CN110287294A - Intellectual property concept answers method and system automatically - Google Patents
Intellectual property concept answers method and system automatically Download PDFInfo
- Publication number
- CN110287294A CN110287294A CN201811610973.1A CN201811610973A CN110287294A CN 110287294 A CN110287294 A CN 110287294A CN 201811610973 A CN201811610973 A CN 201811610973A CN 110287294 A CN110287294 A CN 110287294A
- Authority
- CN
- China
- Prior art keywords
- answer
- question
- semantic feature
- matching degree
- feature vector
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
Abstract
The invention patent discloses intellectual property concept and answers method and system automatically, and the term vector of information extraction serializing is inputted to user；Semantics recognition is carried out using term vector of the semantics recognition model based on machine learning to serializing, and obtains semantic feature vector；The semantic feature vector of problem in the semantic feature vector of input information obtained and default question and answer library is subjected to matching degree calculating by the sequences match algorithm based on machine learning；And corresponding Intellectual Property Rights Issues are exported according to matching degree and are answered.The present invention automatically and efficiently carries out the conceptual answer of intellectual property by the natural language processing to user's input text.
Description
Technical field
The present invention relates to natural language processing fields, and in particular to a kind of intellectual property based on machine learning algorithm is general
Read automatic answer method and system.
Background technique
In order to improve work effectiveness and intelligence, in more and more fields, begin to use artificial intelligence to the mankind from
Right language is understood and is fed back.In particular with the continuous innovation of machine learning algorithm, processing and reason for natural language
Xie Douyou significant progress the plurality of application scenes such as is read in machine translation, speech recognition, machine and is achieved preferably
Effect.
Intelligent chat robots are based on machine learning and construct to natural language processing algorithm.Wherein, in order to solve specific
The chat robots of business, referred to as Task talk with robot.Task humanoid robot usually provides information or clothes under given conditions
Business.Under normal conditions, be in order to meet and have the user that has a definite purpose, such as look into flow, look into telephone expenses, ticket booking, dining reservation, booking,
The Tasks scenes such as consulting.Task chat robots have been widely used for the interactive interface of a variety of smart machines, are applied to
In various interaction scenarios, solve the problems, such as that consulting and service chaining enter.
Intellectual property service field is substantially also identical as other service fields, and there is a large amount of basic interaction scenarios.
And in the routine work of intellectual property service organization, still having the action of a large portion is to client interpretation knowledge
Basis of property right concept helps the corresponding information in customer inquiries intellectual property data library, and to customer knowledge title services project
Selection is suggested.Basis is linked up in scene, is consumed the plenty of time of intellectual property service organization, is increased operation cost.This
Outside, it since intellectual property industry belongs to the cross discipline of science and technology and law, involves a wide range of knowledge, the scope of one's knowledge of attendant inevitably has
It is lacked, client's most fast and accurately answer can not be given in time.
Chat robots are applied to the conceptual question and answer of intellectual property, there is following difficult points: (1) user does not have usually
The problem of standby intellectual property rudimentary knowledge, proposition usual semantic ambiguity, can not accurately correspond to corresponding entitative concept；And know
Know the crossing domain that property right belongs to law and science and technology, most of problem relatively has depth, and the requirement to order of accuarcy is answered
Nor common every-day language robot is satisfiable；(2) it is related to field complexity, it is same to close in different application scenarios
Keyword may refer to and meaning all can different from, same model can all have on different application scenarios and interactive subject
It is distinguished；(3) in the more complicated content of interaction, more wheel dialogues and interaction is needed, just can solve query, existing dialogue
Robot is only capable of answering single-wheel problem, cannot answer more wheel combinatorial problems.
The application is intended to overcome the difficult point of intellectual property service field chat robots, proposes a kind of based on machine learning
The intellectual property concept of algorithm answers method and system automatically.
