Text generation method and apparatus

Abstract

Embodiments of the present specification provide a text generation method and device, wherein the text generation method comprises: obtaining dialogue text associated with project participation behavior, and determining target question text corresponding to the dialogue text; determining to-be-processed text based on the dialogue text and auxiliary text corresponding to the dialogue text, wherein the auxiliary text is associated with the project participation behavior; determining target answer text associated with the target question text based on the to-be-processed text; and generating dialogue summary text corresponding to the dialogue text based on the target question text and the target answer text. According to the dialogue text, the target question text is determined, and the information source is enriched by introducing the auxiliary text corresponding to the dialogue text. According to the auxiliary text and the dialogue text, the to-be-processed text is determined, the target answer text with higher accuracy can be obtained, the dialogue summary text generated finally is more fine-grained, and the project personnel can know the intention of the project participation user based on the dialogue summary text more quickly, thereby providing more convenient services for the user.

Description

Technical Field

[0001] This specification relates to the field of natural language processing, and particularly to text generation methods. One or more embodiments of this specification also relate to text generation apparatus, text generation system, a computing device, and a computer-readable storage medium. Background Technology

[0002] With the rapid development of computer and network technologies, telephone and internet-based customer service centers have become an important channel for businesses to interact with users. Currently, after a customer service representative completes a conversation with a user, the conversation is recorded according to a pre-defined summary, summarizing the user's inquiry. Subsequent calls from the same user do not require repeating the same question; the representative can understand the user's inquiry based on the previous summary. However, the above-mentioned process of summarizing user conversations often encounters the following problems: customer service personnel need to spend a significant amount of time filling out the summary record after the consultation; the summary record is inaccurate and cannot effectively improve the user's experience for future inquiries. Therefore, how to quickly and accurately summarize customer communication content is a problem that urgently needs to be solved. Summary of the Invention

[0003] In view of this, embodiments of this specification provide a text generation method. One or more embodiments of this specification also relate to a text generation apparatus, a text generation system, a computing device, a computer-readable storage medium, and a computer program, to address the technical deficiencies existing in the prior art.

[0004] According to a first aspect of the embodiments of this specification, a text generation method is provided, comprising:

[0005] Obtain the dialogue text of the related project participation behavior, and determine the target question text corresponding to the dialogue text;

[0006] The text to be processed is determined based on the dialogue text and the corresponding auxiliary text, wherein the auxiliary text is associated with the project participation behavior;

[0007] Based on the text to be processed, determine the target answer text associated with the target question text;

[0008] Based on the target question text and the target answer text, generate a dialogue summary text corresponding to the dialogue text.

[0009] According to a second aspect of the embodiments of this specification, a text generation method is provided, comprising:

[0010] Receive a text generation instruction submitted via a summary page for the dialogue text, wherein the summary page is associated with project participation behavior;

[0011] In response to the text generation instruction, the target question text corresponding to the dialogue text is determined, and the text to be processed is determined based on the dialogue text and the auxiliary text corresponding to the dialogue text, wherein the auxiliary text is associated with the project participation behavior;

[0012] Based on the text to be processed, determine the target answer text associated with the target question text;

[0013] Based on the target question text and the target answer text, a dialogue summary text corresponding to the dialogue text is generated, and the dialogue summary text is displayed through the summary page.

[0014] According to a third aspect of the embodiments of this specification, a text generation method is provided, comprising:

[0015] Obtain the text of the communication dialogue between the user and customer service, and determine the target question text corresponding to the communication dialogue text;

[0016] The text to be processed is determined based on the communication dialogue text and the corresponding auxiliary text, wherein the auxiliary text is the communication template text used by the customer service representative;

[0017] Based on the text to be processed, determine the target answer text associated with the target question text;

[0018] Based on the target question text and the target answer text, a customer service communication summary text corresponding to the communication dialogue text is generated.

[0019] According to a fourth aspect of the embodiments of this specification, a text generation apparatus is provided, comprising:

[0020] The acquisition module is configured to acquire the dialogue text of the related project participation behavior and determine the target question text corresponding to the dialogue text;

[0021] The first determining module is configured to determine the text to be processed based on the dialogue text and the auxiliary text corresponding to the dialogue text, wherein the auxiliary text is associated with the project participation behavior;

[0022] The second determining module is configured to determine the target answer text associated with the target question text based on the text to be processed;

[0023] The generation module is configured to generate a dialogue summary text corresponding to the dialogue text based on the target question text and the target answer text.

[0024] According to a fifth aspect of the embodiments of this specification, a text generation apparatus is provided, comprising:

[0025] The receiving module is configured to receive a text generation instruction submitted via a summary page for a dialogue text, wherein the summary page is associated with project participation behavior;

[0026] The first determining module is configured to, in response to the text generation instruction, determine the target question text corresponding to the dialogue text, and determine the text to be processed based on the dialogue text and the auxiliary text corresponding to the dialogue text, wherein the auxiliary text is associated with the project participation behavior;

[0027] The second determining module is configured to determine the target answer text associated with the target question text based on the text to be processed;

[0028] The generation module is configured to generate a dialogue summary text corresponding to the dialogue text based on the target question text and the target answer text, and to display the dialogue summary text through the summary page.

[0029] According to a sixth aspect of the embodiments of this specification, a text generation apparatus is provided, comprising:

[0030] The acquisition module is configured to acquire the text of the communication dialogue between the user and customer service, and determine the target question text corresponding to the communication dialogue text;

[0031] The first determining module is configured to determine the text to be processed based on the communication dialogue text and the auxiliary text corresponding to the communication dialogue text, wherein the auxiliary text is the communication template text used by the customer service representative;

[0032] The second determining module is configured to determine the target answer text associated with the target question text based on the text to be processed;

[0033] The generation module is configured to generate a customer service communication summary text corresponding to the communication dialogue text based on the target question text and the target answer text.

[0034] According to a seventh aspect of the embodiments of this specification, a text generation system is provided, comprising:

[0035] The client is used to store executable instructions for text display, and the server is used to store executable instructions for text generation; when the executable instructions for text display are executed by the client and when the executable instructions for text generation are executed by the server, the steps of the above-mentioned text generation method are implemented.

[0036] According to an eighth aspect of the embodiments of this specification, a computing device is provided, including a memory, a processor, and computer instructions stored in the memory and executable on the processor, wherein the processor executes the computer instructions to implement the steps of the text generation method.

[0037] According to a ninth aspect of an embodiment of this specification, a computer-readable storage medium is provided that stores computer instructions which, when executed by a processor, implement the steps of the text generation method.

[0038] According to a tenth aspect of an embodiment of this specification, a computer program is provided, wherein when the computer program is executed in a computer, it causes the computer to perform the steps of the above-described text generation method.

[0039] The text generation method provided in this specification includes: obtaining dialogue text associated with project participation behavior and determining the target question text corresponding to the dialogue text; determining a text to be processed based on the dialogue text and the auxiliary text corresponding to the dialogue text, wherein the auxiliary text is associated with the project participation behavior; determining a target answer text associated with the target question text based on the text to be processed; and generating a dialogue summary text corresponding to the dialogue text based on the target question text and the target answer text.

