CN109108989A - A kind of legal services special purpose robot of semantics recognition - Google Patents
A kind of legal services special purpose robot of semantics recognition Download PDFInfo
- Publication number
- CN109108989A CN109108989A CN201810805075.5A CN201810805075A CN109108989A CN 109108989 A CN109108989 A CN 109108989A CN 201810805075 A CN201810805075 A CN 201810805075A CN 109108989 A CN109108989 A CN 109108989A
- Authority
- CN
- China
- Prior art keywords
- vocabulary
- law
- legal
- word
- semantics recognition
- 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.)
- Granted
Links
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
- B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
- B25J11/00—Manipulators not otherwise provided for
- B25J11/008—Manipulators for service tasks
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
- B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
- B25J11/00—Manipulators not otherwise provided for
- B25J11/0005—Manipulators having means for high-level communication with users, e.g. speech generator, face recognition means
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
- B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
- B25J5/00—Manipulators mounted on wheels or on carriages
- B25J5/007—Manipulators mounted on wheels or on carriages mounted on wheels
Abstract
The present patent application discloses a kind of legal services special purpose robot of semantics recognition, is related to legal advice field, including robot body and the server being arranged in robot body；The robot body, including being used to acquire the collection terminal asked questions；The server, including being used to the semantics recognition module for carrying out law extraction of semantics is asked questions, the semantics recognition module includes the vocabulary memory module that storage and real-time update have multiple law vocabulary；Semantics recognition module compares each word and the law vocabulary in vocabulary memory module from asking questions middle extraction word, with the law vocabulary replacement of successful match ask questions in word.By the application can the meaning of law expressed by timely learning consultant, and then quickly provide targetedly legal services.
Description
Technical field
The present invention relates to legal advice service fields, and in particular to a kind of legal services special purpose robot of semantics recognition.
Background technique
Artificial intelligence technology develops within the past ten years advances by leaps and bounds, and robot is most representative as artificial intelligence technology
Product be even more closely obtaining biggish breakthrough in two years.Existing robot is mainly used in two general orientation, first is that with
High-precision scientific algorithm humanoid robot is carried out, second is that carry out simple counseling services class robot.First kind robot is main
Applied to industrial environment, price costly, and is had a single function.Although the second class robot gradually extend to service hall,
The public places such as hospital, however it is low to the identifying processing ability of specialized vocabulary, is confined to this kind of robot professional not strong
Field in.
Possess the service industry of a large amount of specialized vocabularies in this way for legal industry, existing service class robot is difficult accurately
The true law meaning that identification consultant expresses, this makes this kind of service robot seem not intelligence.However, because
Legal industry has the general character of service trade, requires to provide service by way of consulting for consultant, to service class machine
Device Man's Demands are growing on and on all the time, therefore are badly in need of releasing a kind of robot for being capable of providing legal services now.
Summary of the invention
The invention is intended to provide a kind of legal services special purpose robot of semantics recognition.
To achieve the above objectives, following scheme is provided:
Scheme one: a kind of legal services special purpose robot of semantics recognition, including robot body and setting are in machine
Server in human body；
The robot body, including being used to acquire the collection terminal asked questions；
The server, including being used to the semantics recognition module for asking questions progress law extraction of semantics, the semanteme
Identification module includes the vocabulary memory module that storage and real-time update have multiple law vocabulary；
Semantics recognition module is from asking questions middle extraction word, and by the law word in each word and vocabulary memory module
Remittance compares, and replaces the word in asking questions with the law vocabulary of successful match.
Explanation of nouns:
Successful match: the meaning phase of a certain word and law vocabulary a certain in vocabulary memory module in asking questions is referred to
Together, then both successful match.
It updates: referring to and increase the old law vocabulary of new law vocabulary and replacement, when the corresponding law of law vocabulary passes through
After revising or thering is new law to promulgate, new law vocabulary will form, the vocabulary of conflict is caused then for new and old law vocabulary
Using new law vocabulary.
The advantages of this programme and effect are:
By the collection terminal being arranged on robot body, what acquisition consultant gave expression to is asked questions for robot, passes through language
Collected ask questions is divided into word one by one by adopted identification module, and by comparing in vocabulary memory module one by one
Word in asking questions is substituted for the law vocabulary of successful match by law vocabulary.It, will be to consulting by semantics recognition module
Problem carries out law extraction of semantics and improves the colloquial legal issue for asking questions and being changed into and being made of law term
Legal profession can carry out professional identification to legal advice problem, be conducive to robot and pointedly provide the law of profession
Service.
Robot in this programme is enabled by the semantics recognition to consultant's law hint expression dedicated for method
Legal services are provided in rule industry.
Scheme two: further, the robot body further includes for showing legal issue and seeking advice from the display of answer
End.