Summary of the invention
The invention patent be designed to provide it is a kind of based on machine learning algorithm realize one kind be based on machine learning calculation
The intellectual property concept of method answers method and system automatically, automatic to carry out by inputting the natural language processing of text to user
The conceptual answer of intellectual property.
To achieve the above object, the invention proposes a kind of intellectual property concepts to answer method automatically, and this method includes such as
Lower step:
Start one wheel question answer dialog, obtain the input information of user, to user input information extraction serializing word to
Amount；
Semantics recognition is carried out using term vector of the semantics recognition model based on machine learning to serializing, and obtains language
Adopted feature vector；
The semantic feature vector of the semantic feature vector of input information obtained and problem in default question and answer library is passed through
Sequences match algorithm based on machine learning carries out matching degree calculating；
As in question and answer library exist higher than matching degree judgment threshold semantic feature vector, then will wherein matching degree it is highest
Answer corresponding to semantic feature vector is as output information；
The semantic feature vector for being higher than matching degree threshold value as matching degree is not present in question and answer library, then according to the semanteme of missing
Eigenvector information returns to guidance information, and supplements input information after user supplements input information, and according to user, returns
Sequences match algorithm steps, execute again；
The semantic feature vector that user inputs information and supplement input information is saved, epicycle question answer dialog is terminated.
Wherein in a preferred embodiment, the question answer dialog has more wheels；The semantic feature vector of each round question answer dialog
Replacement is updated with the semantic feature vector that last round of question answer dialog is saved to obtain.
Wherein in a preferred embodiment, the semantics recognition model based on machine learning is LSTM model, specific to wrap
It includes:
Input layer obtains the term vector of serializing；
LSTM layers, Series Modeling is carried out to the term vector for the serializing that input layer obtains；
Output layer exports semantic feature vector.
Wherein in a preferred embodiment, described LSTM layers is equipped with attention model, which is used for the sequence
The term vector of columnization carries out semantic information weighted filter.
Wherein in a preferred embodiment, the sequences match algorithm is the Seq2Seq algorithm under CNN model, specific to wrap
It includes:
Input layer obtains semantic feature vector；
Convolutional layer carries out convolution to semantic feature vector using convolution kernel；
Pond layer is handled the output of convolutional layer using maximum value pond；
Output layer, it is corresponding to sample space by the output of SOFTMAX function pond layer.
Wherein in a preferred embodiment, it is one of text, image and voice or a variety of that the user, which inputs information,.
Wherein in a preferred embodiment, the question and answer library is constructed according to following steps:
The question and answer pair for obtaining intellectual property conceptual issues are arranged by manual simulation's dialogue and internet public data；
To question and answer to clustering, to cluster the label formed as the tag along sort of question and answer pair, to each contingency table
Setting standard is signed to answer；
Using the question and answer with tag along sort to the training data as semantics recognition model, the semanteme of training acquisition problem
The corresponding relationship of feature vector and standard question and answer.
Wherein in a preferred embodiment, the question and answer library is expanded according to following steps:
When existing question and answer library can not find matching degree and be higher than the semantic feature vector of matching degree threshold value, by manually taking
Business personnel carry out answer；
Save the answer record of manual service personnel and user；
Answer obtained record is sampled；
Be question and answer pair to the answer interpretation of records of sampling, by question and answer to original training data is supplemented to after, re-start language
The training of adopted identification model.
The invention also provides intellectual property concepts to answer system automatically, including input obtains module, term vector extracts mould
Block, semantics recognition module, question and answer library, sequence matching module and matching degree calculate module；
Input obtains module, inputs information for obtaining user；
Term vector extraction module carries out term vector extraction for inputting text to the user of acquisition, and retains serializing
Term vector；
Semantics recognition module, for use the word of serializing of the semantics recognition model based on machine learning to acquisition to
Amount extracts semantic feature vector；
Question and answer library, for storing question and answer pair, and question and answer to the problems in semanteme is extracted by the semantics recognition module
Feature vector, to carry out matching degree calculating for sequence matching module；
Sequence matching module, for using the sequences match algorithm based on machine learning to carry out matching degree calculating, the matching
The semantic feature vector that the user that degree calculates inputs problem in the semantic feature vector sum question and answer library that information generates carries out matching degree
It calculates；
Matching degree threshold control block, for sequence matching module to be calculated to the matching degree and preset matching degree that obtain
Judgment threshold compares, and such as there is the semantic feature vector for being higher than matching degree judgment threshold, then will the wherein highest semanteme of matching degree
Answer corresponding to feature vector is as output information；Semantic feature vector such as there is no matching degree higher than matching degree threshold value,
Guidance information is returned to, to obtain more users input information.