[0040] One embodiment of this specification implements the determination of the corresponding target question text based on the dialogue text, and introduces auxiliary text corresponding to the dialogue text to enrich the information source. The text to be processed is determined based on the auxiliary text and the dialogue text, so as to obtain a more accurate target answer text. The final generated dialogue summary text has a higher granularity, which enables project personnel to understand the intentions of project participants more quickly based on the dialogue summary text, thereby providing users with more convenient services. Attached Figure Description

[0041] Figure 1 This is a schematic diagram illustrating the effect of a text generation method provided in one embodiment of this specification.

[0042] Figure 2 This is a flowchart illustrating a text generation method provided in one embodiment of this specification;

[0043] Figure 3 This is a flowchart of a text generation method provided in another embodiment of this specification;

[0044] Figure 4 This is a flowchart illustrating the processing steps of a text generation method provided in one embodiment of this specification.

[0045] Figure 5 This is a flowchart of a text generation method provided in another embodiment of this specification;

[0046] Figure 6 This is a schematic diagram of the structure of a text generation device provided in one embodiment of this specification;

[0047] Figure 7This is a schematic diagram of the structure of a text generation device provided in another embodiment of this specification;

[0048] Figure 8 This is a schematic diagram of the structure of a text generation device provided in another embodiment of this specification;

[0049] Figure 9 This is a structural block diagram of a computing device provided in one embodiment of this specification. Detailed Implementation

[0050] Many specific details are set forth in the following description to provide a full understanding of this specification. However, this specification can be implemented in many other ways than those described herein, and those skilled in the art can make similar extensions without departing from the spirit of this specification. Therefore, this specification is not limited to the specific implementations disclosed below.

[0051] The terminology used in one or more embodiments of this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the one or more embodiments of this specification. The singular forms “a,” “described,” and “the” as used in one or more embodiments of this specification and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in one or more embodiments of this specification refers to any or all possible combinations including one or more of the associated listed items.

[0052] It should be understood that although the terms first, second, etc., may be used to describe various information in one or more embodiments of this specification, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, first may also be referred to as second without departing from the scope of one or more embodiments of this specification, and similarly, second may also be referred to as first. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to a determination."

[0053] First, the terms and concepts used in one or more embodiments of this specification will be explained.

[0054] Service Summary: After a customer calls in for consultation, customer service representatives need to summarize and describe the consultation process. When the user calls back, a new customer service representative can quickly locate the user's problem and the progress of the process, reducing the need for the user to repeat their description and improving the user experience.

[0055] Service Summary: When a user calls, the service summary describes the user's general problem and request. If there is already a processing progress, it will also include the processing progress of the current user.

[0056] SOP: Standard Operating Procedure. In this article, SOP specifically refers to the problem-solving process performed by customer service personnel.

[0057] ISO: Digital solutions, based on SOP upgrades, for smarter and more convenient operating procedures.

[0058] In the service field, after a user communicates with customer service, the customer service representative summarizes the communication, generating a service summary. This summary is used for the user's next inquiry, allowing the next customer service representative to quickly understand the user's past issues and progress, reducing repetitive descriptions and improving user experience. Currently, generating service summaries faces two main challenges. First, completing service summaries is time-consuming. Service personnel need to summarize and consolidate the communication records after the service ends, enabling the next customer service representative to quickly locate and resolve issues when the user inquires again. Currently, service summaries are manually completed, resulting in significant time and effort required, as customer service representatives need to manually summarize the customer's problem, needs, and provided solutions. Second, the accuracy of service summaries is low. Because existing service summaries are either manually completed or summarized using speech recognition models, the information is often inaccurate, potentially leading to misunderstandings of the user's intent and needs, thus negatively impacting the user's consultation experience. The current customer service system's service summary filling system uses a corpus recognition model to identify enumerated values. This method cannot cover a wide range of real-world business scenarios and the recognition is inaccurate, resulting in coarse-grained filling content that cannot provide accurate information for subsequent customer service personnel.

[0059] Based on this, this specification provides a text generation method for quickly generating more accurate service summaries and abstracts, thereby providing support for customers to quickly locate and resolve problems, and enabling better service to users. This specification also relates to a text generation device, a text generation system, a computing device, a computer-readable storage medium, and a computer program, which will be described in detail in the following embodiments.

[0060] Figure 1 This is a schematic diagram illustrating the effect of a text generation method provided according to an embodiment of this specification. Figure 1 In this context, project personnel can be understood as customer service personnel. After completing communication and service with the user, customer service personnel can use client applications, such as... Figure 1The computer terminal displays the communication record, i.e., the dialogue text. In practical applications, the dialogue text can be obtained by converting the communication audio. The customer service system on the client side, based on the obtained dialogue text, checks whether the customer service representative followed standard operating procedures during the communication. If so, it obtains the corresponding supplementary text, which details the customer's problem, handling method, and processing progress. Subsequently, a dialogue summary text, i.e., a service summary, can be directly generated based on the content of the supplementary text. If no supplementary text is available, speech recognition can be performed on the dialogue text, and a dialogue summary text can be generated based on the recognition results. Figure 1 In this method, if the dialogue summary text is generated based on the original dialogue text, the system can identify the user's original question text, "The merchant refuses to accept returns," and thus generate the target question text, "Quality issue, merchant refuses returns," as well as the target answer text, "Customer service intervened, please wait patiently." Based on ISO and SOP standard operating procedures, the system introduces a more accurate data source for the generation of dialogue summary texts, thereby improving the accuracy of the generated text. This allows subsequent project personnel to quickly understand the user's problem, the handling method, and the progress based on the generated dialogue summary text. In practical applications, the dialogue summary text can also include other information, such as the user's basic account information and communication information between customer service and the merchant. Using the text generation method provided in this manual, dialogue summary texts with higher retention rates can be generated even with the introduction of more accurate auxiliary text. This eliminates the need for project personnel to make significant adjustments to the dialogue summary text later, improving service efficiency and user experience.

[0061] Figure 2 A flowchart of a text generation method according to an embodiment of this specification is shown, including steps 202 to 208.

[0062] Step 202: Obtain the dialogue text of the related project participation behavior, and determine the target question text corresponding to the dialogue text.

[0063] Project participation behavior can be understood as a user's actions in participating in a project provided by the project provider, such as online shopping or online ticket purchase. If a user has questions about their participation after participating in a project, they will consult project personnel, who can be customer service staff of the project platform providing the project or service personnel from each service provider within the platform. Taking online shopping as an example, if a user buys a piece of clothing on platform A but the clothing hasn't been shipped, the project participation behavior is the user's purchase of the item on platform A. If the user wants to inquire about the shipment, they can ask the store's service personnel or directly ask platform A's customer service personnel. A communication record will be kept after both parties communicate. When the user communicates with customer service personnel online via a terminal, the communication record can be in text format. When the user communicates with customer service personnel by phone via a terminal, the communication record can be in audio format, which can then be converted into text through speech recognition. Regardless of the communication method, a dialogue text related to the project participation behavior can be generated, facilitating the subsequent generation of a service summary, i.e., a dialogue summary text. The dialogue summary text may include the target question text, which can be understood as the question the user wanted to ask, such as "the product has not been shipped".

[0064] In practice, to ensure fine-grained and accurate content, the target question text includes both the question type and the original question text, i.e., the user's exact words. In real-world applications, users often communicate with customer service to inquire about relevant issues, such as asking about product details before purchasing (e.g., the phone's camera resolution) or after-sales issues (e.g., why the phone won't turn on). Therefore, a single user's engagement may generate multiple dialogue texts. To improve communication efficiency, customer service personnel can use a service summary to understand the user's intent beforehand, quickly pinpoint the user's problem, and learn about the progress of previous issues. Therefore, generating a dialogue summary text based on the dialogue text improves the user experience.

