Data processing method and device

Abstract

The present application discloses a data processing method and device, which, after outputting a system question corresponding to a user question, obtains user reply content given by the user based on the system question but different from the reply content expected by the system question, and obtains and outputs a system answer corresponding to the user question based on at least the system question and the user reply content. Thus, even if there is a difference between the user reply content and the reply content expected by the system question, the system answer corresponding to the user question can still be output. That is, regardless of whether the user reply content is the reply content expected by the system question, the system answer corresponding to the user question can be output through the intelligent conversation system to solve the user needs corresponding to the user question in this human-computer conversation, thereby reducing the probability of the human-computer conversation being interrupted.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and more specifically, to a data processing method and device. Background Art

[0002] With the development of information technology, intelligent customer service can provide users with increasingly rich online business services. For example, it can communicate with users through online business services to answer questions raised by users. However, during this human-computer conversation, the content input by the user may differ from the input content expected by the intelligent customer service, thereby interrupting the human-computer conversation process. Summary of the Invention

[0003] In view of this, the purpose of this application is to disclose a data processing method and device for reducing the probability of interruption of the human-computer conversation process. The technical solution is as follows:

[0004] The present application discloses a data processing method, which is applied to an intelligent conversation system. The intelligent conversation system can respond to received input information and provide feedback information. The method includes:

[0005] Outputting a system question corresponding to a user question, wherein the user question is a type of input information during a human-computer conversation through the intelligent conversation system;

[0006] Obtaining a user response to the system question, where the user response differs from an expected response to the system question;

[0007] Obtaining a system answer corresponding to the user question based at least on the system question and the user reply content;

[0008] Output the system answer corresponding to the user question, where the system answer corresponding to the user question is a kind of feedback information.

[0009] Preferably, obtaining a system answer corresponding to the user question based at least on the system question and the user reply content includes:

[0010] Based at least on the system question and the user reply, obtaining a relationship between the user reply and a desired reply to the system question;

[0011] Based on the relationship between the user's reply content and the expected reply content of the system question, a system answer corresponding to the user's question is obtained.

[0012] Preferably, obtaining a system answer corresponding to the user question based on the relationship between the user reply content and the expected reply content of the system question includes:

[0013] If the relationship between the user's reply content and the expected reply content of the system question is a one-to-one relationship, obtaining a system answer corresponding to the user's question based on the expected reply content of the system question;

[0014] or

[0015] If there is no correspondence between the user reply content and the expected reply content of the system question, the first prompt content is used as the system answer, and the first prompt content is used to remind the user to enter the expected reply content of the system question.

[0016] Preferably, obtaining a system answer corresponding to the user question based on the relationship between the user reply content and the expected reply content of the system question includes:

[0017] If the relationship between the user's reply content and the expected reply content of the system question is a one-to-many or many-to-many relationship, obtaining a second prompt content for guiding the current human-computer conversation process;

[0018] outputting the second prompt content;

[0019] Obtaining user feedback content given by the user based on the second prompt content;

[0020] Based on the user feedback content and the expected reply content of the system question, a system answer corresponding to the user question is obtained.

[0021] Preferably, if the relationship between the user reply content and the expected reply content of the system question is a one-to-many or many-to-many relationship, obtaining the second prompt content for guiding the human-computer conversation process includes:

[0022] If the relationship between the user reply content and the expected reply content of the system question is a one-to-many or many-to-many relationship, obtaining at least one prompt option corresponding to the user reply content, the at least one prompt option being a second prompt content for guiding the current human-computer conversation process;

[0023] The obtaining of user feedback content given by the user based on the second prompt content includes: obtaining a prompt option selected by the user from the at least one prompt option, the selected prompt option being the user feedback content.

[0024] Preferably, obtaining the relationship between the user reply content and the expected reply content to the system question based at least on the system question and the user reply content includes:

[0025] Obtaining first identification information of a desired reply content to the system question in a preset knowledge base and second identification information of the user reply content in the preset knowledge base, wherein the preset knowledge base is constructed by executing multiple human-computer conversations with the intelligent conversation system;

[0026] The relationship between the first identification information and the second identification information in the preset knowledge base is used as the relationship between the user reply content and the expected reply content of the system question.

[0027] The present application discloses a data processing device, which is applied to an intelligent conversation system. The intelligent conversation system can respond to received input information and provide feedback information. The device includes:

[0028] An output unit, configured to output a system question corresponding to a user question, wherein the user question is a type of input information during a human-computer conversation through the intelligent conversation system;

[0029] A first obtaining unit is configured to obtain a user reply content given by the user based on the system question, wherein the user reply content differs from the reply content expected by the system question;

[0030] A second obtaining unit, configured to obtain a system answer corresponding to the user question based at least on the system question and the user reply content;

[0031] The output unit is further configured to output a system answer corresponding to the user question, where the system answer corresponding to the user question is a type of feedback information.

[0032] Preferably, the second obtaining unit is specifically used to obtain the relationship between the user reply content and the expected reply content of the system question based at least on the system question and the user reply content, and obtain the system answer corresponding to the user question based on the relationship between the user reply content and the expected reply content of the system question.

[0033] Preferably, the second obtaining unit is specifically configured to obtain a system answer corresponding to the user question based on the expected reply content of the system question if the relationship between the user reply content and the expected reply content of the system question is a one-to-one relationship;

[0034] or

[0035] The second obtaining unit is specifically used to use the first prompt content as the system answer if there is no correspondence between the user reply content and the expected reply content of the system question, and the first prompt content is used to remind the user to enter the expected reply content of the system question.

[0036] Preferably, the second obtaining unit is specifically used to obtain second prompt content for guiding the human-computer conversation process if the relationship between the user reply content and the expected reply content of the system question is a one-to-many or many-to-many relationship, and output the second prompt content through the output unit; and the second obtaining unit is specifically used to obtain user feedback content given by the user based on the second prompt content, and obtain a system answer corresponding to the user question based on the user feedback content and the expected reply content of the system question.

