CN106844587B - It is a kind of for talking with the data processing method and device of interactive system - Google Patents
It is a kind of for talking with the data processing method and device of interactive system Download PDFInfo
- Publication number
- CN106844587B CN106844587B CN201710017117.4A CN201710017117A CN106844587B CN 106844587 B CN106844587 B CN 106844587B CN 201710017117 A CN201710017117 A CN 201710017117A CN 106844587 B CN106844587 B CN 106844587B
- Authority
- CN
- China
- Prior art keywords
- answer
- interaction data
- similarity
- answers set
- dialogue interaction
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/332—Query formulation
- G06F16/3329—Natural language query formulation or dialogue systems
Abstract
It is a kind of for talking with the data processing method and device of interactive system, wherein this method comprises: obtaining the dialogue interaction data of user's input and parsing, generate the corresponding candidate answers set of dialogue interaction data；The similarity of each problem in dialogue interaction data and default problem answers set is calculated, and most like problem is determined according to the similarity of each problem, the degree of correlation for presetting the corresponding answer of problem in problem answers set meets the high degree of correlation standard of setting；Obtain the answer of most like problem in default problem answers set, and the similarity of the answer and each answer in candidate answers set of most like problem is calculated, the highest answer of similarity is determined from candidate answers set according to the similarity of answer each in candidate answers set and is exported.The problem of this method is by borrowing high degree of correlation answer set ensure that in dialogue interaction, to the interactive information output highest reply of degree associated therewith, ensure that dialogue interactive quality, improves the user experience in dialogue interactive process.
Description
Technical field
The present invention relates to robotic technology fields, specifically, being related to a kind of for talking with the data processing of interactive system
Method and device.
Background technique
With the continuous development of science and technology, the introducing of information technology, computer technology and artificial intelligence technology, machine
Industrial circle is gradually walked out in the research of people, gradually extends to the neck such as medical treatment, health care, family, amusement and service industry
Domain.And requirement of the people for robot also conform to the principle of simplicity single duplicate mechanical action be promoted to have anthropomorphic question and answer, independence and with
The intelligent robot that other robot interacts, human-computer interaction also just become an important factor for determining intelligent robot development.
Robot generallys use dialogue interactive system to realize the human-computer dialogue between user, talks with the reality of interactive system
Existing framework can substantially be divided into retrieval type model and two kinds of production model.Wherein, retrieval type model is from already existing language
Lookup and problem answer the most matched in material, accuracy rate is higher but adaptability is poor.In contrast, production model is then
It is obtained by a large amount of corpus and machine learning algorithm training, with good system suitability, but can not be protected at present
Demonstrate,prove higher accuracy rate.Due to retrieval type model grammatical and relatively reliable controllable, still to adopt in current industry
Based on retrieval type model.
However, the quality of the quality of question and answer will directly affect it in corpus for retrieval type model
User experience.If corpus quality is bad, it is likely that will cause and interact appearance between dialogue interactive system and user
The problems such as talking incoherently or exporting the property the dealt with answer for not providing any information, this is all to the user experience of conversational system
There are detrimental effects.
Summary of the invention
To solve the above problems, the present invention provides a kind of for talking with the data processing method of interactive system comprising:
Talk with interaction data obtaining step, obtain the dialogue interaction data of user's input and parse, generates the dialogue and hand over
The corresponding candidate answers set of mutual data；
Most like problem determination procedure calculates each problem in the dialogue interaction data and default problem answers set
Similarity, and most like problem is determined according to the similarity of each problem, problem is right with it in the default problem answers set
The degree of correlation of answer is answered to meet the high degree of correlation standard of setting.
Answer generation step, obtains the answer of most like problem described in the default problem answers set, and calculates institute
The similarity for stating each answer in the answer and the candidate answers set of most like problem, according in the candidate answers set
The similarity of each answer is determined the highest answer of similarity and is exported from the candidate answers set.
According to one embodiment of present invention, in the most like problem determination procedure, the dialogue interaction number is calculated
According to include: the step of the similarity of each problem in default problem answers set
To the dialogue interaction data carry out word segmentation processing, obtain it is described dialogue interaction data in each participle word to
Amount；
According to the term vector and its weight of each participle in the dialogue interaction data, the dialogue interaction data is calculated
Sentence vector；
Obtain the sentence vector of each problem in the default problem answers set, calculate the sentence of the dialogue interaction data to
The similarity of amount and the sentence vector of each problem in default problem answers set, obtains the dialogue interaction data and default problem
The similarity of each problem in answer set.