Consulting answer refers to that server being asked questions for what consultant proposed, obtains after providing legal advice service
Legal consequences.The display end being arranged on robot body can not only show the legal issue after asking questions semantics recognition, also
It can show consulting answer, so that consultant is obtained robot and legal services result is provided.
Scheme three: it further, in the vocabulary memory module is equipped with corresponding for carrying out law vocabulary with spoken vocabulary
The synonym dictionary of storage.
By synonym dictionary, the word in asking questions can be carried out with the spoken vocabulary in synonym dictionary one by one
Comparison, finds same spoken vocabulary, and then find the corresponding law vocabulary of the spoken language vocabulary, is replaced with the law vocabulary
The word successively carries out, and colloquial ask questions is converted to law vocabulary.There is provided for the subsequent matching for seeking advice from answer can
Energy.The comparison of synonym dictionary is handled mainly for asking questions for textual.
Scheme four: it further, in the vocabulary memory module is equipped with corresponding for carrying out law vocabulary with spoken vocabulary
The homonym dictionary of storage.
By homonym dictionary, can by saved with audio-frequency information ask questions in word one by one with homonym dictionary
In spoken vocabulary compare, find the spoken vocabulary of same pronunciation, and then find the corresponding law of spoken language vocabulary
Vocabulary, replaces the word with the law vocabulary, successively carries out, and colloquial ask questions is converted to law vocabulary.
Scheme five: further, the server further includes the legal provision memory module for storing legal provision, and
For extracting case processing module of the legal provision as consulting answer from legal provision memory module.
Legal provision refers to be intercepted from the code of written law from by law expert, has the words and phrases completely to look like.Pass through law item
Legal provision is stored in advance in literary memory module, makes to obtain consulting answer to be all that there are laws to abide by, accurate rigorous.Mould is handled by case
Block extracts the legal provision of storage as consulting answer from legal provision memory module, accomplishes providing for answer of consulting all
There are laws to abide by, improves the preciseness and accuracy of answer feedback.
Scheme six: further, a law vocabulary corresponds to multiple spoken vocabulary.
There are many kinds of the expression ways of one word spoken language, and multiple spoken vocabulary are corresponding with a law vocabulary, can
The different word of form of presentation is changed into unified, rigorous law expression.
Scheme seven: further, the robot body further includes for detecting whether the close triggering end of someone, Yi Jiyou
The driving mechanism of triggering end control starting；When triggering end detects someone, triggering end controls driving mechanism and moves towards the direction of people
It is dynamic.
By triggering end, " seeing " is allowed the robot to people, and then by control driving mechanism, allows the robot to " walking "
To in face of people, robot is made to show more intelligent.
Detailed description of the invention
Fig. 1 is the logic diagram of the legal services special purpose robot of semantics recognition of the embodiment of the present invention.
Fig. 2 is the structural schematic diagram of the legal services special purpose robot of semantics recognition of the embodiment of the present invention.
Specific embodiment
Below by the further details of explanation of specific embodiment:
Appended drawing reference in Figure of description includes: all-in-one machine 1, cabinet 2, wheel 3.
Embodiment is substantially as shown in Fig. 1: a kind of legal services special purpose robot of semantics recognition, including robot body
And the server in robot body is set.
Wherein, robot body, including collection terminal, display end, triggering end and the driving mechanism being connect respectively with server.
Collection terminal, for acquiring asking questions for consultant's proposition.Collection terminal can be sound pick-up, at this time collected official communication
Inquire that entitled acoustic information, collection terminal can be keyboard or touch screen, it is collected at this time to ask questions as text information.
Display end, for server is shown from the legal issue that identification extracts is asked questions, meanwhile, also use
To show the consulting answer for being directed to and asking questions.Consulting answer is the knot for the legal services that server provides after semantics recognition
Fruit, consulting answer can be presented in any form, specifically be determined by service content preset in server and service form
Fixed, in the present embodiment, consulting answer occurs in the form of legal opinion.Display end can be only loudspeaker or display screen, when
So it is also possible to that the display of loudspeaker or other electronic equipments with acquisition acoustic information and text information are installed.This reality
It applies in example, collection terminal is the display with loudspeaker, and the display also has touch screen, and collection terminal and display end are just integrated
For an all-in-one machine 1 with touch screen.
Triggering end, for detecting whether that someone is close, when find someone close to when trigger driving mechanism control circuit, make
Driving mechanism driving robot body is located proximate to towards the people being detected.Allow the robot to actively to sense the mankind into
And help is provided for it close to the mankind, so that robot is seemed more intelligent.Triggering end in the present embodiment, using human body infrared
Sensor and the classical serial human body sensing chip of AS08X connected to it, the output pin of human body sensing chip and control drive
The control circuit of mechanism kinematic connects, the on-off of human body sensing chip controls driving mechanisms control circuit, and then controls driving machine
The movement of structure.
