Information processing method, apparatus and electronic device

Abstract

The application provides an information processing method, device and electronic equipment. The information processing method comprises: obtaining a question and answer text pair, the question and answer text pair comprising: a question text and a reply text of the question text; performing multi-dimensional analysis on the question text and the reply text to obtain an evaluation result of the question and answer text pair in each dimension of multiple dimensions; and in a case where the evaluation result meets a preset requirement, configuring the question and answer text pair to feed back the reply text when the question text is received. In the embodiment of the application, by performing multi-dimensional analysis on the question and answer text, and in a case where the evaluation result meets the preset requirement, the question and answer text pair is configured, which can improve the quality of the configured question and answer text pair.

Description

Technical Field

[0001] This application relates to the field of computer technology, and more particularly to an information processing method, apparatus, and electronic device. Background Technology

[0002] In e-commerce, it's common for users to ask merchants questions, such as inquiries about product information before purchase, logistics progress during shipping, and after-sales issues after receiving the goods. To improve service efficiency, merchants typically configure multiple question-and-answer text pairs on the server side, allowing for quick responses to user inquiries.

[0003] In related technologies, when configuring question-and-answer text pairs, it is necessary to first obtain predefined question-and-answer text pairs from the merchant, and then perform semantic matching on these predefined text pairs to obtain the semantic matching degree. Only when the semantic matching degree meets preset requirements will the question-and-answer text pair be configured. However, the above configuration method has the problem of poor quality of question-and-answer text pairs. Summary of the Invention

[0004] This application provides an information processing method, apparatus, and electronic device to address the problem of poor quality in configured question-and-answer text pairs.

[0005] The first aspect of this application provides an information processing method, including: acquiring a question-and-answer text pair, the question-and-answer text pair including: a question text and a response text to the question text; performing multi-dimensional analysis on the question text and the response text to obtain an evaluation result of the question-and-answer text pair in each of the multiple dimensions; and configuring the question-and-answer text pair to provide a response text when the evaluation result meets preset requirements.

[0006] A second aspect of this application provides an information processing apparatus, comprising:

[0007] The acquisition module is used to acquire question-and-answer text pairs, which include the question text and the response text to the question text.

[0008] The analysis module is used to perform multi-dimensional analysis on the question and response texts to obtain the evaluation results of the question and answer text pairs on each of the multiple dimensions.

[0009] The configuration module is used to configure question-and-answer text pairs when the evaluation results meet preset requirements, so as to provide a response text when a question text is received.

[0010] A third aspect of this application provides an electronic device, including: a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the information processing method of the first aspect.

[0011] This application embodiment is applied to a consultation scenario in the e-commerce field. The provided information processing method includes: obtaining question-and-answer text pairs, each including a question text and a response text; performing multi-dimensional analysis on the question text and response text to obtain evaluation results for each dimension of the question-and-answer text pair; and configuring the question-and-answer text pair to provide a response text when the evaluation results meet preset requirements. In this application embodiment, by performing multi-dimensional analysis on the question-and-answer text and configuring the question-and-answer text pair when the evaluation results meet preset requirements, the quality of the configured question-and-answer text pair can be improved. Attached Figure Description

[0012] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments of this application and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0013] Figure 1 A flowchart illustrating the steps of an information processing method provided for an exemplary embodiment of this application;

[0014] Figure 2 A schematic diagram of various dimensions provided for an exemplary embodiment of this application;

[0015] Figure 3 A flowchart illustrating the steps of another information processing method provided as an exemplary embodiment of this application;

[0016] Figure 4 A schematic diagram illustrating yet another information processing method provided for an exemplary embodiment of this application;

[0017] Figure 5 A structural block diagram of an information processing apparatus provided for an exemplary embodiment of this application;

[0018] Figure 6 This is a schematic diagram of the structure of an electronic device provided as an exemplary embodiment of this application. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0020] The information processing method provided in this application is applied to consultation scenarios in the e-commerce field. The server side is configured with an intelligent customer service system to automatically respond to user inquiries, thereby saving costs for merchants while resolving user issues. However, due to differences in merchants' capabilities and understanding of the intelligent customer service system, the quality of the question-and-answer text pairs configured by merchants on the intelligent customer service system varies. This often results in the intelligent customer service system providing users with poor-quality responses, failing to resolve buyer issues and interfering with accurate responses.

