WO2021244574A1 - Procédé et dispositif d'établissement d'une paire de questions-réponses - Google Patents

Procédé et dispositif d'établissement d'une paire de questions-réponses Download PDF

Info

Publication number
WO2021244574A1
WO2021244574A1 PCT/CN2021/097937 CN2021097937W WO2021244574A1 WO 2021244574 A1 WO2021244574 A1 WO 2021244574A1 CN 2021097937 W CN2021097937 W CN 2021097937W WO 2021244574 A1 WO2021244574 A1 WO 2021244574A1
Authority
WO
WIPO (PCT)
Prior art keywords
question
answer pair
reply message
current
customer service
Prior art date
Application number
PCT/CN2021/097937
Other languages
English (en)
Chinese (zh)
Inventor
胡尧
詹泽
娄伟锋
宋长福
Original Assignee
支付宝(杭州)信息技术有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 支付宝(杭州)信息技术有限公司 filed Critical 支付宝(杭州)信息技术有限公司
Publication of WO2021244574A1 publication Critical patent/WO2021244574A1/fr

Links

Images

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
    • G06F16/31—Indexing; Data structures therefor; Storage structures
    • 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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00—Computing arrangements using knowledge-based models
    • G06N5/02—Knowledge representation; Symbolic representation
    • G06N5/022—Knowledge engineering; Knowledge acquisition