The invention also provides a kind of computer equipment, including memory and processor, the memory is stored with calculating
Machine program, which is characterized in that the processor executes the step of above-mentioned intellectual property concept answers method automatically.
The invention also provides a kind of computer readable storage mediums, store computer program thereon, and feature exists
In the step of computer program intellectual property concept above-mentioned when being executed by processor answers method automatically.
By adopting the above technical scheme, the invention patent has the advantages that
(1) semantics recognition is carried out by term vector of the semantics recognition model based on machine learning algorithm to serializing, i.e.,
Association between term vector is obtained by semantic training, and the relationship between different term vectors is not dependent on term vector itself
Semanteme, but based on context connection training obtains.This method can find potential under specific application scene between term vector
Connection.Compared to the algorithm for using Keywords matching, fuzzy matching can be effectively realized.So that user is without accurate description
Intellectual property vocabulary, intention can be understood and be identified by system.Intellectual property service question and answer scene is efficiently solved,
The fuzzy problem of semantic meaning representation.
(2) it is asked by the semantic feature vector sum that the sequences match algorithm based on machine learning inputs in information user
The semantic feature vector for answering problem in library is matched.Semantic feature vector is further matched, and passes through matching degree
Evaluation algorithms assess the matching degree of semantic feature vector.So that the user of different semantic formats is inputted information can pass through
Semantic feature vector is corresponding to pointed practical problem, and all problems, which are answered, to be obtained by matching guidance, has taken into account answer
Accurate and efficiency.
(3) the semantic feature vector talked with by saving each round, and the semantic feature vector as next round dialogue
Basis effectively saves the contextual information in dialogue；Make by take turns more dialogue to a concrete concept carry out put question to become can
Energy.More it is bonded the use demand of this usage scenario of intellectual property concept question and answer.
Invention is further described in detail with reference to the accompanying drawings and embodiments；But it is of the invention based on machine learning
The generation method of user's intent classifier of algorithm is not limited to the embodiment.
Detailed description of the invention
Fig. 1 is the flow diagram that intellectual property concept of the invention answers method automatically；
Fig. 2 is the flow diagram of sequences match algorithm of the invention；
Fig. 3 is the structural schematic diagram of LSTM basic unit；
Fig. 4 is the local structural graph of CNN-LSTM model；
Fig. 5 is the whole process structure chart of CNN-LSTM model.
Specific embodiment
The invention patent is described further in the following with reference to the drawings and specific embodiments.
Shown in Figure 1, the invention proposes a kind of intellectual property concepts to answer method automatically, and this method includes as follows
Step:
S100 starts a wheel question answer dialog, obtains the input information of user, and the word of information extraction serializing is inputted to user
Vector.
It can be one of text, picture and voice or a variety of that the S101 user, which inputs information, wherein text information is logical
It crosses word2vec algorithm and is converted to term vector；Pictorial information is turned the text information in picture by picture text conversion module
Turn to text information, and post-processing approach identical text this information；Voice messaging is turned voice messaging by speech text conversion module
Turn to text information, and post processing mode identical text this information；Picture text conversion module, speech text conversion module have
Disclosed software or open source algorithm are available, no longer do specifically describe explanation herein.