[0021] Furthermore, when configuring question-and-answer text pairs, related technologies mainly evaluate the semantic matching degree of the question and response texts. For example, large neural networks such as BERT (a pre-trained language representation model) are used to model and determine whether the question and response texts are appropriate. However, this approach has two main drawbacks. One is that it ignores whether the question and response texts meet certain criteria across various dimensions, such as conciseness and marketing appeal. The other is that it fails to provide merchants with clear evaluation results, making it difficult for them to identify any inappropriate aspects in the currently configured question-and-answer text pairs.

[0022] In e-commerce consultation scenarios, the aforementioned technologies suffer from poor-quality configured question-and-answer text pairs. This application provides an information processing method that includes: acquiring question-and-answer text pairs, each pair comprising a question text and a response text; performing multi-dimensional analysis on the question text and response text to obtain evaluation results for each dimension of the question-and-answer text pair; and configuring the question-and-answer text pair to provide a response text upon receiving the question text, provided the evaluation results meet preset requirements. In this application embodiment, by performing multi-dimensional analysis on the question-and-answer text and configuring the question-and-answer text pair only when the evaluation results meet preset requirements, the quality of the configured question-and-answer text pairs can be improved.

[0023] In this embodiment, the information processing method can be an overall information processing method implemented using a cloud computing system. Furthermore, the server executing the information processing method can be a cloud server, so as to leverage the advantages of cloud resources to run various neural network models; relative to the cloud, the information processing method can also be applied to conventional servers or server arrays and other server-side devices, without limitation.

[0024] Furthermore, one application scenario of this application embodiment is as follows: a merchant opens a virtual store on the current platform to sell goods, and a user purchases the goods online through the current platform. When purchasing goods, the user needs to send some questions to inquire about product information. Typically, the product information inquired about by users has standardized answers, such as product size, product color, and product usage. To save costs, the merchant configures corresponding answer settings for different questions on the current platform, so that when a user asks a question, the answer can be automatically provided. This application embodiment proposes a multi-dimensional evaluation method for question text and response text in e-commerce customer service, which can provide more targeted improvement suggestions when merchants configure question-and-answer text pairs, further enhancing user satisfaction with the dialogue.

[0025] The technical solutions provided by the various embodiments of this application are described in detail below with reference to the accompanying drawings.

[0026] Figure 1 A flowchart illustrating the steps of an information processing method provided for an exemplary embodiment of this application. Figure 1 The information processing method shown includes the following steps:

[0027] S101, retrieve the question-and-answer text pair.

[0028] The question-and-answer text pairs consist of the question text and the response text to the question text. These text pairs are sent by the merchant and / or mined from historical conversation information.

[0029] In this embodiment of the application, the merchant will configure some question-and-answer text pairs in the server for the products they sell, so that when the server receives the user's question text, it can quickly and efficiently send the corresponding reply text to the user.

[0030] S102, Perform multi-dimensional analysis on the question text and response text to obtain the evaluation results of the question and answer text pair on each of the multiple dimensions.

[0031] The multiple dimensions include: multiple first dimensions corresponding to the question text and multiple second dimensions corresponding to the response text, as shown in the reference. Figure 2The first dimension includes dimension 21, which corresponds only to the question text, and dimension 22, which is shared by both the question and response texts. The second dimension includes dimension 23, which corresponds only to the response text, and dimension 22, which is shared by both the question and response texts. Specifically, the first dimension includes at least one of the following: clarity of intent, singularity of intent, fluency, whether it contains sensitive words, whether it contains special characters, and whether it contains typos. The second dimension includes at least one of the following: fluency, whether it contains sensitive words, whether it contains special characters, whether it contains typos, whether it is colloquial, conciseness, sufficiency, completeness, semantic relevance, timeliness, universality, diversity, accuracy, marketing appeal, professionalism, credibility, and whether it guides users to human assistance.

[0032] Referring to Table 1, this example illustrates whether the evaluation results of the question text meet the preset requirements in multiple first dimensions.

[0033] Table 1

[0034]

[0035] Referring to Table 2, examples illustrate whether the evaluation results of the response text meet the preset requirements in multiple second dimensions. Furthermore, examples of response text performance in the second dimensions—fluency, presence of sensitive words, presence of special characters, and presence of typos—can be found in Table 1; they will not be repeated in Table 2.

[0036] Table 2

[0037]

[0038]

[0039]

[0040]

[0041] In this embodiment of the application, multi-dimensional analysis of the question text and the response text includes: analyzing a portion of the first dimensions of the question text in multiple first dimensions in Table 1, and / or analyzing a portion of the second dimensions of the response text in multiple second dimensions in Table 2.