According to one embodiment of present invention, each participle in the dialogue interaction data is calculated according to tf-idf algorithm
Weight.
According to one embodiment of present invention, in the answer generation step, if the answer of the most like problem
It is respectively less than default similarity threshold with the similarity of each answer in the candidate answers set, then to the candidate answers set
Carry out cluster iteration, determine in the candidate answers set with the maximally related answer of the dialogue interaction data and export.
The present invention also provides a kind of for talking with the data processing equipment of interactive system comprising:
Talk with interaction data and obtain module, be used to obtain the dialogue interaction data of user's input and parse, described in generation
Talk with the corresponding candidate answers set of interaction data；
Most like problem determination module, be used to calculate the dialogue interaction data with it is each in default problem answers set
The similarity of problem, and most like problem, problem in the default problem answers set are determined according to the similarity of each problem
The degree of correlation of corresponding answer meets the high degree of correlation standard of setting.
Answer generation module is used to obtain the answer of most like problem described in the default problem answers set, and
The similarity for calculating each answer in the answer and the candidate answers set of the most like problem, according to the candidate answers
The similarity of each answer is determined the highest answer of similarity and is exported from the candidate answers set in set.
According to one embodiment of present invention, the most like problem determination module is configured to calculate institute according to following steps
State the similarity of each problem in dialogue interaction data and default problem answers set:
To the dialogue interaction data carry out word segmentation processing, obtain it is described dialogue interaction data in each participle word to
Amount；
According to the term vector and its weight of each participle in the dialogue interaction data, the dialogue interaction data is calculated
Sentence vector；
Obtain the sentence vector of each problem in the default problem answers set, calculate the sentence of the dialogue interaction data to
The similarity of amount and the sentence vector of each problem in default problem answers set, obtains the dialogue interaction data and default problem
The similarity of each problem in answer set.
According to one embodiment of present invention, the most like problem determination module is configured to be calculated according to tf-idf algorithm
The weight of each participle in the dialogue interaction data.
According to one embodiment of present invention, if it is each in the answer of the most like problem and the candidate answers set
The similarity of a answer is respectively less than default similarity threshold, and the answer generation module is then configured to the candidate answers set
Carry out cluster iteration, determine in the candidate answers set with the maximally related answer of the dialogue interaction data and export.
It is provided by the present invention for talking with the data processing method of interactive system compared to existing method, pass through borrow
The problem of high degree of correlation answer set, ensure that in dialogue interaction, to the highest reply of interactive information output degree associated therewith, from
And ensure that dialogue interactive quality, improve the user experience in dialogue interactive process.
Other features and advantages of the present invention will be illustrated in the following description, also, partly becomes from specification
It obtains it is clear that understand through the implementation of the invention.The objectives and other advantages of the invention can be by specification, right
Specifically noted structure is achieved and obtained in claim and attached drawing.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is required attached drawing in technical description to do simple introduction:
Fig. 1 is that the implementation process of the data processing method according to an embodiment of the invention for talking with interactive system is shown
It is intended to；
Fig. 2 is that calculating dialogue interaction data according to an embodiment of the invention is asked with each in default problem answers set
The flow diagram of the similarity of topic；
Fig. 3 is in accordance with another embodiment of the present invention for talking with the implementation process of the data processing method of interactive system
Schematic diagram；
Fig. 4 is according to an embodiment of the invention for talking with the structural representation of the data processing equipment of interactive system
Figure.
Specific embodiment
Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings and examples, how to apply to the present invention whereby
Technological means solves technical problem, and the realization process for reaching technical effect can fully understand and implement.It needs to illustrate
As long as not constituting conflict, each feature in each embodiment and each embodiment in the present invention can be combined with each other,
It is within the scope of the present invention to be formed by technical solution.
Meanwhile in the following description, for illustrative purposes and numerous specific details are set forth, to provide to of the invention real
Apply the thorough understanding of example.It will be apparent, however, to one skilled in the art, that the present invention can not have to tool here
Body details or described ad hoc fashion are implemented.