[0065] In one embodiment of this specification, if a user wants to purchase amusement park tickets, they will ask customer service personnel some specific details about the tickets, such as whether the tickets can be used on weekdays and whether the tickets allow access to all attractions in the park. After the user's communication with customer service personnel ends, the customer service system will automatically obtain the dialogue text of this communication and determine the target question text corresponding to the dialogue text, which is "Ticket usage: Can this ticket be used on weekdays?"

[0066] Furthermore, relying solely on enumerated questions to determine a user's question may fail to accurately identify their intent. For instance, if a user wants to inquire about product delivery time, but the enumerated questions don't include any related to delivery time, the identified question might only be the product's manufacturing date, leading to misinterpretation and potentially causing customer service staff to misunderstand the user's intent. Therefore, the question text should include the user's original words to accurately describe the user's inquiry. Specifically, determining the target question text corresponding to the dialogue text includes: selecting an initial question text from a question text database based on the dialogue text; determining the original question text from the dialogue text based on the initial question text; merging the initial question text and the original question text; and obtaining the target question text based on the merged result.

[0067] The initial question text can be understood as a pre-defined question. The question text database stores historically asked questions, essentially an exhaustive list of manually enumerated values. The initial question text represents the question category. However, the initial question text is coarse-grained and may not accurately describe the user's actual problem. For example, if the initial question text is "return," but it doesn't reveal what the user wants to return or why, the original question text can be used to supplement it. A finer-grained question text is generated based on the initial and original question texts, making the information more accurate. The original question text can be understood as the user's exact words describing the problem, such as "I want to return this phone because I bought the wrong model."

[0068] In practical applications, to ensure more accurate content and accurately capture key information from the service process, the user's exact words can be added to the question text, helping customer service improve efficiency. After determining the initial question text, the most suitable descriptive text for the question can be identified from the entire conversation text—this is the original question text—which is then used to generate the subsequent target question text.

[0069] In a specific embodiment of this specification, an initial question text is selected from the question text database based on the dialogue text between the user and customer service. The initial question text is "ticket usage". Based on the determined initial question text, the original question text is selected from the dialogue text and the original question text is determined as "Can this ticket be used on holidays?". The initial question text and the original question text are merged, and the target question text is obtained based on the merging result. The target question text is "ticket usage: Can this ticket be used on holidays?".

[0070] In practical applications, to address the issue of low accuracy in corpus recognition, participation status information can be introduced when determining the initial question text. Specifically, the initial question text is selected from the question text database based on the dialogue text, including: obtaining the participation status information corresponding to the dialogue text; filtering the question text database based on the participation status information, and determining candidate question texts based on the filtering results; calculating the matching degree between each candidate question text and the dialogue text, and selecting the initial question text based on each matching degree.

[0071] Among them, participation status information can be understood as the status information of project participation behavior. For example, if a user consults before purchasing a product, the participation status information is pre-sales status. Based on this participation status information, pre-sales question texts can be filtered out from the question text database. The filtered question texts are candidate question texts. Further selection can be made based on the dialogue text to determine the initial question text.

[0072] In practice, when a user inquires after purchasing a product, the status information can include not only after-sales status information but also specific order information, such as "paid and awaiting shipment," or historical conversation summaries, including summaries of the user's previous communication with customer service. This allows for more accurate filtering of potential question texts. After identifying the potential question texts, the matching degree between each potential question text and the conversation text can be calculated. The matching degree can be understood as the relevance between the question text and the conversation text. A higher relevance indicates that the question text is more likely to match the user's intent, i.e., more relevant to the conversation text. Therefore, the initial question text can be determined based on the matching degree between each potential question text and the conversation text.

[0073] In a specific embodiment of this specification, the participation status information corresponding to the dialogue text is obtained. If the participation status information is a pre-sale status, then pre-sale questions can be filtered in the question text database first. Based on the filtering results, candidate question texts are determined, and the matching degree between each candidate question text and the dialogue text is calculated. The candidate question text with the highest matching degree is selected as the initial question text, which is "ticket usage".

[0074] Based on this, by introducing a user-generated text field to populate the problem, and using a combination of the initial problem text and the original problem text, the problem can be accurately and comprehensively described.

[0075] Step 204: Determine the text to be processed based on the dialogue text and the corresponding auxiliary text, wherein the auxiliary text is associated with the project participation behavior.

[0076] Among them, the text to be processed can be understood as the text that needs to be identified and processed later, and the auxiliary text can be understood as the operation process text of the SOP and ISO system. The auxiliary text includes different process nodes selected by the customer service during the service process. Each process node has a corresponding descriptive text, which can determine the final processing solution given to the user.

[0077] In practical applications, customer service representatives typically utilize an ISO or SOP guidance system during the service process. This can be understood as a lengthy tree structure. Based on the user's different needs and descriptions, the customer service representative selects different branch nodes, and the final leaf node represents the solution provided to the user. Since the auxiliary text is the descriptive text selected by the customer service representative step-by-step according to the user's description during communication, it has a higher priority than the dialogue text. The answer text can be directly determined based on the auxiliary text, eliminating the need to extract the answer text from the dialogue text and improving the efficiency of generating the dialogue summary text.

[0078] In practice, since a project participation behavior may involve multiple dialogue texts, and each dialogue text has a corresponding auxiliary text, the auxiliary text for this dialogue text needs to be determined based on the current dialogue text. Specifically, the determination of the auxiliary text includes: determining the dialogue identifier information of the dialogue text; and determining the auxiliary text corresponding to the dialogue text based on the dialogue identifier information.

[0079] Dialogue identification information can be understood as a unique identifier for dialogue text, such as a unique ID generated based on communication time and user information. The auxiliary text corresponding to the dialogue text can be determined based on the dialogue identification information.

[0080] In practical applications, customer service personnel may not use ISO or SOP operating procedures for communication in order to better serve users. In this case, the dialogue text will not have auxiliary text, which can be understood as the dialogue text being empty for mobile auxiliary text. However, since auxiliary text has higher priority than dialogue text, when determining the answer text, it is first necessary to identify whether the answer text exists in the auxiliary text. Specifically, the text to be processed is determined based on the dialogue text and the auxiliary text corresponding to the dialogue text, including: identifying the auxiliary text, and if the identification result determines that the answer text exists in the auxiliary text, then the auxiliary text is taken as the text to be processed.

[0081] In practice, after determining the dialogue text and its corresponding auxiliary text, the first step is to identify whether the auxiliary text contains answer text. Since the auxiliary text is generated by customer service personnel clicking through process steps, if the auxiliary text contains answer text, this answer text is directly used as the target answer text and filled into the dialogue summary text, without needing to perform further corpus recognition from the dialogue text. Therefore, if the auxiliary text is identified and, based on the identification results, contains answer text, it is treated as text to be processed. Subsequently, the target answer text associated with the target question text is determined based on the auxiliary text.

[0082] In another scenario, if the auxiliary text does not contain the answer text, the dialogue text is then identified. Specifically, the text to be processed is determined based on the dialogue text and the corresponding auxiliary text. This includes: identifying the auxiliary text; and if the identification result determines that the auxiliary text does not contain the answer text, then the dialogue text is taken as the text to be processed.