S200 carries out semantics recognition using term vector of the semantics recognition model based on machine learning to serializing, and obtains
Obtain semantic feature vector；
The semantics recognition model based on machine learning is LSTM model, is specifically included:
S210 input layer obtains the term vector of serializing；
S220LSTM layers, Series Modeling is carried out to the term vector for the serializing that input layer obtains
Wherein the structure of LSTM unit is shown in Figure 2, each LSTM unit has input gate, out gate and forgetting
Door, LSTM unit has the order of connection, and connects and form input gate and the corresponding sequence of out gate.Each LSTM unit has
There is forgetting door, it, therefore can be according to the sequence of reading to hereafter predicting since LSTM structure is formed based on sequence relation training
Judgement, and transmitted according to sequence direction, when information updates, outmoded information is forgotten by forgeing door, and incite somebody to action
Valuable information is transmitted to next LSTM unit by out gate.
By input gate, out gate and forget door, is expressed as it, ot, fiLSTM update method are as follows:
it=σ (Wixt+Uiht-1+bi)
ot=σ (Woxt+Uoht-1+bo)
ft=σ Wfxt+Ufht-1+bf)
The state at the LSTM unit current time is ct, ht。
Described LSTM layers is equipped with attention model, which is used to carry out language to the term vector of the serializing
Adopted information weighting filtering.
S230 output layer exports semantic feature vector.
S300 is by the semantic feature vector of the semantic feature vector of input information obtained and problem in default question and answer library
Matching degree calculating is carried out by the sequences match algorithm based on machine learning.
Referring to fig. 4 shown in -5, the sequences match algorithm is the Seq2Seq algorithm under CNN model, is specifically included:
S310 input layer obtains the semantic feature vector that LSTM model generates；
S320 convolutional layer carries out convolution to semantic feature vector using convolution kernel；
The pond S330 layer is handled the output of convolutional layer using maximum value pond；
S340 output layer, it is corresponding to sample space by the output of SOFTMAX function pond layer.
The sequences match algorithm is realized using rerank scoring model.Semantic feature vector is corresponded to sample space,
Sequences match algorithm carries out matching degree calculating by scoring model to the two semantic feature vectors compared.
There is the semantic feature vector higher than matching degree judgment threshold in S400 such as question and answer library, then it will wherein matching degree be most
Answer corresponding to high semantic feature vector is as output information；
The question and answer library is constructed according to following steps:
S410 obtains basic question and answer data by manual simulation's dialogue and internet public data；
S420 sorts out Questions & Answers data；
Mutually similar basic question and answer data are passed through the identical semantic feature vector of semantics recognition model extraction by S430, and
Using the semantic feature vector as the matching semantic feature vector of such problem.
There is no the semantic feature vectors that matching degree is higher than matching degree threshold value in S500 such as question and answer library, then according to missing
Semantic feature vector information returns to guidance information, and supplements input information after user supplements input information, and according to user,
Sequences match algorithm steps are returned, are executed again.
S510 is preset with the maximum number of times of supplemental information, is more than the maximum number of times as user supplements input information, still not
Matching answer is found, then a possibility that there are answers in judgement system is low.To prevent endless loop, system closure continues to look for mending
Information is filled, and selecting will access manual service interface or return prompt " problem is temporarily answered without matching ".
S600 saves the semantic feature vector that user inputs information and supplement input information, terminates epicycle question answer dialog.
The question answer dialog has more wheels；The semantic feature vector of each round question answer dialog is with last round of question answer dialog
The semantic feature vector saved updates replacement and obtains.
S700 is shown in Figure 2, and the question and answer library constructs in the following manner:
S710 arranges the question and answer for obtaining intellectual property conceptual issues by manual simulation's dialogue and internet public data
It is right；
S720 to question and answer to clustering, to cluster the label formed as the tag along sort of question and answer pair, to each point
Class label is arranged standard and answers；
S730, to the training data as semantics recognition model, trains acquisition problem using the question and answer with tag along sort
The corresponding relationship of semantic feature vector and standard question and answer.