In addition, step shown in the flowchart of the accompanying drawings can be in the department of computer science of such as a group of computer-executable instructions
It is executed in system, although also, logical order is shown in flow charts, and it in some cases, can be to be different from herein
Sequence execute shown or described step.
For the problems of in the prior art, the present invention provides a kind of new for talking with the data of interactive system
Processing method, this method ensure that in dialogue interaction, exported to interactive information by answer set the problem of the borrow high degree of correlation
The highest reply of degree associated therewith.In order to clearly show at the data provided by the present invention for talking with interactive system
The realization principle of reason method realizes process and advantage, comes below in conjunction with different embodiments to the data processing method
It is further described.
Embodiment one:
Fig. 1 is shown provided by the present embodiment for talking with the implementation process signal of the data processing method of interactive system
Figure.
As shown in Figure 1, data processing method provided by the present embodiment obtains user's input in step s101 first
Dialogue interaction data simultaneously parses the dialogue interaction data, to obtain candidate answers corresponding to the dialogue interaction data
Set.
It should be pointed out that in different embodiments of the invention, this method accessed user in step s101
The dialogue interaction data of input is either text data, is also possible to voice data, and the invention is not limited thereto.
After obtaining candidate answers set corresponding to above-mentioned dialogue interaction data, this method can calculate in step s 102
The similarity of each problem in above-mentioned dialogue interaction data and default problem answers set.Wherein, above-mentioned default problem answers collection
The degree of correlation of the corresponding answer of problem meets preset high degree of correlation standard in conjunction, i.e., each in default problem answers set to ask
The degree of correlation for inscribing corresponding answer is all larger than or is equal to default relevance threshold.
Specifically, method provided by the present embodiment preferably calculates dialogue interaction number using method as shown in Figure 2
According to the similarity with each problem in default problem answers set.As shown in Fig. 2, this method in step s 201 can be to above-mentioned right
It talks about interaction data and carries out word segmentation processing, and obtain the term vector of each participle in dialogue interaction data.
In different embodiments of the invention, according to actual needs, this method can be come in different ways to above-mentioned
Talk with interaction data and carries out word segmentation processing.
In obtaining dialogue interaction data after the term vector of each participle, this method can be in step S202 according to above-mentioned right
The term vector and its weight for talking about each participle in interaction data, calculate the sentence vector of the dialogue interaction data.
Specifically, in the present embodiment, this method calculates each in dialogue interaction data advantageously according to tf-idf algorithm
The weight of word.
In the given file of portion, word frequency (term frequency, abbreviation tf) refers to some given word
The number occurred in this document.This number would generally be normalized (molecule is generally less than denominator, is different from idf), to prevent
Only it is biased to long file.
And reverse document-frequency (inverse document frequency, abbreviation idf) is that a word is generally important
The measurement of property.The idf of a certain particular words, can be by general act number divided by the number of the file comprising the word, then incites somebody to action
To quotient take logarithm to obtain.
In the present embodiment, after obtaining the weight of each participle by tf-idf algorithm, this method can also be to each participle
Weight be normalized, thus obtain it is each participle dialogue interaction data in weight.By by the weight and step
The term vector of obtained each participle is multiplied in rapid S201, also can be obtained by a completely new vector, the vector in this way
Exactly talk with the sentence vector of interaction data.
As shown in Fig. 2, this method can obtain default answer in step S203 after the sentence vector for obtaining dialogue interaction data
The sentence vector of each problem in case set, and the sentence vector and default problem answers for talking with interaction data are calculated in step S204
The similarity of the sentence vector of each problem in set is asked to obtain dialogue interaction data with each in default problem answers set
The similarity of topic.
In the present embodiment, each problem and the sentence vector of answer are precomputed and are gone forward side by side in default problem answers set
Corresponding mark is gone.For example, based on the method that above content is illustrated, can be obtained for problem " I increasingly likes you "
Sentence vector to the problem is (0.2,0.7,0.1).And for the answer of the problem " I also likes you ", it is based on above content institute
The method of elaboration, the sentence vector of the available answer are (0.4,0.4,0.2).For default problem answers set,
Each problem and the sentence vector of answer can be known by reading.