[0083] In practice, if customer service personnel do not communicate with users according to standard operating procedures and choose to input their own reply content, the auxiliary text for this dialogue will be empty. In order to determine the answer text, it is necessary to perform corpus recognition on the dialogue text, and the dialogue text will be used as the text to be processed in subsequent recognition.

[0084] Step 206: Determine the target answer text associated with the target question text based on the text to be processed.

[0085] Among them, based on the text to be processed determined under two different conditions, the target answer text associated with the target question text is determined.

[0086] In practical applications, when answer text exists in the auxiliary text, determining the target answer text associated with the target question text based on the text to be processed includes: using the answer text in the auxiliary text as the target answer text associated with the target question text. In other words, the answer text in the auxiliary text is directly used as the target question text.

[0087] In another scenario, when no answer text exists in the auxiliary text, determining the target answer text associated with the target question text based on the text to be processed includes: performing recognition processing on the dialogue text, determining the answer text in the dialogue text based on the recognition result, and using the answer text in the dialogue text as the target answer text associated with the target question text. That is, the dialogue text is used as the text to be processed, and corpus recognition processing is performed on the dialogue text. The answer text is determined based on the recognition result, and subsequently, the answer text identified in the dialogue text is used as the target answer text associated with the target question text.

[0088] To address the issue of coarse granularity in previous recognition methods, a preset recognition rule is employed. Specifically, the dialogue text is processed for recognition, and the answer text within the dialogue text is determined based on the recognition results. This answer text is then used as the target answer text associated with the target question text. The process includes: inputting the dialogue text into a guidance model for guidance recognition processing to obtain guidance answer text; and inputting the dialogue text into an action model for action recognition processing to obtain action answer text. Based on the guidance answer text and the action answer text, the answer text within the dialogue text is determined, and this answer text is used as the target answer text associated with the target question text.

[0089] Both the guidance model and the action model recognize dialogue text, but their focus differs. The guidance model tends to identify what the customer service representative guides the user to do, such as guiding the user to check logistics information. The action model, on the other hand, tends to identify what the customer service representative directly helps the user, such as directly checking logistics information for the user. Combining the two models—that is, combining the two recognition rules—allows for more business scenarios and makes the recognition results more accurate. In practical applications, the results from both models can be further processed and used as the answer text in the dialogue text, or the more accurate result can be selected as the answer text in the dialogue text based on the recognition results.

[0090] In a specific embodiment of this specification, the dialogue text is input into the guidance model and the action model respectively. The guidance model performs guidance recognition processing on the dialogue text, and the action model performs action recognition processing on the dialogue text to obtain the guidance answer text output by the guidance model and the action answer text output by the action model. The guidance answer text and the action answer text are combined to generate the answer text of the dialogue text. The answer text is "This ticket can be used at any time". This answer text is used as the target answer text associated with the target question text.

[0091] Step 208: Generate a dialogue summary text corresponding to the dialogue text based on the target question text and the target answer text.

[0092] Once the target question text and target answer text are determined using the methods described above, a dialogue summary text corresponding to this conversation can be generated. In practical applications, after determining the question asked by the user and the solution provided by customer service, a service summary for this communication service can be generated based on both.

[0093] To enable customer service staff to quickly locate and resolve issues, the question text and answer text can be combined to generate a dialogue summary text. Furthermore, to ensure the standardization of the dialogue summary text, the question text and answer text can be filled into a preset dialogue summary template. Specifically, generating the dialogue summary text corresponding to the dialogue text based on the target question text and the target answer text includes: determining the participation progress information of the project participation behavior; determining the dialogue summary template corresponding to the dialogue text based on the participation progress information; writing the target question text and the target answer text into the dialogue summary template; and generating the dialogue summary text corresponding to the dialogue text.

[0094] Among them, participation progress information can be understood as the progress information of the user's current participation in the project. For example, the order progress information when shopping online includes whether the transaction has been completed, whether the user has left a review, and the content of the review.

[0095] In practical applications, the corresponding dialogue summary templates may differ depending on the user's engagement stage. For example, if a user inquires before the transaction is complete, the summary template might include transaction status text, target question text, target answer text, and merchant status text. Conversely, if a user inquires after the transaction is complete, the summary template might include user review text, target question text, and target answer text. Therefore, for different engagement stages, it's necessary to determine the dialogue summary template for the current stage of the dialogue text to provide a more detailed description of the issues encountered by the user at that stage.

[0096] In a specific embodiment of this specification, the order information of a user's online shopping behavior is determined, it is determined that the user has completed the order and made a review, and a dialogue summary template is determined based on the participation progress information. The target question text and target answer text are written into the dialogue summary template. The dialogue summary template also includes the user's satisfaction rating for this online shopping experience. As a result, customer service can quickly learn about the user's previous inquiry and the solution based on the dialogue summary text generated based on the dialogue summary template, and can accurately understand the user's satisfaction with this online shopping experience, thereby providing better customer service and improving the user experience.

[0097] In another scenario, a summary statement can be automatically generated based on the target question text and the target answer text. Specifically, the target question text and the target answer text are fused together, and a dialogue summary text corresponding to the dialogue text is generated based on the fusion result.

[0098] In practice, the target question text and target answer text can be fused based on a pre-trained model to obtain the dialogue summary text output by the model. Taking the target question text as "Ticket usage, can this ticket be used on holidays?" and the target answer text as "This ticket can be used at any time" as an example, the target question text and target answer text are input into the pre-trained model. The pre-trained model encodes and decodes both, generating the fused dialogue summary text, which reads, "The user inquired about the validity period of the ticket; the user has been replied that the ticket can be used at any time." This method provides customer service with a more convenient service summary function, enabling customer service to quickly and accurately understand the user's historical questions and whether the questions have been resolved, thereby providing better consultation services.

[0099] In practical applications, the text generation method provided in this specification can also be implemented through a pre-trained model. Specifically, the method further includes: inputting dialogue text related to project participation behavior into a text generation model; determining the target question text corresponding to the dialogue text through the text generation model; determining the text to be processed based on the dialogue text and the auxiliary text corresponding to the dialogue text, wherein the auxiliary text is associated with the project participation behavior; determining the target answer text associated with the target question text based on the text to be processed; generating the dialogue summary text corresponding to the dialogue text based on the target question text and the target answer text; and outputting the text generation model.

[0100] The process involves inputting the user's communication dialogue with customer service into a trained text generation model. The model automatically determines the target question text corresponding to the dialogue text, identifies the text to be processed based on the dialogue text and auxiliary text, identifies the target answer text within the processed text, and finally outputs the generated dialogue summary text. In practical applications, order features can also be input into the text generation model for prediction and evaluation, improving the accuracy of the prediction results.

[0101] The text generation method provided in this specification includes: acquiring dialogue text associated with project participation behavior; determining the target question text corresponding to the dialogue text; determining a text to be processed based on the dialogue text and corresponding auxiliary text, wherein the auxiliary text is associated with the project participation behavior; determining a target answer text associated with the target question text based on the text to be processed; and generating a dialogue summary text corresponding to the dialogue text based on the target question text and the target answer text. By determining the corresponding target question text based on the dialogue text and introducing corresponding auxiliary text to enrich the information source, and by determining the text to be processed based on the auxiliary text and the dialogue text, a more accurate target answer text can be obtained. This results in a more granular final dialogue summary text, enabling project personnel to quickly understand the intentions of project participants based on the dialogue summary text, thereby providing users with more convenient services.