Question and answer library described in S740 is expanded according to following steps:
S741 passes through people when existing question and answer library can not find matching degree and be higher than the semantic feature vector of matching degree threshold value
Work attendant carries out answer；
S742 saves the answer record of manual service personnel and user；
S743 samples answer obtained record；Stochastical sampling or root may be selected in the sample mode for answering record
It is sampled according to semantic feature vector；
S744 is question and answer pair to the answer interpretation of records of sampling, by question and answer to original training data is supplemented to after, again into
The training of row semantics recognition model.By the supplement to training data, semantics recognition model can further be updated, be made
It obtains self iteration of system to update, constantly improve in use.
The invention also provides intellectual property concepts to answer system automatically, including input obtains module, term vector extracts mould
Block, semantics recognition module, question and answer library, sequence matching module and matching degree calculate module；
Input obtains module, inputs information for obtaining user；
Term vector extraction module carries out term vector extraction for inputting text to the user of acquisition, and retains serializing
Term vector；
Semantics recognition module, for use the word of serializing of the semantics recognition model based on machine learning to acquisition to
Amount extracts semantic feature vector；
Question and answer library, for storing question and answer pair, and question and answer to the problems in semanteme is extracted by the semantics recognition module
Feature vector, to carry out matching degree calculating for sequence matching module；
Sequence matching module, for using the sequences match algorithm based on machine learning to carry out matching degree calculating, the matching
The semantic feature vector that the user that degree calculates inputs problem in the semantic feature vector sum question and answer library that information generates carries out matching degree
It calculates；
Matching degree threshold control block, for sequence matching module to be calculated to the matching degree and preset matching degree that obtain
Judgment threshold compares, and such as there is the semantic feature vector for being higher than matching degree judgment threshold, then will the wherein highest semanteme of matching degree
Answer corresponding to feature vector is as output information；Semantic feature vector such as there is no matching degree higher than matching degree threshold value,
Guidance information is returned to, to obtain more users input information.
The invention also provides a kind of computer equipment, including memory and processor, the memory is stored with calculating
Machine program, which is characterized in that the processor executes the step of above-mentioned intellectual property concept answers method automatically.
The invention also provides a kind of computer readable storage mediums, store computer program thereon, and feature exists
In the step of computer program intellectual property concept above-mentioned when being executed by processor answers method automatically.
Three application examples will be enumerated below explains the course of work of the invention.
One, text is putd question to
User inputs " how long is registered trademark needs? "
It inputs text to extract by term vector, generates term vector " registration ", " trade mark ", " needs ", " how many ", " time ".
The term vector of serializing carries out the extraction of semantic feature vector by semantics recognition module, and finds out most matched ask
It answers questions: trademark application periodic problem.
System using the answer of the problem as output answer return, and terminate the wheel dialogue, by the semantic feature of problem to
Amount saves and as next round dialog semantics feature base.
Two, picture is putd question to
User sends picture or pdf document (patent accepts notice).System identification picture obtains the text in picture
Information, including application number, patent name, notice issue the time.And with the word content of extraction constituted serializing word to
Amount.
By semantic feature extraction algorithm, the question and answer pair in question and answer library are matched.Judged by matching degree, discovery is not present
Meet the answer of matching degree threshold value.It returns to prompt information: " having been identified as patent flow file, if need to inquire patent most
New dynamic? " after obtaining the further feedback of user's input text, according to the semantic feature vector that user updates, best is found
With the problem of, and make answer.
Three, voice is putd question to
User input voice " priority be what the meaning? " system is converted into text for semantics recognition is inputted, and then carries out
Term vector extracts.Generate term vector " priority ", "Yes", " what ", " meaning ".
The term vector of serializing carries out the extraction of semantic feature vector by semantics recognition module, and finds out most matched ask
Answer questions: priority concept is answered a question.
System using the answer of the problem as output answer return, and terminate the wheel dialogue, by the semantic feature of problem to
Amount saves and as next round dialog semantics feature base.
