receive to deduce module
(102) the problem of sending solves process string, the problem string is formalized, final output agenda.Existing software is certainly
The structure stage (105) of the knowledge base 104 of dynamicization system excavates vocabulary template, structure, it is necessary to define centering from the new term of input
Build out lexical semantic ATL, for understand unit 101 inquiry obtain semantic pattern, knowledge base (lexical semantic ATL or
Lexical semantic storehouse) in vocabulary definitions include semantic pattern, knowledge schema etc.；Neologisms of the serial analysis technology to input can be used
The definition that converges is handled to obtain corresponding with vocabulary perceptual pattern collection, can also to the vocabulary knowledge progress knowledge parsing of input with
Obtain the knowledge related to vocabulary.The knowledge preserved in knowledge base 104 is vocabulary language corresponding with the functional mode of the vocabulary
Connection between justice, and consciousness semanteme set representations knowledge and vocabulary definitions in the use artificial intelligence field of knowledge base 104, and on
Stating the structure of knowledge base 104 needs to be accomplished manually to be combined with being automatically performed.The knowledge base structure of existing software automation system
Build larger with the cost safeguarded, and need business personnel to summarize word language material and knowledge, O＆M needs to be continuously added into new term new
Knowledge；Because knowledge base (general knowledge storehouse) can be increasing, deduce or the process of calculating is deduced or understood to Understanding Module unit
Can be more and more time-consuming, efficiency declines, and existing software automation system need further to improve.
【Claim】
1st, a kind of software automation method and system, it is characterised in that this method or system include step A, by reception
Description information carries out understanding processing, obtains the grammatical and semantic structure of description information；The description information includes the grammer of multiple sentences
Semantic structure, i.e. grammatical and semantic structure collection；This is Understanding Module (natural language understanding system)；B, according to the grammer of description information
The extraction of semantic structure collection determines intractability element；Intractability element is initial problem string.C, formula is deduced using intractability element
Expansion problems string, untill the problem string does not have intractability element, BC is deduction module；D, gone here and there and arranged according to the problem of final
Go out action sequence, and the action sequence is formalized, export program statement sequence sets corresponding with the description information, this is form
Change module.
2nd, the method according to claim 11 or system, it is characterised in that further comprise before the step A A ',
Collection formats or the lexical information of unformatted record, and after implementing structuring processing in advance, generation lexical semantic knowledge base is (often
Know storehouse), this is the method and structure module in structure natural language understanding storehouse.The step A ' includes A ' 1, judges that the vocabulary is determined
Whether right way of conduct method is formatting method, if it is, performing step A ' 2, the lexical information that collection formats, generates lexical semantic
Knowledge base, otherwise, perform step A ' 3, gather the lexical information of unformatted, generate lexical semantic knowledge base.
3rd, according to the method for claim 2, it is characterised in that the vocabulary that the collections of step A ' 2 format
Information, lexical semantic knowledge base is generated, including A ' 21, determination need the vocabulary point of automatic data collection, are the vocabulary point that need to be gathered
Distribute vocabulary ID；A ' 22, according to the vocabulary point for needing automatic data collection, lexical information record is acquired, according to formatting
Mark obtains lexical semantic information corresponding with the vocabulary point for needing automatic data collection and relevant knowledge, is the lexical information bag
The lexical semantic allocated semantics point ID contained；A ' 23, the lexical information stored in the form of five-tuple；Any institute's predicate
The five-tuple of meeting point includes word ID (vocabulary), consciousness semantic, semantic point ID, example and relevant knowledge.
4th, according to the method for claim 2, it is characterised in that in the above method, the collection nonformats of step A ' 3
The lexical information record of change, generation lexical semantic knowledge base includes A ' 31, determines to need the vocabulary point of automatic data collection, need to be adopted to be described
The vocabulary point distribution vocabulary ID of collection；The lexical information that A ' 32, pair determination are acquired implements the lattice such as morphological analysis and syntactic analysis
Formula step；A ' 33, the vocabulary that the lexical information point of the formatted automatic data collection includes is identified, obtain vocabulary point；
The semantic and corresponding knowledge of word is determined, and distributes phrase semantic point ID；A ' 34, select the language that includes from meaning description record
Justice point information simultaneously understands consciousness semantic structure collection corresponding to output；A ' 35, the language to be become more meticulous by example (knowledge) understanding acquisition
Justice point information.A ' 36, identified lexical information stored in the form of five-tuple；The five-tuple of any vocabulary point
Comprising word ID (vocabulary), consciousness semantic, semantic point ID, example and related structured knowledge, the preservation of these quintuple forms
Vocabulary point information aggregate forms the lexical semantic knowledge base (semantics comprehension on natural language storehouse) of the system.