Driving mechanism, the motor including wheel 3, the wheel shaft for driving wheel 3 to rotate and drive shaft rotation, the interior peace of motor
It is powered or is broken by triggering end control equipped with being used to control control circuit control circuit that motor rotation perhaps stops operating
Electricity.
Wherein, human body sensing chip and control circuit are all existing, and motor itself can control it by control circuit
It rotates forward, reversion, and then driving mechanism is enable to change direction of advance.
As shown in Fig. 2, robot body includes the all-in-one machine 1, cabinet 2 and wheel 3 successively installed from top to bottom.Server
It is mounted in cabinet 2.
Server can directly select the server sold on existing market, as Dell PowerEdgeT30 server,
Or ThinkServer TS250 of association etc..
As shown in Figure 1, server, including network transceiving module, problem acquire memory module, semantics recognition module, element
Extraction module, subscriber information storing module, legal provision memory module, relationship storage module and case processing module.
Specifically, network transceiving module, including cable network transceiver module and wireless network transceiver module, cable network are received
Hair module can select green PCI-E network interface card, and wireless network transceiver module can select U6 300M enhanced edition WIFI mould of rising
Block or 4.0 module of bluetooth or radio-frequency module or ZigBee module etc..By network transceiving module, enable robot
The network connection of enough and other equipment, facilitates information sharing and monitoring.
Legal provision memory module corresponds to table for storing legal provision, be used to carry out according to law classification and chapters and sections by
Layer classification storage and all kinds of legal provisions of real-time update；Legal provision corresponds in table, including a plurality of legal provision and correspondence are often
The set of tags of a legal provision, set of tags include all multiple labels for being used to indicate this legal provision feature, are specifically included
For indicating the summary word of the legal provision content, for indicating that the legal provision is subordinate to the classification of some chapters and sections of certain code
Word, for indicate the legal provision in the database the grade word of access privileges sequence, for indicating the legal provision and its
The conjunctive word of his legal provision correlation and other words that can be used to indicate the legal provision feature, for example be used to describe
The word of age level, the gender of region, related party that the legal provision often occurs etc..Here legal provision, refers to
Be effective law, regulation, in code according to the law original text sentence divided that completely looks like.Summarizing word is by law
Expert extracts in the sentence, indicates each emphasis word of its causation of law and conditional relationship.In each legal provision
Summarizing is specifically causality or conditional relationship between word, is determined by the law meaning of each legal provision itself.
Relationship storage module, for storing the corresponding priority of grade word and each method of each classification legal provision
Restrain provision relevance grades corresponding with the conjunctive word of other legal provision；For storing in same portion's law, regulation or code,
Chapters and sections relationship, subordinate relation between each legal provision are created as the relational graph of a similar tree.For storing not
In same law, regulation or code, each legal provision and an other law, regulation, between the legal provision in code
Relationship forms the relational graph of a similar reticular structure.
Problem acquires memory module, for storing asking questions for consultant's proposition；The acquisition mode asked questions has more
Kind, it can be direct text input, be also possible to after being inputted by voice input or image after progress semantics recognition extraction again
It is transformed into text input.The problems in the present embodiment acquisition memory module can connect the touch for carrying out text information input
Screen, or connection are able to carry out the sound pick-up of voice pickup, will directly be asked questions by touch screen and be write as text input transmission
Into problem acquisition memory module, it will be asked questions by sound pick-up and existing Iflytek speech recognition is input to by voice
In device, voice input is changed into after text input by Iflytek speech recognition device to be sent in problem acquisition memory module.
Elements recognition module, definition are used to characterize the element word of keyword, and according to element word from asking questions middle extraction
The keyword asked questions is characterized out；And record the number of keyword appearance.Element refers to law expert according to law article system
Fixed is used to judge various conditions required for a case, and element word is exactly the word for describing these elements, and keyword is just
Be it is each ask questions for each element word, showed in currently asking questions embodied case, meet element
The particular content of word.
The semantics recognition module of similar semantic net is provided between elements recognition module and problem acquisition memory module.One
As for, ordinary populace propose ask questions when, all very colloquial style, the unprofessional inaccuracy of many words pass through
Semantics recognition module can form the colloquial law vocabulary for being converted into and being redescribed by legal profession term that asks questions
Sentence.
It include the vocabulary memory module for storing each legal profession vocabulary, semantics recognition module in semantics recognition module
Problem is acquired in memory module and has been converted into asking questions and the legal profession word in vocabulary memory module for text information
Remittance compares, and will ask questions the keyword for being converted to and being indicated one by one with law specialized vocabulary.It is asked questions to improve
The precision of law conversion, is provided with synonym dictionary and homonym dictionary in vocabulary memory module.