[0037] Preferably, the second obtaining unit is specifically used to obtain at least one prompt option corresponding to the user reply content if the relationship between the user reply content and the expected reply content of the system question is a one-to-many or many-to-many relationship, and the at least one prompt option is the second prompt content used to guide the human-computer conversation process; and the second obtaining unit is specifically used to obtain the prompt option selected by the user from the at least one prompt option, and the selected prompt option is the user feedback content.

[0038] Preferably, the second obtaining unit is specifically configured to obtain first identification information of the expected reply content to the system question in a preset knowledge base and second identification information of the user reply content in the preset knowledge base, wherein the preset knowledge base is constructed by executing multiple human-computer conversations with the intelligent conversation system;

[0039] The relationship between the first identification information and the second identification information in the preset knowledge base is used as the relationship between the user reply content and the expected reply content of the system question.

[0040] The present application also discloses an electronic device, comprising a processor and a display, wherein the processor has an intelligent conversation system, and the intelligent conversation system is capable of responding to received input information and providing feedback information;

[0041] The display is configured to output a system question corresponding to a user question, wherein the user question is a type of input information during a human-computer conversation through the intelligent conversation system;

[0042] The processor is used to obtain the user response content given by the user based on the system question, obtain the system answer corresponding to the user question based on at least the system question and the user response content, and output the system answer corresponding to the user question through the display. The system answer corresponding to the user question is a kind of feedback information, and there is a difference between the user response content and the expected response content of the system question.

[0043] The present application also discloses a storage medium, in which computer program code is stored. When the computer program code is executed, the above-mentioned data processing method is implemented.

[0044] It can be seen from the above technical solution that after outputting the system question corresponding to the user question, the user response content given by the user based on the system question but different from the response content expected by the system question is obtained, and the system answer corresponding to the user question is obtained and output based on at least the system question and the user response content. Therefore, even if there is a difference between the user response content and the response content expected by the system question, the system answer corresponding to the user question can still be output. That is to say, regardless of whether the user response content is the response content expected by the system question, the system answer corresponding to the user question can be output through the intelligent conversation system to solve the user needs corresponding to the user question in this human-computer conversation, thereby reducing the probability of the human-computer conversation being interrupted. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0046] Figure 1 is a schematic diagram of the intelligent conversation system disclosed in an embodiment of the present application;

[0047] Figure 2 is a flow chart of a data processing method disclosed in an embodiment of the present application;

[0048] Figure 3 is a schematic diagram of an expanded display area of ​​an electronic device disclosed in an embodiment of the present application;

[0049] Figure 4 is a flow chart of another data processing method disclosed in an embodiment of the present application;

[0050] Figure 5 is a schematic diagram of an entity relationship network disclosed in an embodiment of the present application;

[0051] Figure 6 This is a schematic diagram of a one-to-one relationship scenario disclosed in an embodiment of the present application;

[0052] Figure 7 This is a schematic diagram of a one-to-many relationship scenario disclosed in an embodiment of the present application;

[0053] Figure 8 This is a schematic diagram of a many-to-many relationship scenario disclosed in an embodiment of the present application;

[0054] Figure 9 This is a schematic diagram of a scenario in which no corresponding relationship exists, as disclosed in the embodiments of this application;

[0055] Figure 10 It is a structural diagram of a data processing device disclosed in an embodiment of the present application. DETAILED DESCRIPTION

[0056] Currently, human-computer conversations can be conducted through intelligent conversation systems. For example, the intelligent conversation system can be applied to electronic devices, such as installing the intelligent conversation system in the electronic device in the form of an APP (application). When the electronic device runs the APP, the user can conduct human-computer conversations through the electronic device. The input information received by the intelligent conversation system (such as user reply content) may be different from the reply content expected by the intelligent conversation system.

[0057] like Figure 1 As shown, when the user asks the user question "How to connect to the network" through the intelligent conversation system (which can also be regarded as a kind of input information), the intelligent conversation system outputs the system question "Please provide the model of the device that needs to be connected to the network" (a kind of feedback information). If the reply input by the user through the intelligent conversation system is the model, the intelligent conversation system will output the network connection method of the model. However, sometimes the user does not know what the model is. In this case, the input through the intelligent conversation system may not be the model, for example Figure 1 The user's reply was "My phone is from the A series, but I don't know the specific model." This user reply was significantly different from the reply expected by the intelligent conversation system, which resulted in the intelligent conversation system being unable to obtain a system answer that met the user's needs (the needs represented by the input information of this human-computer conversation). In other words, it was unable to obtain the network connection method for model B, which resulted in the interruption of this human-computer conversation.

[0058] To this end, the present application discloses a data processing method and device, which, after outputting a system question corresponding to a user question, obtains user response content given by the user based on the system question but different from the response content expected by the system question, and obtains and outputs a system answer corresponding to the user question based at least on the system question and the user response content. Thus, even if there is a difference between the user response content and the response content expected by the system question, the system answer corresponding to the user question can still be output. That is to say, regardless of whether the user response content is the response content expected by the system question, the system answer corresponding to the user question can be output through the intelligent conversation system to solve the user needs corresponding to the user question in this human-computer conversation, thereby reducing the probability of the human-computer conversation being interrupted.

[0059] To make the purpose, technical solutions, and advantages of the embodiments of this application more clear, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0060] See also Figure 2 , which illustrates a data processing method disclosed in an embodiment of the present application. The data processing method is applied to an intelligent conversation system. The intelligent conversation system can respond to received input information and provide feedback information. The method may include the following steps:

[0061] 201: Outputting a system question corresponding to the user question.

[0062] In this embodiment, user questions can be considered as a type of input information during the human-computer conversation process through the intelligent conversation system. In addition to the above-mentioned user questions, the input information during the human-computer conversation process also includes the user's response content to the system's questions (that is, the user's feedback on the system's questions).

[0063] Among them, user questions can indicate user needs, that is, user questions are questions that users ask to the intelligent conversation system in response to a user need. For example, for an electronic device, the needs include but are not limited to how to use a certain function of the electronic device, after-sales maintenance, etc., so users can ask questions to the intelligent conversation system in response to these needs. Figure 1 As shown, the user question "How to connect to the network" asked by the intelligent conversation system is a question raised by the user regarding the user demand for the network connection function of the electronic device being used.