[0042] Furthermore, for example, the specific dimensions for analyzing question and response texts can be set according to actual needs. For instance, all question texts can be analyzed for clarity of intent, singularity of intent, fluency, presence of sensitive words, presence of special characters, and presence of typos. Response texts can be analyzed for fluency, presence of sensitive words, presence of special characters, presence of typos, conciseness, sufficiency, completeness, semantic relevance, accuracy, professionalism, and credibility. During major promotional events, timeliness analysis can be added to response texts, and marketing analysis can be added to response texts targeting specific products. This application's embodiment improves analysis efficiency by selectively performing multi-dimensional analysis on question and response texts.

[0043] Furthermore, after conducting multi-dimensional analysis, each dimension has an evaluation result. See Tables 1 and 2 for details.

[0044] Furthermore, multi-dimensional analysis can be conducted using pre-trained dimensional evaluation models. For example, the question and response texts can be input into the dimensional evaluation model, which outputs the evaluation results of the question-and-answer text pairs on each of the multiple dimensions. The dimensional evaluation model can employ both rule-based and algorithmic analysis methods to perform multi-dimensional analysis on the question and response texts. Specifically, rule-based analysis is used for dimensions such as intent singularity, whether sensitive words are included, whether special characters are included, timeliness, diversity, and whether it guides users to human assistance. Algorithmic analysis is used for dimensions such as intent clarity, fluency, whether there are typos, fluency, whether the expression is colloquial, conciseness, sufficiency, completeness, semantic relevance, universality, accuracy, marketing relevance, professionalism, and credibility.

[0045] The rule analysis method specifically involves comparison. For example, regarding the singularity of intent, it can be determined by comparing whether there are multiple aspects to the question text. For instance, in the question text "Hello, my mother is over 50 years old, and I want to buy this product for her. Is this product suitable for middle-aged and elderly users? Is it suitable for someone over 50 years old?", the question text includes two aspects: one is "Is this product suitable for middle-aged and elderly users?" and the other is "Is it suitable for someone over 50 years old?". Therefore, it can be determined that the intent of the question text is not singular.

[0046] Furthermore, the algorithmic analysis method includes analyzing different dimensions using corresponding algorithms. For example, analyzing conciseness involves extracting keywords from the question text, inputting these keywords, the response text, and the dialogue scenario into a GBDT (Ensemble Tree Decision Tree) model for analysis, and obtaining an evaluation result on whether the response text is concise. The GBDT model is pre-trained. In this embodiment, the dialogue scenario can be a pre-sales scenario, a transportation scenario, or a after-sales scenario, etc.

[0047] Furthermore, for dimensions that require rule-based analysis, rule-based analysis methods can be pre-configured; for dimensions that require algorithmic analysis, different sub-evaluation models can be pre-trained for different dimensions. These rule-based analysis methods and sub-evaluation models can then be integrated into a dimensional evaluation model to perform multi-dimensional analysis on question and response texts.

[0048] In this embodiment of the application, by performing multi-dimensional analysis on the question text and the response text, the evaluation results of the question-and-answer text pair in each of the multiple dimensions are obtained, which can achieve white-boxing, so that merchants can clearly know in which dimension the question-and-answer text pair has a problem, and then make modifications to improve the quality of the configured question-and-answer text pair.

[0049] S103, if the evaluation results meet the preset requirements, configure question-answer text pairs to provide a response text when a question text is received.

[0050] Specifically, question-and-answer text pairs can be configured when the evaluation results in each dimension meet the preset requirements, or when more than a preset proportion of the evaluation results meet the preset requirements.

[0051] In this embodiment, the question-and-answer text pairs are configured in the server to provide automated services to users. For example, when the server receives a question text sent by a user, it sends the corresponding pre-configured reply text to the user.

[0052] The question-and-answer text pairs are sent by the merchant. The configuration of these text pairs, assuming the evaluation results meet preset requirements, includes: if the evaluation results of the question-and-answer text pairs on the target dimension do not meet the preset requirements, then sending the evaluation results for the target dimension to the merchant to obtain question-and-answer text pairs that meet the preset requirements. The evaluation results are used to instruct adjustments to the question-and-answer text pairs based on the evaluation results of the target dimension, which is one of multiple dimensions; if the evaluation results of the question-and-answer text pairs on each dimension meet the preset requirements, then configuring the question-and-answer text pairs.