By synonym dictionary, the word in asking questions can be carried out with the spoken vocabulary in synonym dictionary one by one
Comparison, finds same spoken vocabulary, and then find the corresponding law vocabulary of the spoken language vocabulary, is replaced with the law vocabulary
The word successively carries out, and colloquial ask questions is converted to law vocabulary.There is provided for the subsequent matching for seeking advice from answer can
Energy.The comparison of synonym dictionary is handled mainly for asking questions for textual.
By homonym dictionary, can by saved with audio-frequency information ask questions in word one by one with homonym dictionary
In spoken vocabulary compare, find the spoken vocabulary of same pronunciation, and then find the corresponding law of spoken language vocabulary
Vocabulary, replaces the word with the law vocabulary, successively carries out, and colloquial ask questions is converted to law vocabulary.
In synonym dictionary and homonym dictionary, a law vocabulary corresponds to multiple spoken vocabulary.Generally, a word
There are many kinds of the expression ways of language spoken language, and multiple spoken vocabulary are corresponding with a law vocabulary, can be different by form of presentation
Word be changed into unified, rigorous law expression.
Equally there is the attribute word label for indicating preferred selection order on each law vocabulary, makes semantics recognition module right
When than the word in asking questions with spoken vocabulary, the high law vocabulary pair of those attribute word priority can be preferentially pushed
The spoken vocabulary answered.The time of semantics recognition can not only be saved in this way, additionally it is possible to which help solves those one spoken vocabulary pair
The case where answering multiple law vocabulary.Because Chinese of extensive knowledge and profound scholarship, look like more, along with the differentiation of various spoken languages, make one
Word may contain there are many meaning or the word of pronunciation of the same race agree can there are many meanings, therefore in synonym dictionary and unisonance
In word dictionary, the case where corresponding to multiple law vocabulary there are a spoken vocabulary, at this time, the attribute word of law vocabulary is just played
Identify the effect of judgement.The higher law vocabulary of attribute word priority-labeled, it is easier be selected as replacement ask questions middle correspondence
The law of word is expressed.And when extracting keyword below, law vocabulary its attribute word in the same sentence it is preferential
The grade the high also easier to be extracted as keyword.The number that the priority of attribute word and each vocabulary are selected is positively correlated, some
Vocabulary is selected more, and attribute priority is higher.In the present embodiment, semantics recognition module can be directly with semantic net
Form exists.
Subscriber information storing module, for storing and updating user information；The user information mainly includes consultant's information
One of them is only recorded if consultant and party are a people with party's information.By collecting record user's letter
Breath, can record related personnel's information of a certain specific case as much as possible, so occurred by the case region, case kind
Class distinguishes different crowds, for relevant department's legislation, amends the law and the adjustment of relevant policies and decision provide data and support.Its
In, party's information includes at least essential informations, these essential informations and the methods such as age, gender, birthplace, residence, nationality
Other words in rule provision memory module in the label of each legal provision mutually echo, and enable using party's information in law
Provision selection is upper more accurate.
Case processing module, is stored with consulting model, the consulting model include multiple legal opinion templates and with
The corresponding judgment models of each legal opinion template；Each judgment models are all made of decision-tree model, and each judgment models are equal
Including multiple input factors and output factor.Input factor is made of multiple element words.Element word in input factor is to be directed to
For asking questions.Output factor is made of multiple pre-set introducers, and the introducer in output factor is for method
For rule provision.Different judgment models, input factor are different with the output set content of factor.
Firstly, consulting model by one by one by according to the element word of all judgment models from semantics recognition module law
Language is specialized to ask questions extraction keyword, in all element words of a judgment models can be asked questions from this
Keyword is extracted, then the judgment models is selected to carry out model calculating as currently used model.If there is multiple judgement moulds simultaneously
Type is all met the requirements, then is calculated separately simultaneously using multiple judgment models, obtain multiple consulting answers.Such case is fitted
The case where one case of contract is related to different laws.
Judgment models, according to the decision tree that the judgment models are formed, are selected defeated when calculating from all keywords
Enter the corresponding keyword of factor to be input in decision-tree model, according to each node of decision tree and corresponding secondary input because
Element, successively selection corresponds to the keyword of each node from remaining keyword, is finally exported according to the structure of decision tree
Each label of legal provision each in the introducer for the factor that exports, with legal provision memory module is compared, is looked for by factor
Out with the matched label of introducer, the corresponding legal provision of the label is extracted from legal provision memory module, is answered as consulting
Case output.In order to make consultant more intuitively obtain specialty legal advisory opinion, by the legal provision of extraction and collect
Input factor and all secondary input factors seek union together, all information are all filled up to the correspondence of legal advice consulting book
On position, complete legal advice consulting book is formed.
In order to guarantee the accurate of legal advice consulting book, when specifically forming legal advice consulting book, adopted from problem
Collection memory module and the same timing node of subscriber information storing module extraction, the same case of same consultant, and according to element
Extraction module and case processing module extract this from legal provision module and ask questions corresponding legal provision, integrate shape
At legal advice consultant service.