[0064] The system questions corresponding to user questions are questions that the intelligent conversation system outputs to users. The purpose of these questions is to obtain key entity information, which is the information necessary to solve user problems. That is to say, in order to solve user problems, the intelligent conversation system needs to provide some key entity information in the user questions (such as words that can correctly and comprehensively reflect user needs). Only then can it output the system answer corresponding to the user question based on the obtained key entity information (that is, the answer to the user question given by the intelligent conversation system). However, user questions usually do not actively provide key entity information, or the user does not know what information is needed to solve their own needs, resulting in the lack of key entity information in the user questions. Therefore, in order to solve the user needs indicated by the user questions, the intelligent conversation system needs to output the corresponding system questions for the user questions to obtain key entity information from the user's reply content.

[0065] For example Figure 1 The user question shown is "How to connect to the network?" However, different electronic devices currently have different network connection methods. In order to solve this user problem, the intelligent conversation system needs to obtain the key entity information of the device model. Therefore, the intelligent conversation system needs to output the system questions corresponding to the user question, so as to obtain the key entity information required to solve the user problem through the system questions, such as Figure 1 The system asks the question "Please provide the model that needs to connect to the network" through the intelligent conversation system to obtain the key entity information of the device model that can solve the user's problem.

[0066] In this embodiment, the output system question can be output to the electronic device through the intelligent conversation system, and then the electronic device displays the system question, such as displaying it on a display area of ​​the electronic device. The display area can belong to the electronic device or be an extended display area of ​​the electronic device, such as Figure 3 As shown, the projection area of ​​the projector is used as the display area of ​​the electronic device, the electronic device is connected to the projector, the electronic device outputs the system questions to the projector, and the system questions are displayed with the help of the projection function of the projector, so that the user can view the system questions.

[0067] 202: Obtaining the user's response to the system question. The user's response differs from the expected response to the system question.

[0068] It is understood that the expected response content to a system question refers to the response content that includes the key entity information required to solve the user's question. Therefore, if the user's response content is exactly the expected response content to the system question, the intelligent conversation system can extract the key entity information from the user's response content, and thus obtain and output the correct system answer based on the key entity information. The intelligent conversation system does not need to perform the method provided in this embodiment. Therefore, this embodiment is intended for situations where the user's response content differs from the expected response content to the system question.

[0069] The discrepancy between the user's response and the system's expected response means that the intelligent conversational system cannot extract key entity information from the user's response. This key entity information is the information in the system's expected response. This discrepancy specifically includes the following two aspects:

[0070] The first aspect is: the entity information extracted from the user's reply content is different from the key entity information in the reply content expected by the system question, so that the correct system answer cannot be directly obtained based on the entity information extracted from the user's reply content. For example, the entity information from the user's reply content is part of the key entity information, and a unique system answer cannot be obtained based on the part of the key entity information, or the entity information extracted from the user's reply content is another expression of the key entity information. The second aspect is: no entity information can be extracted from the user's reply content, or the user's reply content does not contain any entity information. Therefore, by comparing the entity information extracted from the user's reply content with the key entity information extracted from the reply content expected by the system question, it is determined whether the user's reply content has a gap with the reply content expected by the system question. If no entity information can be extracted from the user's reply content, it means that the entity information extracted from the user's reply content is empty, and the so-called entity information is a word that cannot correctly and comprehensively reflect the user's needs.

[0071] Usually, user reply content is provided in text form. For user reply content in text form, text analysis tools can be used to analyze the user reply content to extract entity information. The specific process will not be elaborated in detail. Of course, it is not ruled out that the user reply content is provided in other forms. For example, the user reply content may also appear in the form of audio. That is to say, the user may reply to a voice message in response to the system question. For user reply content in audio form, the voice information can be converted into text information using a speech recognition engine to obtain the user reply content in text form. Then, the user reply content in text form can be analyzed to extract entity information. The specific process will not be elaborated in detail. The following still uses Figure 1 Take the user question "How to connect to the network" as an example:

[0072] In order to solve this user problem, the key entity information required by the intelligent conversation system is the model, so the intelligent conversation system outputs Figure 1 The system question shown is "Please provide the model of the phone that needs to connect to the network". At this time, the intelligent conversation system may receive user reply content as described in the above two aspects for this system question. For example, the user reply content described in the first aspect may be "I only know that mine is an A series mobile phone, but I don't know the specific model". From this user reply content, an entity information can be extracted: the model series, which is different from the model in the reply content expected by the system question. This shows that there is a gap between the user reply content and the reply content expected by the system question; the user reply content described in the second aspect may be "How do I know what model the mobile phone is", that is, the user reply content is just pure complaining, and the intelligent conversation system cannot obtain any entity information from it. This shows that there is a gap between the user reply content and the reply content expected by the system question.

[0073] 203: Based on at least the system question and the user's reply, obtain a system answer corresponding to the user's question. The system answer corresponding to the user's question includes a complete solution that can solve the user's question, which can be an operating step description of the operation targeted by the user's question, or an explanation of the function targeted by the user's question. For example Figure 1 The user question shown is "How to connect to the Internet?", and the system's answer is the multiple operating steps involved in network connection. By referring to these operating steps, the electronic device can be successfully connected to the Internet.

[0074] In this embodiment, the implementation methods for obtaining the system answer corresponding to the user question include but are not limited to the following two:

[0075] The first implementation method is to analyze the user's reply content based on the big data model, infer the key entity information in the reply content expected by the system question, and then obtain the system answer corresponding to the user's question based on the inferred key entity information.

[0076] For example Figure 1 In the scenario shown, the output system question is "Please provide the model that needs to connect to the network." The user's reply provides the purchase price and model series. The intelligent conversation system can then use the big data model to determine several models based on the price and model series. Then, based on the big data model, the intelligent conversation system can select the model with the highest sales volume from the determined models as the model corresponding to the user's question in this human-computer conversation, and then obtain the system answer based on the inferred model.