[0053] Specifically, after performing multi-dimensional analysis on the question-and-answer text pairs, if the evaluation result in a certain dimension (target dimension) does not meet the requirements, the evaluation result for the target dimension is sent to the merchant. The merchant displays the evaluation result for that target dimension so that the merchant can adjust the question-and-answer text pairs. The merchant sends the adjusted question-and-answer text pairs to the server, and the server performs multi-dimensional analysis on the adjusted question-and-answer text pairs to obtain evaluation results. This process is repeated multiple times until the evaluation result of the adjusted question-and-answer text pairs meets the preset requirements. Then, the question-and-answer text pairs that meet the preset requirements are configured. In an optional embodiment, if the merchant does not adjust the question-and-answer text pairs that do not meet the preset requirements, they can directly configure the question-and-answer text pairs.

[0054] Furthermore, if the evaluation results of the question and answer text in the target dimension do not meet the preset requirements, the evaluation results of the target dimension are sent to the merchant, including: if the evaluation results of the question text in the first dimension do not meet the preset requirements, the evaluation results of the first dimension are sent to the merchant to obtain question text that meets the preset requirements, and the evaluation results are used to instruct the adjustment of the question text based on the evaluation results of the first dimension; if the evaluation results of the response text in the second dimension do not meet the preset requirements, the evaluation results of the second dimension are sent to the merchant to obtain response text that meets the preset requirements, and the evaluation results are used to instruct the adjustment of the response text based on the evaluation results of the second dimension.

[0055] In this embodiment of the application, if the evaluation result of the question text does not meet the preset requirements, the question text is adjusted; if the evaluation result of the response text does not meet the preset requirements, the response text is adjusted.

[0056] This application primarily addresses the problem in the field of intelligent e-commerce customer service where users often fail to receive adequate service during inquiries due to the large volume and inconsistent quality of question-and-answer text pairs configured by merchants. This application employs a white-box approach, proposing a multi-dimensional evaluation scheme for e-commerce customer service question-and-answer text pairs. This scheme provides merchants with more targeted adjustment suggestions, thereby further improving user satisfaction during conversations.

[0057] Figure 3 A flowchart illustrating the steps of training an image recognition model, provided as an exemplary embodiment of this application. Figure 3 As shown, the specific steps include:

[0058] S301, retrieve historical dialogue information.

[0059] Among them, historical dialogue information refers to the dialogue information between users and customer service. Specifically, historical dialogue information is the dialogue information between users and customer service in the past.

[0060] S302, Analyze historical dialogue information to obtain question-and-answer text pairs.

[0061] The question-and-answer text pair includes: the question text and multiple response texts to the question text.

[0062] In this embodiment of the application, historical dialogue information is summarized and analyzed to obtain multiple question-and-answer text pairs. Each question-and-answer text pair includes a question text and multiple response texts for that question text.

[0063] For example: For the question text "I am very prone to allergies, is this product suitable for me?", reply text A is "Yes, it is suitable, you can buy it and try it out", reply text B is "Dear, the ingredients of this product are all plant extracts, pure natural and additive-free, without allergenic ingredients, suitable for all skin types", reply text C is "Suitable", and reply text D is "Dear, the ingredients of this product are very gentle and not likely to cause allergies".

[0064] S303, perform multi-dimensional analysis on the question text and response text to obtain the evaluation results of the question and answer text pair on each of the multiple dimensions.

[0065] In the embodiments of this application, when the question-and-answer text pair includes the question text and multiple response texts, multi-dimensional analysis can be performed on the question text and each response text. The specific analysis method is as described above and will not be repeated here.

[0066] S304, for each of the multiple response texts, determine the number of evaluation results that meet the preset requirements.

[0067] Here, we determine the evaluation results of each response text across multiple dimensions. Each dimension corresponds to one evaluation result, so each response text corresponds to multiple evaluation results. We then determine the number of evaluation results that meet the preset requirements among these multiple evaluation results.

[0068] For example, for the five response texts A, B, C, and D, each response text has corresponding evaluation results across multiple dimensions. These dimensions include: fluency, presence of sensitive words, presence of special characters and typos, conciseness, sufficiency, and credibility (7 dimensions in total). Response text A meets the preset requirements in 4 out of 7 aspects, response text B in 6 out of 7 aspects, response text C in 2 out of 7 aspects, and response text D in 3 out of 7 aspects.

[0069] S305, Sort the multiple reply texts according to their number to obtain the sort number of each reply text.

[0070] Among them, the more items there are, the smaller the sorting number, and the smaller the sorting number, the better the quality of the reply text.

[0071] For example, referring to the above, the reply text with the sort number 1 is B, the reply text with the sort number 2 is A, the reply text with the sort number 3 is D, and the reply text with the sort number 4 is C.

[0072] Furthermore, based on the number of responses, the multiple response texts are sorted to obtain the sorting number of each response text, including: obtaining the dialogue scenario corresponding to each response text; and sorting the multiple response texts belonging to the same dialogue scenario based on the number of responses to obtain the sorting number of each response text in the corresponding dialogue scenario.