It should be pointed out that in different embodiments of the invention, this method can be according to actual needs using different
The similarity of the sentence vector of each problem in sentence vector and default problem answers set of the mode to calculate dialogue interaction data, this
It invents without being limited thereto.Such as in one embodiment of the invention, this method can use such as vector space cosine similarity
Algorithm, Pearson correlation coefficients algorithm, Jaccard similarity factor algorithm, adjustment cosine similarity algorithm or Euclidean distance algorithm
The similarity of the sentence vector of each problem in sentence vector and default problem answers set to calculate dialogue interaction data.
Again as shown in Figure 1, obtaining the similarity of dialogue interaction data with each problem in default problem answers set
Afterwards, this method can in step s 103 according to dialogue interaction data and default problem answers set in each problem similarity from
Most like problem is determined in default problem answers set.That is, extracted from all problems of default problem answers collection with
The problem of talking with interaction data similarity highest, which is most like problem.
After the most like problem for obtaining corresponding to dialogue interaction data, this method can be in step S104 from default problem
The answer of the most like problem is obtained in answer set, and calculates the answer and candidate answers of most like problem in step s105
The similarity of each answer in set.In the phase for the answer and each answer in candidate answers set for obtaining above-mentioned most like problem
After degree, this method also can in step s 106 according to the answer of above-mentioned most like problem with it is each in candidate answers set
The similarity of answer, determination provide the highest answer of similarity and export as final result to user.
For example, for the dialogue interaction data " I like well your (0.21,0.67,0.12) " of user's input, in corpus
In there may be three candidate answers (such as: " I what if also liked (0.38,0.43,0.19) ", " you say what (0.1,
0.3,0.6) ", " today, weather was fine (0.5,0.2,0.3) ").For above-mentioned dialogue interaction data, the present embodiment is provided
Method can calculate the similarity of each problem in the dialogue interaction data and default problem answers set first, and determine most
Similar Problems.For example, the most like problem determined through the above way is " I increasingly likes you ".Therefore also
Think that " I likes you well " and expressing the meaning for " I increasingly liking you " are identical or at least semantic similar.
The reply (i.e. answer) for talking with interaction data and above-mentioned most like problem can be different, but if there is identical
Or if similar answer, then can also be considered data to be output.Therefore, this method accounting is counted in stating most like ask
The similarity of each answer in the answer and candidate answers set of topic, and the highest candidate answers of similarity are selected as most
Whole answer is exported to user.For example, exporting " I what if also liked " as final result to user.
It is provided by the present invention for talking with interactive system from foregoing description as can be seen that compared to existing method
Data processing method improves interactive information and its reply in dialogue interaction by answer set the problem of the borrow high degree of correlation
Similarity improve the user experience in dialogue interactive process to ensure that dialogue interactive quality.
Embodiment two:
Fig. 3 is outputed provided by the present embodiment for talking with the implementation process signal of the data processing method of interactive system
Figure.
As shown in figure 3, data processing method provided by the present embodiment obtains user's input first in step S301
Dialogue interaction data simultaneously parses the dialogue interaction data, to obtain candidate answers corresponding to the dialogue interaction data
Set.
In step s 302, this method accounting is counted in stating each problem in dialogue interaction data and default problem answers set
Similarity, and most like problem (i.e. problem answers set is determined according to the similarity of each problem in step S303
In the problem most like with above-mentioned dialogue interaction data).
After determining most like problem, this method can be obtained from above-mentioned default problem answers set in step s 304
The answer of most like problem, and in step S305 calculate step S304 in obtained most like problem answer and step
Talk with the similarity of each candidate answers in the candidate answers set of interaction data generated in S301.
It should be pointed out that in the present embodiment, above-mentioned steps S301 to the specific implementation principle of step S305 and realization
Process is similar to the realization principle of step S101 in above-described embodiment one to step S105 and realization process, therefore no longer right herein
The particular content of step S301 to step S305 are repeated.
As shown in figure 3, in the present embodiment, when each answer in the answer and candidate answers set for obtain most like problem
After similarity, this method can judge each answer in the answer and candidate answers set of above-mentioned most like problem in step S306
Similarity whether be respectively less than default similarity threshold.If it is lower, so party's rule can answer candidate in step S307
Case set carries out cluster iteration, so that it is determined that out in candidate answers set with the maximally related answer of dialogue interaction data and export.