[0077] The second implementation method is to obtain the relationship between the user's reply and the expected reply to the system question based on the system question and the user's reply content, and then obtain the system answer corresponding to the user's question based on the relationship between the user's reply and the expected reply to the system question. This implementation method specifically includes the following three cases:

[0078] The first case is a one-to-one relationship between the user's reply and the expected reply to the system's question. This means that the user's reply can be used to obtain the expected reply to the system's question. Based on the expected reply to the system's question, the system can then obtain the answer to the user's question.

[0079] In the second case, if there is no correspondence between the user's reply and the expected reply to the system question, the first prompt content is output to the user as the system answer, wherein the first prompt content is used to remind the user to enter the expected reply to the system question, so as to guide the current human-computer conversation process through the first prompt content;

[0080] In the third scenario, if the relationship between the user's reply and the expected response to the system question is one-to-many or many-to-many, a second prompt is obtained based on the user's reply, the second prompt is output, user feedback based on the second prompt is obtained, and based on the user feedback and the expected response to the system question, a system answer corresponding to the user's question is obtained. The second prompt is used to guide the current human-computer conversation.

[0081] That is to say, in the second and third cases, it is necessary to guide the human-computer conversation process based on the first prompt content and the second prompt content respectively, so as to obtain user response content that has a one-to-one relationship with the expected response content of the system question through multiple human-computer interactions, and then obtain the system answer corresponding to the user question.

[0082] 204: Outputting a system answer corresponding to the user question. The system answer corresponding to the user question is a kind of feedback information.

[0083] For example, the system answer can be output in the same way as the system question, such as outputting the system answer to the electronic device through the intelligent conversation system, and then the electronic device displays the system question.

[0084] The output system answer can be expressed in a variety of forms, such as text, voice, video and picture. Figure 1 For the user question "How to connect to the network?", the system answer can be a pre-recorded operation video corresponding to the user's model, a flowchart composed of real pictures, or a text describing the specific operation method of connecting to the network. This embodiment does not limit its form.

[0085] It can be seen from the above technical solution that after outputting the system question corresponding to the user question, the user response content given by the user based on the system question but different from the response content expected by the system question is obtained, and the system answer corresponding to the user question is obtained and output based on at least the system question and the user response content. Therefore, even if there is a difference between the user response content and the response content expected by the system question, the system answer corresponding to the user question can still be output. That is to say, regardless of whether the user response content is the response content expected by the system question, the system answer corresponding to the user question can be output through the intelligent conversation system to solve the user needs corresponding to the user question in this human-computer conversation, thereby reducing the probability of the human-computer conversation being interrupted.

[0086] The following is a method for obtaining the relationship between the user's reply content and the expected reply content of the system question provided by this embodiment, and in this method, the system answer based on different relationships is explained, such as Figure 4As shown, the process of obtaining the relationship between the user's reply content and the expected reply content of the system question may include the following steps:

[0087] 401: Obtain first identification information of the expected reply content to the system question in the preset knowledge base and second identification information of the user reply content in the preset knowledge base.

[0088] In this embodiment, the preset knowledge base is used to record the relationship between multiple entities. In the preset knowledge base, an entity is used as a node, and the relationship between entities is used as an edge to connect the entities, so that the preset knowledge base forms an entity relationship network. Figure 5 The entity relationship network is a form of entity relationship network. The relationship between each entity in the entity relationship network is represented by arrows, and the identification information of each entity is represented by its own content. If there is no connection between two entities, it means that there is no corresponding relationship between the two entities. Of course, the entity relationship network includes but is not limited to Figure 5 The entities displayed and the relationships between them, Figure 5 The entity relationship network shown should be understood as an optional specific implementation of multiple entity relationship networks. For example, in other entity relationship networks with a wider scope, in addition to Figure 5 In addition to the entity information shown, it can also include purchase price, screen size, camera pixels, etc. The constructed entity relationship network can be pre-set by analyzing user logs and records of human-computer conversations performed by the intelligent conversation system, and combined with the advice of business experts. This embodiment does not describe in detail the construction process of the entity relationship network.

[0089] The relationships between entities in the preset knowledge base can be stored using a relational database, for example, MySQL. Each entity in the preset knowledge base can be regarded as a table in the relational database, and the relationships between entities can be expressed in the form of primary keys and / or foreign keys of the tables. For example Figure 5 The entity relationship network shown as the preset knowledge base, in which the five entities can be regarded as five tables in the relational database. Among them, the entity information of the model can be stored in the relational database in the form of the following Table 1.

[0090] Table 1 Storage table of models

[0091] model SN number Screen size Model Series Order IMEI A SN number 1 5.5 inches First series …… …… B SN number 2 5.5 inches First series …… …… C SN number 3 5.7 inches First series …… ……

[0092] In Table 1, the SN number can be regarded as the primary key. One SN number can uniquely identify a record in Table 1 (a row in the table is called a record. In this embodiment, a record can be regarded as the representation of a device in the database). One record corresponds to a unique model, indicating that there is an equivalent relationship between the SN number and the model. A model can be uniquely identified based on an SN number.

[0093] And it can be seen from Table 1 above that the content of a certain node in the preset knowledge base, such as the above-mentioned SN number, IMEI (International Mobile Equipment Identity) and model, can indicate a unique object. When the user's demand is for this unique object, the intelligent conversation system can get a unique answer, and the key entity information in the expected reply content of the system question can also get a unique answer. Therefore, the identification information of the node in the preset knowledge base that can represent the unique object can be considered as key entity information. Similarly, the identification information of the node in the preset knowledge base that cannot indicate a unique object can be considered as entity information in the user's reply content.

[0094] Based on this, a human-computer conversation is conducted through the intelligent conversation system, and the first identification information of the expected reply content of the system question in the preset knowledge base can be obtained in combination with the preset knowledge base. If the first identification information obtained is information consistent with the key entity information extracted from the expected reply content of the system question, after obtaining the first identification information, the node in the preset knowledge base with the first identification information as content can also be determined; similarly, the first identification information of the user's reply content in the preset knowledge base can be obtained in combination with the preset knowledge base. If the second identification information obtained is information consistent with the entity information extracted from the user's reply content, after obtaining the second identification information, the node in the preset knowledge base with the second identification information as content can also be determined.