The foregoing is merely present pre-ferred embodiments, therefore, it cannot be limited according to technical scope of the invention, therefore Fan Yiben
Equivalent changes and modifications made by the technical spirit and description of invention, in the range of should all belonging to technical solution of the present invention.
Claims (10)
1. intellectual property concept answers method automatically, which is characterized in that this method comprises the following steps:
Start a wheel question answer dialog, obtains the input information of user, the term vector of information extraction serializing is inputted to user；
Semantics recognition is carried out using term vector of the semantics recognition model based on machine learning to serializing, and obtains semantic feature
Vector；
The semantic feature vector of the semantic feature vector of input information obtained and problem in default question and answer library is passed through and is based on
The sequences match algorithm of machine learning carries out matching degree calculating；
As there is the semantic feature vector higher than matching degree judgment threshold in question and answer library, then will wherein matching degree it is highest semantic special
Answer corresponding to vector is levied as output information；
If in question and answer library be not present matching degree be higher than matching degree threshold value semantic feature vector, then according to the semantic feature of missing to
Information is measured, guidance information is returned to, and after user supplements input information, and input information is supplemented according to user, executes sequence again
Column matching algorithm, and recalculate matching degree；
The semantic feature vector that user inputs information and supplement input information is saved, epicycle question answer dialog is terminated.
2. intellectual property concept according to claim 1 answers method automatically, it is characterised in that: the question answer dialog has
More wheels；The semantic feature vector update that the semantic feature vector of each round question answer dialog is saved with last round of question answer dialog is replaced
Change acquisition.
3. intellectual property concept according to claim 1 answers method automatically, it is characterised in that: described to be based on machine learning
Semantics recognition model be LSTM model, specifically include:
Input layer obtains the term vector of serializing；
LSTM layers, Series Modeling is carried out to the term vector for the serializing that input layer obtains；
Output layer exports semantic feature vector.
4. intellectual property concept according to claim 3 answers method automatically, it is characterised in that: described LSTM layers is equipped with note
Meaning power model, the attention model are used to carry out semantic information weighted filter to the term vector of the serializing.
5. intellectual property concept according to claim 1 answers method automatically, which is characterized in that the sequences match algorithm
For the Seq2Seq algorithm under CNN model, specifically include:
Input layer obtains semantic feature vector；
Convolutional layer carries out convolution to semantic feature vector using convolution kernel；
Pond layer is handled the output of convolutional layer using maximum value pond；
Output layer, it is corresponding to sample space by the output of SOFTMAX function pond layer.
6. intellectual property concept according to claim 1 answers method automatically, it is characterised in that: the user inputs information
For one of text, image and voice or a variety of.
7. intellectual property concept according to claim 1 answers method automatically, it is characterised in that: the sequences match algorithm
It is realized using rerank scoring model.
8. intellectual property concept according to claim 1 answers method automatically, which is characterized in that the question and answer library is according to such as
Lower step building:
The question and answer pair for obtaining intellectual property conceptual issues are arranged by manual simulation's dialogue and internet public data；
To question and answer to clustering, to cluster the label formed as the tag along sort of question and answer pair, each tag along sort is set
Set standard answer；
Using the question and answer with tag along sort to the training data as semantics recognition model, the semantic feature of training acquisition problem to
The corresponding relationship of amount and standard question and answer.
9. intellectual property concept according to claim 8 answers method automatically, which is characterized in that the question and answer library is according to such as
Lower step expands:
When existing question and answer library can not find matching degree and be higher than the semantic feature vector of matching degree threshold value, pass through manual service personnel
Carry out answer；
Save the answer record of manual service personnel and user；
Answer obtained record is sampled；
Be question and answer pair to the answer interpretation of records of sampling, by question and answer to original training data is supplemented to after, re-start semantic knowledge
The training of other model.