[0102] Figure 3 A flowchart of a text generation method according to another embodiment of this specification is shown, including steps 302 to 308.

[0103] Step 302: Receive a text generation instruction submitted via the summary page for the dialogue text, wherein the summary page is associated with project participation behavior.

[0104] The summary page can be understood as the page displayed by the customer service system to customer service personnel. After a customer service personnel connects with a user, they can generate a service summary text for this communication on the summary page. After the customer service personnel submit the text generation command for the dialogue text through the summary page, the customer service system will automatically start generating a service summary for this communication for the customer service personnel.

[0105] Step 304: In response to the text generation instruction, determine the target question text corresponding to the dialogue text, and determine the text to be processed based on the dialogue text and the auxiliary text corresponding to the dialogue text, wherein the auxiliary text is associated with the project participation behavior.

[0106] In practical applications, in response to the text generation command, the target question text corresponding to the dialogue text is determined. The target question text is the question that the user wants to ask in this communication. The target question text includes the question type and the user's original voice field. If the auxiliary text corresponding to the dialogue text contains an answer text, the auxiliary text is directly used as the text to be processed and subsequent text generation operations are performed. If the auxiliary text does not contain an answer text, the dialogue text is used as the text to be processed and subsequent text generation operations are performed.

[0107] Step 306: Determine the target answer text associated with the target question text based on the text to be processed.

[0108] In practical applications, when the text to be processed is auxiliary text, the answer text in the auxiliary text is directly used as the target answer text associated with the target question text. When the text to be processed is dialogue text, corpus recognition is performed on the dialogue text, and the answer text in the dialogue text is determined based on the recognition results, and then used as the target answer text.

[0109] Step 308: Generate a dialogue summary text corresponding to the dialogue text based on the target question text and the target answer text, and display the dialogue summary text through the summary page.

[0110] In practical applications, after combining the target question text and the target answer text to generate the dialogue summary text corresponding to the dialogue text, the dialogue summary text is displayed to customer service personnel through the summary page, which is the service summary.

[0111] The text generation method provided in this specification includes receiving a text generation instruction submitted via a summary page for a dialogue text, wherein the summary page is associated with project participation behavior; responding to the text generation instruction, determining the target question text corresponding to the dialogue text, and determining a text to be processed based on the dialogue text and the corresponding auxiliary text, wherein the auxiliary text is associated with the project participation behavior; determining a target answer text associated with the target question text based on the text to be processed; generating a dialogue summary text corresponding to the dialogue text based on the target question text and the target answer text, and displaying the dialogue summary text through the summary page. Determining the corresponding target question text based on the dialogue text, and introducing the corresponding auxiliary text to enrich the information source, and determining the text to be processed based on the auxiliary text and the dialogue text, enables the acquisition of a more accurate target answer text, resulting in a more granular final dialogue summary text. This allows project personnel to quickly understand the intentions of project participants based on the dialogue summary text, thereby providing users with more convenient services.

[0112] The following is in conjunction with the appendix Figure 4 Taking the text generation method provided in this specification as an example in an online shopping project, the text generation method will be further explained. Figure 4 The flowchart of a text generation method provided in one embodiment of this specification is shown, with specific steps including steps 402 to 410.

[0113] Step 402: Obtain the dialogue text of the related project participation behavior, and select the initial question text from the question text database based on the dialogue text.

[0114] In one feasible approach, the project participation behavior is the user's online shopping behavior, and the associated dialogue text is the dialogue text between the user and customer service regarding this online shopping behavior. The participation status information corresponding to this dialogue text is obtained. If the participation status information is after-sales status, then the problem text database is filtered according to the after-sales status to select candidate problem texts. Then, the matching degree between each candidate problem text and the dialogue text is calculated, and the one with the highest matching degree is selected as the initial problem text, which is "quality problem".

[0115] Step 404: Determine the original question text in the dialogue text based on the initial question text, merge the initial question text and the original question text, and obtain the target question text based on the merge result.

[0116] In one feasible approach, based on the determined initial question text, an original question text describing the initial question text is selected from the dialogue text. The original question text is "The product has a problem, and the merchant refuses to accept returns." The initial question text and the original question text are merged, and the target question text is obtained based on the merged result. The target question text is "Quality problem, the product has a problem, and the merchant refuses to accept returns."

[0117] Step 406: Determine the text to be processed based on the dialogue text and the corresponding auxiliary text, wherein the auxiliary text is associated with the project participation behavior.

[0118] In one feasible approach, auxiliary text corresponding to the dialogue text is determined based on the dialogue identifier of the dialogue text. The auxiliary text is the description text of the ISO or SOP standard operating procedure. The auxiliary text is identified, and if the answer text exists in the auxiliary text, the auxiliary text is used as the text to be processed.

[0119] In another possible approach, auxiliary text corresponding to the dialogue text is determined based on the dialogue identifier of the dialogue text. The auxiliary text is then identified, and if no answer text exists in the auxiliary text, the dialogue text is treated as the text to be processed.

[0120] Step 408: Determine the target answer text associated with the target question text based on the text to be processed.

[0121] In one feasible approach, when the text to be processed is auxiliary text, the answer text, i.e. the problem solution, in the auxiliary text is directly used as the target answer text associated with the target problem text.

[0122] In another feasible approach, when the text to be processed is dialogue text, the dialogue text is input into the guidance model and the action model respectively for combined recognition processing. Based on the recognition results, the answer text in the dialogue text is determined as the problem solution, and the answer text in the dialogue text is used as the target answer text associated with the target problem text.

[0123] Step 410: Generate a dialogue summary text corresponding to the dialogue text based on the target question text and the target answer text.

[0124] In one feasible approach, a dialogue summary template corresponding to the dialogue text is determined, the target question text and the target answer text are written into the dialogue summary template, and a dialogue summary text as a service summary is generated.

[0125] In another possible approach, the target question text and the target answer text are fused together, and a dialogue summary text corresponding to the dialogue text is generated based on the fusion result. The dialogue summary text is "The user wants to return the product due to quality issues, and a return service has been provided to the user."

[0126] This specification provides a text generation method that determines the corresponding target question text based on the dialogue text and introduces auxiliary text corresponding to the dialogue text to enrich the information source. Based on the auxiliary text and the dialogue text, the text to be processed is determined, which enables the acquisition of a more accurate target answer text. The final generated dialogue summary text has a higher granularity, allowing project personnel to quickly understand the intentions of project participants based on the dialogue summary text, thereby providing users with more convenient services.

[0127] Figure 5 A flowchart of a text generation method according to another embodiment of this specification is shown, including steps 502 to 508.

[0128] Step 502: Obtain the communication dialogue text between the user and customer service, and determine the target question text corresponding to the communication dialogue text.

[0129] The communication dialogue text can be understood as the text of the conversation between the user and customer service regarding the online shopping behavior. The target question text can be understood as the main question the user wants to ask in this communication dialogue, including the question type and the user's original voice. The target question text can be used to fill in the service summary so that customer service staff providing services to the user later can understand the user's intentions more quickly, thereby improving communication efficiency and enhancing the user experience.

[0130] In one embodiment of this specification, a user purchases a piece of clothing and inquires with customer service about after-sales issues related to the clothing. After communication between customer service and the user, the text of this communication dialogue is obtained, and the target question text of the communication dialogue text is determined. The target question text is the main question that the user wants to ask in this communication.

[0131] Step 504: Determine the text to be processed based on the communication dialogue text and the corresponding auxiliary text, wherein the auxiliary text is the communication template text used by the customer service representative.

[0132] In this context, the auxiliary text can be understood as the dialogue text provided by customer service representatives during communication with users, following a communication template. Therefore, the solutions to user problems recorded in the auxiliary text are the most accurate. When the auxiliary text contains the answer text, it can be directly used as the text to be processed, and the target answer text can be determined based on the auxiliary text. If the auxiliary text does not contain the answer text, then speech recognition can be performed on the dialogue text, and the target answer text can be determined based on the recognition results.

[0133] In one embodiment of this specification, the customer service representative replies to the user according to the communication template, obtains the auxiliary text corresponding to the communication dialogue text, and uses the auxiliary text as the text to be processed. Subsequently, the target answer text can be determined from the auxiliary text.

[0134] Step 506: Determine the target answer text associated with the target question text based on the text to be processed.

[0135] In practical applications, when the text to be processed is auxiliary text, the answer text in the auxiliary text is directly used as the target answer text associated with the target question text. When the text to be processed is dialogue text, corpus recognition is performed on the dialogue text, and the answer text in the dialogue text is determined based on the recognition results, and then used as the target answer text.

[0136] In one embodiment of this specification, following the example above, when the text to be processed is auxiliary text, the target answer text associated with the target question text can be directly determined from the auxiliary text.

[0137] Step 508: Based on the target question text and the target answer text, generate a customer service communication summary text corresponding to the communication dialogue text.

[0138] In one embodiment of this specification, the target question text and the target answer text are combined to generate a customer service communication summary text corresponding to the current communication dialogue text. The customer service communication summary text is a service summary of the service provided by the customer service representative to the user. Subsequently, the customer service representative can quickly understand the user's intent based on this service summary and provide the user with more efficient consultation services.

[0139] The text generation method provided in this specification includes: acquiring the communication dialogue text between a user and customer service, and determining the target question text corresponding to the communication dialogue text; determining the text to be processed based on the communication dialogue text and the corresponding auxiliary text, wherein the auxiliary text is the communication template text used by the customer service; determining the target answer text associated with the target question text based on the text to be processed; and generating a customer service communication summary text corresponding to the communication dialogue text based on the target question text and the target answer text. This method achieves the goal of determining the corresponding target question text based on the communication dialogue text, enriching the information source by introducing the corresponding auxiliary text, and determining the text to be processed based on the auxiliary text and the dialogue text. This results in a more accurate target answer text, a more granular final customer service communication summary text, and allows project personnel to quickly understand the intent of the users involved in the project based on the dialogue summary text, thereby providing users with more convenient services.

[0140] Corresponding to the above method embodiments, this specification also provides embodiments of a text generation apparatus. Figure 6 A schematic diagram of a text generation apparatus according to an embodiment of this specification is shown. Figure 6 As shown, the device includes:

[0141] The acquisition module 602 is configured to acquire the dialogue text of the related project participation behavior and determine the target question text corresponding to the dialogue text;

[0142] The first determining module 604 is configured to determine the text to be processed based on the dialogue text and the auxiliary text corresponding to the dialogue text, wherein the auxiliary text is associated with the project participation behavior;

[0143] The second determining module 606 is configured to determine the target answer text associated with the target question text based on the text to be processed;

[0144] The generation module 608 is configured to generate a dialogue summary text corresponding to the dialogue text based on the target question text and the target answer text.

[0145] Optionally, the acquisition module 602 is further configured to:

[0146] Based on the dialogue text, select an initial question text from the question text database;

[0147] The original question text is determined from the dialogue text based on the initial question text;

[0148] The initial question text and the original question text are merged, and the target question text is obtained based on the merging result.

[0149] Optionally, the acquisition module 602 is further configured to:

[0150] Obtain the participation status information corresponding to the dialogue text;

[0151] Based on the participation status information, the problem text database is filtered, and the candidate problem texts are determined according to the filtering results.

[0152] Calculate the matching degree between each candidate question text and the dialogue text, and select the initial question text based on each matching degree.

[0153] Optionally, the first determining module 604 is further configured to:

[0154] The auxiliary text is identified, and if the identification result determines that the auxiliary text contains answer text, the auxiliary text is used as the text to be processed.

[0155] Accordingly, determining the target answer text associated with the target question text based on the text to be processed includes:

[0156] The answer text in the auxiliary text is used as the target answer text associated with the target question text.

[0157] Optionally, the first determining module 604 is further configured to:

[0158] If the auxiliary text is identified and the identification result determines that there is no answer text in the auxiliary text, the dialogue text is taken as the text to be processed.

[0159] Accordingly, determining the target answer text associated with the target question text based on the text to be processed includes:

[0160] The dialogue text is processed for recognition, and the answer text in the dialogue text is determined based on the recognition result. The answer text in the dialogue text is then used as the target answer text associated with the target question text.

[0161] Optionally, the first determining module 604 is further configured to:

[0162] The dialogue text is input into the guidance model for guidance recognition processing to obtain the guidance answer text, and the dialogue text is input into the action model for action recognition processing to obtain the action answer text;

[0163] The answer text in the dialogue text is determined based on the guided answer text and the action answer text, and the answer text in the dialogue text is used as the target answer text associated with the target question text.

[0164] Optionally, the generation module 608 is further configured to:

[0165] Determine the participation progress information of the project participation behavior, determine the dialogue summary template corresponding to the dialogue text based on the participation progress information, write the target question text and the target answer text into the dialogue summary template, and generate the dialogue summary text corresponding to the dialogue text; or...

[0166] The target question text and the target answer text are fused together, and a dialogue summary text corresponding to the dialogue text is generated based on the fusion result.

[0167] Optionally, the device further includes an input module configured to:

[0168] Input the dialogue text related to project participation behavior into the text generation model;

[0169] The text generation model determines the target question text corresponding to the dialogue text, determines the text to be processed based on the dialogue text and the auxiliary text corresponding to the dialogue text, wherein the auxiliary text is associated with the project participation behavior, determines the target answer text associated with the target question text based on the text to be processed, generates the dialogue summary text corresponding to the dialogue text based on the target question text and the target answer text, and outputs the text generation model.

[0170] Optionally, the first determining module 604 is further configured to:

[0171] Determine the dialogue identification information of the dialogue text;

[0172] The auxiliary text corresponding to the dialogue text is determined based on the dialogue identifier information.

[0173] This specification provides a text generation device, comprising: an acquisition module configured to acquire dialogue text associated with project participation behavior and determine a target question text corresponding to the dialogue text; a first determination module configured to determine a text to be processed based on the dialogue text and corresponding auxiliary text, wherein the auxiliary text is associated with the project participation behavior; a second determination module configured to determine a target answer text associated with the target question text based on the text to be processed; and a generation module configured to generate a dialogue summary text corresponding to the dialogue text based on the target question text and the target answer text. By determining the corresponding target question text based on the dialogue text and introducing corresponding auxiliary text to enrich the information source, and by determining the text to be processed based on the auxiliary text and the dialogue text, a more accurate target answer text can be obtained, resulting in a more granular final dialogue summary text. This allows project personnel to quickly understand the intentions of project participants based on the dialogue summary text, thereby providing users with more convenient services.

[0174] The above is an illustrative scheme of a text generation device according to this embodiment. It should be noted that the technical solution of this text generation device and the technical solution of the above-described text generation method belong to the same concept. For details not described in detail in the technical solution of the text generation device, please refer to the description of the technical solution of the above-described text generation method.

[0175] Corresponding to the above method embodiments, this specification also provides embodiments of a text generation apparatus. Figure 7 A schematic diagram of a text generation apparatus according to another embodiment of this specification is shown. Figure 7 As shown, the device includes:

[0176] The receiving module 702 is configured to receive a text generation instruction submitted via a summary page for a dialogue text, wherein the summary page is associated with project participation behavior;

[0177] The first determining module 704 is configured to, in response to the text generation instruction, determine the target question text corresponding to the dialogue text, and determine the text to be processed based on the dialogue text and the auxiliary text corresponding to the dialogue text, wherein the auxiliary text is associated with the project participation behavior;

[0178] The second determining module 706 is configured to determine the target answer text associated with the target question text based on the text to be processed;

[0179] The generation module 708 is configured to generate a dialogue summary text corresponding to the dialogue text based on the target question text and the target answer text, and to display the dialogue summary text through the summary page.

[0180] Optionally, the first determining module 702 is further configured to:

[0181] Based on the dialogue text, select an initial question text from the question text database;

[0182] The original question text is determined from the dialogue text based on the initial question text;

[0183] The initial question text and the original question text are merged, and the target question text is obtained based on the merging result.

[0184] Optionally, the first determining module 702 is further configured to:

[0185] Obtain the participation status information corresponding to the dialogue text;

[0186] Based on the participation status information, the problem text database is filtered, and the candidate problem texts are determined according to the filtering results.

[0187] Calculate the matching degree between each candidate question text and the dialogue text, and select the initial question text based on each matching degree.

[0188] Optionally, the first determining module 702 is further configured to:

[0189] The auxiliary text is identified, and if the identification result determines that the auxiliary text contains answer text, the auxiliary text is used as the text to be processed.

[0190] Accordingly, determining the target answer text associated with the target question text based on the text to be processed includes:

[0191] The answer text in the auxiliary text is used as the target answer text associated with the target question text.

[0192] Optionally, the first determining module 702 is further configured to:

[0193] If the auxiliary text is identified and the identification result determines that there is no answer text in the auxiliary text, the dialogue text is taken as the text to be processed.

[0194] Accordingly, determining the target answer text associated with the target question text based on the text to be processed includes:

[0195] The dialogue text is processed for recognition, and the answer text in the dialogue text is determined based on the recognition result. The answer text in the dialogue text is then used as the target answer text associated with the target question text.

[0196] Optionally, the first determining module 702 is further configured to:

[0197] The dialogue text is input into the guidance model for guidance recognition processing to obtain the guidance answer text, and the dialogue text is input into the action model for action recognition processing to obtain the action answer text;

[0198] The answer text in the dialogue text is determined based on the guided answer text and the action answer text, and the answer text in the dialogue text is used as the target answer text associated with the target question text.

[0199] Optionally, the generation module 708 is further configured to:

[0200] Determine the participation progress information of the project participation behavior, determine the dialogue summary template corresponding to the dialogue text based on the participation progress information, write the target question text and the target answer text into the dialogue summary template, and generate the dialogue summary text corresponding to the dialogue text; or...

[0201] The target question text and the target answer text are fused together, and a dialogue summary text corresponding to the dialogue text is generated based on the fusion result.

[0202] Optionally, the device further includes an input module configured to:

[0203] Input the dialogue text related to project participation behavior into the text generation model;

[0204] The text generation model determines the target question text corresponding to the dialogue text, determines the text to be processed based on the dialogue text and the auxiliary text corresponding to the dialogue text, wherein the auxiliary text is associated with the project participation behavior, determines the target answer text associated with the target question text based on the text to be processed, generates the dialogue summary text corresponding to the dialogue text based on the target question text and the target answer text, and outputs the text generation model.

[0205] Optionally, the first determining module 702 is further configured to:

[0206] Determine the dialogue identification information of the dialogue text;

[0207] The auxiliary text corresponding to the dialogue text is determined based on the dialogue identifier information.

[0208] This specification provides a text generation device, comprising: a receiving module configured to receive a text generation instruction submitted via a summary page for a dialogue text, wherein the summary page is associated with project participation behavior; a first determining module configured to, in response to the text generation instruction, determine a target question text corresponding to the dialogue text, and determine a text to be processed based on the dialogue text and corresponding auxiliary text, wherein the auxiliary text is associated with the project participation behavior; a second determining module configured to determine a target answer text associated with the target question text based on the text to be processed; and a generation module configured to generate a dialogue summary text corresponding to the dialogue text based on the target question text and the target answer text, and display the dialogue summary text through the summary page. By determining the corresponding target question text based on the dialogue text and introducing corresponding auxiliary text to enrich the information source, and by determining the text to be processed based on the auxiliary text and the dialogue text, a more accurate target answer text can be obtained, resulting in a more granular final dialogue summary text. This allows project personnel to quickly understand the intentions of project participants based on the dialogue summary text, thereby providing users with more convenient services.

[0209] The above is an illustrative scheme of a text generation device according to this embodiment. It should be noted that the technical solution of this text generation device and the technical solution of the above-described text generation method belong to the same concept. For details not described in detail in the technical solution of the text generation device, please refer to the description of the technical solution of the above-described text generation method.

[0210] Corresponding to the above method embodiments, this specification also provides embodiments of a text generation apparatus. Figure 8 A schematic diagram of a text generation apparatus according to another embodiment of this specification is shown. Figure 8 As shown, the device includes:

[0211] The acquisition module 802 is configured to acquire the communication dialogue text between the user and customer service, and determine the target question text corresponding to the communication dialogue text;

[0212] The first determining module 804 is configured to determine the text to be processed based on the communication dialogue text and the auxiliary text corresponding to the communication dialogue text, wherein the auxiliary text is the communication template text used by the customer service representative;

[0213] The second determining module 806 is configured to determine the target answer text associated with the target question text based on the text to be processed;

[0214] The generation module 808 is configured to generate a customer service communication summary text corresponding to the communication dialogue text based on the target question text and the target answer text.

[0215] Optionally, the acquisition module 802 is further configured to:

[0216] Based on the dialogue text, select an initial question text from the question text database;

[0217] The original question text is determined from the dialogue text based on the initial question text;

[0218] The initial question text and the original question text are merged, and the target question text is obtained based on the merging result.

[0219] Optionally, the acquisition module 802 is further configured to:

[0220] Obtain the participation status information corresponding to the dialogue text;

[0221] Based on the participation status information, the problem text database is filtered, and the candidate problem texts are determined according to the filtering results.

[0222] Calculate the matching degree between each candidate question text and the dialogue text, and select the initial question text based on each matching degree.

[0223] Optionally, the first determining module 804 is further configured to:

[0224] The auxiliary text is identified, and if the identification result determines that the auxiliary text contains answer text, the auxiliary text is used as the text to be processed.

[0225] Accordingly, determining the target answer text associated with the target question text based on the text to be processed includes:

[0226] The answer text in the auxiliary text is used as the target answer text associated with the target question text.