[0095] Step 401 can also be considered as a process of extracting a conversation group set from the system question and the user reply content corresponding to the system question.

[0096] A conversation group set includes at least one conversation group. A conversation group can be considered a combination of key entity information and entity information of user reply content. If the key entity information is considered the first identification information and the entity information of the user reply content is considered the second identification information, then the conversation group can also be considered the combination of the first identification information and the second identification information.

[0097] For ease of understanding, an example of extracting conversation groups is given, such as Figure 1As shown, the system question is "Please provide the model that needs to connect to the network." Correspondingly, the key entity information extracted is the model. For the above system question, if the user's reply is "The serial number of the mobile phone is xxx," then the entity information extracted from the user's reply is the SN number (i.e., the product serial number). The conversation group extracted from the above system question and the corresponding user reply is model-SN number.

[0098] 402: The relationship between the first identification information and the second identification information is used as the relationship between the expected reply content of the system question and the user reply content.

[0099] Since the first identification information and the second identification information can correspond to nodes in the preset knowledge base, and the preset knowledge base records the corresponding relationship between the nodes, the relationship between the two nodes corresponding to the first identification information and the second identification information in the preset knowledge base can be used as the relationship between the first identification information and the second identification information. In addition, the first identification information corresponds to the expected reply content of the system question, and the second identification information corresponds to the user's reply content. Therefore, the relationship between the first identification information and the second identification information can be used as the relationship between the expected reply content of the system question and the user's reply content, that is, the relationship between the two nodes with the first identification information and the second identification information is regarded as the relationship between the expected reply content of the system question and the user's reply content.

[0100] For example, for the conversation group of model-SN number extracted in the example of step 401, by querying the preset knowledge base, it can be obtained that the nodes with these two identification information are SN number and model, and there is a one-to-one relationship between SN number and model, thereby determining that the relationship between the expected reply content of the system question and the user's reply content is a one-to-one relationship.

[0101] Based on the relationship between the expected reply content of the system question and the user reply content obtained in step 402, the following describes the process of obtaining the system answer corresponding to the user question based on different relationships in this embodiment.

[0102] like Figure 5 As shown in the figure, the relationships between entities in the preset knowledge base include one-to-one, one-to-many, many-to-many, and non-corresponding relationships. The relationships between the expected responses to the corresponding system questions and the user responses also include the above four. For each different relationship, the process of obtaining the system answer corresponding to the user question can be implemented in the following four different ways. The following describes these four implementation methods in detail based on the corresponding scenarios:

[0103] The first implementation method: corresponds to the situation where there is a one-to-one relationship between the expected reply content of the system question and the user's reply content. In this case, the preset knowledge base can be used to directly determine the expected reply content of the system question based on the user's reply content, and then obtain the system answer corresponding to the user's question based on the expected reply content of the system question.

[0104] For example Figure 6 In the scenario shown, the user's question is "How do I connect to the network?" To obtain a system answer corresponding to this user's question, the intelligent conversation system needs to obtain the model as key entity information. Therefore, the output system question is "Please provide the model required to connect to the network." In this case, the user may not provide the model to the intelligent conversation system, but instead provide information related to the IMEI. After extracting the IMEI entity information from the user's reply, the intelligent conversation system queries the preset knowledge base and finds that there is a one-to-one relationship between IMEI and model. It then determines the model based on the IMEI in the user's reply, and then obtains the system answer corresponding to the user's question "How do I connect to the network?", that is, the specific operation method for connecting to the network for the model determined by the IMEI in the user's reply, and then outputs it in text form.

[0105] The second implementation method: corresponding to the situation where there is a one-to-many relationship between the expected reply content of the system question and the user's reply content, in this case, the intelligent conversation system obtains the second prompt content based on the user's reply content, outputs the second prompt content and obtains the user feedback content given by the user based on the second prompt content, and then obtains the system answer corresponding to the user question based on the user feedback content and the expected reply content of the system question, and outputs the system answer corresponding to the user question.

[0106] The second prompt content is used to guide the current human-computer conversation process. The second prompt content can be output in various forms, including but not limited to:

[0107] Obtaining at least one prompt option corresponding to the user's reply content, wherein the at least one prompt option is a second prompt content for guiding the current human-computer conversation process, that is, outputting the second prompt content in the form of options, and the user provides feedback by clicking on one of the options, such as using the prompt option selected by the user as the user feedback content;

[0108] The second prompt content is output in the form of text list, voice, video and image, and the user gives feedback through text, voice and other means.

[0109] The process of obtaining the system answer corresponding to the user question based on the user feedback content and the expected reply content of the system question is similar to the first implementation method. If the preset knowledge base can be used to directly determine the expected reply content of the system question based on the user feedback content, the system answer corresponding to the user question can be obtained based on the expected reply content of the system question. If the preset knowledge base cannot directly determine the expected reply content of the system question based on the user feedback content, such as the user feedback content obtained using the preset knowledge base and the expected reply content of the system question are still not in a one-to-one relationship, it is necessary to give a second prompt content based on the user feedback content to guide the human-computer conversation process.

[0110] For example, in Figure 7 In the one-to-many scenario shown, the user question is "How to connect to the network". In order to obtain the system answer corresponding to the user question, the intelligent conversation system needs to obtain the model as the key entity information, so the output system asks "Please provide the model that needs to be connected to the network". At this time, the user may not feedback the model to the intelligent conversation system, but instead feedback relevant information about the model series. After the intelligent conversation system extracts the entity information of the model series from the user's reply content, it knows through the preset knowledge base that the model series and the model are in a one-to-many relationship. Then, it obtains multiple models corresponding to the model series extracted from the user's reply content, and outputs the multiple models in the form of text. After the user replies to one of the models, the intelligent conversation system obtains the operation method of connecting to the network of the model based on the model replied by the user, and outputs it as the system answer corresponding to the user question.