Problem, which acquires memory module, can pass through elements recognition module and relationship to obtain complete case information input
Memory module forms multiple problems relevant to consultant's presentation content, is formed with consultant and interact communication, until all and official communication
Inquiry person describes after associated all problems have all asked, by all description informations of consultant, formation is completely asked questions.
Specifically, when problem acquisition memory module receive the initial description content of consultant after, semantics recognition module from
According to all elements recognition keywords in elements recognition module, and by these keywords directly with it is each in legal provision module
The label of a legal provision compares, and corresponding legal provision is searched out in legal provision module, and store according to relationship
Other legal provisions of the legal provision degree of association from high to low are found in module, because being directed to each method in legal provision module
Rule provision is previously provided with corresponding booting problem, these booting problems are successively fed back to consultant, and guidance consultant carries out
It answers, problem acquisition memory module is enable to obtain more fully case information, until the description content of consultant will not be again
Cause new relative laws provision, and the related booting problem of institute is all inquired and finished, the memory module of problem acquisition at this time is received
All description contents collected are exactly asking questions for the case.
When constructing legal advice consulting Database, firstly, all legal provisions are intercepted by law expert, are torn open
Classification storage carries out remarks mark into law striped memory module, and for each legal provision sentence completely explained after point
Note forms the set of tags that is made of multiple labels, in these labels, including at least showing which law the legal provision belongs to
The classificating word of which chapters and sections, indicates each label priority of the legal provision at the summary word for indicating legal provision content summary
Grade word, indicate the conjunctive word of the legal provision Yu other legal provision degrees of association, and for indicating that the legal provision goes out
Often there are other words of other attributes such as region in occurrence number.
Then, all labels of all legal provisions are closed according to the subordinate relation of legal provision, chapters and sections relationship, association
System etc. carries out label contextual definition, and all labels is made to be capable of forming a network of personal connections from top to bottom.
Third, the initial value by problem acquisition memory module is zero, when user terminal starts to acquire asking questions for consultant
Afterwards, it records and stores and asks questions, extract the keyword in asking questions, progress will be asked questions keyword same number more
Classification storage and update.
4th, in elements recognition module, store original factor word and these corresponding priority of element word and association
Degree obtains the most element word of frequency of occurrence automatically from all ask questions and mentions after user terminal starts to acquire and ask questions
Its high priority, when making carry out keyword extraction next time, the advantage distillation corresponding keyword of element word.
5th, zero is set by the original storage of subscriber information storing module, after user terminal is started to work, according to element
The element word for belonging to user information in extraction module carries out user information collection and update.
6th, in case processing module, be stored in advance multiple judgment models and with the one-to-one law of judgment models
Position paper template, with same timing node, based on what same consultant proposed asks questions, according in elements recognition module
Element word extracts the keyword in asking questions, and according to the judgment models of selection, it is defeated will to meet its keyword for inputting factor
Enter into judgment models, by the judgement of decision tree, obtains output factor.According to the introducer in output factor, in law item
Label identical with these introducers is found in literary memory module, extracts the corresponding legal provision of these labels and consultant's consulting
Ask questions together, fill in as content into legal opinion template, formed legal opinion.Certainly, processing law meaning
See outside book, by being used in combination with different templates, this system can be made to generate more feedback appearance forms to user terminal.
For example, there is a consultant that can adopt to a description as described in divorce, consultant is taught before collection terminal
With oral account or perhaps pen is write to carry out event description, and while description, the description information that collection terminal will acquire is sent to
These description informations are divided into word one by one by problem acquisition module, semantics recognition module, and by these word synonyms
Spoken vocabulary of the dictionary in other words in homonym dictionary compares, and finds same or unisonance spoken vocabulary, and will
The corresponding law vocabulary of the spoken language vocabulary replaces corresponding word, if a spoken vocabulary corresponds to multiple law vocabulary, selects
Select that law vocabulary of attribute word highest priority.All description informations that consultant provides to collection terminal are all gradually stored
Into problem acquisition memory module, until acquisition terminal is terminated to the acquisition of the case description information, the same case
All case description informations of the same consultant in range at the same time constitute one and ask questions.It is same in asking questions
One time range and same consultant and same case, these information are all that collection terminal is direct using the prior art in acquisition
Obtain, for example, using direct access inquiry consultant user information or be directly that collection terminal generates when communicating
The existing means such as self-clocking, unique access username, IP address on physical layer are defined.