10. intellectual property concept answers system automatically, it is characterised in that: obtain module, term vector extraction module, language including input
Adopted identification module, question and answer library, sequence matching module and matching degree calculate module；
Input obtains module, inputs information for obtaining user；
Term vector extraction module carries out term vector extraction for inputting text to the user of acquisition, and retain the word of serializing to
Amount；
Semantics recognition module, for using the term vector of serializing of the semantics recognition model based on machine learning to acquisition to extract
Semantic feature vector；
Question and answer library, for storing question and answer pair, and question and answer to the problems in semantic feature is extracted by the semantics recognition module
Vector, to carry out matching degree calculating for sequence matching module；
Sequence matching module, for using the sequences match algorithm based on machine learning to carry out matching degree calculating, the matching degree meter
The semantic feature vector that the user of calculation inputs problem in the semantic feature vector sum question and answer library that information generates carries out matching degree calculating；
Matching degree threshold control block judges threshold with preset matching degree for sequence matching module to be calculated the matching degree obtained
Value compares, such as exist be higher than matching degree judgment threshold semantic feature vector, then will wherein the highest semantic feature of matching degree to
The corresponding answer of amount is as output information；Such as there is no the semantic feature vector that matching degree is higher than matching degree threshold value, then return
Guidance information, to obtain more users input information.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811610973.1A CN110287294A (en) | 2018-12-27 | 2018-12-27 | Intellectual property concept answers method and system automatically |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811610973.1A CN110287294A (en) | 2018-12-27 | 2018-12-27 | Intellectual property concept answers method and system automatically |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110287294A true CN110287294A (en) | 2019-09-27 |
Family
ID=68000922
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811610973.1A Pending CN110287294A (en) | 2018-12-27 | 2018-12-27 | Intellectual property concept answers method and system automatically |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110287294A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111813909A (en) * | 2020-06-24 | 2020-10-23 | 泰康保险集团股份有限公司 | Intelligent question answering method and device |
| CN111966840A (en) * | 2020-08-18 | 2020-11-20 | 北京猿力未来科技有限公司 | Man-machine interaction management method and management system for language teaching |
| CN112163075A (en) * | 2020-09-27 | 2021-01-01 | 北京乐学帮网络技术有限公司 | Information recommendation method and device, computer equipment and storage medium |
| CN116932723A (en) * | 2023-07-28 | 2023-10-24 | 世优（北京）科技有限公司 | Man-machine interaction system and method based on natural language processing |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040030556A1 (en) * | 1999-11-12 | 2004-02-12 | Bennett Ian M. | Speech based learning/training system using semantic decoding |
| CN104461525A (en) * | 2014-11-27 | 2015-03-25 | 韩慧健 | Intelligent user-defined consulting platform generating system |
| CN107562792A (en) * | 2017-07-31 | 2018-01-09 | 同济大学 | A kind of question and answer matching process based on deep learning |
| CN107688667A (en) * | 2017-09-30 | 2018-02-13 | 平安科技（深圳）有限公司 | Intelligent robot client service method, electronic installation and computer-readable recording medium |
| CN108804654A (en) * | 2018-06-07 | 2018-11-13 | 重庆邮电大学 | A kind of collaborative virtual learning environment construction method based on intelligent answer |
| CN108829757A (en) * | 2018-05-28 | 2018-11-16 | 广州麦优网络科技有限公司 | A kind of intelligent Service method, server and the storage medium of chat robots |
| CN108829756A (en) * | 2018-05-25 | 2018-11-16 | 杭州知智能科技有限公司 | A method of more wheel video question and answer are solved using layering attention context network |
| CN108846077A (en) * | 2018-06-08 | 2018-11-20 | 泰康保险集团股份有限公司 | Semantic matching method, device, medium and the electronic equipment of question and answer text |
-
2018