[0111] The third implementation method: corresponding to the situation where there is a many-to-many relationship between the expected reply content of the system question and the user's reply content, in this case, the intelligent conversation system obtains the second prompt content based on the user's reply content, outputs the second prompt content and obtains the user feedback content given by the user based on the second prompt content, and then obtains the system answer corresponding to the user question based on the user feedback content and the expected reply content of the system question, and outputs the system answer corresponding to the user question.

[0112] The second prompt content is used to guide this human-computer conversation, and the output form may be the same as the output form of the second prompt content in the second implementation manner described above.

[0113] The process of obtaining the system answer corresponding to the user question based on the user feedback content and the expected reply content of the system question is similar to the first implementation method. If the preset knowledge base can be used to directly determine the expected reply content of the system question based on the user feedback content, the system answer corresponding to the user question can be obtained based on the expected reply content of the system question. If the preset knowledge base cannot directly determine the expected reply content of the system question based on the user's feedback on the second prompt content, such as the user feedback content obtained using the preset knowledge base and the expected reply content of the system question are still not in a one-to-one relationship, it is necessary to give the second prompt content based on the user feedback content to guide the human-computer conversation process.

[0114] For example, in Figure 8 In the schematic diagram of the many-to-many scenario shown, the user question is "Query the repair order". In order to obtain the system answer corresponding to the user question, the intelligent conversation system needs to obtain the order number as the key entity information. Therefore, the output system asks "Please provide the order number to be queried". At this time, the user may not feedback the order number, but feedback the relevant information of the device model. After the intelligent conversation system extracts the entity information of the model from the user's reply content, it knows through the preset knowledge base that the model and the order number are in a many-to-many relationship. Then, it obtains the multiple order numbers of the models extracted from the user's reply content, and outputs the multiple order numbers in the form of text. After the user replies to one of the order numbers, the intelligent conversation system obtains the repair order corresponding to the order number based on the order number replied by the user, and outputs it as the system answer corresponding to the user's question.

[0115] The fourth implementation method: In the case where there is no correspondence between the expected reply content of the system question and the user's reply content, the intelligent conversation system obtains the first prompt content and outputs the first prompt content as the system answer corresponding to the user question.

[0116] The first prompt is used to prompt the user to enter the desired response to the system question. Optionally, the first prompt can be the same as the system question output in this embodiment, that is, the system question can be output again, or content related to the desired response to the system question can be provided as the first prompt to guide the user. Other forms of the first prompt are not described in this embodiment.

[0117] For example, in Figure 9In the diagram of the scenario where no corresponding relationship exists, the user's question is "How do I connect to the network?" To obtain a system answer corresponding to this user question, the intelligent conversational system needs to obtain the model as key entity information. Therefore, it outputs the system question to the user, "Please provide the model of the device required to connect to the network." In this case, the user's reply may not include any relevant information about the entity, and the entity information extracted by the intelligent conversational system from the user's reply is empty. In this case, it can be assumed that there is no corresponding relationship between the entity information in the user's reply and the key entity information in the preset knowledge base. Therefore, the intelligent conversational system obtains the first prompt content and outputs the first prompt content as the system answer corresponding to the user question.

[0118] In this embodiment, after obtaining the user's feedback on the first prompt content, the intelligent conversation system can use the user's feedback on the first prompt content as the user reply content, and execute step 401 of this embodiment again. If there is a correspondence between the user reply content this time and the reply content expected by the system question, the process of obtaining the system answer corresponding to the user question of this embodiment is executed in one of the above-mentioned first implementation method, second implementation method and third implementation method according to the specific correspondence, and then the system answer corresponding to the user question is output.

[0119] That is to say, in the second, third and fourth implementation methods, it is necessary to guide the human-computer conversation process based on the first prompt content and the second prompt content respectively, so as to obtain user response content that has a one-to-one relationship with the expected response content of the system question through multiple human-computer interactions, and then obtain the system answer corresponding to the user question.

[0120] Corresponding to the above method embodiment, the embodiment of the present application discloses a data processing device, which is applied to an intelligent conversation system. The intelligent conversation system can respond to the received input information and provide feedback information. Please refer to Figure 10 The data processing device may include: an output unit 101, a first obtaining unit 102 and a second obtaining unit 103.

[0121] The output unit 101 is used to output a system question corresponding to a user question. The user question is a type of input information during a human-computer conversation through the intelligent conversation system.

[0122] The first obtaining unit 102 is configured to obtain a user response to a system question, where the user response differs from the expected response to the system question. The discrepancy between the user response and the expected response to the system question means that the intelligent conversation system cannot extract key entity information from the user response, which is information in the expected response to the system question. This discrepancy specifically includes the following two aspects:

[0123] The first aspect is: the entity information extracted from the user's reply content is different from the key entity information in the reply content expected by the system question, so that the correct system answer cannot be directly obtained based on the entity information extracted from the user's reply content. For example, the entity information from the user's reply content is part of the key entity information, and a unique system answer cannot be obtained based on the part of the key entity information, or the entity information extracted from the user's reply content is another expression of the key entity information. The second aspect is: no entity information can be extracted from the user's reply content, or the user's reply content does not contain any entity information. Therefore, by comparing the entity information extracted from the user's reply content with the key entity information extracted from the reply content expected by the system question, it is determined whether the user's reply content has a gap with the reply content expected by the system question. If no entity information can be extracted from the user's reply content, it means that the entity information extracted from the user's reply content is empty, and the so-called entity information is a word that cannot correctly and comprehensively reflect the user's needs.

[0124] The second obtaining unit 103 is configured to obtain a system answer corresponding to the user question based at least on the system question and the user reply content.

[0125] In this embodiment, the second obtaining unit 103 obtains the system answer corresponding to the user question in the following two ways:

[0126] The first implementation method is to analyze the user's reply content based on the big data model, infer the key entity information in the reply content expected by the system question, and then obtain the system answer corresponding to the user's question based on the inferred key entity information.