[0073] The dialogue scenarios include pre-sales, shipping, and after-sales scenarios. For example, a dialogue scenario with reply texts A and B is a pre-sales scenario, while a dialogue scenario with reply texts C and D is an after-sales scenario. In the pre-sales scenario, reply text A has a sorting number of 2, and reply text B has a sorting number of 1. In the after-sales scenario, reply text D has a sorting number of 1, and reply text C has a sorting number of 2.

[0074] S306, Configure the question text, multiple response texts, and the sorting number corresponding to each response text, so that when the question text is received, the response text is fed back according to the sorting number.

[0075] If a question text is received from a user, the system can determine the reply text with the smallest sort number from among multiple reply texts and send it to the user.

[0076] In one optional embodiment, the question text and the reply text with the smallest sorting number can be configured so that when the question text sent by the user is received, the reply text with the smallest sorting number can be sent to the user.

[0077] Furthermore, the configuration includes the question text, multiple response texts, and the sorting number corresponding to each response text. This includes configuring the question text, multiple response texts, the dialogue scenario corresponding to each response text, and the sorting number of each response text in the corresponding dialogue scenario. When the question text is received in the dialogue scenario, the response text is fed back according to the sorting number in the dialogue scenario.

[0078] If a user sends a question text, and the current dialogue scenario is determined to be a pre-sales scenario, then the reply text with the smallest sorting number is sent to the user within the pre-sales scenario.

[0079] In one optional embodiment, the question text, the reply text with the smallest sorting number in each dialogue scenario, and the corresponding dialogue scenario can be configured. For example, the question text, reply text B, and the pre-sales scenario can be configured, and the reply text D and the after-sales scenario can be configured. Then, when a question text sent by a user is received in the pre-sales scenario, reply text B can be sent to the user; when a question text sent by a user is received in the after-sales scenario, reply text D can be sent to the user.

[0080] In this embodiment of the application, dialogue text pairs mined from historical dialogue information and dialogue text pairs sent by merchants can be uniformly configured on the server for online use.

[0081] The proposed implementation addresses the problem in the field of intelligent e-commerce customer service where users often fail to receive adequate service due to the large volume and inconsistent quality of question-and-answer text pairs configured by merchants. This implementation uses a white-box approach to propose a multi-dimensional evaluation scheme for e-commerce customer service question-and-answer text pairs, providing merchants with more targeted adjustment suggestions and ultimately improving user satisfaction.

[0082] Reference Figure 4 After configuring the question text, multiple response texts, the dialogue scenario corresponding to each response text, and the sorting number of each response text in its corresponding dialogue scenario, the application steps for the configured question-answer text pairs are as follows:

[0083] S401, Receive the current problem text sent by the user.

[0084] If a user needs to consult with a merchant about a product, the user sends the current question text to the server.

[0085] For example, the current question text is "I am prone to allergies, is this product suitable for me?"

[0086] S402, Determine the target dialogue scenario corresponding to the current question text.

[0087] The server determines whether the target dialogue scenario corresponding to the current question text is a pre-sales scenario, a transportation scenario, or an after-sales scenario.

[0088] S403, Based on the current question text, determine at least one current response text corresponding to the current question text in the target dialogue scenario, as well as the sorting number of each current response text.

[0089] Specifically, if the target dialogue scenario is a pre-sales scenario, then in this scenario, the current question text corresponds to two current response texts: one is response text A, with a sorting number of 2, and the other is response text B, with a sorting number of 1. If the current question text corresponds to only one current response text, which is response text B, then that current response text is determined, and its corresponding sorting number does not need to be determined.

[0090] S404, determine the current reply text with the smallest sorting number as the target reply text.

[0091] S405, send the target response text to the user.

[0092] In this embodiment of the application, the current reply text with the smallest sorting number is considered to be of higher quality, and this current reply text is then sent to the user as the target reply text.

[0093] In one alternative embodiment, if there is only one current reply text, then that current reply text is sent to the user as the target reply text.

[0094] S406 If there is no corresponding reply text for the current problem text, send the current problem text to the merchant.

[0095] In this embodiment of the application, if the response text for the current problem text has not been configured in advance, the system switches from automated mode to manual mode and sends the current problem text to the merchant so that customer service can respond to the current problem text.

[0096] S407: Receive real-time response text from the merchant in response to the current issue text.

[0097] S408 performs multi-dimensional analysis on the real-time response text to obtain the current evaluation result of the real-time response text in each of the multiple dimensions.