- 2018-12-27 CN CN201811610973.1A patent/CN110287294A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040030556A1 (en) * | 1999-11-12 | 2004-02-12 | Bennett Ian M. | Speech based learning/training system using semantic decoding |
| CN104461525A (en) * | 2014-11-27 | 2015-03-25 | 韩慧健 | Intelligent user-defined consulting platform generating system |
| CN107562792A (en) * | 2017-07-31 | 2018-01-09 | 同济大学 | A kind of question and answer matching process based on deep learning |
| CN107688667A (en) * | 2017-09-30 | 2018-02-13 | 平安科技（深圳）有限公司 | Intelligent robot client service method, electronic installation and computer-readable recording medium |
| CN108829756A (en) * | 2018-05-25 | 2018-11-16 | 杭州知智能科技有限公司 | A method of more wheel video question and answer are solved using layering attention context network |
| CN108829757A (en) * | 2018-05-28 | 2018-11-16 | 广州麦优网络科技有限公司 | A kind of intelligent Service method, server and the storage medium of chat robots |
| CN108804654A (en) * | 2018-06-07 | 2018-11-13 | 重庆邮电大学 | A kind of collaborative virtual learning environment construction method based on intelligent answer |
| CN108846077A (en) * | 2018-06-08 | 2018-11-20 | 泰康保险集团股份有限公司 | Semantic matching method, device, medium and the electronic equipment of question and answer text |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111813909A (en) * | 2020-06-24 | 2020-10-23 | 泰康保险集团股份有限公司 | Intelligent question answering method and device |
| CN111966840A (en) * | 2020-08-18 | 2020-11-20 | 北京猿力未来科技有限公司 | Man-machine interaction management method and management system for language teaching |
| CN112163075A (en) * | 2020-09-27 | 2021-01-01 | 北京乐学帮网络技术有限公司 | Information recommendation method and device, computer equipment and storage medium |
| CN116932723A (en) * | 2023-07-28 | 2023-10-24 | 世优（北京）科技有限公司 | Man-machine interaction system and method based on natural language processing |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109493166B (en) | Construction method for task type dialogue system aiming at e-commerce shopping guide scene | |
| CN110287294A (en) | Intellectual property concept answers method and system automatically | |
| CN108932342A (en) | A kind of method of semantic matches, the learning method of model and server | |
| CN109753565A (en) | Intellectual Property intelligent service method and system | |
| CN109658271A (en) | A kind of intelligent customer service system and method based on the professional scene of insurance | |
| CN109783666A (en) | A kind of image scene map generation method based on iteration fining | |
| CN111274790B (en) | Chapter-level event embedding method and device based on syntactic dependency graph | |
| Brabra et al. | Dialogue management in conversational systems: a review of approaches, challenges, and opportunities | |
| CN113254610B (en) | Multi-round conversation generation method for patent consultation | |
| CN111325571B (en) | Automatic generation method, device and system for commodity comment labels for multitask learning | |
| CN114722839B (en) | Man-machine cooperative dialogue interaction system and method | |
| CN112989761B (en) | Text classification method and device | |
| CN112100375A (en) | Text information generation method and device, storage medium and equipment | |
| CN111651571A (en) | Man-machine cooperation based session realization method, device, equipment and storage medium | |
| CN110210036A (en) | A kind of intension recognizing method and device | |
| CN117056531A (en) | Domain knowledge driven large language model fine tuning method, system, equipment and storage medium | |
| CN113268561A (en) | Problem generation method based on multi-task joint training | |
| CN112528136A (en) | Viewpoint label generation method and device, electronic equipment and storage medium | |
| CN114020897A (en) | Conversation emotion recognition method and related device | |
| CN117056451A (en) | New energy automobile complaint text aspect-viewpoint pair extraction method based on context enhancement | |
| CN114491265B (en) | Construction method of operation service system of business space platform | |
| CN116341519A (en) | Event causal relation extraction method, device and storage medium based on background knowledge | |
| CN115688758A (en) | Statement intention identification method and device and storage medium | |
| CN112328774A (en) | Method for realizing task type man-machine conversation task based on multiple documents | |
| CN113946649A (en) | Providing method of mediation plan, training method, related device and storage medium |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190927 |