[0127] The second implementation method is to obtain the relationship between the user's reply and the expected reply to the system question based on the system question and the user's reply content, and then obtain the system answer corresponding to the user's question based on the relationship between the user's reply and the expected reply to the system question. This implementation method specifically includes the following three cases:

[0128] In the first case, the relationship between the user's reply content and the expected reply content of the system question is a one-to-one relationship, which means that the expected reply content of the system question can be obtained through the user's reply content. Therefore, the second obtaining unit 103 can obtain the system answer corresponding to the user's question based on the expected reply content of the system question;

[0129] Case 2: There is no correspondence between the user's reply and the expected reply to the system question. The second obtaining unit 103 outputs the first prompt as the system answer to the user. The first prompt is used to remind the user to input the expected reply to the system question, thereby guiding the human-computer conversation process through the first prompt.

[0130] In the third scenario, the relationship between the user's reply and the expected reply to the system question is a one-to-many or many-to-many relationship. The second obtaining unit 103 obtains a second prompt based on the user's reply, outputs the second prompt, obtains user feedback based on the second prompt, and obtains a system answer corresponding to the user's question based on the user feedback and the expected reply to the system question. The second prompt is used to guide the current human-computer conversation process.

[0131] An optional method for the second prompt content in the third case is: obtain at least one prompt option corresponding to the user's reply content, use the at least one prompt option as the second prompt content for guiding this human-computer conversation process, and the corresponding user feedback content is the prompt option selected by the user from the at least one prompt option.

[0132] In a second implementation method, the second acquisition unit 103 obtains the relationship between the user's reply content and the expected reply content of the system question by: obtaining the first identification information of the expected reply content of the system question in the preset knowledge base and the second identification information of the user's reply content in the preset knowledge base, and using the relationship between the first identification information and the second identification information in the preset knowledge base as the relationship between the user's reply content and the expected reply content of the system question. The preset knowledge base is constructed by executing multiple human-computer conversations through the intelligent conversation system.

[0133] The output unit 101 is further configured to output a system answer corresponding to the user question. The system answer corresponding to the user question is a type of feedback information.

[0134] The specific working process of each unit in the data processing device disclosed in this embodiment can refer to the data processing method disclosed in the above method embodiment, and will not be elaborated in this embodiment.

[0135] It can be seen from the above technical solution that after outputting the system question corresponding to the user question, the user response content given by the user based on the system question but different from the response content expected by the system question is obtained, and the system answer corresponding to the user question is obtained and output based on at least the system question and the user response content. Therefore, even if there is a difference between the user response content and the response content expected by the system question, the system answer corresponding to the user question can still be output. That is to say, regardless of whether the user response content is the response content expected by the system question, the system answer corresponding to the user question can be output through the intelligent conversation system to solve the user needs corresponding to the user question in this human-computer conversation, thereby reducing the probability of the human-computer conversation being interrupted.

[0136] In addition, an embodiment of the present application further discloses an electronic device, including a processor and a display, wherein the processor has an intelligent conversation system, and the intelligent conversation system can respond to received input information and provide feedback information;

[0137] A display for outputting system questions corresponding to user questions, wherein user questions are a type of input information during a human-computer conversation through the intelligent conversation system;

[0138] The processor is used to obtain the user reply content given by the user based on the system question, obtain the system answer corresponding to the user question based on at least the system question and the user reply content, and output the system answer corresponding to the user question through the display. The system answer corresponding to the user question is a kind of feedback information. There is a difference between the user reply content and the expected reply content of the system question. For the description of the specific functions of the processor, please refer to the data processing method disclosed in the above method embodiment, which will not be elaborated in this embodiment.

[0139] The embodiment of the present application further discloses a storage medium having computer program code stored thereon, and the above-mentioned data processing method is implemented when the computer program code is executed.

[0140] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similarities between the various embodiments can be referred to in conjunction with each other. For device embodiments, since they are generally similar to method embodiments, their description is relatively simple, and for relevant details, reference can be made to the description of the method embodiments.

[0141] Finally, it should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or device comprising the element.

[0142] The above description of the disclosed embodiments will enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

[0143] The above is only a preferred embodiment of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A data processing method, applied to an intelligent conversation system, wherein the intelligent conversation system is capable of responding to received input information and providing feedback information, the method comprising: Outputting a system question corresponding to a user question, wherein the user question is a type of input information during a human-computer conversation through the intelligent conversation system; Obtaining a user response content given by the user in response to the system question, wherein the user response content differs from the response content expected by the system question; wherein the difference between the user response content and the response content expected by the system question means that the intelligent conversation system cannot extract key entity information from the user response content, the key entity information being information in the response content expected by the system question; Obtaining a system answer corresponding to the user question based at least on the system question and the user reply content; Outputting a system answer corresponding to the user question, wherein the system answer corresponding to the user question is a kind of feedback information; The obtaining of a system answer corresponding to the user question based at least on the system question and the user reply content includes: obtaining entity information in the user reply content based on an analysis of the user reply content, the entity information being different from key entity information in the reply content expected by the system question, the key entity information being entity information used to obtain the system answer when there is a difference between the user reply content and the reply content expected by the system question; guiding the user to provide feedback again based on a relationship between the entity information in the user reply content and the key entity information, and obtaining a system answer corresponding to the user question based on the user's feedback again; The obtaining of a system answer corresponding to the user question based at least on the system question and the user reply content includes: Based at least on the system question and the user reply, obtaining a relationship between the user reply and a desired reply to the system question; Obtaining a system answer corresponding to the user question based on the relationship between the user's reply content and the expected reply content of the system question; The obtaining, based at least on the system question and the user reply content, of a relationship between the user reply content and the expected reply content to the system question includes: Obtaining first identification information of a desired reply content to the system question in a preset knowledge base and second identification information of the user reply content in the preset knowledge base, wherein the preset knowledge base is constructed by executing multiple human-computer conversations with the intelligent conversation system; The relationship between the first identification information and the second identification information in the preset knowledge base is used as the relationship between the user reply content and the expected reply content of the system question.

2. The method according to claim 1, wherein obtaining a system answer corresponding to the user question based on the relationship between the user reply content and the expected reply content of the system question comprises: If the relationship between the user's reply content and the expected reply content of the system question is a one-to-one relationship, obtaining a system answer corresponding to the user's question based on the expected reply content of the system question; or If there is no correspondence between the user reply content and the expected reply content of the system question, the first prompt content is used as the system answer, and the first prompt content is used to remind the user to enter the expected reply content of the system question.