5th, the method according to claim 2-4, it is characterised in that semanteme and knowledge described in step A ' 2 and A ' 3,
Its storage form is all structuring.The step A ' 35 understands the semantic point information for obtaining and becoming more meticulous, these letters by example
Breath includes part of speech consciousness, and the semantic collection of consciousness of vocabulary definitions is fulfilled.
6th, the method and system according to claim 4 and described in claim 1, it is characterised in that the step A ' 32 with
Step A vocabulary understands, including the matching to lexicon, the description information that will be received gradually perform matching to lexicon,
And effect is understood with the basis of the understanding of consciousness collection, carrying out understanding semantic corresponding to vocabulary and judging according to what-why, and handle
Understand that result is output to parsing unit, this understands unit for the word (vocabulary) of the Understanding Module.The step A ' 32
Matched with step A syntactic analysis, including the knowledge of word sequence set pair knowledge base or semantic template, corresponding knowledge semantic
Understanding be that on the basis of understanding that effect and consciousness collection understand based on what-why, knowledge semantic corresponding to Lexical collocation is carried out
Understand what is judged, i.e., by understanding that (true value of semantic level is about so as to carry out the why factor constraints of semantic level to collocation
Beam), this is the parsing unit of the Understanding Module；According to the syntactic analysis based on understanding, to morphology cutting (morphology point
Analysis) it is adjusted, this is the morphological analysis adjustment unit of the Understanding Module.Described semanteme and semantic knowledge are in order to understand mesh
, applied to the morphological analysis and syntactic analysis on the basis of understanding, effect is understood to determine morphology according to the what-why of semanteme
The reasonability of analysis and syntactic analysis.
7. according to the method and system described in claim 1-6, it is characterised in that the step A ' 36 and A ' 23 are by described in
Lexical information is stored in the form of five-tuple, and it will carry out vocabulary playback, semantic playback before storing；Vocabulary playback includes looking into
General knowledge storehouse is ask, sees whether the vocabulary is existing, the semantic similarities and differences should be differentiated if existing, then consistently increase definition；Language
Justice playback includes inquiring about the semantic component in general knowledge storehouse, sees whether the semanteme is existing, and consistently increase semantic interlink.Knowledge
Playback includes inquiring about general knowledge storehouse, sees whether the knowledge is existing, and consistently increase the knowledge semantic.The step A ' 23 with
Semanteme and knowledge described in A ' 36, its storage form are all structurings；Semanteme therein is consciousness semanteme.
8. according to the method and system described in claim any one of 1-7, it is characterised in that the step C, which is deduced, to be included
Intractability element is replaced to insert and rewrite with semantic or knowledge, the intractability mark of the element removes, and this is intractability element
Rewriting unit；If it was found that not in known conditions, for intractability element, this finds for the intractability element of the deduction module
Unit；Replacement policy, to reduce the approach of solution, avoids the multiple shot array of solution approach also including the use of heuristic knowledge.
9. according to the method and system described in claim any one of 1-8, it is characterised in that the step D-shaped formula bag
Include D1, the preceding variable word of problem string is identified the machine value sensing feature in the corresponding semanteme of word and form
Change, this is the preceding variable word form unit for formalizing module；D2, the word in solution string and corresponding programming
Corresponding relation and the programming language rule of the semantic phase same sex of language, semantic feature is carried out to language elements such as oeprators
Identification and translation corresponding with programming language term, this is translation unit；The result exported after formalization forms the behaviour of description information
Make semanteme, this is formalization output unit；D3, the tissue for passing through variable arrangement and statement sequence of being gone here and there to the problem of expansion, output
Program statement arrangement set, this is program circuit organizational unit.
10th, the method and system according to claim 1-9, it is characterised in that step A and step A ' 32 forms nature
The method of language understanding；Understanding Module and structure module composition natural language understanding system, including：(1) lexical semantic definition collection
(dictionary or general knowledge storehouse), for storing the semanteme of word corresponding thereto, and related knowledge；(2) description information word method
Cutter (word parsing unit), description information are sent phrase semantic to define collection and matched, by relevant knowledge semanteme
Grammatical and semantic structure collection that is corresponding, being formed after understanding, while analyze each syntactic information.