In elements recognition module, be previously stored with some classification words for defining requirement, be such as related to law name,
Bridegroom's or husband's side name, wife's side name, mistake description etc..When asking questions when continuous collecting is not over, elements recognition mould
Block will carry out keyword extraction to the legal issue after the translation of semantics recognition module, and legal issue at this time is because of consulting
Problem is not acquired and is finished, and legal issue is also finished without conversion, is all the state carried out simultaneously, is conducive to quickly carry out in this way
Information identification, the case description information for also facilitating control collection terminal to collect sufficiently complete completely ask questions to constitute.?
Consultant describes a case about divorce in this example, then according to element word law name to should belong to such word
It, quickly can will be in " marriage law " in " marriage " two words and set in case description information that consultant provides in set
Two word successful match of marriage, so that two words of marriage is come out as the corresponding keyword extraction of element word, with such
It pushes away, can also extract mistake and be described as overstepping the limit, during collection terminal acquisition, by having the keyword being extracted
Compared with the label in legal provision memory module, find corresponding legal provision, then found by relationship module and this
The asked questions of the corresponding storage of each legal provision being found are sent to by the associated other legal provision of a little legal provisions
Consultant answer, wherein asked questions refer to it is pre-stored be used to receiving asking questions and mutually coping with for consultant
After words, it is used to put question to consultant, so that consultant provides the default problem of more relevant informations.Make consultant in description case
When incomplete, consultant is inquired in a manner of problem by collection terminal and display end.It is closed from relationship module
When joining legal provision and problem selection, pushed from big to small according to the grade of the degree of association.Until the method duplicated
It restrains provision and when asked questions, illustrates that all relative laws provisions and correspondence problem have been asked, at this time collection terminal and aobvious
Show end stop with regard to this case and consultant continue interact communication, formed completely to seek advice from problem acquisition memory module and be asked
Topic, and then complete legal issue is obtained by the identification of semantics recognition module.
Keyword is extracted in the continuation of elements recognition module from legal issue, until the legal issue is according to case processing module
In until the corresponding element word extraction of all input factors finishes keyword.Because consulting model includes multiple judgment models, and
The element word that the input factor of each judgment models includes is different, if general one asks questions independent case, then its quilt
The keyword extracted can only fully meet the input factor an of judgment models, this judgment models being satisfied is i.e. at this time
Asking questions for this needs judgment models to be used.It such as, is divorce case in this example, then its crucial base for being extracted
The relevant word of corresponding to marriage law judgment models on this, for example, overstep the limit, be small three, family is sudden and violent, get married, divorce etc., and very
It will appear murder less, set fire, steal etc. in other types law and judging the word that will appear.
After selected judgment models, which is to pre-set the input of input factor and secondary input factor is defeated
Enter, judges the decision-tree model constructed according to decision Tree algorithms finally by continuous selection.For example, marriage law in this example
In judgment models, the element word of the decision tree of the first order is that both sides are the servicemen or both sides are not the servicemen, should
It is soldier that the corresponding descriptor of element word, which has, be all soldier, be not soldier etc., by all keys for comparing the legal issue
Word is selected, and the corresponding element word of the input factor is that soldier enters if having selected not to be soldier after comparison
The operation of next stage, the bridegroom's or husband's side whether there is bigamy, which is bigamy, twice wedding etc., then at case
Reason module continuation looks for whether the keyword same or similar with these descriptors in all keywords, and then under entrance
The calculating of level-one.In addition to the factor that inputs in the first order is we term it other than input factor, our other all habitual addresses
For secondary input factor.The judgment models of each method type of law are the bodies constructed in strict accordance with the judgement thinking of its relevant law
The decision-tree model of existing statutory rules.After last calculate, obtained result is pre-stored certain introducers.This example
In ask questions after calculating obtained introducer be " the 32nd article of marriage law ", case processing module is according to this guidance
Word is compared with each label in legal provision memory module, finds matching legal provision, by marriage law the 32nd
The legal provision content of item feeds back to consultant as a part of consulting answer.Because being directed to each judgement mould in consulting model
Type is stored with legal opinion template, and each composition legal opinion is provided on the automatic filling column in legal opinion template
Element word, according to the prior art, the corresponding key words content of element word that will be extracted in judgment models deterministic process automatically
And the result generated after each decision node in judgment models is filled on corresponding automatic filling column, forms one completely
Divorce legal opinion.
It is true that the generation of legal opinion and the reasonable judgement of case all rely on asking questions for consultant's offer
Reliably, it however, if being difficult to identify whether consultant's those set forth is true not by proof, and during deciding a case, requires
Juristic fact is proved with evidence, and in these evidences, invoice is the most frequently used evidence for being also easiest to obtain.On collection terminal
It is provided with camera and ultraviolet light, camera opens ultraviolet light when shooting invoice, by existing image comparison technology, will clap
The invoice photo to be formed is taken the photograph to compare with pre-stored fluorescence photo, for identifying that invoice photo whether there is fluorescence, if
There are fluorescence then to prove that the invoice is true.So the corresponding juristic fact of the invoice is also identified as really, before this
The legal opinion formed under mentioning has probative effect advantageously.