Because being clustered by all answer datas to same problem, if the data that a certain answer categorical clusters go out are more, that
The dialogue interaction data that this group of answer should have a possibility that bigger to be inputted to user forms one group of relevant dialogue.
In this way by mostly wheel iteration also just to obtain ideal answer.
And if the similarity of the answer of above-mentioned most like problem and each answer in candidate answers set is not small
In default similarity threshold, then this method also can in step S308 according in candidate answers set by similarity highest
Answer export as final answer to user.
The present invention also provides a kind of for talking with the data processing equipment of interactive system, and Fig. 4 is shown in the present embodiment
The structural schematic diagram of the data processing equipment.
As shown in figure 4, being preferably included provided by the present embodiment for talking with the data processing equipment of interactive system: right
It talks about interaction data and obtains module 401, most like problem determination module 402 and answer generation module 403.Wherein, dialogue interaction
Data acquisition module 401 is used to obtain the dialogue interaction data of user's input, and parses to the dialogue interaction data, thus
Determine candidate answers set corresponding to the dialogue interaction data.
Most like problem determination module 402 for calculate above-mentioned dialogue interaction data with it is each in default problem answers set
The similarity of problem, and determined most according to the similarity of each problem in above-mentioned dialogue interaction data and default problem set
Similar Problems.Wherein, the degree of correlation of each corresponding answer of problem meets preset height in above-mentioned default problem answers set
Degree of correlation standard.
After obtaining most like problem, most like problem determination module 402 can be raw to answer by the most like Question Transmission
At module 403.In the present embodiment, answer generation module 403, can be from default problem answers after receiving above-mentioned most like problem
The answer of the most like problem is obtained in set, and calculate separately the answer of the most like problem with it is each in candidate answers set
The similarity of answer (i.e. candidate answers), and according to the similarity of answer each in candidate answers set from candidate answers set
It determines the highest answer of similarity and exports.
It should be pointed out that in different embodiments of the invention, dialogue interaction data obtains module 401, most like asks
Inscribe determining module 402 and answer generation module 403 realize its respectively the concrete principle of function and process both can with it is above-mentioned
The content that step S101 is illustrated to step S106 in embodiment one is identical, can also be with step S301 in above-described embodiment two extremely
The content that step S3008 is illustrated is identical, therefore it is determining no longer to obtain module 401, most like problem to dialogue interaction data herein
Module 402 and the related content of answer generation module 403 are repeated.
It should be understood that disclosed embodiment of this invention is not limited to specific structure disclosed herein or processing step
Suddenly, the equivalent substitute for these features that those of ordinary skill in the related art are understood should be extended to.It should also be understood that
It is that term as used herein is used only for the purpose of describing specific embodiments, and is not intended to limit.
" one embodiment " or " embodiment " mentioned in specification means the special characteristic described in conjunction with the embodiments, structure
Or characteristic is included at least one embodiment of the present invention.Therefore, the phrase " reality that specification various places throughout occurs
Apply example " or " embodiment " the same embodiment might not be referred both to.
Although above-mentioned example is used to illustrate principle of the present invention in one or more application, for the technology of this field
For personnel, without departing from the principles and ideas of the present invention, hence it is evident that can in form, the details of usage and implementation
It is upper that various modifications may be made and does not have to make the creative labor.Therefore, the present invention is defined by the appended claims.
Claims (8)
1. a kind of for talking with the data processing method of interactive system characterized by comprising
Talk with interaction data obtaining step, obtain the dialogue interaction data of user's input and parse, generates the dialogue interaction number
According to corresponding candidate answers set；
Most like problem determination procedure, calculate it is described dialogue interaction data in default problem answers set each problem it is similar
Degree, and most like problem is determined according to the similarity of each problem, problem is corresponding in the default problem answers set answers
The degree of correlation of case meets the high degree of correlation standard of setting；
Answer generation step obtains the answer of most like problem described in the default problem answers set, and calculating is described most
The similarity of each answer in the answer of Similar Problems and the candidate answers set, according to each in the candidate answers set
The similarity of answer is determined the highest answer of similarity and is exported from the candidate answers set.