Definitions

  • This specification relates to the field of Internet technology, and in particular to a method and device for establishing question and answer pairs.
  • this specification provides a method and device for establishing question and answer pairs.
  • a method for establishing a question and answer pair includes: monitoring a paste operation of a human customer service; and using the text corresponding to the paste operation as the current reply message to the current user question ; Determine the number of historical reply messages matching the current reply message in the current period, the historical reply message is input by the manual customer service; if the number of historical reply messages meets a preset condition, according to the current reply The message and the current user question generate a question and answer pair; the question and answer pair is stored in the knowledge base.
  • a device for establishing question and answer pairs comprising: a monitoring unit for monitoring a paste operation of a human customer service; an acquisition unit for taking the text corresponding to the pasting operation as a current reply message to a current user question; and a counting unit , Used to determine the number of historical reply messages matching the current reply message in the current period, the historical reply message being input by the manual customer service; the judging unit, used to determine if the number of historical reply messages meets a preset condition , The question and answer pair is generated according to the current reply message and the current user question; the storage unit is used to store the question and answer pair in the knowledge base.
  • a device for establishing question and answer pairs includes: a processor; a memory for storing machine executable instructions; wherein, by reading and executing the machine executable instructions corresponding to the method for establishing question and answer pairs stored in the memory, the The processor is prompted to: monitor the paste operation of the human customer service; use the text corresponding to the paste operation as the current reply message of the current user question; determine the number of historical reply messages matching the current reply message in the current period, the The historical reply message is input by the manual customer service; if the number of historical reply messages meets a preset condition, a question-and-answer pair is generated based on the current reply message and the current user question; the question-and-answer pair is stored in the knowledge base .
  • An embodiment of this specification realizes that the pasting operation of the human customer service is monitored, and the text corresponding to the pasting operation is used as the current reply message, and then the reply message is matched with the historical reply message in the current period.
  • a question-and-answer pair is generated according to the reply message and the corresponding user question, and the question-and-answer pair is stored in the knowledge base.
  • it if it is monitored that the human customer service enters similar reply messages multiple times in a cycle, it means that the reply message is used more frequently. Subsequent questions and answers can be generated and saved based on the reply message and the corresponding user question. right.
  • the method of this embodiment saves labor costs and is more efficient; compared with the establishment of question and answer pairs through machine learning algorithms, the method of this embodiment not only reduces the pressure on server performance, It also has better timeliness.
  • Fig. 1 is a schematic flowchart of a method for establishing question and answer pairs according to an exemplary embodiment of the present specification.
  • Fig. 2 is a schematic flowchart of another method for establishing question and answer pairs according to an exemplary embodiment of the present specification.
  • Fig. 3 is a schematic structural diagram of an apparatus for establishing question and answer pairs according to an exemplary embodiment of the present specification.
  • Fig. 4 is a block diagram of a device for establishing question and answer pairs according to an exemplary embodiment of the present specification.
  • first, second, third, etc. may be used in this specification to describe various information, the information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other.
  • first information may also be referred to as second information, and similarly, the second information may also be referred to as first information.
  • word “if” as used herein can be interpreted as "when” or “when” or “in response to a certainty”.
  • the customer service system refers to a system for responding to user inquiries. In order to improve the response efficiency, it is mainly aimed at the response efficiency of repeated questions.
  • a question and answer pair can be established, and the question and answer pair is usually a mapping relationship between the question and the answer.
  • the customer service system receives the user's question, it can query the question and answer pair, and if the corresponding answer is inquired, it can automatically reply to the user, or give a reference to the reply message.
  • how to screen them and establish effective question and answer pairs in a timely manner has become an urgent problem to be solved.
  • manual input can be used to establish question-and-answer pairs.
  • a human customer service can, based on experience, enter a user question with a higher frequency and a reply message corresponding to the user question into the knowledge base that stores the question and answer pairs.
  • intelligent algorithms can also be used for data mining to establish question-and-answer pairs.
  • the customer service system usually needs to obtain a large number of user questions and corresponding reply messages, and then build a model based on these data to dig out effective question and answer pairs.
  • this method requires a large amount of data to be collected in advance, making the establishment of question and answer pairs lagging behind, and time-sensitive user questions cannot be added in time.
  • the question-and-answer pairs established by this method have low accuracy, require a lot of calculations, and consume equipment performance.
  • this manual proposes a method for establishing question-and-answer pairs.
  • you can monitor the reply message entered by the manual customer service for the user's question, and then match the reply message with the historical reply message in the current period. When the number of matching historical reply messages meets the preset condition, a question-and-answer pair is generated according to the reply message and the corresponding user question.
  • the method of this manual has less manual participation and less pressure on the server. While establishing question and answer pairs efficiently, it can also have higher accuracy and timeliness.
  • the method of this manual is applied to a customer service system.
  • the client of the customer service system can display user questions, reply messages or other prompt messages through the conversation window.
  • the server of the customer service system stores a knowledge base and performs calculations and searches. deal with.
  • the client and server can interact through the network.
  • the knowledge base is a database or data table used to store question and answer pairs.
  • users can manually input question and answer pairs that are frequently used to the knowledge base. Then, question and answer pairs can be generated according to the method in this manual and stored in the knowledge base.