Meanwhile further proving that the juristic fact of consultant's offer is true and reliable, pass through the pickup being arranged on collection terminal
Device complete documentation consultant it is all ask questions and it is corresponding be described, enable consultant afterwards and can not deny from
The juristic fact that oneself once described forms the constraint of a kind of couple of consultant, and consultant is enable to demonstrate,prove it certainly by the behavior of itself
The content of description is true.The audio that collection terminal enrolls sound pick-up saves as former state, and collection terminal is by system in interacting with consultant
All information provided are converted into the synchronous preservation of text information.It in this way can be by the interacting Question-Answer between system and consultant
The complete dialogue of record, understands the true intention of consultant's expression by talking with when helping to inquire afterwards.Meanwhile by right
Lossless, the preservation as former state of consultant's audio, avoid the fraud operation for being possible to occur below, make the audio of preservation as former state more
With probative effect.
What has been described above is only an embodiment of the present invention, and the common sense such as well known specific structure and characteristic are not made herein in scheme
Excessive description, technical field that the present invention belongs to is all before one skilled in the art know the applying date or priority date
Ordinary technical knowledge can know the prior art all in the field, and have using routine experiment hand before the date
The ability of section, one skilled in the art can improve and be implemented in conjunction with self-ability under the enlightenment that the application provides
This programme, some typical known features or known method should not become one skilled in the art and implement the application
Obstacle.It should be pointed out that for those skilled in the art, without departing from the structure of the invention, can also make
Several modifications and improvements out, these also should be considered as protection scope of the present invention, these all will not influence the effect that the present invention is implemented
Fruit and patent practicability.The scope of protection required by this application should be based on the content of the claims, the tool in specification
The records such as body embodiment can be used for explaining the content of claim.
Claims (7)
1. a kind of legal services special purpose robot of semantics recognition, it is characterised in that: including robot body and be arranged in machine
Server in device human body；
The robot body, including being used to acquire the collection terminal asked questions；
The server, including being used to the semantics recognition module for asking questions progress law extraction of semantics, the semantics recognition
Module includes the vocabulary memory module that storage and real-time update have multiple law vocabulary；
Semantics recognition module from asking questions middle extraction word, and by the law vocabulary in each word and vocabulary memory module into
Row comparison replaces the word in asking questions with the law vocabulary of successful match.
2. the legal services special purpose robot of semantics recognition according to claim 1, it is characterised in that: the robot sheet
Body further includes for showing legal issue and seeking advice from the display end of answer.
3. the legal services special purpose robot of semantics recognition according to claim 1, it is characterised in that: the vocabulary storage
The synonym dictionary for being used to that law vocabulary is carried out to corresponding storage with spoken vocabulary is equipped in module.
4. the legal services special purpose robot of semantics recognition according to claim 1, it is characterised in that: the vocabulary storage
The homonym dictionary for being used to that law vocabulary is carried out to corresponding storage with spoken vocabulary is equipped in module.
5. the legal services special purpose robot of semantics recognition according to claim 1, it is characterised in that: the server is also
Legal provision memory module including being used to store legal provision, and be used to extract law item from legal provision memory module
Case processing module of the text as consulting answer.
6. the legal services special purpose robot of semantics recognition according to claim 3 or 4, it is characterised in that: a law
Vocabulary corresponds to multiple spoken vocabulary.