Claims (10)
1. a kind of software automation method and system, it is characterised in that this method or system include step A, the description by reception
Information carries out understanding processing, obtains the grammatical and semantic structure of description information；The description information includes the grammatical and semantic of multiple sentences
Structure, i.e. grammatical and semantic structure collection；This is Understanding Module (natural language understanding system)；B, according to the grammatical and semantic of description information
The extraction of structure collection determines intractability element；Intractability element is initial problem string.C, formula expansion is deduced using intractability element
Problem string, untill the problem string does not have intractability element, BC is deduction module；D, arrangement of being gone here and there according to the problem of final is set out
Make sequence, and the action sequence is formalized, export program statement sequence sets corresponding with the description information, this is formalization mould
Block.
2. the method according to claim 11 or system, it is characterised in that further comprise A ', collection before the step A
Format or the lexical information of unformatted records, after implementing structuring processing in advance, generate lexical semantic knowledge base (general knowledge
Storehouse), this is the method and structure module in structure natural language understanding storehouse.The step A ' includes A ' 1, judges the vocabulary definitions
Whether method is formatting method, if it is, performing step Α ' 2, the lexical information that collection formats, generates lexical semantic
Knowledge base, otherwise, perform step A ' 3, gather the lexical information of unformatted, generate lexical semantic knowledge base.
3. according to the method for claim 2, it is characterised in that the lexical information that the collections of step A ' 2 format,
Lexical semantic knowledge base is generated, including A ' 21, determination need the vocabulary point of automatic data collection, and word is distributed for the vocabulary point that need to be gathered
Remittance ID；A ' 22, according to the vocabulary point for needing automatic data collection, lexical information record is acquired, obtained according to mark is formatted
Lexical semantic information corresponding with the vocabulary point for needing automatic data collection and relevant knowledge, the word included for the lexical information
Converge semantic allocated semantics point ID；A ' 23, the lexical information stored in the form of five-tuple；Any vocabulary point
Five-tuple includes word ID (vocabulary), consciousness semantic, semantic point ID, example and relevant knowledge.
4. according to the method for claim 2, it is characterised in that in the above method, the collection unformatteds of step A ' 3
Lexical information records, and generation lexical semantic knowledge base includes A ' 31, determines to need the vocabulary point of automatic data collection, described to gather
Vocabulary point distribution vocabulary ID；The lexical information that A ' 32, pair determination are acquired is implemented morphological analysis and syntactic analysis etc. and formatted
Step；A ' 33, the vocabulary that the lexical information point of the formatted automatic data collection includes is identified, obtain vocabulary point；It is determined that
The semantic and corresponding knowledge of word, and distribute phrase semantic point ID；Α ' 34, select the semanteme that includes from meaning description record
Point information simultaneously understands consciousness semantic structure collection corresponding to output；Α ' 35, the semanteme to be become more meticulous by example (knowledge) understanding acquisition
Point information.A ' 36, identified lexical information stored in the form of five-tuple；The five-tuple bag of any vocabulary point
ID containing word (vocabulary), consciousness semantic, semantic point ID, example and related structured knowledge, the word that these quintuple forms preserve
Meeting point information aggregate forms the lexical semantic knowledge base (semantics comprehension on natural language storehouse) of the system.
5. according to the method described in claim 2-4, it is characterised in that semanteme and knowledge described in step A ' 2 and A ' 3, it is stored up
It is all structuring to deposit form.The step Α ' 35 understands the semantic point information for obtaining and becoming more meticulous, these packets by example
Part of speech consciousness is included, and the semantic collection of consciousness of vocabulary definitions is fulfilled.
6. according to the method and system described in claim 4 and claim 1, it is characterised in that the step A ' 32 and step A
Vocabulary understand that including the matching to lexicon, the description information that will be received gradually performs matching to lexicon, and according to
On the basis of what-why understands that effect understands with consciousness collection, carry out understanding semantic corresponding to vocabulary and judge, and tied understanding
Fruit is output to parsing unit, and this understands unit for the word (vocabulary) of the Understanding Module.The step A ' 32 and step A
Syntactic analysis, including the knowledge of word sequence set pair knowledge base or semantic template matched, the understanding of corresponding knowledge semantic
It is on the basis of understanding that effect understands with consciousness collection based on what-why, understanding to be carried out to knowledge semantic corresponding to Lexical collocation and sentenced
Fixed, i.e., the why factor constraints of semantic level (true value of semantic level constrains) are carried out to collocation by understanding, this is
The parsing unit of the Understanding Module；According to the syntactic analysis based on understanding, morphology cutting (morphological analysis) is adjusted
Whole, this is the morphological analysis adjustment unit of the Understanding Module.Described semanteme and semantic knowledge are applied to understand purpose
Morphological analysis and syntactic analysis on the basis of understanding, effect is understood to determine morphological analysis and syntax according to the what-why of semanteme
The reasonability of analysis.