  • the question-and-answer pair in the knowledge base can use the key words of the question as the user's question, and the complete sentence as the reply message.
  • the server of the customer service system can use a preset algorithm to obtain one or more keywords in the user question, and then query the response corresponding to the keyword in the knowledge base information.
  • extracting keywords reference may be made to related technologies, which will not be repeated in this embodiment.
  • the question and answer pair in the knowledge base can also save both the user's question and the reply message as a complete sentence.
  • the server of the customer service system can use a preset algorithm to match the received user question with the user question in the knowledge base question and answer pair one by one. If the matching is successful, then get The reply message of the matched question and answer pair.
  • the server can query whether there is a question-and-answer pair in the knowledge base that matches the user’s question, such as extracting keywords in the user’s question and storing it in the knowledge base. Query the keyword in the database; or directly match the user question with the user question in the knowledge base question and answer pair.
  • the reply message of the queried question and answer pair can be returned to the user.
  • This process does not require manual customer service to participate, and it can also be called intelligent customer service.
  • the customer service system when the customer service system receives a user question, on the one hand, it assigns a human customer service to the user question, and on the other hand, it can use the server to query whether there is a matching user question in the knowledge base. Q&A right.
  • the server will return the reply message in the queried Q&A pair to the client, and the client will display the received reply message through floating windows or embedded interfaces, and then the human customer service can determine whether the reply message is suitable for use as The reply message of the user's question, and operations such as editing and sending can be performed on the reply message.
  • the server informs the client that the query result is a query failure, and the client can notify the manual customer service of the query failure through a pop-up window, and listen to the reply message entered by the manual customer service in the conversation window, and the follow-up can be based on the reply message of the manual customer service To generate question and answer pairs.
  • Step S102 monitor the paste operation of the human customer service
  • Step S104 use the text corresponding to the paste operation as the current reply message of the current user question.
  • the customer service system can assign manual customer service to the user's problem through the server, and create a new conversation window on the client of the designated manual customer service to display the user's problem, and then monitor the conversation window through the client to obtain the manual customer service
  • the reply message monitored by the client can be called the current reply message.
  • a manual customer service such as randomly selecting a manual customer service that is in an idle state, or assigning a corresponding type of manual customer service to the user's problem according to the category to which the user belongs, etc., which can be assigned according to the actual situation.
  • a manual customer service such as randomly selecting a manual customer service that is in an idle state, or assigning a corresponding type of manual customer service to the user's problem according to the category to which the user belongs, etc., which can be assigned according to the actual situation.
  • assign a manual customer service such as randomly selecting a manual customer service that is in an idle state, or assigning a corresponding type of manual customer service to the user's problem according to the category to which the user belongs, etc., which can be assigned according to the actual situation.
  • the human customer service can manually enter the corresponding reply message in the input box of the conversation window.
  • the client can monitor the input operation in the conversation window and use the input as the current reply message.
  • the human customer service can also copy the reply message from other materials and paste it into the input box.
  • the client can monitor the paste operation in the conversation window, and use the text corresponding to the paste operation as the current reply message.
  • the pasted text is generally more fixed and used more frequently.
  • Monitoring the pasting operation of the client can increase the threshold for generating question and answer pairs, reduce low-quality question and answer pairs, and reduce customer service personnel
  • the workload of the review and at the same time, when the customer service staff does not conduct the review and automatically puts the question and answer into the database, the quality of the question and answer pair can be guaranteed to a certain extent.
  • Step S106 Determine the number of historical reply messages matching the current reply message in the current period, and the historical reply messages are input by the manual customer service.
  • the historical reply message may be manually input by a human customer service, or it may be input by a manual customer service by pasting.
  • the customer service system may store historical reply messages that the human customer service has replied on the server side, for example, maintain a list of historical reply messages.
  • the historical reply message and the timestamp of the historical reply message may be recorded, and the timestamp may be the reply time of the historical reply message.
  • the historical reply messages can be cleaned up regularly.
  • each historical reply message can be polled regularly, and the corresponding historical reply message can be judged whether the corresponding historical reply message has been longer than one month based on the timestamp. If it exceeds, the historical reply message can be deleted.
  • historical reply messages can be recorded in the historical reply message list, and the aging time can be set for the historical reply messages.
  • the aging time is the length of a cycle, such as 1 day, 1 week, or 1 month.
  • start timing when the historical reply message is stored in the list and delete the historical reply message when the aging time is exceeded. It is understandable that if the method of this embodiment is adopted, all historical reply messages in the list are in the "current period".
  • the reply message saved in the history reply message list is a complete sentence.
  • the reply message entered by the manual customer service is "today can be shipped"
  • the reply message is saved in the history reply message list.
  • the historical reply message list may also store one or more keywords of the reply message. For example, in the above example, it can be determined that the keywords of the reply message are "today” and "delivery", and then these two keywords can be saved.
  • the historical reply message list may also store the correspondence between reply messages and reply message keywords.
  • the entries shown in Table 1 can be saved.
  • the user question corresponding to the reply message can also be saved in the historical reply message list, which is determined according to actual needs and is not limited here.