7. the legal services special purpose robot of semantics recognition according to claim 1, it is characterised in that: the robot sheet
Body further includes for detecting whether the close triggering end of someone, and the driving mechanism by triggering end control starting；Work as triggering end
When detecting someone, it is mobile towards the direction of people that triggering end controls driving mechanism.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810805075.5A CN109108989B (en) | 2018-07-20 | 2018-07-20 | Semantic recognition's legal service special purpose robot |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810805075.5A CN109108989B (en) | 2018-07-20 | 2018-07-20 | Semantic recognition's legal service special purpose robot |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109108989A true CN109108989A (en) | 2019-01-01 |
| CN109108989B CN109108989B (en) | 2021-02-02 |
Family
ID=64862360
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810805075.5A Active CN109108989B (en) | 2018-07-20 | 2018-07-20 | Semantic recognition's legal service special purpose robot |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109108989B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110053055A (en) * | 2019-03-04 | 2019-07-26 | 平安科技（深圳）有限公司 | A kind of robot and its method answered a question, storage medium |
| CN111300443A (en) * | 2020-02-29 | 2020-06-19 | 重庆百事得大牛机器人有限公司 | Emotional placating method based on legal consultation robot |
| CN111360843A (en) * | 2020-03-09 | 2020-07-03 | 重庆百事得大牛机器人有限公司 | Legal consultation method based on behavior prediction |
| CN112151019A (en) * | 2019-06-26 | 2020-12-29 | 阿里巴巴集团控股有限公司 | Text processing method and device and computing equipment |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106503239A (en) * | 2016-11-07 | 2017-03-15 | 上海智臻智能网络科技股份有限公司 | A kind of method and apparatus of legal information inquiry |
| CN107220912A (en) * | 2017-06-12 | 2017-09-29 | 上海市高级人民法院 | Litigation services intelligence system and robot |
-
2018
- 2018-07-20 CN CN201810805075.5A patent/CN109108989B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106503239A (en) * | 2016-11-07 | 2017-03-15 | 上海智臻智能网络科技股份有限公司 | A kind of method and apparatus of legal information inquiry |
| CN107220912A (en) * | 2017-06-12 | 2017-09-29 | 上海市高级人民法院 | Litigation services intelligence system and robot |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110053055A (en) * | 2019-03-04 | 2019-07-26 | 平安科技（深圳）有限公司 | A kind of robot and its method answered a question, storage medium |
| CN112151019A (en) * | 2019-06-26 | 2020-12-29 | 阿里巴巴集团控股有限公司 | Text processing method and device and computing equipment |
| CN111300443A (en) * | 2020-02-29 | 2020-06-19 | 重庆百事得大牛机器人有限公司 | Emotional placating method based on legal consultation robot |
| CN111300443B (en) * | 2020-02-29 | 2020-11-13 | 重庆百事得大牛机器人有限公司 | Emotional placating method based on legal consultation robot |
| CN111360843A (en) * | 2020-03-09 | 2020-07-03 | 重庆百事得大牛机器人有限公司 | Legal consultation method based on behavior prediction |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109108989B (en) | 2021-02-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109086371A (en) | A kind of semantic net interactive system and exchange method for legal advice | |
| US10977452B2 (en) | Multi-lingual virtual personal assistant | |
| US11740863B2 (en) | Search and knowledge base question answering for a voice user interface | |
| CN109002538A (en) | Legal advice cloud platform and method based on database | |
| JP7022062B2 (en) | VPA with integrated object recognition and facial expression recognition | |
| CN108920706A (en) | A kind of legal advice consulting Database and its construction method | |
| US11302330B2 (en) | Clarifying questions for rewriting ambiguous user utterance | |
| US11256868B2 (en) | Architecture for resolving ambiguous user utterance | |
| CN109002431A (en) | Legal opinion generates system and production method | |
| CN109416816B (en) | Artificial intelligence system supporting communication | |
| CN109108989A (en) | A kind of legal services special purpose robot of semantics recognition | |
| CN109033083A (en) | A kind of legal advice system based on semantic net | |
| CN110096577A (en) | From the intention of abnormal profile data prediction user | |
| CN106503239A (en) | A kind of method and apparatus of legal information inquiry | |
| Griol et al. | Combining speech-based and linguistic classifiers to recognize emotion in user spoken utterances | |
| WO2020233386A1 (en) | Intelligent question-answering method and device employing aiml, computer apparatus, and storage medium | |
| CN113468302A (en) | Combining parameters of multiple search queries sharing an interrogation line | |
| CN112133406A (en) | Multi-mode emotion guidance method and system based on emotion maps and storage medium | |
| CN109165281A (en) | A kind of shared legal advice system of multirobot terminal based on semantic net | |
| CN114360678A (en) | Information processing method, device, equipment and storage medium | |
| JP2007148118A (en) | Voice interactive system | |
| CN109063060A (en) | A kind of semantic net legal advice service robot | |
| CN107153685B (en) | Cognitive method and device based on timeline memory in man-machine conversation system | |
| CN115576975A (en) | Data query method, device, equipment and medium | |
| CN109145106A (en) | A kind of self-learning type legal advice robot |
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 | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20200108 Address after: 401147 Office 1, building 18, xiangnai mansion, Zhongyu, No. 101, Xinji Avenue, Longxi street, Yubei District, Chongqing Applicant after: Chongqing best Daniel robot Co., Ltd Address before: 401147 Nineteen Floors, 7 Chennai Mansion, No. 99 Xinji Avenue, Yubei District, Chongqing Applicant before: Wu Yi |
|
| TA01 | Transfer of patent application right | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP01 | Change in the name or title of a patent holder |
Address after: Office 1, 18, building 7, Zhongyu xiangnai mansion, No. 101, Xinji Avenue, Longxi street, Yubei District, Chongqing Patentee after: Chongqing Daniu Cognitive Technology Co.,Ltd. Address before: Office 1, 18, building 7, Zhongyu xiangnai mansion, No. 101, Xinji Avenue, Longxi street, Yubei District, Chongqing Patentee before: Chongqing best Daniel robot Co.,Ltd. |
|
| CP01 | Change in the name or title of a patent holder |