3. The method according to claim 1, wherein obtaining a system answer corresponding to the user question based on the relationship between the user reply content and the expected reply content of the system question comprises: If the relationship between the user's reply content and the expected reply content of the system question is a one-to-many or many-to-many relationship, obtaining a second prompt content for guiding the current human-computer conversation process; outputting the second prompt content; Obtaining user feedback content given by the user based on the second prompt content; Based on the user feedback content and the expected reply content of the system question, a system answer corresponding to the user question is obtained.

4. The method according to claim 3, wherein if the relationship between the user's reply content and the expected reply content of the system question is a one-to-many or many-to-many relationship, obtaining the second prompt content for guiding the human-computer conversation process comprises: If the relationship between the user reply content and the expected reply content of the system question is a one-to-many or many-to-many relationship, obtaining at least one prompt option corresponding to the user reply content, the at least one prompt option being a second prompt content for guiding the current human-computer conversation process; The obtaining of user feedback content given by the user based on the second prompt content includes: obtaining a prompt option selected by the user from the at least one prompt option, the selected prompt option being the user feedback content.

5. A data processing device, applied to an intelligent conversation system, wherein the intelligent conversation system is capable of responding to received input information and providing feedback information, the device comprising: An output unit, configured to output a system question corresponding to a user question, wherein the user question is a type of input information during a human-computer conversation through the intelligent conversation system; a first obtaining unit, configured to obtain a user reply content given by the user based on the system question, wherein the user reply content differs from the reply content expected by the system question; wherein the difference between the user reply content and the reply content expected by the system question means that the intelligent conversation system cannot extract key entity information from the user reply content, the key entity information being information in the reply content expected by the system question; A second obtaining unit, configured to obtain a system answer corresponding to the user question based at least on the system question and the user reply content; The output unit is further configured to output a system answer corresponding to the user question, wherein the system answer corresponding to the user question is a type of feedback information; The obtaining of a system answer corresponding to the user question based at least on the system question and the user reply content includes: obtaining entity information in the user reply content based on an analysis of the user reply content, the entity information being different from key entity information in the reply content expected by the system question, the key entity information being entity information used to obtain the system answer when there is a difference between the user reply content and the reply content expected by the system question; guiding the user to provide feedback again based on a relationship between the entity information in the user reply content and the key entity information, and obtaining a system answer corresponding to the user question based on the user's feedback again; The second obtaining unit is specifically configured to obtain, based at least on the system question and the user's reply content, a relationship between the user's reply content and the expected reply content to the system question, and obtain a system answer corresponding to the user question based on the relationship between the user's reply content and the expected reply content to the system question; The obtaining, based at least on the system question and the user reply content, of a relationship between the user reply content and the expected reply content to the system question includes: Obtaining first identification information of a desired reply content to the system question in a preset knowledge base and second identification information of the user reply content in the preset knowledge base, wherein the preset knowledge base is constructed by executing multiple human-computer conversations with the intelligent conversation system; The relationship between the first identification information and the second identification information in the preset knowledge base is used as the relationship between the user reply content and the expected reply content of the system question.

6. The apparatus according to claim 5, wherein the second obtaining unit is configured to obtain a system answer corresponding to the user question based on the expected reply content of the system question if the relationship between the user reply content and the expected reply content of the system question is a one-to-one relationship; or The second obtaining unit is specifically configured to use the first prompt content as the system answer if there is no correspondence between the user reply content and the expected reply content of the system question, wherein the first prompt content is used to remind the user to enter the expected reply content of the system question; or The second obtaining unit is specifically used to obtain second prompt content for guiding the human-computer conversation process if the relationship between the user reply content and the expected reply content of the system question is a one-to-many or many-to-many relationship, and output the second prompt content through the output unit; and the second obtaining unit is specifically used to obtain user feedback content given by the user based on the second prompt content, and obtain a system answer corresponding to the user question based on the user feedback content and the expected reply content of the system question.

7. An electronic device comprising a processor and a display, wherein the processor has an intelligent conversation system capable of responding to received input information and providing feedback information; The display is configured to output a system question corresponding to a user question, wherein the user question is a type of input information during a human-computer conversation through the intelligent conversation system; The processor is configured to obtain a user reply content given by the user based on the system question, obtain a system answer corresponding to the user question based at least on the system question and the user reply content, and output the system answer corresponding to the user question through the display, wherein the system answer corresponding to the user question is a type of feedback information, and there is a difference between the user reply content and the expected reply content of the system question; wherein, The discrepancy between the user's reply and the expected reply to the system question means that the intelligent conversation system cannot extract key entity information from the user's reply, where the key entity information is information in the expected reply to the system question; The obtaining of a system answer corresponding to the user question based at least on the system question and the user reply content includes: obtaining entity information in the user reply content based on an analysis of the user reply content, the entity information being different from key entity information in the reply content expected by the system question, the key entity information being entity information used to obtain the system answer when there is a difference between the user reply content and the reply content expected by the system question; guiding the user to provide feedback again based on a relationship between the entity information in the user reply content and the key entity information, and obtaining a system answer corresponding to the user question based on the user's feedback again; The obtaining of a system answer corresponding to the user question based at least on the system question and the user reply content includes: Based at least on the system question and the user reply, obtaining a relationship between the user reply and a desired reply to the system question; Obtaining a system answer corresponding to the user question based on the relationship between the user's reply content and the expected reply content of the system question; The obtaining, based at least on the system question and the user reply content, of a relationship between the user reply content and the expected reply content to the system question includes: Obtaining first identification information of a desired reply content to the system question in a preset knowledge base and second identification information of the user reply content in the preset knowledge base, wherein the preset knowledge base is constructed by executing multiple human-computer conversations with the intelligent conversation system; The relationship between the first identification information and the second identification information in the preset knowledge base is used as the relationship between the user reply content and the expected reply content of the system question.

Images