CN110188183A - A kind of update method, device, equipment and the storage medium of intelligent answer knowledge base - Google Patents
A kind of update method, device, equipment and the storage medium of intelligent answer knowledge base Download PDFInfo
- Publication number
- CN110188183A CN110188183A CN201910482105.8A CN201910482105A CN110188183A CN 110188183 A CN110188183 A CN 110188183A CN 201910482105 A CN201910482105 A CN 201910482105A CN 110188183 A CN110188183 A CN 110188183A
- Authority
- CN
- China
- Prior art keywords
- answer
- user
- knowledge base
- feedback information
- access request
- 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
- 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
- G06F16/3325—Reformulation based on results of preceding query
- G06F16/3326—Reformulation based on results of preceding query using relevance feedback from the user, e.g. relevance feedback on documents, documents sets, document terms or passages
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Theoretical Computer Science (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Human Computer Interaction (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses update method, device, equipment and the storage mediums of a kind of intelligent answer knowledge base, this method comprises: obtaining original intelligent answer knowledge base to be updated；Active user's access request is received, is searched from original intelligent answer knowledge base according to active user's access request and the matched answer of active user's access request；User is received to the feedback information of answer, adjust answer confidence level in real time according to feedback information, to update original intelligent answer knowledge base, it realizes after user every time feeds back answer, answer confidence level is directly adjusted according to feedback information, ensure that the real-time update of data information in intelligent answer knowledge base.
Description
Technical field
The present embodiments relate to intelligent answer technology more particularly to a kind of update methods of intelligent answer knowledge base, dress
It sets, equipment and storage medium.
Background technique
For newly established enterprise, there is no any experience on foundation road, needs hatching mechanism-crowd's wound of pioneering enterprise
Space provides accurate and quick response counseling services, it is contemplated that time and cost of labor, crowd's wound spaces are difficult to accomplish
The one-to-one real-time instruction of tutor causes pioneering enterprise to be difficult to handle relevant issues in time, influences the development of enterprise.
It is difficult to handle the relevant issues of pioneering enterprise in time for many wound spaces in the prior art, is currently suggested a kind of intelligence
Energy question answering system, realizes the one-to-many online guidance to newly established enterprise.But existing intelligent Answer System, in artificial intelligence system
During system engages in the dialogue with user, the confidence level of answer in knowledge base can not be adjusted in real time, but runs and ties in program
It is updated after beam, real-time is not strong enough, thus greatly reduces the usage experience of user.
Summary of the invention
In view of this, the present invention provides update method, device, equipment and the storage medium of a kind of intelligent answer knowledge base,
Adjustment answer confidence level in real time, to improve the usage experience of user.
In a first aspect, the embodiment of the invention provides a kind of update methods of intelligent answer knowledge base, comprising:
Obtain original intelligent answer knowledge base to be updated；
Active user's access request is received, according to active user's access request from the original intelligent answer knowledge
It is searched and the matched answer of active user's access request in library；
User is received to the feedback information of the answer, answer confidence level is adjusted according to the feedback information in real time, with more
The new original intelligent answer knowledge base.
Second aspect, the embodiment of the invention also provides a kind of updating devices of intelligent answer knowledge base, comprising:
First obtains module, for obtaining original intelligent answer knowledge base to be updated；
Searching module, for receiving active user's access request, according to active user's access request from the original
It is searched and the matched answer of active user's access request in beginning intelligent answer knowledge base；
The first adjustment update module, for receiving user to the feedback information of the answer, according to the feedback information reality
When adjust answer confidence level, to update the original intelligent answer knowledge base.
The third aspect, the embodiment of the invention also provides a kind of equipment, which includes:
One or more processors；
Storage device, for storing one or more programs；
When one or more of programs are executed by one or more of processors, so that one or more of processing
Device realizes the update method of the intelligent answer knowledge base as described in any of the above-described.
A kind of fourth aspect, computer readable storage medium, is stored thereon with computer program, which is held by processor
The update method of the intelligent answer knowledge base as described in any of the above-described is realized when row.
The present invention is receiving active user's access request, root by obtaining original intelligent answer knowledge base to be updated
It is searched from original intelligent answer knowledge base according to active user's access request and the matched answer of active user's access request；Then
User is received to the feedback information of answer, answer confidence level is adjusted in real time according to feedback information, is known with updating original intelligent answer
Know library, realizes after user every time feeds back answer, answer confidence level is directly adjusted according to feedback information, ensure that
The real-time update of data information in intelligent answer knowledge base.
Detailed description of the invention
Fig. 1 is a kind of flow chart of the update method of intelligent answer knowledge base provided in an embodiment of the present invention；
Fig. 2 is the process of the update method of another intelligent answer knowledge base provided in an embodiment of the present invention；
Fig. 3 is the flow chart of the update method of another intelligent answer knowledge base provided in an embodiment of the present invention；
Fig. 4 is the flow chart of the update method of another intelligent answer knowledge base provided in an embodiment of the present invention；
Fig. 5 is the flow chart of the update method of another intelligent answer knowledge base provided in an embodiment of the present invention；
Fig. 6 is a kind of structural block diagram of the updating device of intelligent answer knowledge base provided in an embodiment of the present invention；
Fig. 7 is a kind of hardware structural diagram of equipment provided in an embodiment of the present invention.
Specific embodiment
The present invention is described in further detail with reference to the accompanying drawings and examples.It is understood that this place is retouched
The specific embodiment stated is used only for explaining the present invention rather than limiting the invention.It also should be noted that in order to just
Only the parts related to the present invention are shown in description, attached drawing rather than entire infrastructure.
Fig. 1 is a kind of flow chart of the update method of intelligent answer knowledge base provided in an embodiment of the present invention, the present embodiment
It is applicable to the case where real-time update is carried out to intelligent answer knowledge base, this method can be by the more new clothes of intelligent answer knowledge base
It sets to execute, wherein this method can be realized by the mode of hardware and/or software, and can be generally integrated in configured with artificial intelligence
In the equipment of (Artificial Intelligence, AI) system.It should be noted that being configured with intelligence in AI system
Energy question and answer knowledge base, so that user interacts with AI system.
As shown in Figure 1, this method specifically comprises the following steps:
S110, original intelligent answer knowledge base to be updated is obtained.
It should be noted that needing to create original intelligent answer first before obtaining original intelligent answer knowledge base
Knowledge base.Wherein, original intelligent answer knowledge base configuration is in the equipment for being integrated with AI system.It need to be to original intelligent answer knowledge
Knowledge is filled in library, and carries out AI training.
In embodiment, to the filling of original intelligent answer knowledge base, knowledge can there are two types of modes, one is passing through network
Crawler technology acquires open data and saves into the knowledge base of AI system corresponding to original intelligent answer knowledge base.Its
In, data collecting field including but not limited to the knowledge relevant to innovation undertaking such as double wound policies, salon's activity, foundation contest,
Activity and policy.While data acquisition, AI system can identify the corresponding field of institute's collecting data or classification, according to different
Field or classification carry out classification storage；The second is by many foundation tutors created in space, according to neck belonging to problem and knowledge
Domain, artificial inputs problem and its corresponding answer to the knowledge base category in AI system.
Currently, the mode of above two filling knowledge can save the corresponding keyword of problem and the difference of same problem mentions
It asks mode, and answer confidence level is set to initial value；It also, is the equal configuration association problem of each problem, and will be between problem
The degree of association be set to initial value.If AI system can not answer the problem of user proposes in system operating, foundation tutor is needed
After settling a dispute by the parties concerned themselves, the problem and answer that timely update in original intelligent answer knowledge base.
In the actual mechanical process of AI training, the mode of AI training includes but is not limited to: foundation tutor carries out simulation and mentions
Ask judge whether the answer that AI system provides is consistent with expection.If repeatedly can achieve the desired results after test, then pass through training；
Conversely, then inputting answer or the solution party in the intelligent answer knowledge base for improving AI system by crawler technology or tutor's knowledge
Method is filled a vacancy with leakage detection, and finds the knowledge being not directed in the intelligent answer knowledge base in AI system in time, and improve and answer questions
Efficiency and answer accuracy rate.
After completing intelligent answer knowledge base, know the intelligent answer knowledge base as original intelligent answer to be updated
Library is known, to be updated according to the practical operation of user to original intelligent answer knowledge base.
S120, active user's access request is received, according to active user's access request from original intelligent answer knowledge base
Middle lookup and the matched answer of active user's access request.
Wherein, the current problem of user's enquirement is carried in active user's access request.In embodiment, user proposes to ask
After topic, AI system can be according to the keyword and main body judges problem category in problem, in conjunction with the original intelligent answer of semantic matches
The problems in knowledge base, then the associated optimum answer of problem is showed into user, to achieve the purpose that solve customer problem.Herein
It should be noted that a problem can correspond to multiple answers, each answer has corresponding confidence level score value, uses in each answer
When the problem of family, only recommend the highest answer of confidence level, if the highest answer of confidence level does not solve customer problem, recommends confidence level
Second answer, and so on.
Certainly, if the problem of AI system hair is retrieved in original intelligent answer knowledge base less than matching with customer problem,
Or AI system provide answer do not solve user the problem of, then AI system needs to preserve the current problem of user, and
And the foundation tutor in field corresponding to current problem classification is matched, it is linked up by tutor and user.If AI system can not solve
The current problem of user then carries out online question-answering by foundation tutor；If foundation tutor fails to solve currently asking for user on line
Topic, then the tutor that starts an undertaking is solved under needing online by multiple sources and method, and passes through the modes such as phone, mailbox in time
To reply user.After solving the problems, such as, foundation tutor needs timely to fill knowledge to AI system and carries out AI training.
S130, user is received to the feedback information of answer, adjust answer confidence level, in real time according to feedback information to update original
Beginning intelligent answer knowledge base.
Wherein, answer confidence level refers to that the answer in intelligent answer knowledge base solves the problems, such as its associated degree.
AI system dynamically increases and decreases answer confidence level according to the feedback of user.In embodiment, it is assumed that answer confidence level is initial
Score value is c0, and minimum score value is c1.When answer is with a low credibility when minimum score value c1, AI system will no longer recommend the answer, directly
The answer is updated to by crawler technology (master) or foundation tutor (auxiliary).Point phenomenon is brushed in order to avoid there is user, in embodiment
In, AI system is limited for the feedback of same user, the answer in current intelligent answer knowledge base can only be carried out primary credible
Spend the bonus point or deduction operation of score value.
In embodiment, after each user completes the question and answer of the problem, the answer to the problem is fed back, so that
AI system adjusts answer confidence level according to feedback information, and updates in original intelligent answer knowledge base about the credible of the answer
Degree.
The technical solution of the present embodiment is receiving current use by obtaining original intelligent answer knowledge base to be updated
Family access request is searched from original intelligent answer knowledge base and active user's access request according to active user's access request
The answer matched；Then user is received to the feedback information of answer, adjusts answer confidence level, in real time according to feedback information to update original
Beginning intelligent answer knowledge base, realizes after user every time feeds back answer, directly adjusts answer according to feedback information
Confidence level ensure that the real-time update of data information in intelligent answer knowledge base.
On the basis of the above embodiments, it is searched from original intelligent answer knowledge base according to active user's access request
Make further embody with the matched answer of active user's access request.Fig. 2 is another intelligence provided in an embodiment of the present invention
The flow chart of the update method of energy question and answer knowledge base, as shown in Fig. 2, this method comprises:
S210, original intelligent answer knowledge base to be updated is obtained.
S220, active user's access request is received.
S230, identification extract the generic and main body of current problem in active user's access request.
In embodiment, after user's proposition problem, AI system segments current problem by participle technique, with
Obtain the type and theme included in current problem.
S240, judge whether comprising main body in current problem, if it is not, thening follow the steps S250；If so, thening follow the steps
S260。
S250, using the main body of current problem in last user access request as currently being asked in active user's access request
The main body of topic.
In embodiment, when user proposes to current problem, if AI system is unidentified to included in current problem
Main body, then using included in last user access request the problem of main body as the main body of current problem.For example, user first
A problem are as follows: " what be salon activity? " main body are as follows: salon's activity；User's Second Problem are as follows: " which thing should be paid attention to
? " the problem is without main body, then AI system brings the main body of first problem " salon's activity " into Second Problem automatically, that is, manages
Solution are as follows: " which item salon's activity should pay attention to? "；If the lower problem of user has main body, new main body is saved, it is deleted
The main body of preceding preservation.In addition, AI system enters dormant state if user is reactionless within the T1 time；If user T2 (T2 >
T1) reactionless in the time, then AI system finishing current sessions.
S260, problem source corresponding to current problem is matched according to the generic and main body of current problem, search problem source
Corresponding answer.
Wherein, problem source refers to that problem included in active user's access request is corresponding in intelligent answer knowledge base
Problem.In embodiment, each problem can correspond to multiple question formulations, be proposed in user by the question formulation of itself
When one problem, in order to enable AI system to go out corresponding optimum answer according to the problem identification, need to the problem into
Row conversion, to match problem source corresponding to the problem from intelligent answer knowledge base, and knows according to problem source from intelligent answer
Know in library and finds corresponding answer.
S270, user is received to the feedback information of answer, adjust answer confidence level, in real time according to feedback information to update original
Beginning intelligent answer knowledge base.
The technical solution of the present embodiment, by retaining the main body and keyword of current problem, to cope with, user is next to be asked
Topic does not have the case where main body, improves the usage experience of user.
On the basis of the above embodiments, when being updated to original intelligent answer knowledge base, it is also necessary to be closed to problem
Connection degree is updated.Fig. 3 is the flow chart of the update method of another intelligent answer knowledge base provided in an embodiment of the present invention, such as
Shown in Fig. 3, this method comprises:
S310, original intelligent answer knowledge base to be updated is obtained.
S320, active user's access request is received.
S330, it is searched from original intelligent answer knowledge base according to active user's access request and active user's access request
Matched answer.
S340, user is received to the feedback information of answer, adjust answer confidence level, in real time according to feedback information to update original
Beginning intelligent answer knowledge base.
S350, identification extract the current problem in active user's access request.
In embodiment, active user's access request is segmented by participle technique, current use is extracted with identification
The current problem that family access request is included.
S360, acquisition and the highest target problem of the current problem degree of association.
Wherein, target problem, the problem of referring to current problem degree of association highest.For the ease of true according to current problem
Problem, can be divided into process problem and non-process problem by the fixed and highest target problem of the current problem degree of association.For stream
Journey problem can directly determine target problem from current problem.For example, user has currently done a thing, then can determine that in next step
Have to do b thing such issues that.Crowd wound spaces in, there are many processes, for example, Ru Fu enterprise pay into
Process in expense participates in every race or activity process etc., and this kind of item often has a set of standardized process.For this
Standardized process problem is planted, needs to lay in the problem of each node of the process is likely encountered in intelligent answer knowledge base and answers
Case, and the problem of set between these nodes and node between degree of association score value it is higher than initial degree of association score value.In this way, if with
The problem of family, belongs to the problem of this process, then AI system can be according to the problem, and anti-release user is in the process
Which node, then the problems in latter node in the process and answer are elected to user.Therefore, asking for process
Topic, the precision of AI system prediction can be relatively high.
It, can be by the way of fuzzy calculate for non-process problem.Specifically, according in user's current problem
The problem of main body, search is relevant to the main body in intelligent answer knowledge base, user's access times highest (certainly, excludes root
According to user's current problem the problem of intelligent answer knowledge base is matched to) and corresponding answer.Wherein, user's access times refer to
Be number that the problems in intelligent answer knowledge base is checked by user.In embodiment, intelligent answer knowledge is checked in user
Behind the problems in library, the access times of the problem are carried out plus one operates.Certainly, access of the same position user to same problem,
AI system only carries out a primary plus operation to access times.
S370, using target problem as the problems in user access request next time.
In embodiment, after determining the corresponding target problem of current problem, AI system is using target problem as next
The problems in secondary user access request.
S380, user is received to the feedback information of target problem.
Certainly, the target problem of AI system recommendation, it is possible that the unsatisfied situation of user.If user thinks that target is asked
It is invalid that topic is recommended, then its next step of the feedback of user's selectivity is allowed to be intended.Intend in next step if user does not feed back, only by problem
Between the degree of association carry out deduction, terminate this question and answer；If user feedback is intended in next step, recommend corresponding problem and solution
Method.Such as: AI system solves the problems, such as the Q0 that user proposes, but is not previously predicted out the next step behavior of user, but root
According to the plan of user feedback, Q1 problem is elected, then by the degree of association score value between Q1 problem and Q0 problem in initial value
On the basis of plus G0 point.In this way, failing the deficiency that user's next step behavior is effectively predicted to make up AI system.
S390, the problem degree of association is adjusted according to the feedback information of target problem in real time, to update original intelligent answer knowledge
Library.
In embodiment, the problem degree of association refers to the correlation degree in intelligent answer knowledge base between each problem, can
To be interpreted as, there is the problem of a chain of property.For example, having asked that A problem, AI system can speculate that user in next step may in user
B operation can be done, then recommends C problem relevant to B operation to user, describing the problem C and A has relevance.AI system according to
The feedback at family dynamically increases and decreases the problem degree of association.In embodiment, it is assumed that problem degree of association initial value is b0, most
Low score value is b1.When the problem degree of association is lower than minimum score value, AI system will no longer recommend the related question, until AI system is more
New related question.Point phenomenon is brushed in order to avoid there is user, AI system is limited for the feedback of same user, intelligence can only be asked
Answer bonus point or deduction operation that the problems in knowledge base carries out a problem degree of association score value.
On the basis of the above embodiments, in order to guarantee the accuracy of feedback information, feedback information need to be screened, with
Obtain effective Feedback information.Fig. 4 is the process of the update method of another intelligent answer knowledge base provided in an embodiment of the present invention
Figure, as shown in figure 4, this method comprises:
S410, original intelligent answer knowledge base to be updated is obtained.
S420, active user's access request is received, according to active user's access request from original intelligent answer knowledge base
Middle lookup and the matched answer of active user's access request.
S430, user is received to the feedback information of answer.
S440, the current adjustment rate for obtaining answer confidence level.
In embodiment, it is assumed that answer confidence level score value is T0 at the time of being set to initial value, and AI system saves the T0 moment
Answer confidence level score value G0, and since the T0 moment, at regular intervals, add deduct to the increasing of current answer confidence level score value
Few rate is calculated.For example, having arrived the T1 moment, current credibility score value is G1, then currently adjustment rate is v=(G1-
G0)/(T1-T0)。
S450, the validity that feedback information is determined according to current adjustment rate.
If currently adjustment rate reaches default adjustment rate-valve value, it is determined that feedback information is invalid；If current adjustment speed
The not up to default adjustment rate-valve value of rate, it is determined that feedback information is effective.Illustratively, if answer confidence level in this period
Current adjustment rate be more than default adjustment rate-valve value, then determine that current answer confidence level has brush point phenomenon, automatically will
The answer confidence level is set to initial value.If by the operation, the number which is set to initial value is more than
N0 times, then AI system notify foundation tutor, which is assessed, to avoid AI system to brush divide phenomenon into
Row erroneous judgement, while also avoiding the operation that user carries out brush point.
Certainly, when user determines the answer or recommendation that AI system provides the problem of whether effectively after (no matter effective or nothing
Effect), allow the feedback user of user's selectivity exchanged with AI system in the problem, have what recommendation on improvement etc. to AI system,
The feedback information of user is periodically collected, is arranged in crowd's wound space, and for reasonable, useful, rationality feedback, space is corresponding to user
Certain reward is given by enterprise；Suggest for not filling in feedback, and repeatedly determines the problem of AI system is not answered a question or recommended
Useless user, crowd's wound space can be linked up with the enterprise customer, to go and find out what's going in time.
S460, answer confidence level is adjusted according to effective feedback information in real time, to update original intelligent answer knowledge base.
On the basis of the above embodiments, by taking intelligent answer knowledge base is the intelligent answer of incubation mechanism as an example, to intelligence
The update method of question and answer knowledge base is illustrated.Fig. 5 be another intelligent answer knowledge base provided in an embodiment of the present invention more
The flow chart of new method, as shown in figure 5, this method comprises:
S501, user propose problem.
Whether the problem of S502, user currently propose be reasonable.
In embodiment, judge whether the problem of user currently proposes be reasonable, i.e. whether the problem of judgement currently proposition belongs to
In innovation undertaking (double wounds) the problem of.Wherein, the whether reasonable method of decision problem are as follows: pass through the key in acquisition customer problem
Whether word and main body match with key, the problem about double wounds in intelligent answer knowledge base, if mismatching, then it represents that do not conform to
Reason then enters S503；Otherwise rationally, then enter S504.
S503, prompt user the input problem related with double wounds.
In embodiment, prompt user's current problem is unrelated with double wounds, need to propose the problem related with double wounds.Then, it ties
Line journey waits put question to user's next time into S521.
S504, AI system receive current problem.
In embodiment, AI system is after receiving current problem, be extracted and preserved the keyword in current problem and
Main body, and current problem generic is determined according to keyword, in conjunction with the problems in semantic matches knowledge base.
S505, with the presence or absence of problem source corresponding to current problem, if so, executing S506；If it is not, then executing S511.
In embodiment, AI system feedback problem matching result.If not found in intelligent answer knowledge base with user's
The problem of problem matches source then enters S511；If being matched to problem source, enter S506.
S506, AI system answer problem.
In embodiment, the problem of showing matching degree highest to user and the highest answer of the corresponding confidence level of problem；It will
Matching degree is more than to be consulted the problem of recommending threshold value and the highest answer of the corresponding confidence level of problem alternately item for user.
Whether S507, AI system solve current problem.If AI system solves customer problem, enter S515；If problem
It is unresolved, then enter S508.
S508, the confidence level minus fifteen by current answer.The answer is no longer recommended in this dialogue, and prepares to recommend next
The highest answer of confidence level.
S509, whether recommend there are also answer and recommend whether the number of answer is greater than k times.
In embodiment, AI system judge whether there are also answer recommend and for current problem recommend answer number whether
Greater than k times.If thering is answer to recommend and number being recommended to be less than or equal to k times, continue to recommend answer, into S506；If no answer pushes away
It recommends, no matter then recommending number, into S510.It determine whether there are also the methods that answer is recommended are as follows: if being associated with current problem
The confidence levels of all answers be below minimum, then it represents that current problem is without recommending answer, and AI will not recommend confidence level
Lower than the answer of minimum, until by crawler technology (master) or tutor (auxiliary) find new suitable answer (new answer it is credible
Degree is set to initial value).
S510, the current problem that user proposes, and the foundation tutor of matching treatment such problem are saved.
In embodiment, AI system saves current an open question, and matching treatment such problem is led
Teacher.