2. the method as described in claim 1, which is characterized in that in the most like problem determination procedure, it is described right to calculate
It talks about interaction data and includes: the step of the similarity of each problem in default problem answers set
Word segmentation processing is carried out to the dialogue interaction data, obtains the term vector of each participle in the dialogue interaction data；
According to it is described dialogue interaction data in each participle term vector and its weight, calculate it is described dialogue interaction data sentence to
Amount；
Obtain the sentence vector of each problem in the default problem answers set, calculate the sentence vector of the dialogue interaction data with
The similarity of the sentence vector of each problem, obtains the dialogue interaction data and default problem answers in default problem answers set
The similarity of each problem in set.
3. method according to claim 2, which is characterized in that calculated according to tf-idf algorithm each in the dialogue interaction data
The weight of a participle.
4. method according to any one of claims 1 to 3, which is characterized in that in the answer generation step, if institute
The similarity for stating each answer in the answer and the candidate answers set of most like problem is respectively less than default similarity threshold, then
Cluster iteration is carried out to the candidate answers set, is determined most related to the dialogue interaction data in the candidate answers set
Answer and output.
5. a kind of for talking with the data processing equipment of interactive system characterized by comprising
Talk with interaction data and obtain module, be used to obtain the dialogue interaction data of user's input and parse, generates the dialogue
The corresponding candidate answers set of interaction data；
Most like problem determination module is used to calculate each problem in the dialogue interaction data and default problem answers set
Similarity, and most like problem is determined according to the similarity of each problem, in the default problem answers set problem and its
The degree of correlation of corresponding answer meets the high degree of correlation standard of setting；
Answer generation module, is used to obtain the answer of most like problem described in the default problem answers set, and calculates
The similarity of each answer in the answer of the most like problem and the candidate answers set, according to the candidate answers set
In the similarity of each answer determine the highest answer of similarity from the candidate answers set and export.
6. device as claimed in claim 5, which is characterized in that the most like problem determination module is configured to according to following step
Suddenly the similarity of each problem in the dialogue interaction data and default problem answers set is calculated:
Word segmentation processing is carried out to the dialogue interaction data, obtains the term vector of each participle in the dialogue interaction data；
According to it is described dialogue interaction data in each participle term vector and its weight, calculate it is described dialogue interaction data sentence to
Amount；
Obtain the sentence vector of each problem in the default problem answers set, calculate the sentence vector of the dialogue interaction data with
The similarity of the sentence vector of each problem, obtains the dialogue interaction data and default problem answers in default problem answers set
The similarity of each problem in set.
7. device as claimed in claim 6, which is characterized in that the most like problem determination module is configured to according to tf-idf
Algorithm calculates the weight of each participle in the dialogue interaction data.
8. the device as described in any one of claim 5~7, which is characterized in that if the answer of the most like problem with
The similarity of each answer is respectively less than default similarity threshold in the candidate answers set, and the answer generation module then configures
To carry out cluster iteration to the candidate answers set, determine in the candidate answers set with the dialogue interaction data most phase
The answer and output of pass.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710017117.4A CN106844587B (en) | 2017-01-11 | 2017-01-11 | It is a kind of for talking with the data processing method and device of interactive system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710017117.4A CN106844587B (en) | 2017-01-11 | 2017-01-11 | It is a kind of for talking with the data processing method and device of interactive system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106844587A CN106844587A (en) | 2017-06-13 |
| CN106844587B true CN106844587B (en) | 2019-11-08 |
Family
ID=59117186