  • the number of historical reply messages matching the current reply message in the current period can be determined through the steps shown in FIG. 2.
  • Step S1062 Obtain one or more keywords in the current reply message.
  • Step S1064 Judging whether the current reply message matches each historical reply message in the current period according to the keywords in the current reply message.
  • first search for historical reply messages in the current period corresponds to the method of recording historical reply messages. Taking the method of recording historical reply messages in the above two embodiments as an example, the following method can be used to find historical reply messages in the current period.
  • the method of recording the timestamp of historical reply messages for each historical reply message, it is judged whether it is in the current cycle according to the timestamp. For example, the time difference between the current time and the time stamp of the historical reply message can be calculated, and if the time difference is less than the duration of one period, it is determined that the historical reply message is in the current period.
  • the current reply messages can be compared with the historical reply messages one by one.
  • the keywords of the historical reply message can be compared with the keywords of the current reply message.
  • the ratio of the two keywords that are the same or similar is determined as the keyword matching rate. If the keyword matching rate exceeds a preset threshold, it can be considered that the current reply message matches the historical reply message.
  • the method for judging whether keywords are similar may include whether the distance between the word vectors of the keywords meets a preset condition; or a word database of synonyms may be established in advance. The specific method is determined according to actual needs, and is not limited here.
  • a historical reply message list is used to save historical reply messages, and the time stamp of each historical reply message is also recorded in the list, as shown in Table 2.
  • Timestamp 1 Can be shipped today delivery today 2020/1/1 7:00 2 The product is out of stock, please apply for a return Sold Out 2020/1/1 8:00 3 Delivery location is Guangzhou Guangzhou, delivery 2020/1/1 9:00
  • the current reply message is "The warehouse is already packing and will be shipped today", and the current time is 10:00 on 1/1, 2020.
  • the threshold of the keyword matching rate is 90%, it can be determined that the historical reply message with sequence number 1 matches the current reply message, and other historical reply messages do not match the current reply message.
  • reply message keywords are not saved, it is also necessary to first extract the keywords of the historical reply message to be compared, and then the extracted The keywords are compared with the keywords of the current reply message.
  • Step S1066 Count the number of historical reply messages matching the current reply message.
  • Step S108 If the number of historical reply messages meets a preset condition, a question-and-answer pair is generated according to the current reply message and the current user question.
  • the preset conditions may include a quantity threshold and a judgment condition. For example, when the number of historical reply messages is greater than the number threshold, the preset condition is met; or, when the number of historical reply messages is not less than the number threshold, the preset condition is met.
  • a relatively high number threshold can be set, for example, when the number of historical reply messages matching the current reply message is not less than 10 Next, generate question and answer pairs.
  • the number of question and answer pairs in the knowledge base is relatively small.
  • the requirement for the establishment of question and answer pairs is relatively low.
  • a relatively low number threshold can be set, such as the number of historical reply messages that match the current reply message. If there is no less than 1 item, a question-and-answer pair is generated.
  • this embodiment does not generate all user questions and reply messages to generate question and answer pairs, but uses preset conditions to set the threshold for generating question and answer pairs, and achieve The question-and-answer pairs of this generation threshold generally have relatively good reproducibility and generalizability.
  • the established question and answer pair may include the corresponding relationship between the keywords of the user’s question and the current reply message.
  • the specific method is as follows: First, the keyword in the user’s question is determined according to the user’s question, and the specific method of determining the keyword can be referred to Related technologies are not limited here.
  • step S104 if the user's question is "when will it be shipped", it can be determined that the keyword of the user's question is "delivery time".
  • the generated question and answer pairs can be as shown in Table 3.
  • Step S110 Store the question and answer pair in the knowledge base.
  • the customer service system may directly store the established question and answer pairs in the knowledge base for subsequent use in replying to messages.
  • the customer service system stores the question and answer pairs approved by the manual customer service in the knowledge base.
  • the client displays the question and answer pair through a floating window, and provides options such as pass and reject; then, the client obtains the manual customer service
  • the audit result is sent to the server; then, the server judges the audit result received from the client, and if the audit result indicates that the question and answer pair passed the audit, then the question and answer pair Add to the knowledge base; if the review result indicates that the question and answer pair has not passed the review, the question and answer pair will be deleted.
  • the customer service system can preset two knowledge bases, one is a personal knowledge base for each manual customer service, which can be used by the corresponding manual customer service personnel, and the other is a public knowledge base shared by the manual customer service All manual customer service use.
  • the customer service system can first display the Q&A pair to the human customer service who sent the current reply message in the Q&A pair, and the manual customer service will review the Q&A pair. If the review is passed, the Q&A pair will be stored To the personal knowledge base of the human customer service.
  • the Q&A pair can be displayed to multiple manual customer service personnel in a specific review team, and the review results of the multiple manual customer service personnel can be integrated. If the review is passed, the Q&A pair will be stored in public knowledge In the library.
  • the customer service system may store the question and answer pairs that have not been reviewed by the manual customer service in the knowledge base.
  • the customer service system can preset a timeout period, such as 3 days. If the manual customer service does not review the question and answer pair within the timeout period, the customer service system automatically adds the question and answer pair to the knowledge base.
  • the steps in Figure 1 can be applied to the server side of the customer service system.
  • the server side can receive the reply message monitored by the client, then perform processing such as matching the reply message, and finally generate and save the question and answer pair.
  • the steps in Figure 1 can also be applied to the client.
  • the client monitors the current reply message, and then determines the number of historical reply messages, and processes such as the generation and storage of question and answer pairs.
  • each step in FIG. 1 can also be performed by the client and the server respectively, which will not be repeated here.
  • a new question and answer pair can be generated.
  • the method in this embodiment can efficiently screen out reply messages with a high frequency of use, and has less human involvement, saving labor costs.
  • the method in this embodiment has good timeliness and consumes energy on the server. Also less.
  • this embodiment can greatly improve the efficiency of the customer service system in replying to messages.
  • the customer service system can extract the keywords in the user's question, and then search the knowledge base for question and answer pairs that match the keyword. If the query is successful, the reply message in the question and answer pair can be sent to the user to increase the automatic reply rate and reduce the manual participation rate; or the reply message in the question and answer pair can be displayed in the conversation window for reference by manual customer service.
  • the customer service can directly refer to the question and answer pair to reply, instead of manually inputting or consulting data, etc., greatly improving the reply speed.
  • this embodiment may also delete the question and answer pairs in the knowledge base.
  • question and answer pairs that are used less frequently can be deleted from the knowledge base.
  • the customer service system can preset a timeout period, such as 30 days. For each question and answer pair in the knowledge base, if a user question that matches the question and answer pair is not received within the timeout period, it will be deleted from the knowledge base. Q&A right.
  • this specification also provides an embodiment of a device for establishing a question and answer pair.
  • the embodiment of the device for establishing question-and-answer pairs in this specification can be applied to the server.
  • the device embodiments can be implemented by software, or can be implemented by hardware or a combination of software and hardware. Taking software implementation as an example, as a logical device, it is formed by reading the corresponding computer program instructions in the non-volatile memory into the memory through the processor of the server where it is located. From a hardware perspective, as shown in FIG. 3, an exemplary embodiment of this specification shows a structural schematic diagram of a device for establishing question and answer pairs, which is used to indicate a hardware of the server where the device for establishing question and answer pairs is located in this specification.
  • the server where the device is located in the embodiment usually includes other hardware according to the actual function of the server. Go into details.
  • Fig. 4 is a block diagram of a device for establishing question and answer pairs according to an exemplary embodiment of the present specification.
  • the device 400 for establishing question and answer pairs can be applied to the server shown in FIG.
  • the text corresponding to the pasting operation is used as the current reply message of the current user question;
  • the counting unit 406 is used to determine the number of historical reply messages matching the current reply message in the current period, and the historical reply message is input by the manual customer service Judging unit 408, if the number of historical reply messages meets a preset condition, generate question and answer pairs according to the current reply message and the current user question; storage unit 410, used to store the question and answer pairs in In the knowledge base.
  • the counting unit 406 is configured to: obtain one or more keywords in the current reply message; determine whether the current reply message matches each historical reply message in the current period according to the keywords; The number of historical reply messages matching the current reply message.
  • the generating a question-and-answer pair according to the current reply message and the current user question includes: obtaining one or more keywords in the current user question; saving the keyword and the current reply message The corresponding relationship of is used as the generated question and answer pair.
  • the storage unit 410 displays the question and answer pair to a human customer service through a visual interface; receives the review result sent by the manual customer service for the question and answer pair; and indicates that the question and answer pair has passed the review in the review result.
  • the question and answer pair is stored in the knowledge base; if the review result sent by the manual customer service for the question and answer pair is not received within the first preset time, the question and answer pair is stored in the knowledge base.
  • the device further includes: a first assigning unit 412, configured to: when a user question is received, query whether there is a question and answer pair matching the user question in the knowledge base; In the case of a matching question and answer pair, the reply message in the matching question and answer pair is returned as the reply message of the user question; in the case that there is no question and answer pair matching the user question, the user question is assigned Manual customer service.
  • a first assigning unit 412 configured to: when a user question is received, query whether there is a question and answer pair matching the user question in the knowledge base; In the case of a matching question and answer pair, the reply message in the matching question and answer pair is returned as the reply message of the user question; in the case that there is no question and answer pair matching the user question, the user question is assigned Manual customer service.
  • the device further includes: a second assigning unit 414, configured to: when a user problem is received, assign a manual customer service to the user problem, and query whether there is a problem in the knowledge base that is related to the user problem. Matched question and answer pair; in the case that there is a question and answer pair that matches the user's question, display the reply message of the matched question and answer pair to the human customer service.
  • a second assigning unit 414 configured to: when a user problem is received, assign a manual customer service to the user problem, and query whether there is a problem in the knowledge base that is related to the user problem. Matched question and answer pair; in the case that there is a question and answer pair that matches the user's question, display the reply message of the matched question and answer pair to the human customer service.
  • the device further includes: a deleting unit 416, configured to: for each question and answer pair in the knowledge base, in the case that a user question matching the question and answer pair is not received within a second preset time , Delete the question and answer pair from the knowledge base.
  • a deleting unit 416 configured to: for each question and answer pair in the knowledge base, in the case that a user question matching the question and answer pair is not received within a second preset time , Delete the question and answer pair from the knowledge base.
  • the device embodiment since it basically corresponds to the method embodiment, please refer to the part of the description of the method embodiment for related parts.
  • the device embodiments described above are merely illustrative, where the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in One place, or it can be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution in this specification. Those of ordinary skill in the art can understand and implement without creative work.
  • a typical implementation device is a computer.
  • the specific form of the computer can be a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email receiving and sending device, and a game control A console, a tablet computer, a wearable device, or a combination of any of these devices.
  • this specification also provides a device for establishing question and answer pairs.
  • the device includes a processor and a memory for storing machine executable instructions.
  • the processor and the memory are usually connected to each other via an internal bus.
  • the device may also include an external interface to be able to communicate with other devices or components.
  • the processor is prompted to: monitor the paste operation of the human customer service; and correspond to the paste operation
  • the text of is used as the current reply message of the current user question; the number of historical reply messages matching the current reply message in the current period is determined, and the historical reply message is input by the manual customer service; if the number of historical reply messages meets With a preset condition, a question-and-answer pair is generated according to the current reply message and the current user question; the question-and-answer pair is stored in the knowledge base.
  • the determining the number of historical reply messages matching the current reply message in the current period includes: acquiring one or more keywords in the current reply message; judging the current reply message according to the keywords Whether the reply message matches each historical reply message in the current period; count the number of historical reply messages matching the current reply message.
  • the generating a question-and-answer pair according to the current reply message and the current user question includes: obtaining one or more keywords in the current user question; saving the keyword and the current reply message The corresponding relationship of is used as the generated question and answer pair.
  • the storing the question and answer pair in the knowledge base includes: displaying the question and answer pair to a human customer service through a visual interface; receiving the review result sent by the manual customer service regarding the question and answer pair; if the review is The result indicates that the question and answer pair is approved, and the question and answer pair is stored in the knowledge base.
  • the storing the question-and-answer pair in the knowledge base further includes: if the review result sent by the manual customer service for the question-and-answer pair is not received within the first preset time, storing the question-and-answer pair The pair is stored in the knowledge base.
  • the following steps are also implemented: when a user question is received, query whether there is a question and answer pair that matches the user question in the knowledge base; if it exists, the reply message in the matched question and answer pair is sent Return as a reply message to the user question; if it does not exist, assign a human customer service to the user question.
  • the following steps are also implemented: when a user question is received, assign a human customer service to the user question, and query whether there is a question-and-answer pair matching the user question in the knowledge base;
  • the manual customer service displays the reply message in the matching question and answer pair.
  • the following step is further implemented: for each question and answer pair in the knowledge base, if a user question that matches the question and answer pair is not received within a second preset time, delete the question and answer pair from the knowledge base Q&A right.
  • this specification also provides a computer-readable storage medium on which a computer program is stored, and when the program is executed by a processor, the following steps are implemented: Paste operation of manual customer service; use the text corresponding to the paste operation as the current reply message of the current user question; determine the number of historical reply messages matching the current reply message in the current period, and the historical reply message is determined by the manual Customer service input; if the number of historical reply messages meets a preset condition, a question-and-answer pair is generated according to the current reply message and the current user question; the question-and-answer pair is stored in the knowledge base.
  • the determining the number of historical reply messages matching the current reply message in the current period includes: acquiring one or more keywords in the current reply message; judging the current reply message according to the keywords Whether the reply message matches each historical reply message in the current period; count the number of historical reply messages matching the current reply message.
  • the generating a question-and-answer pair according to the current reply message and the current user question includes: obtaining one or more keywords in the current user question; saving the keyword and the current reply message The corresponding relationship of is used as the generated question and answer pair.
  • the storing the question and answer pair in the knowledge base includes: displaying the question and answer pair to a human customer service through a visual interface; receiving the review result sent by the manual customer service regarding the question and answer pair; if the review is The result indicates that the question and answer pair is approved, and the question and answer pair is stored in the knowledge base.
  • the storing the question-and-answer pair in the knowledge base further includes: if the review result sent by the manual customer service for the question-and-answer pair is not received within the first preset time, storing the question-and-answer pair The pair is stored in the knowledge base.
  • the following steps are further implemented: when a user question is received, query whether there is a question and answer pair that matches the user question in the knowledge base; if there is, the reply message in the matched question and answer pair is sent Return as a reply message to the user question; if it does not exist, assign a human customer service to the user question.
  • the following steps are also implemented: when a user question is received, assign a human customer service to the user question, and query whether there is a question-and-answer pair matching the user question in the knowledge base;
  • the manual customer service displays the reply message in the matching question and answer pair.
  • the following step is further implemented: for each question and answer pair in the knowledge base, if a user question that matches the question and answer pair is not received within a second preset time, delete the question and answer pair from the knowledge base Q&A right.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • Databases & Information Systems (AREA)
  • Software Systems (AREA)
  • Computational Linguistics (AREA)
  • Human Computer Interaction (AREA)
  • Evolutionary Computation (AREA)
  • Computing Systems (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

Procédé et dispositif d'établissement d'une paire de questions-réponses. Le procédé comprend : la surveillance d'une opération de collage d'un service de client manuel (S102) ; la prise d'un texte correspondant à l'opération de collage en tant que message de réponse actuel de la question d'utilisateur actuelle (S104) ; la détermination du nombre de messages de réponse historiques correspondant au message de réponse actuel à l'intérieur de la période actuelle, les messages de réponse historiques étant entrés par le service de client manuel (S106) ; si le nombre de messages de réponse historiques satisfait une condition prédéfinie, la génération d'une paire de questions-réponses en fonction du message de réponse actuel et de la question d'utilisateur actuelle (S108) ; et le stockage de la paire de questions-réponses dans une base de connaissances (S110).
PCT/CN2021/097937 2020-06-05 2021-06-02 Procédé et dispositif d'établissement d'une paire de questions-réponses WO2021244574A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010504143.1 2020-06-05
CN202010504143.1A CN111400482B (zh) 2020-06-05 2020-06-05 一种建立问答对的方法及装置

Publications (1)

Publication Number Publication Date
WO2021244574A1 true WO2021244574A1 (fr) 2021-12-09

Family

ID=71429940

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/097937 WO2021244574A1 (fr) 2020-06-05 2021-06-02 Procédé et dispositif d'établissement d'une paire de questions-réponses

Country Status (2)

Country Link
CN (2) CN112131338B (fr)
WO (1) WO2021244574A1 (fr)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114490975A (zh) * 2021-12-31 2022-05-13 马上消费金融股份有限公司 用户问题的标注方法及装置
CN115907785A (zh) * 2022-11-22 2023-04-04 深圳市米糠云科技有限公司 一种基于关联关键词的客服问答智能辅助方法及装置

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112131338B (zh) * 2020-06-05 2024-02-09 支付宝(杭州)信息技术有限公司 一种建立问答对的方法及装置
CN112632232B (zh) * 2021-03-09 2022-03-15 北京世纪好未来教育科技有限公司 一种文本匹配方法、装置、设备及介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150088998A1 (en) * 2013-09-26 2015-03-26 International Business Machines Corporation Automatic Question Generation and Answering Based on Monitored Messaging Sessions
CN109308240A (zh) * 2017-07-26 2019-02-05 腾讯科技（深圳）有限公司 操作处理方法及装置、存储介质
CN110162603A (zh) * 2018-11-30 2019-08-23 腾讯科技（深圳）有限公司 一种智能对话方法、动态存储方法和装置
CN110445707A (zh) * 2019-07-23 2019-11-12 精硕科技（北京）股份有限公司 一种消息处理方法及装置
CN111400482A (zh) * 2020-06-05 2020-07-10 支付宝(杭州)信息技术有限公司 一种建立问答对的方法及装置

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105335400B (zh) * 2014-07-22 2018-11-23 阿里巴巴集团控股有限公司 针对用户的提问意图获取答案信息的方法及装置
CN105786851A (zh) * 2014-12-23 2016-07-20 北京奇虎科技有限公司 问答知识库的构建方法、提供搜索的方法及装置
US9858335B2 (en) * 2015-04-21 2018-01-02 International Business Machines Corporation Providing searching strategy in connection with answering question in message
CN108763494B (zh) * 2018-05-30 2020-02-21 苏州思必驰信息科技有限公司 对话系统间的知识分享方法、对话方法及装置
CN109408624B (zh) * 2018-11-06 2020-11-27 江西师范大学 一种视觉聊天机器人会话生成方法
CN110457453A (zh) * 2019-07-12 2019-11-15 平安普惠企业管理有限公司 用户问题与客服对接方法、装置、介质、电子设备

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150088998A1 (en) * 2013-09-26 2015-03-26 International Business Machines Corporation Automatic Question Generation and Answering Based on Monitored Messaging Sessions
CN109308240A (zh) * 2017-07-26 2019-02-05 腾讯科技（深圳）有限公司 操作处理方法及装置、存储介质
CN110162603A (zh) * 2018-11-30 2019-08-23 腾讯科技（深圳）有限公司 一种智能对话方法、动态存储方法和装置
CN110445707A (zh) * 2019-07-23 2019-11-12 精硕科技（北京）股份有限公司 一种消息处理方法及装置
CN111400482A (zh) * 2020-06-05 2020-07-10 支付宝(杭州)信息技术有限公司 一种建立问答对的方法及装置

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114490975A (zh) * 2021-12-31 2022-05-13 马上消费金融股份有限公司 用户问题的标注方法及装置
CN115907785A (zh) * 2022-11-22 2023-04-04 深圳市米糠云科技有限公司 一种基于关联关键词的客服问答智能辅助方法及装置
CN115907785B (zh) * 2022-11-22 2023-09-19 深圳市米糠云科技有限公司 一种基于关联关键词的客服问答智能辅助方法及装置

Also Published As

Publication number Publication date
CN111400482A (zh) 2020-07-10
CN112131338B (zh) 2024-02-09
CN111400482B (zh) 2020-09-01
CN112131338A (zh) 2020-12-25

Similar Documents

Publication Publication Date Title
WO2021244574A1 (fr) Procédé et dispositif d'établissement d'une paire de questions-réponses
US11455599B2 (en) Systems and methods for improved meeting engagement
TWI615792B (zh) 會話內容合倂方法和系統
US8296372B2 (en) Method and system for merging electronic messages
CN106095842B (zh) 在线课程搜索方法和装置
CN111353693A (zh) 一种工单派发、处理、分析系统及其方法
CN114124861A (zh) 消息群发方法、装置、计算机设备和存储介质
US20190042950A1 (en) Learning computing activities and relationships using graphs
US20130339082A1 (en) Contextual information retrieval for groupware integration
US20220318319A1 (en) Focus Events
US11593741B2 (en) Personal data fusion
US20090083221A1 (en) System and Method for Estimating and Storing Skills for Reuse
CN111444287A (zh) 图数据库构建方法、关联信息查询方法、装置及计算设备
CN113127307A (zh) 溯源请求的处理方法、相关装置、系统及存储介质
WO2023222109A1 (fr) Procédé et appareil de gestion d'activation de réseau, dispositif électronique et support de stockage
US20200186371A1 (en) Apparatus and method
US11429877B2 (en) Unified logging of actions for labeling
JP2016021178A (ja) 関連トピック表示制御装置、関連トピック表示制御方法、及びプログラム
US11650998B2 (en) Determining authoritative documents based on implicit interlinking and communication signals
US20190042951A1 (en) Analysis of computing activities using graph data structures
WO2017024966A1 (fr) Procédé et dispositif de classification de table
CN115062676B (zh) 数据处理方法、装置及计算机可读存储介质
CN112818223B (zh) 用户画像的查询处理方法、装置、设备、程序产品及介质
US20200026923A1 (en) System for automatically detecting new home construction progress
EP1515255A2 (fr) Segmentation des participants pour la gestion de communications

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21818412

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21818412

Country of ref document: EP

Kind code of ref document: A1