[0098] In this embodiment of the application, the analysis method for multi-dimensional analysis of real-time reply text is the same as described above, and will not be repeated here.

[0099] S409: If the current evaluation result in the target dimension does not meet the preset requirements, send the current evaluation result in the target dimension to the merchant to obtain a real-time response text that meets the preset requirements.

[0100] The current evaluation result is used to indicate how to adjust the real-time response text based on the current evaluation result of the target dimension.

[0101] S410, if the current evaluation results in each dimension meet the preset requirements, then send a real-time response text to the user and configure the current question text and the current response text.

[0102] In this embodiment, real-time multi-dimensional analysis of the real-time response text can be performed, thereby improving the quality of the real-time response text and further enhancing user satisfaction with the conversation.

[0103] The proposed implementation addresses the problem in the field of intelligent e-commerce customer service where users often fail to receive adequate service due to the large volume and inconsistent quality of question-and-answer text pairs configured by merchants. This implementation uses a white-box approach to propose a multi-dimensional evaluation scheme for e-commerce customer service question-and-answer text pairs, providing merchants with more targeted adjustment suggestions and ultimately improving user satisfaction.

[0104] In the embodiments of this application, reference is made to Figure 5 In addition to providing an information processing method, an information processing apparatus 50 is also provided, which includes:

[0105] The acquisition module 51 is used to acquire question-and-answer text pairs, which include: question text and response text to the question text;

[0106] Analysis module 52 is used to perform multi-dimensional analysis on question text and response text to obtain the evaluation results of question and answer text pairs on each of the multiple dimensions;

[0107] Configuration module 53 is used to configure question-answer text pairs when the evaluation results meet preset requirements, so as to provide a response text when a question text is received.

[0108] In one optional embodiment, the question-and-answer text pair is sent by the merchant, and the configuration module 53 is specifically used to: if the evaluation result of the question-and-answer text pair on the target dimension does not meet the preset requirements, then send the evaluation result of the target dimension to the merchant to obtain a question-and-answer text pair that meets the preset requirements. The evaluation result is used to indicate the adjustment of the question-and-answer text pair according to the evaluation result of the target dimension, where the target dimension is a dimension among multiple dimensions; if the evaluation result of the question-and-answer text pair on each dimension meets the preset requirements, then configure the question-and-answer text pair.

[0109] In one optional embodiment, the multiple dimensions include: multiple first dimensions corresponding to the question text and multiple second dimensions corresponding to the response text. When the configuration module 53 sends the evaluation result of the target dimension to the merchant if the evaluation result of the question and answer text does not meet the preset requirements, it is specifically used to: send the evaluation result of the first dimension to the merchant if the evaluation result of the question text does not meet the preset requirements, so as to obtain the question text that meets the preset requirements. The evaluation result is used to indicate the adjustment of the question text according to the evaluation result of the first dimension; and send the evaluation result of the second dimension to the merchant if the evaluation result of the response text does not meet the preset requirements, so as to obtain the response text that meets the preset requirements. The evaluation result is used to indicate the adjustment of the response text according to the evaluation result of the second dimension.

[0110] In an optional embodiment, the first dimension includes at least one of the following: clarity of intent, singularity of intent, fluency, whether it contains sensitive words, whether it contains special characters, and whether it contains typos; the second dimension includes at least one of the following: fluency, whether it contains sensitive words, whether it contains special characters, whether it contains typos, whether it is a colloquial expression, conciseness, sufficiency, completeness, semantic relevance, timeliness, universality, diversity, accuracy, whether it is marketing-oriented, professionalism, credibility, and whether it guides users to human assistance.

[0111] In an optional embodiment, the acquisition module 51 is specifically used to: acquire historical dialogue information, which is the dialogue information between the user and customer service; analyze the historical dialogue information to obtain question-and-answer text pairs, which include: question text and multiple response texts to the question text.

[0112] In an optional embodiment, the configuration module 53 is specifically used to: determine the number of evaluation results that meet preset requirements for each of the multiple reply texts; sort the multiple reply texts according to the number to obtain the sorting number of each reply text, wherein the more reply texts there are, the smaller the sorting number is, and the smaller the sorting number is, the better the quality of the reply text; and configure the question text, the multiple reply texts and the sorting number corresponding to each reply text so that when the question text is received, the reply text is fed back according to the sorting number.