7. according to the method and system described in claim 1-6, it is characterised in that the step A ' 36 and A ' 23 are by the vocabulary
Information is stored in the form of five-tuple, and it will carry out vocabulary playback, semantic playback before storing；It is normal that vocabulary playback includes inquiry
Know storehouse, see whether the vocabulary is existing, the semantic similarities and differences should be differentiated if existing, then consistently increase definition；Semanteme is returned
Position includes the semantic component in inquiry general knowledge storehouse, sees whether the semanteme is existing, and consistently increase semantic interlink.Knowledge playbacks
Including inquiry general knowledge storehouse, see whether the knowledge is existing, and consistently increase the knowledge semantic.The step A ' 23 and A ' 36
Described semanteme and knowledge, its storage form are all structurings；Semanteme therein is consciousness semanteme.
8. according to the method and system described in claim any one of 1-7, it is characterised in that the step C, which is deduced, includes term
Justice or knowledge replace intractability element to insert and rewrite, and the intractability mark of the element removes, and this rewrites for intractability element
Unit；If it was found that not in known conditions, for intractability element, this finds single for the intractability element of the deduction module
Member；Replacement policy, to reduce the approach of solution, avoids the multiple shot array of solution approach also including the use of heuristic knowledge.
9. according to the method and system described in claim any one of 1-8, it is characterised in that the step D-shaped formulaization include Dl,
Machine value in the corresponding semanteme of word points to feature and the preceding variable word of problem string is identified and formalized, and this is
The preceding variable word form unit of the formalization module；D2, the word in solution string and corresponding programming language art
The semantic phase same sex of language corresponding relation and programming language rule, to the language elements such as oeprator carry out semantic feature identification and
Translation corresponding with programming language term, this is translation unit；The result exported after formalization forms the operational semantics of description information,
This is formalization output unit；D3, the tissue for passing through variable arrangement and statement sequence of being gone here and there to the problem of expansion, output program sentence
Arrangement set, this is program circuit organizational unit.
10. according to the method and system described in claim 1-9, it is characterised in that step A and step A ' 32 forms natural language
The method of understanding；Understanding Module and structure module composition natural language understanding system, including：(1) lexical semantic definition collection (dictionary
Or general knowledge storehouse), for storing the semanteme of word corresponding thereto, and related knowledge；(2) description information word method cutter
(word parsing unit), description information are sent phrase semantic to define collection and matched, by the semantic correspondence of relevant knowledge, shape
Into the grammatical and semantic structure collection after understanding, while analyze each syntactic information.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710013071.9A CN107340999A (en) | 2017-01-09 | 2017-01-09 | Software automation method and system and the method in structure natural language understanding storehouse |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710013071.9A CN107340999A (en) | 2017-01-09 | 2017-01-09 | Software automation method and system and the method in structure natural language understanding storehouse |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107340999A true CN107340999A (en) | 2017-11-10 |
Family
ID=60222480
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710013071.9A Pending CN107340999A (en) | 2017-01-09 | 2017-01-09 | Software automation method and system and the method in structure natural language understanding storehouse |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107340999A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108021559A (en) * | 2018-02-05 | 2018-05-11 | 威盛电子股份有限公司 | Natural language understanding system and lexical analysis method |
| CN110399597A (en) * | 2018-04-24 | 2019-11-01 | 西门子股份公司 | Template extraction systems, devices and methods |
| CN111158648A (en) * | 2019-12-18 | 2020-05-15 | 西安电子科技大学 | Interactive help system development method based on live-action semantic understanding and platform thereof |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1794240A (en) * | 2006-01-09 | 2006-06-28 | 北京大学深圳研究生院 | Computer information retrieval system based on natural speech understanding and its searching method |