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710017117.4A Active CN106844587B (en) | 2017-01-11 | 2017-01-11 | It is a kind of for talking with the data processing method and device of interactive system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106844587B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108959387B (en) * | 2018-05-31 | 2020-09-11 | 科大讯飞股份有限公司 | Information acquisition method and device |
| CN110941706A (en) * | 2018-09-21 | 2020-03-31 | 华为技术有限公司 | Answer determining method and system |
| CN109657038B (en) * | 2018-10-10 | 2023-04-18 | 创新先进技术有限公司 | Question and answer pair data mining method and device and electronic equipment |
| CN109658114A (en) * | 2018-12-21 | 2019-04-19 | 万达信息股份有限公司 | The high efficiency smart client service method of large corpora |
| CN110263141A (en) * | 2019-06-25 | 2019-09-20 | 杭州微洱网络科技有限公司 | A kind of customer service question answering system based on BERT |
| CN110750629A (en) * | 2019-09-18 | 2020-02-04 | 平安科技（深圳）有限公司 | Robot dialogue generation method and device, readable storage medium and robot |
| CN112182159B (en) * | 2020-09-30 | 2023-07-07 | 中国人民大学 | Personalized search type dialogue method and system based on semantic representation |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101086843A (en) * | 2006-06-07 | 2007-12-12 | 中国科学院自动化研究所 | A sentence similarity recognition method for voice answer system |
| CN104933152A (en) * | 2015-06-24 | 2015-09-23 | 北京京东尚科信息技术有限公司 | Named entity recognition method and device |
| CN105989040A (en) * | 2015-02-03 | 2016-10-05 | 阿里巴巴集团控股有限公司 | Intelligent question-answer method, device and system |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7191175B2 (en) * | 2004-02-13 | 2007-03-13 | Attenex Corporation | System and method for arranging concept clusters in thematic neighborhood relationships in a two-dimensional visual display space |
| US10318987B2 (en) * | 2014-02-18 | 2019-06-11 | International Business Machines Corporation | Managing cookie data |
-
2017
- 2017-01-11 CN CN201710017117.4A patent/CN106844587B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101086843A (en) * | 2006-06-07 | 2007-12-12 | 中国科学院自动化研究所 | A sentence similarity recognition method for voice answer system |
| CN105989040A (en) * | 2015-02-03 | 2016-10-05 | 阿里巴巴集团控股有限公司 | Intelligent question-answer method, device and system |
| CN104933152A (en) * | 2015-06-24 | 2015-09-23 | 北京京东尚科信息技术有限公司 | Named entity recognition method and device |
Non-Patent Citations (1)
| Title |
|---|
| 协同过滤推荐算法的改进——帕累托最优视角;焦媛媛 等;《工业工程与管理》;20150123;第20卷(第06期);34-41 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106844587A (en) | 2017-06-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106844587B (en) | It is a kind of for talking with the data processing method and device of interactive system | |
| CN107862000B (en) | A kind of agricultural technology consulting interactive method | |
| CN110347792B (en) | Dialog generation method and device, storage medium and electronic equipment | |
| CN106294854B (en) | Man-machine interaction method and device for intelligent robot | |
| CN109063174B (en) | Query answer generation method and device, computer storage medium and electronic equipment | |
| CN110288980A (en) | Audio recognition method, the training method of model, device, equipment and storage medium | |
| CN111090736B (en) | Question-answering model training method, question-answering method, device and computer storage medium | |
| CN107870994A (en) | Man-machine interaction method and system for intelligent robot | |
| CN112100383B (en) | Meta-knowledge fine tuning method and platform for multitask language model | |
| CN109102809A (en) | A kind of dialogue method and system for intelligent robot | |
| CN108062388A (en) | Interactive reply generation method and device | |
| CN106776713A (en) | It is a kind of based on this clustering method of the Massive short documents of term vector semantic analysis | |
| CN106959839A (en) | A kind of human-computer interaction device and method | |
| CN109271493A (en) | A kind of language text processing method, device and storage medium | |
| JP2023535709A (en) | Language expression model system, pre-training method, device, device and medium | |
| CN108228576B (en) | Text translation method and device | |
| CN108959531A (en) | Information search method, device, equipment and storage medium | |
| CN109635080A (en) | Acknowledgment strategy generation method and device | |
| CN108846138A (en) | A kind of the problem of fusion answer information disaggregated model construction method, device and medium | |
| CN112214591A (en) | Conversation prediction method and device | |
| CN112417127A (en) | Method, device, equipment and medium for training conversation model and generating conversation | |
| CN110851650B (en) | Comment output method and device and computer storage medium | |
| CN111274412A (en) | Information extraction method, information extraction model training device and storage medium | |
| CN113505198A (en) | Keyword-driven generating type dialogue reply method and device and electronic equipment | |
| Huang et al. | Applied to mobile multimedia intelligent speech system interactive topic guiding model |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |