CN111475632A

CN111475632A - Question processing method and device, electronic equipment and storage medium

Info

Publication number: CN111475632A
Application number: CN202010274365.9A
Authority: CN
Inventors: 黄玲
Original assignee: Shenzhen Zhuiyi Technology Co Ltd
Current assignee: Shenzhen Zhuiyi Technology Co Ltd
Priority date: 2020-04-09
Filing date: 2020-04-09
Publication date: 2020-07-31

Abstract

The application discloses a question processing method, a question processing device, electronic equipment and a storage medium, which relate to the technical field of Internet, and the method comprises the following steps: acquiring an input current question and parameter information of a user corresponding to the current question; acquiring a standard question database corresponding to the parameter information, wherein the standard question database comprises a plurality of standard questions; respectively comparing the similarity of the current question with a plurality of standard questions in the standard question database to obtain the similarity between the current question and each standard question; acquiring a standard question with the highest similarity between the current question and the plurality of standard questions as a target standard question; and outputting the target standard question sentence. According to the method and the device, the intention of the user can be accurately identified, the smooth completion of the conversation task can be guaranteed, the associated question sentence can be accurately and efficiently recommended to the user, and the user experience is improved.

Description

Question processing method and device, electronic equipment and storage medium

Technical Field

The present application relates to the field of internet technologies, and in particular, to a question processing method and apparatus, an electronic device, and a storage medium.

Background

With the rapid development of science and technology, man-machine interaction technology has penetrated aspects of daily life, for example, various customer service robots have been popularized in daily life, and the customer service robots can provide services such as question answering, information inquiry, flight booking and the like for users through conversation with the users.

Generally, when a customer service robot has a conversation with a user, a question input by the user needs to be received to identify the intention of the user, so that a conversation task is completed. However, the question input by the user may be fuzzy in semantics, and the customer service robot cannot process the question well, so that the user experience is reduced.

Disclosure of Invention

In view of the above problems, the present application provides a question processing method, device, electronic device and storage medium.

In a first aspect, an embodiment of the present application provides a question processing method, including: acquiring an input current question and parameter information of a user corresponding to the current question; acquiring a standard question database corresponding to the parameter information, wherein the standard question database comprises a plurality of standard questions; respectively comparing the similarity of the current question with a plurality of standard questions in a standard question database to obtain the similarity between the current question and each standard question; acquiring a standard question with the highest similarity between the current question and the plurality of standard questions as a target standard question; and outputting the target standard question sentence.

Further, outputting a target standard question, comprising: acquiring a similarity threshold corresponding to the parameter information; and if the similarity between the target standard question and the current question is not less than the similarity threshold, outputting the target standard question.

Further, after outputting the target standard question if the similarity between the target standard question and the current question is not less than the similarity threshold, the method further includes: acquiring a target associated question corresponding to the target standard question; and outputting the target association question.

Further, acquiring a target associated question corresponding to the target standard question includes: performing semantic recognition on the target standard question to obtain a semantic recognition result; acquiring a dialogue scene corresponding to a semantic recognition result from a dialogue scene database, wherein the dialogue scene comprises a plurality of first associated question sentences; and acquiring a first associated question corresponding to the target standard question from the plurality of first associated questions as a target associated question.

Further, acquiring a first related question corresponding to the target standard question from the plurality of first related questions as a target related question, including: and if the plurality of first associated questions comprise the target standard question, taking the first associated questions except the target standard question in the plurality of first associated questions as the target associated questions.

Further, the dialog scenario includes a plurality of first associated questions sorted according to a preset question order, and the method for acquiring a first associated question corresponding to the target standard question from the plurality of first associated questions as a target associated question includes: if the plurality of first associated question sentences comprise the target standard question sentence, acquiring the sequence numbers of the target standard question sentence in the plurality of first associated question sentences as first target sequence numbers; and acquiring a first associated sentence with the sequence number arranged behind the first target sequence number in the plurality of first associated question sentences as a target associated question sentence.

Further, acquiring a target associated question corresponding to the target standard question includes: obtaining customer service log data, wherein the customer service log data comprises a plurality of question and answer paragraphs; acquiring a question-answer paragraph comprising a target standard question from the question-answer paragraphs as a target question-answer paragraph, wherein the target question-answer paragraph comprises a plurality of second associated question, and the target standard question is one of the second associated question; and acquiring second associated question sentences except the target standard question sentences in the plurality of second associated question sentences as target associated question sentences.

Further, the question-answer paragraph includes a plurality of second related question sentences ordered according to a specified question order, and obtains, as target related question sentences, second related question sentences other than the target standard question sentences from the plurality of second related question sentences, including: acquiring the sequence numbers of the target standard question in the plurality of second associated question as second target sequence numbers; and acquiring a second associated sentence with the sequence number arranged behind the second target sequence number in the plurality of second associated question sentences as a target associated question sentence.

Further, the method for outputting the target related question includes the following steps: acquiring the number and types of the associated question sentences corresponding to the parameter information; and outputting the target associated question corresponding to the number of the associated questions or/and the type of the associated questions.

Further, the method for outputting the target associated question includes the following steps: receiving an input associated question display instruction; and responding to the associated question display instruction, and outputting the target associated question with the type and the number corresponding to the associated question display instruction.

Further, the number of the target related question sentences is plural, and after the target related question sentences are output, the method further includes: receiving an input selected instruction; responding to the selected instruction, and acquiring a target associated question corresponding to the selected instruction from the target associated question as an associated question to be answered; and acquiring reply information corresponding to the associated question to be replied and outputting the reply information.

Further, the parameter information comprises one or more combinations of channel labels, role labels and grade labels.

In a second aspect, an embodiment of the present application provides a question processing apparatus, including: the system comprises an information acquisition module, a standard question database acquisition module, a similarity comparison module, a target standard question determination module and an output module. The information acquisition module is used for acquiring an input current question and user parameter information corresponding to the current question; the standard question database acquisition module is used for acquiring a standard question database corresponding to the parameter information, and the standard question database comprises a plurality of standard questions; the similarity comparison module is used for respectively comparing the similarity of the current question with a plurality of standard questions in the standard question database to obtain the similarity between the current question and each standard question; the target standard question determining module is used for acquiring a standard question with the highest similarity between the current question and the plurality of standard questions as a target standard question; the output module is used for outputting the target standard question sentence.

In a third aspect, an embodiment of the present application provides an electronic device, which includes: memory, one or more processors, and one or more applications. Wherein the one or more processors are coupled with the memory. One or more application programs are stored in the memory and configured to be executed by the one or more processors, the one or more application programs configured to perform the method of the first aspect as described above.

In a fourth aspect, embodiments of the present application provide a computer-readable storage medium, in which program code is stored, and the program code can be called by a processor to execute the method according to the first aspect.

According to the question processing method, the device, the electronic equipment and the storage medium provided by the embodiment of the application, the input current question and the parameter information of the user corresponding to the current question are obtained, and the standard question database corresponding to the parameter information is obtained, wherein the standard question database comprises a plurality of standard questions, so that different standard question databases can be used according to different parameter information of the user, and the standard question selected from the standard question databases can have higher relevance with the user. And then the similarity comparison is carried out on the current question and a plurality of standard questions in a standard question database respectively to obtain the similarity between the current question and each standard question, the standard question with the highest similarity between the current question and the plurality of standard questions is obtained to be used as a target standard question, and the target standard question is output, so that the target standard question most similar to the question input by the user can be accurately hit according to the similarity, the target standard question can be output to replace the question input by the user to complete the dialogue with the user, the processing capacity of the question input by the user is improved, and the user experience is improved.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.

Fig. 1 shows a schematic diagram of an application environment suitable for the embodiment of the present application.

Fig. 2 is a schematic flow chart illustrating a question processing method according to a first embodiment of the present application.

Fig. 3 is a flowchart illustrating a question processing method according to a second embodiment of the present application.

Fig. 4 is a flowchart illustrating a question processing method according to a third embodiment of the present application.

Fig. 5 shows a schematic diagram of a dialog interaction interface provided in the third embodiment of the present application.

Fig. 6 is a flowchart illustrating a question processing method according to a fourth embodiment of the present application.

Fig. 7 is a flowchart illustrating a question processing method according to a fifth embodiment of the present application.

Fig. 8 is a flowchart illustrating a question processing method according to a sixth embodiment of the present application.

Fig. 9 is a flowchart illustrating a question processing method according to a seventh embodiment of the present application.

Fig. 10 is a flowchart illustrating a question processing method according to an eighth embodiment of the present application.

Fig. 11 is a block diagram showing a question processing apparatus according to a ninth embodiment of the present application.

Fig. 12 is a block diagram of an electronic device for executing a question processing method according to a tenth embodiment of the present application.

Fig. 13 is a storage unit according to an eleventh embodiment of the present application for storing or carrying program code for implementing a question processing method according to an embodiment of the present application.

Detailed Description

The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.

With the development of science and technology, artificial intelligence technology is more and more popular, and services such as appointments, consultations and the like in daily life are changed from artificial services to machine services. Compared with manual service, the service efficiency is greatly improved through machine service, and great convenience is brought to daily life of people.

Generally, the customer service robot provides services for the user by means of a dialogue with the user. In the conversation process, the service robot is inevitably involved in identifying the question input by the user, so whether the question identification result is accurate or not directly determines whether the service robot can process the question input by the user or not. For example, the semantics of a question input by a user is not clear, so that the customer service robot cannot determine the intention of the user, the customer service robot cannot continue to perform a conversation task, the service cannot be provided for the user, and the user experience is reduced.

The inventor finds that if a plurality of standard question sentences are recorded in advance to form a standard question sentence database, then the standard question sentence with the highest similarity to the question sentence input by the user is found out from the standard question sentence database according to the question sentence input by the user to be used as the target standard question sentence, and the target standard question sentence is used for replacing the question sentence input by the user to carry out the conversation task, the intention of the user can be accurately identified, and meanwhile, the smooth completion of the conversation task can be ensured.

The inventor finds in practical research that if some associated questions related to the target standard question are output to the user at the same time when the customer service robot outputs the target standard question, the user can have more choices, so that the user can select the question that the user wants to perform the next conversation, and the diversity and flexibility of the conversation can be effectively improved. However, at present, the adding manner of the associated questions is realized based on manual adding of business personnel, and for the user served by the robot customer service, many questions may be asked in the actual question answering, and business personnel are required to add associated questions associated with the business personnel as much as possible, so that the workload of the business personnel is increased, and the adding efficiency and accuracy cannot be guaranteed.

In order to improve the above problem, the inventors propose a question processing method, device, electronic device, and storage medium in the embodiments of the present application. The method and the device can accurately identify the user intention, ensure that the conversation task can be completed smoothly, accurately and efficiently recommend the associated question to the user, and improve the user experience.

The question processing method, device, electronic device and storage medium provided by the embodiments of the present application will be described in detail by specific embodiments.

Referring to fig. 1, fig. 1 is a schematic diagram illustrating an application environment suitable for the embodiment of the present application. The question processing method provided by the embodiment of the application can be applied to the interactive system 100 shown in fig. 1. The interactive system 100 comprises a terminal device 101 and a server 102, wherein the server 102 is in communication connection with the terminal device 101. The server 102 may be a conventional server or a cloud server, and is not limited herein.

The terminal device 101 may be various electronic devices that have a display screen, a data processing module, a camera, an audio input/output function, and the like, and support data input, including but not limited to a smart phone, a tablet computer, a laptop portable computer, a desktop computer, a self-service terminal, a wearable electronic device, and the like. Specifically, the data input may be inputting voice based on a voice module provided on the electronic device, inputting characters based on a character input module, and the like.

The terminal device 101 may have a client application installed thereon, and the user may be based on the client application (for example, APP, wechat applet, etc.), where the conversation robot in this embodiment is also a client application configured in the terminal device 101. A user may register a user account in the server 102 based on the client application program, and communicate with the server 102 based on the user account, for example, the user logs in the user account in the client application program, inputs information through the client application program based on the user account, and may input text information or voice information, and the like, after receiving information input by the user, the client application program may send the information to the server 102, so that the server 102 may receive the information, process and store the information, and the server 102 may also receive the information and return a corresponding output information to the terminal device 101 according to the information.

In some embodiments, the apparatus for processing the data to be recognized may also be disposed on the terminal device 101, so that the terminal device 101 can interact with the user without relying on the server 102 to establish communication, and in this case, the interactive system 100 may only include the terminal device 101.

First embodiment

Referring to fig. 2, fig. 2 is a schematic flow chart illustrating a question processing method according to an embodiment of the present application. As shown in fig. 2, the method may include:

s101, acquiring the input current question and the parameter information of the user corresponding to the current question.

Among them, the terminal device 101 (hereinafter, may be referred to as an electronic device) configured with the conversation robot can acquire the current question sentence input by the user. Optionally, the user may input the current question through a key input module configured on the electronic device, may also input the current question through a voice input module configured on the electronic device, and may also upload the current question through a mobile terminal in communication with the electronic device. Meanwhile, the electronic device may further obtain parameter information of a user corresponding to the current question, that is, parameter information of the user who inputs the current question, where the parameter information may be a tag of a channel through which the user inputs the current information. Alternatively, the channel may be various types of network platforms, such as Taobao, Baidu, etc. Alternatively, the parameter information may be membership grade tags of the user at a certain platform, such as VIP3, VIP4, and the like.

S102, a standard question database corresponding to the parameter information is obtained, and the standard question database comprises a plurality of standard questions.

After the electronic device acquires the parameter information, the electronic device may acquire a standard question database corresponding to the parameter information according to the parameter information. Optionally, different parameter information may establish a corresponding relationship with different standard question databases in advance, and generate a standard question database corresponding relationship table, so that a standard question database corresponding to the parameter information may be found by querying the standard question database corresponding relationship table, where one parameter information may correspond to one or more standard question databases. For example, when the parameter information is a channel tag, specifically, when the channel tag is a shopping platform such as panning and kyoto, the parameter information may respectively establish a correspondence relationship with a database of standard questions for commodity reservation and a database of standard questions for commodity query, and generate a standard question database correspondence table.

Alternatively, the standard query database may be stored locally in the electronic device, or may be stored in a cloud server connected to the electronic device through a network, and may be called by the electronic device.

The standard question may have a standard format and standard content, and the semantic meaning of the standard question is clear and can be accurately recognized by a conversation robot in the electronic device.

S103, respectively comparing the similarity of the current question with a plurality of standard questions in a standard question database to obtain the similarity between the current question and each standard question.

Optionally, when the electronic device performs semantic similarity scoring, the electronic device may perform scoring by using various similarity algorithms and Natural language Processing (N L P) techniques.

And S104, acquiring the standard question with the highest similarity between the current question and the plurality of standard questions as a target standard question.

As an example, after obtaining the similarity between the current question and each standard question, the electronic device may sort each standard question in order of high similarity to low similarity, for example. For example, the standard question includes a first standard question, a second standard question, a third standard question and a fourth standard question, wherein the similarity between the first standard question and the current question, the similarity between the second standard question and the current question and the similarity between the third standard question and the current question are 90, 92, 85 and 75. Therefore, the standard question is sorted into the second standard question, the first standard question, the third standard question and the fourth standard question. The second standard question may be selected as the target standard question.

And S105, outputting a target standard question.

In some embodiments, the dialog robot in the electronic device may output and display the target standard question, or may output the target standard question to the reply machine learning model, and obtain and display reply information corresponding to the target standard question output by the reply machine model. The answer machine model can be obtained by training in advance based on a plurality of sample standard question sentences and a plurality of sample answer information.

In this embodiment, by acquiring the input current question and the parameter information of the user corresponding to the current question, and acquiring the standard question database corresponding to the parameter information, where the standard question database includes a plurality of standard questions, different standard question databases can be used for different parameter information of users, so that the standard question selected from the standard question database can have a higher association with the user. And then the similarity comparison is carried out on the current question and a plurality of standard questions in a standard question database respectively to obtain the similarity between the current question and each standard question, the standard question with the highest similarity between the current question and the plurality of standard questions is obtained to be used as a target standard question, and the target standard question is output, so that the target standard question most similar to the question input by the user can be accurately hit according to the similarity, the target standard question can be output to replace the question input by the user to complete the dialogue with the user, the processing capacity of the question input by the user is improved, and the user experience is improved.

Second embodiment

Referring to fig. 3, fig. 3 is a schematic flow chart illustrating a question processing method according to an embodiment of the present application. The method may comprise the steps of:

s201, acquiring the input current question and the parameter information of the user corresponding to the current question.

Optionally, the parameter information includes one or more combinations of channel tags, role tags, and level tags.

S202, a standard question database corresponding to the parameter information is obtained, and the standard question database comprises a plurality of standard questions.

And S203, respectively comparing the similarity of the current question with a plurality of standard questions in the standard question database to obtain the similarity between the current question and each standard question.

And S204, acquiring the standard question with the highest similarity between the current question and the multiple standard questions as a target standard question.

The specific implementation of S201 to S204 can refer to S101 to S104, and therefore is not described herein.

And S205, acquiring a similarity threshold corresponding to the parameter information.

In some embodiments, the electronic device may set a similarity threshold for each of the plurality of pieces of parameter information in advance, as an example, for example, the parameter information is a label of a channel, and the channel may include a shopping platform, a reservation platform, and a search platform, where different platforms may correspond to different similarity thresholds, for example, the similarity threshold of the shopping platform is 80, the similarity threshold of the reservation platform is 90, and the similarity threshold of the search platform is 60.

And S206, if the similarity between the target standard question and the current question is not less than the similarity threshold, outputting the target standard question.

As an example, for example, when the current channel is a reservation platform, and the electronic device detects that the similarity between the target standard question and the current question is 95, the target standard question may be output. When the similarity between the target standard question and the current question is 80, it can be stated that the target standard question matched with the current question does not exist in the standard question database at this time, the target standard question may not be output, and then the electronic device may send out a prompt message to prompt the user to input the question of the information.

In this embodiment, by obtaining the similarity threshold corresponding to the parameter information and outputting the target standard question when the similarity between the target standard question and the current question is not less than the similarity threshold, it can be ensured that the output target standard question has a higher similarity with the current question.

Third embodiment

Referring to fig. 4, fig. 4 is a schematic flow chart illustrating a question processing method according to an embodiment of the present application. The method may comprise the steps of:

s301, acquiring the input current question and the parameter information of the user corresponding to the current question.

S302, a standard question database corresponding to the parameter information is obtained, and the standard question database comprises a plurality of standard questions.

And S303, respectively comparing the similarity of the current question with the similarity of a plurality of standard questions in the standard question database to obtain the similarity between the current question and each standard question.

S304, the standard question with the highest similarity between the current question and the multiple standard questions is obtained and serves as the target standard question.

And S305, acquiring a similarity threshold corresponding to the parameter information.

And S306, if the similarity between the target standard question and the current question is not less than the similarity threshold, outputting the target standard question.

The specific implementation of S301 to S306 can refer to S201 to S206, and therefore is not described herein.

S307, acquiring a target associated question corresponding to the target standard question.

In some embodiments, after the electronic device respectively obtains the similarities (semantic similarities) between the multiple standard question sentences in the standard question sentence database and the current question sentence, the multiple standard question sentences may be sorted according to the sequence of the similarities from high to low to obtain a standard question sequence. And taking the first standard question in the standard question sequence as a target standard question, and taking a specified number of standard questions behind the target standard question in the standard question sequence as target associated questions corresponding to the target standard question, wherein the associated questions can be called as semantic-based associated questions.

As an example, the similarity between the first standard question and, the second standard question, the third standard question, and the fourth standard question and the current question is 90, 92, 85, 75, for example. Therefore, the standard question is sorted into the second standard question, the first standard question, the third standard question and the fourth standard question. The second standard question may be selected as the target standard question. If the specified number is 2, the first standard question and the third standard question may be selected as the target related question. Thereby providing a target associated question corresponding to the target standard question according to the semantic similarity. Alternatively, when the number of standard question sentences exceeds 10, the 10 standard question sentences before the standard question sentence sequence may be used as the target associated question sentences.

In other embodiments, a dialog scenario of the target standard question may also be obtained, and a target associated question may be given according to the dialog scenario of the target standard question, and such an associated question may be referred to as a scenario-based associated question. .

And S308, outputting the target association question.

The electronic device may display the target related question on an interface displaying the target standard question or an interface displaying the reply information corresponding to the target standard question. As an example, as shown in fig. 5, the target related question may be displayed under the reply information corresponding to the target standard question with a font size and a color different from those of the target standard question, so that the conversation robot can supplement the related question to remind the user to complete the conversation. Specifically, when the input information of the user is "how to apply for a meeting", the conversation robot may give a target standard question "ask whether to make a meeting room reservation" and simultaneously display, below the target standard question, an associated question 1 "ask a time to reserve a meeting room" and an associated question 2 "ask a place to reserve a meeting room".

In this embodiment, by obtaining the target associated question corresponding to the target standard question and outputting the target associated question, the user can be optionally supplemented with questions according to semantic similarity, so as to improve the hit effect of the question.

In some embodiments, the types of the target related question sentences are multiple, the number of the target related question sentences is multiple, and the implementation of S308 may be: acquiring the number and types of the associated question sentences corresponding to the parameter information; and outputting the target associated question corresponding to the number of the associated questions or/and the type of the associated questions.

As an example, when the conversation robot performs real-time question answering, for example, the conversation robot may cut off the number or/and type of the target associated questions according to the associated question setting corresponding to the channel label. No matter the target associated questions are given based on semantics or scenes, the corresponding channel associated question number needs to be intercepted before answers are given, and then final robot reply content can be given based on the labels of the matched channels, the target standard question, the associated question types and the associated question number. For example, when the label of the channel is a shopping platform, the number of the associated questions corresponding to the label may be 2, and the corresponding type of the associated questions may be a type recommended for the product. Specifically, when the conversation robot detects that the user is inputting the current question using a search-class platform such as Baidu, the conversation robot may output 2 associated questions, where the 2 associated questions may be commodity information associated with the target standard question. For example, if the target standard question is "know basketball, then the 2-item association question recommended to the telephone robot may be" need basketball shoes "and" need football clothes ".

As another example, for example, the parameter information is a rating label of the user, and the conversation robot may set different numbers of related question sentences and types of related question sentences in advance for different rating labels. For example, the number of the associated question sentences corresponding to the level 1 of the user is 3, the types of the corresponding target associated question sentences include 1, and for example, the number of the associated question sentences corresponding to the level 2 of the user is 5, and the types of the corresponding target associated questions include 2. The associated question types of the present embodiment may include a semantic-based associated question and a scene-based associated question. Optionally, when the level corresponding to the level tag of the user is higher, the number of the target association questions may be larger, and the types related to the target association questions may also be larger.

In the embodiment, by acquiring the number and the type of the associated question corresponding to the parameter information and outputting the target associated question corresponding to the number or/and the type of the associated question, personalized target associated questions are accurately provided for different users, diversity of the target associated question is ensured, and user experience is improved.

In some embodiments, S309 may be further included after S308, wherein a specific embodiment of S309 may be that an input selected instruction is received. And responding to the selected instruction, and acquiring a target associated question corresponding to the selected instruction from the target associated question as an associated question to be answered. And acquiring reply information corresponding to the associated question to be replied and outputting the reply information.

As an example, after displaying one or more target related question sentences, the electronic device may detect whether a selection instruction input by a user is received in real time, for example, the user may click a certain target related question sentence on a touch screen of the electronic device to generate the selection instruction, and for example, each target related question sentence has a number corresponding to the target related question sentence, and the user may input the number through a numeric keypad of the electronic device to generate the selection instruction. And then the electronic equipment responds to the selected instruction, and acquires a target associated question corresponding to the selected instruction from the target associated question as an associated question to be answered. And finally, the electronic equipment can find the reply information corresponding to the question sentence to be replied from the reply information database in a mode of inquiring the reply information relation table and output the reply information. The reply information relation table may be generated by associating a plurality of reply-associated question sentences with a plurality of reply information in advance. Alternatively, the reply information may be output by displaying on a screen of the electronic device or by voice broadcasting.

Fourth embodiment

Referring to fig. 6, fig. 6 is a schematic flow chart illustrating a question processing method according to an embodiment of the present application. The method may comprise the steps of:

s401, acquiring the input current question and the parameter information of the user corresponding to the current question.

S402, a standard question database corresponding to the parameter information is obtained, and the standard question database comprises a plurality of standard questions.

And S403, respectively comparing the similarity of the current question with a plurality of standard questions in the standard question database to obtain the similarity between the current question and each standard question.

S404, the standard question with the highest similarity between the current question and the plurality of standard questions is obtained and serves as the target standard question.

And S405, acquiring a similarity threshold corresponding to the parameter information.

And S406, if the similarity between the target standard question and the current question is not less than the similarity threshold, outputting the target standard question.

The specific implementation of S401 to S406 can refer to S301 to S306, and therefore will not be described herein.

S407, performing semantic recognition on the target standard question to obtain a semantic recognition result.

The dialogue robot can input the target standard question into the intention recognition model, and then obtains an intention result output by the intention recognition model and corresponding to the target standard question as a semantic recognition result. The intention recognition model can be obtained by training in advance according to a plurality of sample standard question sentences and a plurality of sample intention results. Alternatively, the semantic recognition result, i.e., the intention recognition result, may include an intention result such as ticket reservation, merchandise purchase, weather inquiry, and the like.

S408, obtaining a dialogue scene corresponding to the semantic recognition result from the dialogue scene database, wherein the dialogue scene comprises a plurality of first associated question sentences.

In some embodiments, the dialog scene may be obtained from the dialog scene database through the dialog scene correspondence table and the semantic recognition result, where the dialog scene correspondence table may be obtained by establishing a correspondence relationship in advance according to a plurality of dialog scenes and a plurality of semantic recognition results.

S409, obtaining a first related question corresponding to the target standard question from the plurality of first related questions as a target related question.

Optionally, the dialog robot may optionally select one or more of the plurality of first related question sentences as the target related question sentences, or may select all of the plurality of first related question sentences as the target related question sentences.

And S410, outputting a target association question.

In this embodiment, the target associated question is obtained in the dialog scene corresponding to the target standard question, so that the target associated question can be better close to the context possibly associated with the target standard question, and the user can conveniently perform the dialog task.

Fifth embodiment

Referring to fig. 7, fig. 7 is a schematic flowchart illustrating a question processing method according to an embodiment of the present application. The method may comprise the steps of:

s501, the input current question and the parameter information of the user corresponding to the current question are obtained.

S502, a standard question database corresponding to the parameter information is obtained, and the standard question database comprises a plurality of standard questions.

And S503, respectively comparing the similarity of the current question with a plurality of standard questions in the standard question database to obtain the similarity between the current question and each standard question.

S504, the standard question with the highest similarity between the current question and the multiple standard questions is obtained and serves as the target standard question.

And S505, acquiring a similarity threshold corresponding to the parameter information.

And S506, if the similarity between the target standard question and the current question is not less than the similarity threshold, outputting the target standard question.

And S507, performing semantic recognition on the target standard question to obtain a semantic recognition result.

And S508, acquiring a dialogue scene corresponding to the semantic recognition result from the dialogue scene database, wherein the dialogue scene comprises a plurality of first associated question sentences.

The specific implementation of S501 to S508 can refer to S401 to S408, and therefore is not described herein.

S509, if the plurality of first related questions include the target standard question, regarding the first related question except the target standard question as the target related question.

And S510, outputting a target association question.

In this embodiment, when the plurality of first related questions include the target standard question, the first related question other than the target standard question among the plurality of first related questions is used as the target related question, so that the same question can be effectively prevented from being repeatedly output.

Sixth embodiment

Referring to fig. 8, fig. 8 is a schematic flow chart illustrating a question processing method according to an embodiment of the present application. The method may comprise the steps of:

s601, acquiring the input current question and the parameter information of the user corresponding to the current question.

S602, a standard question database corresponding to the parameter information is obtained, and the standard question database comprises a plurality of standard questions.

S603, the similarity of the current question is compared with the similarity of a plurality of standard questions in the standard question database respectively to obtain the similarity between the current question and each standard question.

S604, the standard question with the highest similarity between the current question and the multiple standard questions is obtained and used as the target standard question.

And S605, acquiring a similarity threshold corresponding to the parameter information.

And S606, if the similarity between the target standard question and the current question is not less than the similarity threshold, outputting the target standard question.

S607, carrying out semantic recognition on the target standard question to obtain a semantic recognition result.

S608, a dialogue scene corresponding to the semantic recognition result is obtained from the dialogue scene database, wherein the dialogue scene comprises a plurality of first associated question sentences which are ordered according to a preset question ordering.

The specific implementation of S601 to S608 can refer to S501 to S508, and therefore, is not described herein.

As an example, for example, the plurality of first associated questions included in the dialog scenario are: 1. asking whether to make a meeting room reservation. 2. Asking for the time of the booking room. 3. Asking for the location of the booking room.

And S609, if the plurality of first associated question sentences include the target standard question sentence, acquiring the sequence number of the target standard question sentence in the plurality of first associated question sentences as the first target sequence number.

As an example, for example, if the target standard question is "time to ask for reservation of a conference room", the sequence number of the target standard question may be 2, and the sequence number may be used as the first target sequence number.

S610, a first associated sentence with a sequence number arranged behind the first target sequence number in the plurality of first associated question sentences is obtained and used as a target associated question sentence.

As an example, for example, when the target standard question is "ask for a time to reserve a conference room", a "ask for a place to reserve a conference room" question with a sequence number of 3 may be used as the target associated question without proposing a question with a sequence number of 1 "ask for whether or not to reserve a conference room".

S611, outputting the target association question.

In this embodiment, when the dialog scene includes a plurality of first related questions sorted according to the preset question ordering, and the plurality of first related questions includes the target standard question, the sequence numbers of the target standard question in the plurality of first related questions are obtained as the first target sequence numbers, and the first related sentences whose sequence numbers are arranged behind the first target sequence numbers in the plurality of first related questions are obtained as the target related questions, so that unnecessary questions in the dialog scene are prevented from being re-presented, and the efficiency of the dialog is improved.

Seventh embodiment

Referring to fig. 9, fig. 9 is a schematic flow chart illustrating a question processing method according to an embodiment of the present application. The method may comprise the steps of:

s701, acquiring the input current question and the parameter information of the user corresponding to the current question.

S702, a standard question database corresponding to the parameter information is obtained, and the standard question database comprises a plurality of standard questions.

And S703, respectively comparing the similarity of the current question with a plurality of standard questions in the standard question database to obtain the similarity between the current question and each standard question.

S704, the standard question with the highest similarity between the current question and the multiple standard questions is obtained and serves as the target standard question.

S705, a similarity threshold corresponding to the parameter information is obtained.

And S706, if the similarity between the target standard question and the current question is not less than the similarity threshold, outputting the target standard question.

The specific implementation of S701 to S706 may refer to S401 to S406, and therefore is not described herein.

S707, obtaining customer service log data, wherein the customer service log data comprises a plurality of question and answer paragraphs.

The customer service log data may include historical conversation records of completed conversation tasks, wherein the conversation records may be divided into a plurality of question and answer paragraphs.

S708, a question-answer paragraph including a target standard question is obtained from the plurality of question-answer paragraphs as the target question-answer paragraph, where the target question-answer paragraph includes a plurality of second related question, and the target standard question is one of the second related question.

When the dialog robot detects that the plurality of question-answer paragraphs include the target standard question, the dialog robot may acquire the question-answer paragraphs including the target standard question from the plurality of question-answer paragraphs, and use the question-answer paragraphs as the target question-answer paragraphs, and use the question in the target question-answer paragraphs except the target standard question as the second associated question.

S709, obtain a second related question in the plurality of second related questions except the target standard question as a target related question.

As an example, for example, the target question-answering paragraph includes a second related question: 1. asking whether to make a meeting room reservation. 2. Asking for the time of the booking room. 3. Asking for the location of the booking room. If the target standard question is "asking whether to make a meeting room reservation", both "time to ask for a reservation of the meeting room" and "place to ask for a reservation of the meeting room" may be used as the target associated question.

And S710, outputting the target association question.

In the embodiment, by obtaining the customer service log data and obtaining the context information related to the target standard question from the customer service log data, the next question of the user can be predicted and prompted according to the conversation context of the historical conversation task, and the user experience is further improved.

Eighth embodiment

Referring to fig. 10, fig. 10 is a schematic flow chart illustrating a question processing method according to an embodiment of the present application. The method may comprise the steps of:

s801, acquiring the input current question and the parameter information of the user corresponding to the current question.

S802, a standard question database corresponding to the parameter information is obtained, and the standard question database comprises a plurality of standard questions.

And S803, respectively comparing the similarity of the current question with the similarity of a plurality of standard questions in the standard question database to obtain the similarity between the current question and each standard question.

S804, the standard question with the highest similarity between the current question and the multiple standard questions is obtained and serves as the target standard question.

And S805, acquiring a similarity threshold corresponding to the parameter information.

And S806, if the similarity between the target standard question and the current question is not less than the similarity threshold, outputting the target standard question.

S807, obtaining customer service log data, wherein the customer service log data comprises a plurality of question and answer paragraphs.

And S808, obtaining a question and answer paragraph including a target standard question from the plurality of question and answer paragraphs as the target question and answer paragraph, wherein the question and answer paragraph includes a plurality of second associated question which are ordered according to a specified question ordering, and the target standard question is one of the second associated question.

The embodiments of S801 to S808 refer to S701 to S708, and therefore are not described herein.

And S809, acquiring the sequence numbers of the target standard question in the plurality of second related questions as second target sequence numbers.

And S810, acquiring a second associated sentence with the sequence number arranged behind the second target sequence number in the plurality of second associated question sentences as a target associated question sentence.

The specific implementation of S809 to S810 can refer to S609 to S610, and therefore is not described herein.

S811, a target-related question is output.

In this embodiment, by acquiring the sequence numbers of the target standard question in the plurality of second related questions as the second target sequence numbers, and acquiring the second related sentences, of which the sequence numbers are arranged after the second target sequence numbers, in the plurality of second related questions as the target related questions, unnecessary questions in the context of the dialog are prevented from being re-presented, and the efficiency of the dialog is improved.

In some embodiments, after the dialog robot outputs the target standard question, the service log data may be pulled, the associated question most likely to be asked by the next sentence of the user is calculated in combination with the context of the target standard question hit in the service log data, and then the associated question is screened in combination with the channel tag information and displayed according to the sequence of the associated question in the service log data.

In some embodiments, in addition to recommending target associated questions based on semantic and customer service log data, a knowledge point edit page may be provided by the electronic device for the user to select associated questions, thereby facilitating the user to manually add associated questions, and when adding, the user may select two ways to add. On one hand, data can be pulled based on channel label information, and associated questions can be selected according to the type (semantic associated questions & scene associated questions) of intelligent recommendation. After the knowledge points are on line, the knowledge points can be synchronized to the robot to ask for answers in real time. On the other hand, real-time question-answering intelligent associated questions can be selected, and associated question recommendation and display modes can be selected in advance, for example, the associated question priority based on semantic recommendation is higher than the association based on scene recommendation.

Ninth embodiment

Referring to fig. 11, fig. 11 is a block diagram illustrating a question processing apparatus according to an embodiment of the present application. The device 900 is applied to an electronic device with a display screen or other image output devices, and the electronic device may be an electronic device such as a smart phone, a tablet computer, a projector, a wearable intelligent terminal, and the like.

As will be explained below with respect to the block diagram of the module shown in fig. 12, the question processing apparatus 900 includes: an information acquisition module 910, a standard question database acquisition module 920, a similarity comparison module 930, a target standard question sentence determination module 940 and an output module 950. The information obtaining module 910 is configured to obtain an input current question and parameter information of a user corresponding to the current question; the standard question database obtaining module 920 is configured to obtain a standard question database corresponding to the parameter information, where the standard question database includes a plurality of standard questions; the similarity comparison module 930 is configured to compare similarity between the current question and a plurality of standard questions in the standard question database, respectively, to obtain similarity between the current question and each standard question; the target standard question determining module 940 is configured to obtain a standard question with the highest similarity between the current question and the multiple standard questions as a target standard question; the output module 950 is used for outputting the target standard question sentence.

Optionally, the output module 950 includes:

and a similarity threshold acquisition unit for acquiring a similarity threshold corresponding to the parameter information.

And the target standard question output unit is used for outputting the target standard question if the similarity between the target standard question and the current question is not less than the similarity threshold value.

Optionally, the output module 950 further includes:

and the target associated question acquiring unit is used for acquiring a target associated question corresponding to the target standard question.

And the target associated question output unit is used for outputting the target associated question.

Optionally, the target-related question acquiring unit includes:

and the semantic recognition subunit is used for performing semantic recognition on the target standard question to obtain a semantic recognition result.

And the dialogue scene acquisition subunit is used for acquiring a dialogue scene corresponding to the semantic recognition result from the dialogue scene database, wherein the dialogue scene comprises a plurality of first associated question sentences.

And a first target related question acquiring subunit, configured to acquire, from the plurality of first related questions, a first related question corresponding to the target standard question as a target related question.

Optionally, the first target related question acquiring subunit is further configured to, if the plurality of first related questions include the target standard question, use, as the target related question, a first related question of the plurality of first related questions, which is other than the target standard question.

Optionally, the dialog scenario includes a plurality of first associated questions sorted according to a preset question ordering, and the first target associated question acquiring subunit is further configured to acquire, if the plurality of first associated questions include the target standard question, a sequence number of the target standard question in the plurality of first associated questions as the first target sequence number; and acquiring a first associated sentence with the sequence number arranged behind the first target sequence number in the plurality of first associated question sentences as a target associated question sentence.

Optionally, the target-related question acquiring unit includes:

and the customer service log data acquisition subunit is used for acquiring customer service log data, and the customer service log data comprises a plurality of question and answer paragraphs.

And the target question-answer paragraph acquiring subunit is used for acquiring a question-answer paragraph including a target standard question from the plurality of question-answer paragraphs as a target question-answer paragraph, wherein the target question-answer paragraph includes a plurality of second associated question, and the target standard question is one of the second associated question.

And a second target related question acquiring subunit, configured to acquire, as the target related question, a second related question other than the target standard question from among the plurality of second related questions.

Optionally, the question-answer paragraph includes a plurality of second associated question sentences ordered according to a specified question order, and the second target associated question sentence acquisition subunit is further configured to acquire sequence numbers of the target standard question sentences in the plurality of second associated question sentences as second target sequence numbers; and acquiring a second associated sentence with the sequence number arranged behind the second target sequence number in the plurality of second associated question sentences as a target associated question sentence.

Optionally, the types of the target associated question sentences are multiple, the number of the target associated question sentences is multiple, and the target associated question sentence output unit is further configured to obtain the number of the associated question sentences and the types of the associated question sentences corresponding to the parameter information; and outputting the target associated question corresponding to the number of the associated questions or/and the type of the associated questions.

Optionally, the types of the target associated question sentences are multiple, the number of the target associated questions is multiple, and the target associated question sentence output unit is further configured to receive an input associated question sentence display instruction; and responding to the associated question display instruction, and outputting the target associated question with the type and the number corresponding to the associated question display instruction.

Optionally, the number of target associated question sentences is multiple, and the output module 950 further includes:

and the receiving unit is used for receiving the input selected instruction.

And the response unit is used for responding to the selected instruction, and acquiring the target associated question corresponding to the selected instruction from the target associated question as the associated question to be replied.

And the reply unit is used for acquiring reply information corresponding to the associated question sentence to be replied and outputting the reply information.

The question processing device provided in the embodiment of the present application is used for implementing the corresponding question processing method in the foregoing method embodiment, and has the beneficial effects of the corresponding method embodiment, which are not described herein again.

As will be clear to those skilled in the art, the question processing apparatus provided in the embodiment of the present application can implement each process in the foregoing method embodiments, and for convenience and brevity of description, the specific working processes of the above-described apparatus and modules may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.

In the embodiments provided in the present application, the coupling or direct coupling or communication connection between the modules shown or discussed may be through some interfaces, and the indirect coupling or communication connection between the devices or modules may be in an electrical, mechanical or other form.

In addition, each functional module in the embodiments of the present application may be integrated into one processing module, or each module may exist alone physically, or two or more modules may be integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode.

Tenth embodiment

Referring to fig. 12, a block diagram of an electronic device 1000 according to an embodiment of the present disclosure is shown. The electronic device 1000 may be an electronic device capable of running an application, such as a smart phone or a tablet computer. The electronic device 1000 in the present application may include one or more of the following components: a processor 1010, a memory 1020, and one or more applications, wherein the one or more applications may be stored in the memory 1020 and configured to be executed by the one or more processors 1010, the one or more programs configured to perform a method as described in the aforementioned method embodiments.

The processor 1010 may be implemented in the form of at least one hardware of Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), Programmable logic Array (Programmable logic Array, P L A), the processor 1010 may be implemented in the form of at least one of a Central Processing Unit (CPU), Graphics Processing Unit (GPU), and modem, wherein the CPU is primarily responsible for Processing operating systems, user interfaces, and applications, etc., the processor 1010 may be implemented in a single piece of hardware for rendering and rendering content, the modem may be implemented in a single piece of hardware for rendering and rendering content, and the modem 1010 may be implemented in a separate piece of wireless communication.

The Memory 1020 may include a Random Access Memory (RAM) or a Read-Only Memory (Read-Only Memory). The memory 1020 may be used to store instructions, programs, code, sets of codes, or sets of instructions. The memory 1020 may include a stored program area and a stored data area, wherein the stored program area may store instructions for implementing an operating system, instructions for implementing at least one function (such as a touch function, a sound playing function, an image playing function, etc.), instructions for implementing various method embodiments described below, and the like. The data storage area may also store data created by the electronic device 1000 during use (e.g., phone book, audio-video data, chat log data), and the like.

Eleventh embodiment

Referring to fig. 13, a block diagram of a computer-readable storage medium according to an embodiment of the present application is shown. The computer-readable storage medium 1100 has stored therein program code that can be invoked by a processor to perform the methods described in the method embodiments above.

The computer-readable storage medium 1100 may be an electronic memory such as a flash memory, an EEPROM (electrically erasable programmable read only memory), an EPROM, a hard disk, or a ROM. Alternatively, the computer-readable storage medium 1100 includes a non-volatile computer-readable storage medium. The computer readable storage medium 1100 has storage space for program code 1110 for performing any of the method steps of the method described above. The program code can be read from or written to one or more computer program products. The program code 1110 may be compressed, for example, in a suitable form.