| CN102681982A (en) * | 2012-03-15 | 2012-09-19 | 上海云叟网络科技有限公司 | Method for automatically recognizing semanteme of natural language sentences understood by computer |
| CN103440234A (en) * | 2013-07-25 | 2013-12-11 | 清华大学 | Natural language understanding system and method |
| US20150356200A1 (en) * | 2014-06-09 | 2015-12-10 | Cognitive Scale, Inc. | Composite Cognitive Insights |
-
2017
- 2017-01-09 CN CN201710013071.9A patent/CN107340999A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1794240A (en) * | 2006-01-09 | 2006-06-28 | 北京大学深圳研究生院 | Computer information retrieval system based on natural speech understanding and its searching method |
| CN102681982A (en) * | 2012-03-15 | 2012-09-19 | 上海云叟网络科技有限公司 | Method for automatically recognizing semanteme of natural language sentences understood by computer |
| CN103440234A (en) * | 2013-07-25 | 2013-12-11 | 清华大学 | Natural language understanding system and method |
| US20150356200A1 (en) * | 2014-06-09 | 2015-12-10 | Cognitive Scale, Inc. | Composite Cognitive Insights |
Non-Patent Citations (3)
| Title |
|---|
| 黄培红: "基于自然语言理解的认知系统原理与算法初探", 《计算机科学》 * |
| 黄培红: "探索语言理解与融汇贯通的因果律", 《现代计算机》 * |
| 黄培红: "自然语言加工中What-Why理解效应", 《现代计算机》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108021559A (en) * | 2018-02-05 | 2018-05-11 | 威盛电子股份有限公司 | Natural language understanding system and lexical analysis method |
| CN110399597A (en) * | 2018-04-24 | 2019-11-01 | 西门子股份公司 | Template extraction systems, devices and methods |
| CN110399597B (en) * | 2018-04-24 | 2023-11-17 | 西门子股份公司 | Template extraction system, device and method |
| CN111158648A (en) * | 2019-12-18 | 2020-05-15 | 西安电子科技大学 | Interactive help system development method based on live-action semantic understanding and platform thereof |
| CN111158648B (en) * | 2019-12-18 | 2023-04-07 | 西安电子科技大学 | Interactive help system development method based on live-action semantic understanding and platform thereof |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109213870A (en) | Document process | |
| CN109918489A (en) | A kind of knowledge question answering method and system of more strategy fusions | |
| CN107340999A (en) | Software automation method and system and the method in structure natural language understanding storehouse | |
| CN106325969A (en) | Requirement change reverse tracking system | |
| CN107291450A (en) | A kind of quick code automatic generation method for programming friendly | |
| CN108345686A (en) | A kind of data analysing method and system based on search engine technique | |
| CN109858042A (en) | A kind of determination method and device of translation quality | |
| CN104391881A (en) | Word segmentation algorithm-based log parsing method and word segmentation algorithm-based log parsing system | |
| EP3349131A1 (en) | Method and system for extracting user-specific content | |
| CN109885823A (en) | A kind of distributed semantic recognition methods of financial industry and system and device | |
| CN115438198B (en) | Interpretable medical data structuring method and system based on knowledge base | |
| Markovtsev et al. | STYLE-ANALYZER: fixing code style inconsistencies with interpretable unsupervised algorithms | |
| Kamalabalan et al. | Tool support for traceability of software artefacts | |
| CN117331730A (en) | Fault diagnosis method, device, electronic equipment and storage medium | |
| CN109299179A (en) | Structural data extraction element, method and storage medium | |
| CN107818082A (en) | With reference to the semantic role recognition methods of phrase structure tree | |
| CN112579444A (en) | Text cognition-based automatic analysis modeling method, system, device and medium | |
| CN110400173A (en) | Market sentiment monitoring system method for building up and system | |
| CN116225455A (en) | Method for rapidly generating statistical analysis report | |
| Jha et al. | Wang notation tool: Layout independent representation of tables | |
| CN115795046A (en) | Data processing method, device, system, electronic device and storage medium | |
| Alalfi et al. | Variability identification and representation for automotive Simulink models | |
| Jiang et al. | Fine-tuning BERT-based models for plant health bulletin classification | |
| CN110085287A (en) | The method that records of values carries out quality control in a kind of pair of case history | |
| CN109858027A (en) | One tool method for identifying and classifying of internet four product of electric business merchandise news |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20171110 |
|
| WD01 | Invention patent application deemed withdrawn after publication |