[0113] In an optional embodiment, when the configuration module 53 sorts multiple reply texts according to the number of replies to obtain the sorting number of each reply text, it is specifically used to obtain the dialogue scenario corresponding to each reply text; sort multiple reply texts belonging to the same dialogue scenario according to the number of replies to obtain the sorting number of each reply text in the corresponding dialogue scenario; and configure the question text, multiple reply texts, and the sorting number corresponding to each reply text, including configuring the question text, multiple reply texts, the dialogue scenario corresponding to each reply text, and the sorting number of each reply text in the corresponding dialogue scenario, so that when the question text is received in the dialogue scenario, the reply text is fed back according to the sorting number in the dialogue scenario.

[0114] In an optional embodiment, the information processing device 50 further includes:

[0115] A processing module (not shown) is used to receive the current question text sent by the user; determine the target dialogue scenario corresponding to the current question text; determine at least one current reply text corresponding to the current question text in the target dialogue scenario and the sorting number of each current reply text based on the current question text; determine the current reply text with the smallest sorting number as the target reply text; and send the target reply text to the user.

[0116] In an optional embodiment, the processing module is further configured to: send the current question text to the merchant if no corresponding response text exists; receive the real-time response text sent by the merchant for the current question text; perform multi-dimensional analysis on the real-time response text to obtain the current evaluation result of the real-time response text in each of the multiple dimensions; if the current evaluation result in the target dimension does not meet the preset requirements, send the current evaluation result of the target dimension to the merchant to obtain a real-time response text that meets the preset requirements, wherein the current evaluation result is used to indicate the adjustment of the real-time response text according to the current evaluation result of the target dimension; and if the current evaluation result in each dimension meets the preset requirements, send the real-time response text to the user and configure the current question text and the current response text.

[0117] The information processing apparatus provided in this application embodiment can improve the quality of configured question-and-answer text pairs by performing multi-dimensional analysis on the question-and-answer texts and configuring question-and-answer text pairs when the evaluation results meet preset requirements.

[0118] Furthermore, in some of the processes described in the above embodiments and accompanying drawings, multiple operations appear in a specific order. However, it should be clearly understood that these operations may not be executed in the order they appear herein, or may be executed in parallel. The sequence numbers are merely used to distinguish different operations, and the sequence numbers themselves do not represent any execution order. Additionally, these processes may include more or fewer operations, and these operations may be executed sequentially or in parallel. It should be noted that the descriptions such as "first," "second," etc., in this document are used to distinguish different messages, devices, modules, etc., and do not represent a sequential order, nor do they limit "first" and "second" to different types.

[0119] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an example embodiment of this application. Figure 6 As shown, the electronic device 60 includes a processor 61 and a memory 62 communicatively connected to the processor 61, the memory 62 storing computer-executed instructions.

[0120] The processor executes computer execution instructions stored in the memory to implement the information processing method provided in any of the above method embodiments. The specific functions and technical effects that can be achieved will not be elaborated here.

[0121] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the information processing method provided in any of the above method embodiments.

[0122] This application also provides a computer program product, which includes a computer program stored in a readable storage medium. At least one processor of an electronic device can read the computer program from the readable storage medium, and the at least one processor executes the computer program to cause the electronic device to perform the information processing method provided in any of the above method embodiments.

[0123] In the embodiments provided in this application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between systems or units may be electrical, mechanical, or other forms.

[0124] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0125] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in a combination of hardware and software functional units.

[0126] The integrated units implemented as software functional units described above can be stored in a computer-readable storage medium. These software functional units, stored in a storage medium, include several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute some steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0127] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional modules is merely an example. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the system can be divided into different functional modules to complete all or part of the functions described above. The specific working process of the system described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0128] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this application are indicated by the following claims.

[0129] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. An information processing method, characterized in that, include: Obtain question-and-answer text pairs, which are provided by the merchant; The question-and-answer text pair includes: the question text and the response text to the question text; The question text and the response text are analyzed in multiple dimensions to obtain the evaluation results of the question and answer text pair in each of the multiple dimensions; the multiple dimensions include: multiple first dimensions corresponding to the question text and multiple second dimensions corresponding to the response text. If the evaluation result meets the preset requirements, configure the question-and-answer text pair to provide the response text when a question text is received; The step of configuring the question-and-answer text pair when the evaluation result meets preset requirements includes: If the evaluation result of the question text in the first dimension does not meet the preset requirements, the evaluation result of the first dimension is sent to the merchant to obtain the question text that meets the preset requirements. The evaluation result is used to indicate that the question text should be adjusted according to the evaluation result of the first dimension. If the evaluation result of the response text in the second dimension does not meet the preset requirements, the evaluation result of the second dimension is sent to the merchant to obtain a response text that meets the preset requirements. The evaluation result is used to indicate that the response text should be adjusted according to the evaluation result of the second dimension. If the evaluation results of the question-and-answer text pair in each dimension meet the preset requirements, then the question-and-answer text pair is configured.

2. The information processing method according to claim 1, characterized in that, The first dimension includes at least one of the following: clarity of intent, singularity of intent, fluency, whether it contains sensitive words, whether it contains special characters, and whether it contains typos; the second dimension includes at least one of the following: fluency, whether it contains sensitive words, whether it contains special characters, whether it contains typos, whether it is a colloquial expression, conciseness, sufficiency, completeness, semantic relevance, timeliness, universality, diversity, accuracy, professionalism, and credibility.

3. The information processing method according to any one of claims 1 or 2, characterized in that, The acquisition of question-and-answer text pairs includes: Obtain historical dialogue information, which is the dialogue information between the user and customer service; Analyzing the historical dialogue information yields the question-and-answer text pairs, which include the question text and multiple response texts to the question text.

4. The information processing method according to claim 3, characterized in that, The step of configuring the question-and-answer text pair when the evaluation result meets preset requirements includes: For each of the multiple response texts, determine the number of evaluation results that meet the preset requirements; Based on the number of responses, the multiple response texts are sorted to obtain a sorting number for each response text. The more responses there are, the smaller the sorting number becomes, and the smaller the sorting number, the better the quality of the response text. Configure the question text, the multiple response texts, and the sorting number corresponding to each response text so that when the question text is received, the response text is fed back according to the sorting number.

5. The information processing method according to claim 4, characterized in that, The step of sorting the multiple reply texts according to the number to obtain a sorting number for each reply text includes: Get the dialogue scenario corresponding to each reply text; Based on the number, sort the multiple reply texts belonging to the same dialogue scenario to obtain the sorting number of each reply text in the corresponding dialogue scenario; The configuration of the question text, the plurality of reply texts, and the sorting number corresponding to each reply text includes: configuring the question text, the plurality of reply texts, the dialogue scenario corresponding to each reply text, and the sorting number of each reply text in the corresponding dialogue scenario, so that when the question text is received in the dialogue scenario, the reply text is fed back according to the sorting number in the dialogue scenario.

6. The information processing method according to claim 5, characterized in that, After configuring the question-and-answer text pair when the evaluation result meets the preset requirements, the method further includes: Receive the current problem text sent by the user; Determine the target dialogue scenario corresponding to the current question text; Based on the current question text, determine at least one current response text corresponding to the current question text in the target dialogue scenario and the sorting number of each current response text; The current reply text with the smallest sorting number is determined as the target reply text; Send the target response text to the user.

7. The information processing method according to claim 6, characterized in that, Also includes: If there is no corresponding reply text for the current question text, then send the current question text to the merchant; Receive real-time response text from the merchant in response to the current issue text; The real-time response text is analyzed in multiple dimensions to obtain the current evaluation result of the real-time response text in each of the multiple dimensions. If the current evaluation result in the target dimension does not meet the preset requirements, the current evaluation result in the target dimension is sent to the merchant to obtain a real-time response text that meets the preset requirements. The current evaluation result is used to instruct the real-time response text to be adjusted according to the current evaluation result in the target dimension. If the current evaluation results in each dimension meet the preset requirements, then the real-time response text is sent to the user, and the current question text and the current response text are configured.

8. An information processing device, characterized in that, include: The acquisition module is used to acquire question-and-answer text pairs, which are provided by the merchant; The question-and-answer text pair includes: the question text and the response text to the question text; The analysis module is used to perform multi-dimensional analysis on the question text and the response text to obtain the evaluation result of the question-answer text pair on each of the multiple dimensions; the multiple dimensions include: multiple first dimensions corresponding to the question text and multiple second dimensions corresponding to the response text. A configuration module is used to configure the question-and-answer text pair when the evaluation result meets preset requirements, so as to provide the response text when a question text is received; The configuration module is specifically used to send the evaluation result of the first dimension to the merchant if the evaluation result of the problem text in the first dimension does not meet the preset requirements, so as to obtain the problem text that meets the preset requirements. The evaluation result is used to indicate that the problem text should be adjusted according to the evaluation result of the first dimension. If the evaluation result of the response text in the second dimension does not meet the preset requirements, the evaluation result of the second dimension is sent to the merchant to obtain a response text that meets the preset requirements. The evaluation result is used to indicate that the response text should be adjusted according to the evaluation result of the second dimension. If the evaluation results of the question-and-answer text pair in each dimension meet the preset requirements, then the question-and-answer text pair is configured.

9. An electronic device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the information processing method according to any one of claims 1 to 7.

Images