To sum up, the question processing method, the apparatus, the electronic device, and the storage medium provided in the embodiments of the present application use a mechanism for intelligently recommending questions by using associated questions: such as a mechanism that makes recommendations based on context in conjunction with big data context and a mechanism that makes recommendations based on semantic similarity. Therefore, the next question of the user can be predicted and prompted according to the conversation context, the user experience is improved, the supplement of the question given to the user according to the semantic similarity can be selected, and the question hitting effect is improved. And intelligent recommendation can be performed according to multi-dimensional label information through the association questions, so that association knowledge of different quantities and different contents can be given according to different channels and users of different labels, and the robot is more intelligent and has more service properties. And then the user experience is improved.

Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit the same; although the present application has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; such modifications and substitutions do not necessarily depart from the spirit and scope of the corresponding technical solutions in the embodiments of the present application.

Claims

1. A question processing method, characterized in that the method comprises:

acquiring an input current question and parameter information of a user corresponding to the current question;

acquiring a standard question database corresponding to the parameter information, wherein the standard question database comprises a plurality of standard questions;

respectively comparing the similarity of the current question with a plurality of standard questions in the standard question database to obtain the similarity between the current question and each standard question;

acquiring a standard question with the highest similarity between the current question and the plurality of standard questions as a target standard question;

and outputting the target standard question sentence.

2. The method of claim 1, wherein the outputting the target standard question comprises:

acquiring a similarity threshold corresponding to the parameter information;

and if the similarity between the target standard question and the current question is not less than the similarity threshold, outputting the target standard question.

3. The method according to claim 2, wherein after the outputting the target standard question if the similarity between the target standard question and the current question is not less than the similarity threshold, further comprising:

acquiring a target associated question corresponding to the target standard question;

and outputting the target association question.

4. The method according to claim 3, wherein the obtaining of the target associated question corresponding to the target standard question comprises:

performing semantic recognition on the target standard question to obtain a semantic recognition result;

acquiring a dialogue scene corresponding to the semantic recognition result from a dialogue scene database, wherein the dialogue scene comprises a plurality of first associated question sentences;

and acquiring a first associated question corresponding to the target standard question from the plurality of first associated questions as the target associated question.

5. The method according to claim 4, wherein the obtaining, as the target related question, a first related question corresponding to the target standard question from the plurality of first related questions, comprises:

and if the plurality of first associated questions comprise the target standard question, taking the first associated questions except the target standard question in the plurality of first associated questions as the target associated questions.

6. The method according to claim 4, wherein the dialog scenario includes a plurality of first related questions sorted according to a preset question ordering, and the obtaining, from the plurality of first related questions, a first related question corresponding to the target standard question as the target related question includes:

if the plurality of first associated question sentences comprise the target standard question sentence, acquiring the sequence numbers of the target standard question sentence in the plurality of first associated question sentences as first target sequence numbers;

and acquiring a first associated sentence with the sequence number arranged behind the first target sequence number in the plurality of first associated question sentences as the target associated question sentence.

7. The method according to claim 3, wherein the obtaining of the target associated question corresponding to the target standard question comprises:

obtaining customer service log data, wherein the customer service log data comprises a plurality of question and answer paragraphs;

acquiring a question-answer paragraph including the target standard question from the question-answer paragraphs as a target question-answer paragraph, wherein the target question-answer paragraph includes a plurality of second associated question, and the target standard question is one of the second associated question;

and acquiring second associated question sentences except the target standard question sentences in the plurality of second associated question sentences as the target associated question sentences.

8. The method according to claim 7, wherein the question-answering paragraph includes a plurality of second related question sentences ordered in a specified question order, and the obtaining, as the target related question sentence, a second related question sentence other than the target standard question sentence from among the plurality of second related question sentences includes:

acquiring the sequence numbers of the target standard question in the plurality of second associated question as second target sequence numbers;

and acquiring a second associated sentence with the sequence number arranged behind the second target sequence number in the plurality of second associated question sentences as the target associated question sentence.

9. The method according to any one of claims 3 to 8, wherein the types of the target related question sentences are multiple, the number of the target related question sentences is multiple, and the outputting the target related question sentences includes:

acquiring the number and types of the associated question sentences corresponding to the parameter information;

and outputting the target related question sentences corresponding to the related question sentence quantity or/and the related question sentence types.

10. The method according to any one of claims 3 to 8, wherein the types of the target associated question sentences are multiple, the number of the target associated questions is multiple, and the outputting the target associated question sentences includes:

receiving an input associated question display instruction;

and responding to the associated question display instruction, and outputting the target associated question with the type and the number corresponding to the associated question display instruction.

11. The method according to any one of claims 3 to 8, wherein the number of the target related question sentences is plural, and after the outputting the target related question sentences, the method further comprises:

receiving an input selected instruction;

responding to the selected instruction, and acquiring a target associated question corresponding to the selected instruction from the target associated question as an associated question to be answered;

and acquiring reply information corresponding to the question associated with the to-be-replied question and outputting the reply information.

12. The method of any one of claims 1 to 11, wherein the parameter information comprises one or more of a channel tag, a role tag, and a level tag.

13. A question processing apparatus characterized by comprising:

the information acquisition module is used for acquiring an input current question and the parameter information of a user corresponding to the current question;

a standard question database acquisition module, configured to acquire a standard question database corresponding to the parameter information, where the standard question database includes a plurality of standard questions;

the similarity comparison module is used for respectively comparing the similarity of the current question with a plurality of standard questions in the standard question database to obtain the similarity between the current question and each standard question;

a target standard question determining module, configured to obtain a standard question with the highest similarity between the current question and the standard question among the multiple standard questions, as a target standard question;

and the output module is used for outputting the target standard question sentence.

14. An electronic device, comprising:

a memory;

one or more processors coupled with the memory;

one or more programs, wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, the one or more applications configured to perform the method of any of claims 1-12.

15. A computer-readable storage medium, having stored thereon program code that can be invoked by a processor to perform the method according to any one of claims 1 to 12.