S511, foundation tutor's online question-answering carry out real-time communication with user.
S512, judge whether foundation tutor solves current problem, if solving, enter S514；If unresolved, enter
S513。
It is handled under S513, foundation tutor's line, and mail or the user that sends an answer by telephone.
In embodiment, foundation tutor solves customer problem by multiple sources and mode, and passes through phone, mail in time
Etc. modes reply user.
S514, foundation tutor provide the keyword of problem in knowledge-based classification storage problem and the corresponding answer of problem
The different question formulations with same problem；The confidence level of answer is set to+1 point of initial value；Configuration association problem, and by the degree of association
It is set to initial value；Periodically carry out AI training.
S515, user feedback solve the answer of current problem.
In embodiment, since AI system is other than providing the highest answer of the degree of association, alternative answer can be also provided,
Therefore user is needed to feed back answer used by its own.It is to be understood which answer that AI system provides solves
Current problem, user just feed back the answer to AI system.
S516, the confidence level of current answer is increased by one point.
In embodiment, if current answer is determined useful by user, AI system is credible by the current answer of user feedback
The score value of degree adds one.
S517, the next step behavior for predicting user, and provide highest next with active user's problem degree of association score value
The problems and solutions.
In embodiment, the problem of AI system is proposed by user, predicts the next step behavior of user, and provide with it is current
The highest next the problems and solutions of customer problem degree of association score value.Wherein predict the method for user's next step behavior are as follows:
It is the problem of matching next degree of association score value highest, reversed to release according to the logical order between the degree of association and problem of problem
The applicable scene of the problem predicts what user means to do in next step with this.
S518, user judge AI system predict whether it is useful, if useful, enter S520；If useless, enter
S519。
S519, by the degree of association minus fifteen between problem；If hereafter onrelevant question recommending updates related question, lay equal stress on
Set degree of association score value.
In embodiment, if current predictive determined by user it is useless, by the degree of association minus fifteen between problem；If hereafter without
Related question is recommended, then updates related question, and reset degree of association score value.Process terminates, and into S521, user is waited to mention next time
It asks.Where it is determined that whether there is or not the methods that related question is recommended are as follows: if the problems in the associated all knowledge bases of active user's problem
The degree of association be below minimum, then it represents that current onrelevant question recommending, and AI will not recommend the degree of association lower than minimum
The problem of, but match new problem in knowledge base and be associated, or find new problem by crawler technology, tutor,
It stores into knowledge base, and is associated (degree of association of new problem is set to initial value).
S520, the degree of association between problem is added one point.
In embodiment, if current predictive is determined useful by user, the degree of association between problem is added one point.Process knot
Beam waits put question to user's next time into S521.
S521, process terminate, and wait put question to user's next time.
Fig. 6 is a kind of structural block diagram of the updating device of intelligent answer knowledge base provided in an embodiment of the present invention, the device
The case where suitable for carrying out real-time update to intelligent answer knowledge base, which can be realized by hardware/software, and can generally be collected
At in the equipment configured with AI system.As shown in fig. 6, the device includes: the first acquisition module 610, searching module 620 and
One adjustment update module 630.
Wherein, first module 610 is obtained, for obtaining original intelligent answer knowledge base to be updated；
Searching module 620, for receiving active user's access request, according to active user's access request from original intelligence
It is searched and the matched answer of active user's access request in question and answer knowledge base；
The first adjustment update module 630 is adjusted for receiving user to the feedback information of answer according to feedback information in real time
Answer confidence level, to update original intelligent answer knowledge base.
The technical solution of the present embodiment is receiving current use by obtaining original intelligent answer knowledge base to be updated
Family access request is searched from original intelligent answer knowledge base and active user's access request according to active user's access request
The answer matched；Then user is received to the feedback information of answer, adjusts answer confidence level, in real time according to feedback information to update original
Beginning intelligent answer knowledge base, realizes after user every time feeds back answer, directly adjusts answer according to feedback information
Confidence level ensure that the real-time update of data information in intelligent answer knowledge base.
On the basis of the above embodiments, searched from original intelligent answer knowledge base according to active user's access request with
The matched answer of active user's access request, is specifically used for:
The generic and main body of current problem in active user's access request are extracted in identification；
If current problem includes main body, according to the generic of current problem and main body matching, current problem is corresponding asks
Topic source, the corresponding answer in the source that searches problem；
If current problem does not include main body, used using the main body of current problem in last user access request as current
The main body of current problem in the access request of family.
On the basis of the above embodiments, the updating device of intelligent answer knowledge base, further includes:
Extraction module is identified, for after receiving active user's access request, identification to be extracted active user's access and asked
Current problem in asking；
Second obtains module, for obtaining and the highest target problem of the current problem degree of association；
Determining module, for using target problem as the problems in user access request next time.
On the basis of the above embodiments, the updating device of intelligent answer knowledge base, further includes:
Receiving module, for using target problem as the problems in user access request next time after, reception user
To the feedback information of target problem；
Second adjustment update module, for adjusting the problem degree of association in real time according to the feedback information of target problem, to update
Original intelligent answer knowledge base.
On the basis of the above embodiments, the updating device of intelligent answer knowledge base, further includes:
Screening module, for adjusting answer in real time according to feedback information after receiving user to the feedback information of answer
Before confidence level, feedback information is screened, to obtain effective Feedback information.
On the basis of the above embodiments, screening module, comprising:
Acquiring unit, for obtaining the current adjustment rate of answer confidence level；
Determination unit, for determining the validity of feedback information according to current adjustment rate.
On the basis of the above embodiments, determination unit, comprising:
First determines subelement, if reaching default adjustment rate-valve value for currently adjusting rate, it is determined that feedback information
It is invalid；
Second determines subelement, if for currently adjusting the not up to default adjustment rate-valve value of rate, it is determined that feedback letter
Breath is effective.
The updating device of above-mentioned intelligent answer knowledge base can be performed intelligent answer provided by any embodiment of the invention and know
The update method for knowing library, has the corresponding functional module of execution method and beneficial effect.
Fig. 7 is a kind of hardware structural diagram of equipment provided in an embodiment of the present invention.Equipment in the embodiment of the present invention
It is illustrated by taking the equipment configured with AI system as an example.As shown in fig. 7, equipment provided in an embodiment of the present invention, comprising: processor
710 and memory 720, input unit 730 and output device 740.Processor 710 in the computer equipment can be one or
Multiple, in Fig. 7 by taking a processor 710 as an example, in computer equipment processors 710, memory 720,730 and of input unit
Output device 740 can be connected by bus or other modes, in Fig. 7 for being connected by bus.
Memory 720 in the computer equipment is used as a kind of computer readable storage medium, can be used for storing one or
Multiple programs, program can be software program, computer executable program and module, as the above embodiment of the present invention provides
Corresponding program instruction/the module of the update method of intelligent answer knowledge base is (for example, intelligent answer knowledge base shown in Fig. 7 is more
Module in new equipment, comprising: first obtains module 610, searching module 620 and the first adjustment update module 630).Processor
710 software program, instruction and the modules being stored in memory 720 by operation, thereby executing the various of computer equipment
Functional application and data processing, i.e., the update method of intelligent answer knowledge base in realization above method embodiment.
Memory 720 may include storing program area and storage data area, wherein storing program area can storage program area,
Application program needed at least one function；Storage data area, which can be stored, uses created data etc. according to equipment.In addition,
Memory 720 may include high-speed random access memory, can also include nonvolatile memory, for example, at least a disk
Memory device, flush memory device or other non-volatile solid state memory parts.In some instances, memory 720 can be wrapped further
The memory remotely located relative to processor 710 is included, these remote memories can pass through network connection to equipment.Above-mentioned net
The example of network includes but is not limited to internet, intranet, local area network, mobile radio communication and combinations thereof.
Input unit 730 can be used for receiving the number or character information of user's input, to generate the user with terminal device
Setting and the related key signals input of function control.Output device 740 may include that display screen etc. shows equipment.
Also, when one or more included program of above-mentioned computer equipment is held by one or more processor 710
When row, program is proceeded as follows:
Obtain original intelligent answer knowledge base to be updated；Active user's access request is received, is visited according to active user
Ask that request is searched and the matched answer of active user's access request from original intelligent answer knowledge base；User is received to answer
Feedback information adjusts answer confidence level according to feedback information in real time, to update original intelligent answer knowledge base.
The embodiment of the invention also provides a kind of computer readable storage mediums, are stored thereon with computer program, the journey
The update method of intelligent answer knowledge base provided in an embodiment of the present invention is realized when sequence is executed by processor, this method comprises:
Obtain original intelligent answer knowledge base to be updated；Active user's access request is received, is visited according to active user
Ask that request is searched and the matched answer of active user's access request from original intelligent answer knowledge base；User is received to answer
Feedback information adjusts answer confidence level according to feedback information in real time, to update original intelligent answer knowledge base.
The computer storage medium of the embodiment of the present invention, can be using any of one or more computer-readable media
Combination.Computer-readable medium can be computer-readable signal media or computer readable storage medium.It is computer-readable
Storage medium for example can be -- but being not limited to -- electricity, magnetic, optical, electromagnetic, infrared ray or semiconductor system, device or device,
Or any above combination.The more specific example (non exhaustive list) of computer readable storage medium includes: with one
The electrical connection of a or multiple conducting wires, portable computer diskette, hard disk, random access memory (RAM), read-only memory
(ROM), erasable programmable read only memory (EPROM or flash memory), optical fiber, portable compact disc read-only memory (CD-
ROM), light storage device, magnetic memory device or above-mentioned any appropriate combination.In this document, computer-readable storage
Medium can be any tangible medium for including or store program, which can be commanded execution system, device or device
Using or it is in connection.
Computer-readable signal media may include in a base band or as carrier wave a part propagate data-signal,
Wherein carry computer-readable program code.The data-signal of this propagation can take various forms, including but unlimited
In electromagnetic signal, optical signal or above-mentioned any appropriate combination.Computer-readable signal media can also be that computer can
Any computer-readable medium other than storage medium is read, which can send, propagates or transmit and be used for
By the use of instruction execution system, device or device or program in connection.
The program code for including on computer-readable medium can transmit with any suitable medium, including --- but it is unlimited
In wireless, electric wire, optical cable, RF etc. or above-mentioned any appropriate combination.
The computer for executing operation of the present invention can be write with one or more programming languages or combinations thereof
Program code, programming language include object oriented program language-such as Java, Smalltalk, C++, are also wrapped
Include conventional procedural programming language-such as " C " language or similar programming language.Program code can be complete
Ground executes on the user computer, partly executes on the user computer, executing as an independent software package, partially existing
Part executes on the remote computer or executes on a remote computer or server completely on subscriber computer.It is being related to
In the situation of remote computer, remote computer can pass through the network of any kind --- including local area network (LAN) or wide area
Net (WAN)-be connected to subscriber computer, or, it may be connected to outer computer (such as utilize ISP
To be connected by internet).
Note that the above is only a better embodiment of the present invention and the applied technical principle.It will be appreciated by those skilled in the art that
The present invention is not limited to specific embodiments here, be able to carry out for a person skilled in the art it is various it is apparent variation, again
Adjustment and substitution are without departing from protection scope of the present invention.Therefore, although by above embodiments to the present invention carried out compared with
For detailed description, but the present invention is not limited to the above embodiments only, without departing from the inventive concept, can be with
Including more other equivalent embodiments, and the scope of the invention is determined by the scope of the appended claims.
Claims (10)
1. a kind of update method of intelligent answer knowledge base characterized by comprising
Obtain original intelligent answer knowledge base to be updated；
Active user's access request is received, according to active user's access request from the original intelligent answer knowledge base
It searches and the matched answer of active user's access request；
User is received to the feedback information of the answer, answer confidence level is adjusted according to the feedback information in real time, to update
State original intelligent answer knowledge base.
2. the method according to claim 1, wherein it is described according to active user's access request from the original
It is searched and the matched answer of active user's access request in beginning intelligent answer knowledge base, comprising:
The generic and main body of current problem in active user's access request are extracted in identification；
If the current problem includes main body, the current problem is matched according to the generic of the current problem and main body
The corresponding answer in described problem source is searched in corresponding problem source；
If the current problem does not include main body, the main body of current problem in last user access request is worked as described in
The main body of current problem in preceding user access request.
3. the method according to claim 1, wherein it is described receive active user's access request after, also
Include:
The current problem in active user's access request is extracted in identification；
It obtains and the highest target problem of the current problem degree of association；
Using the target problem as the problems in user access request next time.
4. according to the method described in claim 3, it is characterized in that, being visited described using the target problem as user next time
Ask request the problems in after, further includes:
User is received to the feedback information of the target problem；
The problem degree of association is adjusted in real time according to the feedback information of the target problem, to update the original intelligent answer knowledge
Library.
5. the method according to claim 1, wherein the reception user to the feedback information of the answer it
Afterwards, it is described answer confidence level is adjusted according to the feedback information in real time before, further includes:
The feedback information is screened, to obtain effective Feedback information.
6. according to the method described in claim 5, it is characterized in that, described screen the feedback information, to be had
Imitate feedback information, comprising:
Obtain the current adjustment rate of answer confidence level；
The validity of feedback information is determined according to the current adjustment rate.
7. according to the method described in claim 6, it is characterized in that, described determine feedback information according to the current adjustment rate
Validity, comprising:
If the current adjustment rate reaches default adjustment rate-valve value, it is determined that the feedback information is invalid；
If the not up to default adjustment rate-valve value of the current adjustment rate, it is determined that the feedback information is effective.
8. a kind of updating device of intelligent answer knowledge base characterized by comprising
First obtains module, for obtaining original intelligent answer knowledge base to be updated；
Searching module, for receiving active user's access request, according to active user's access request from the original intelligence
It can be searched and the matched answer of active user's access request in question and answer knowledge base；
The first adjustment update module is adjusted for receiving user to the feedback information of the answer according to the feedback information in real time
Whole answer confidence level, to update the original intelligent answer knowledge base.
9. a kind of equipment, which is characterized in that the equipment includes:
One or more processors；
Storage device, for storing one or more programs；
When one or more of programs are executed by one or more of processors, so that one or more of processors are real
The now update method of the intelligent answer knowledge base as described in any in claim 1-7.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the program is by processor
The update method of the intelligent answer knowledge base as described in any in claim 1-7 is realized when execution.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910482105.8A CN110188183A (en) | 2019-06-04 | 2019-06-04 | A kind of update method, device, equipment and the storage medium of intelligent answer knowledge base |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910482105.8A CN110188183A (en) | 2019-06-04 | 2019-06-04 | A kind of update method, device, equipment and the storage medium of intelligent answer knowledge base |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110188183A true CN110188183A (en) | 2019-08-30 |
Family
ID=67720284
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910482105.8A Pending CN110188183A (en) | 2019-06-04 | 2019-06-04 | A kind of update method, device, equipment and the storage medium of intelligent answer knowledge base |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110188183A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111353017A (en) * | 2020-02-20 | 2020-06-30 | 上海智臻智能网络科技股份有限公司 | Intelligent interaction method and device |
| CN111563150A (en) * | 2020-04-30 | 2020-08-21 | 广东美的制冷设备有限公司 | Air conditioner knowledge base updating method, air conditioner, server and system |
| CN111767380A (en) * | 2020-06-29 | 2020-10-13 | 北京百度网讯科技有限公司 | Model adaptive retraining method and device, electronic equipment and storage medium |
| CN113779206A (en) * | 2020-11-11 | 2021-12-10 | 北京沃东天骏信息技术有限公司 | Data processing method and device, electronic equipment and storage medium |
| CN113779050A (en) * | 2020-06-23 | 2021-12-10 | 北京沃东天骏信息技术有限公司 | Method and device for managing knowledge base of customer service robot |
| CN113868538A (en) * | 2021-10-19 | 2021-12-31 | 北京字跳网络技术有限公司 | Information processing method, device, equipment and medium |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102158484A (en) * | 2011-03-17 | 2011-08-17 | 同济大学 | Dynamic car sharing system and method in mobile social network |
| CN103177092A (en) * | 2013-03-08 | 2013-06-26 | 深圳先进技术研究院 | Data updating method and system of knowledge base and knowledge base |
| US20160371263A1 (en) * | 2013-07-08 | 2016-12-22 | Information Extraction Systems, Inc. | Apparatus, system and method for a semantic editor and search engine |
| CN106844686A (en) * | 2017-01-26 | 2017-06-13 | 武汉奇米网络科技有限公司 | Intelligent customer service question and answer robot and its implementation based on SOLR |
| CN107025297A (en) * | 2017-04-18 | 2017-08-08 | 上海斐讯数据通信技术有限公司 | A kind of chat robots and its automatic chatting method |
| CN107343053A (en) * | 2017-08-01 | 2017-11-10 | 骆德轩 | A kind of content recommendation method |
| CN107798123A (en) * | 2017-11-10 | 2018-03-13 | 上海智臻智能网络科技股份有限公司 | Knowledge base and its foundation, modification, intelligent answer method, apparatus and equipment |
| CN107908699A (en) * | 2017-11-03 | 2018-04-13 | 福建中金在线信息科技有限公司 | Content display method and server |
| CN109033262A (en) * | 2018-07-09 | 2018-12-18 | 北京寻领科技有限公司 | Question and answer knowledge base update method and device |
| CN109242573A (en) * | 2018-09-26 | 2019-01-18 | 湖南人文科技学院 | Evaluation method, device, equipment and the storage medium of APP |
-
2019
- 2019-06-04 CN CN201910482105.8A patent/CN110188183A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102158484A (en) * | 2011-03-17 | 2011-08-17 | 同济大学 | Dynamic car sharing system and method in mobile social network |
| CN103177092A (en) * | 2013-03-08 | 2013-06-26 | 深圳先进技术研究院 | Data updating method and system of knowledge base and knowledge base |
| US20160371263A1 (en) * | 2013-07-08 | 2016-12-22 | Information Extraction Systems, Inc. | Apparatus, system and method for a semantic editor and search engine |
| CN106844686A (en) * | 2017-01-26 | 2017-06-13 | 武汉奇米网络科技有限公司 | Intelligent customer service question and answer robot and its implementation based on SOLR |
| CN107025297A (en) * | 2017-04-18 | 2017-08-08 | 上海斐讯数据通信技术有限公司 | A kind of chat robots and its automatic chatting method |
| CN107343053A (en) * | 2017-08-01 | 2017-11-10 | 骆德轩 | A kind of content recommendation method |
| CN107908699A (en) * | 2017-11-03 | 2018-04-13 | 福建中金在线信息科技有限公司 | Content display method and server |
| CN107798123A (en) * | 2017-11-10 | 2018-03-13 | 上海智臻智能网络科技股份有限公司 | Knowledge base and its foundation, modification, intelligent answer method, apparatus and equipment |
| CN109033262A (en) * | 2018-07-09 | 2018-12-18 | 北京寻领科技有限公司 | Question and answer knowledge base update method and device |
| CN109242573A (en) * | 2018-09-26 | 2019-01-18 | 湖南人文科技学院 | Evaluation method, device, equipment and the storage medium of APP |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111353017A (en) * | 2020-02-20 | 2020-06-30 | 上海智臻智能网络科技股份有限公司 | Intelligent interaction method and device |
| CN111563150A (en) * | 2020-04-30 | 2020-08-21 | 广东美的制冷设备有限公司 | Air conditioner knowledge base updating method, air conditioner, server and system |
| CN111563150B (en) * | 2020-04-30 | 2024-04-16 | 广东美的制冷设备有限公司 | Air conditioner knowledge base updating method, air conditioner, server and system |
| CN113779050A (en) * | 2020-06-23 | 2021-12-10 | 北京沃东天骏信息技术有限公司 | Method and device for managing knowledge base of customer service robot |
| CN111767380A (en) * | 2020-06-29 | 2020-10-13 | 北京百度网讯科技有限公司 | Model adaptive retraining method and device, electronic equipment and storage medium |
| CN113779206A (en) * | 2020-11-11 | 2021-12-10 | 北京沃东天骏信息技术有限公司 | Data processing method and device, electronic equipment and storage medium |
| CN113868538A (en) * | 2021-10-19 | 2021-12-31 | 北京字跳网络技术有限公司 | Information processing method, device, equipment and medium |
| WO2023065825A1 (en) * | 2021-10-19 | 2023-04-27 | 北京字跳网络技术有限公司 | Information processing method and apparatus, device, and medium |
| CN113868538B (en) * | 2021-10-19 | 2024-04-09 | 北京字跳网络技术有限公司 | Information processing method, device, equipment and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110188183A (en) | A kind of update method, device, equipment and the storage medium of intelligent answer knowledge base | |
| WO2017213857A1 (en) | System for iteratively training an artificial intelligence using cloud-based metrics | |
| CN110168578A (en) | Multitask neural network with task particular path | |
| CN109299458A (en) | Entity recognition method, device, equipment and storage medium | |
| CN108962238A (en) | Dialogue method, system, equipment and storage medium based on structural neural networks | |
| CN108363745A (en) | The method and apparatus that robot customer service turns artificial customer service | |
| EP3553710B1 (en) | Artificial intelligence optimized telecommunications systems | |
| CN104869581B (en) | It is propagated via the label in wireless sensor network and determines malfunctioning node | |
| CN107016595B (en) | Information recommendation method, device and system | |
| Kell et al. | Operational management procedures: an introduction to the use of evaluation frameworks | |
| CN108664222A (en) | A kind of block catenary system and its application process | |
| RU2016139465A (en) | SYSTEM AND METHOD OF OPERATION OF AN ARTIFICIAL SOCIAL NETWORK | |
| KR102163412B1 (en) | Method for providing business establishing information report | |
| Saha et al. | Towards integrated dialogue policy learning for multiple domains and intents using hierarchical deep reinforcement learning | |
| CN108595609A (en) | Generation method, system, medium and equipment are replied by robot based on personage IP | |
| CN112925888A (en) | Method and device for training question-answer response and small sample text matching model | |
| CN114896899A (en) | Multi-agent distributed decision method and system based on information interaction | |
| Zhuang et al. | The relay satellite scheduling based on artificial bee colony algorithm | |
| WO2021077834A1 (en) | Dialogue system-based method for asking follow-up question in response to user question, and device | |
| CN116501592B (en) | Man-machine interaction data processing method and server | |
| JP2019209797A (en) | System and method of preparing operation arrangement plan | |
| CN112733552A (en) | Machine translation model construction method, device and equipment | |
| Nashed et al. | Selecting the partial state abstractions of MDPs: A metareasoning approach with deep reinforcement learning | |
| US11847614B2 (en) | Method and system for determining collaboration between employees using artificial intelligence (AI) | |
| CN108415983B (en) | Intelligent problem solving method and device based on interaction |
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: 20190830 |