[0227] Optionally, the first determining module 804 is further configured to:

[0228] If the auxiliary text is identified and the identification result determines that there is no answer text in the auxiliary text, the dialogue text is taken as the text to be processed.

[0229] Accordingly, determining the target answer text associated with the target question text based on the text to be processed includes:

[0230] The dialogue text is processed for recognition, and the answer text in the dialogue text is determined based on the recognition result. The answer text in the dialogue text is then used as the target answer text associated with the target question text.

[0231] Optionally, the first determining module 804 is further configured to:

[0232] The dialogue text is input into the guidance model for guidance recognition processing to obtain the guidance answer text, and the dialogue text is input into the action model for action recognition processing to obtain the action answer text;

[0233] The answer text in the dialogue text is determined based on the guided answer text and the action answer text, and the answer text in the dialogue text is used as the target answer text associated with the target question text.

[0234] Optionally, the generation module 808 is further configured to:

[0235] Determine the participation progress information of the project participation behavior, determine the dialogue summary template corresponding to the dialogue text based on the participation progress information, write the target question text and the target answer text into the dialogue summary template, and generate the dialogue summary text corresponding to the dialogue text; or...

[0236] The target question text and the target answer text are fused together, and a dialogue summary text corresponding to the dialogue text is generated based on the fusion result.

[0237] Optionally, the device further includes an input module configured to:

[0238] Input the dialogue text related to project participation behavior into the text generation model;

[0239] The text generation model determines the target question text corresponding to the dialogue text, determines the text to be processed based on the dialogue text and the auxiliary text corresponding to the dialogue text, wherein the auxiliary text is associated with the project participation behavior, determines the target answer text associated with the target question text based on the text to be processed, generates the dialogue summary text corresponding to the dialogue text based on the target question text and the target answer text, and outputs the text generation model.

[0240] Optionally, the first determining module 804 is further configured to:

[0241] Determine the dialogue identification information of the dialogue text;

[0242] The auxiliary text corresponding to the dialogue text is determined based on the dialogue identifier information.

[0243] This specification provides a text generation device, comprising: an acquisition module configured to acquire communication dialogue text between a user and customer service, and determine a target question text corresponding to the communication dialogue text; a first determination module configured to determine a text to be processed based on the communication dialogue text and corresponding auxiliary text, wherein the auxiliary text is a communication template text used by the customer service; a second determination module configured to determine a target answer text associated with the target question text based on the text to be processed; and a generation module configured to generate a customer service communication summary text corresponding to the communication dialogue text based on the target question text and the target answer text. By determining the corresponding target question text based on the communication dialogue text and introducing the corresponding auxiliary text to enrich the information source, and by determining the text to be processed based on the auxiliary text and the dialogue text, a more accurate target answer text can be obtained, resulting in a more granular final customer service communication summary text. This allows project personnel to more quickly understand the intentions of project participants based on the customer service communication summary text, thereby providing users with more convenient services.

[0244] Figure 9 A structural block diagram of a computing device 900 according to an embodiment of this specification is shown. The components of the computing device 900 include, but are not limited to, a memory 910 and a processor 920. The processor 920 is connected to the memory 910 via a bus 930, and a database 950 is used to store data.

[0245] The computing device 900 also includes an access device 940, which enables the computing device 900 to communicate via one or more networks 960. Examples of these networks include a Public Switched Telephone Network (PSTN), a Local Area Network (LAN), a Wide Area Network (WAN), a Personal Area Network (PAN), or a combination of communication networks such as the Internet. The access device 940 may include one or more of any type of wired or wireless network interface (e.g., a Network Interface Card (NIC)), such as an IEEE 902.11 Wireless Local Area Network (WLAN) interface, a Wi-MAX interface, an Ethernet interface, a Universal Serial Bus (USB) interface, a cellular network interface, a Bluetooth interface, a Near Field Communication (NFC) interface, and so on.

[0246] In one embodiment of this specification, the aforementioned components of the computing device 900 and Figure 9 Other components, not shown, can also be connected to each other, for example, via a bus. It should be understood that... Figure 9 The block diagram of the computing device shown is for illustrative purposes only and is not intended to limit the scope of this specification. Those skilled in the art can add or replace other components as needed.

[0247] The computing device 900 can be any type of stationary or mobile computing device, including mobile computers or mobile computing devices (e.g., tablet computers, personal digital assistants, laptop computers, notebook computers, netbooks, etc.), mobile phones (e.g., smartphones), wearable computing devices (e.g., smartwatches, smart glasses, etc.) or other types of mobile devices, or stationary computing devices such as desktop computers or PCs. The computing device 900 can also be a mobile or stationary server.

[0248] The processor 920 implements the text generation method when executing the computer instructions.

[0249] The above is an illustrative scheme of a computing device according to this embodiment. It should be noted that the technical solution of this computing device and the technical solution of the above-described text generation method belong to the same concept. For details not described in detail in the technical solution of the computing device, please refer to the description of the technical solution of the above-described text generation method.

[0250] An embodiment of this specification also provides a text generation system that stores computer instructions. The system includes a server and a client. The client is used to store executable instructions for text display, and the server is used to store executable instructions for text generation. When the executable instructions for text display are executed by the client and when the executable instructions for text generation are executed by the server, the steps of the text generation method described above are implemented.

[0251] An embodiment of this specification also provides a computer-readable storage medium storing computer instructions that, when executed by a processor, implement the steps of the text generation method as described above.

[0252] The above is an illustrative scheme of a computer-readable storage medium according to this embodiment. It should be noted that the technical solution of this storage medium and the technical solution of the text generation method described above belong to the same concept. For details not described in detail in the technical solution of the storage medium, please refer to the description of the technical solution of the text generation method described above.

[0253] An embodiment of this specification also provides a computer program, wherein when the computer program is executed in a computer, it causes the computer to perform the steps of the above-described text generation method.

[0254] The above is an illustrative example of a computer program according to this embodiment. It should be noted that the technical solution of this computer program belongs to the same concept as the technical solution of the text generation method described above. Details not described in detail in the technical solution of the computer program can be found in the description of the technical solution of the text generation method described above.

[0255] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0256] The computer instructions include computer program code, which may be in the form of source code, object code, executable file, or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium may be appropriately added to or subtracted according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media may not include electrical carrier signals and telecommunication signals.

[0257] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the embodiments in this specification are not limited to the described order of actions, because according to the embodiments in this specification, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in this specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to the embodiments in this specification.

[0258] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0259] The preferred embodiments disclosed above are merely illustrative of this specification. The optional embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the embodiments described herein. These embodiments are selected and specifically described in this specification to better explain the principles and practical applications of the embodiments, thereby enabling those skilled in the art to better understand and utilize this specification. This specification is limited only by the claims and their full scope and equivalents.

Claims

1. A text generation method, comprising: Obtain the dialogue text of the related project participation behavior, and determine the target question text corresponding to the dialogue text, wherein the target question text consists of a preset initial question text and the original question text of the related user; The text to be processed is determined based on the dialogue text and the corresponding auxiliary text, wherein the auxiliary text is associated with the project participation behavior; Based on the text to be processed, a target answer text associated with the target question text is determined, wherein, when the text to be processed is the dialogue text, the target answer text is obtained by processing the dialogue text through a guidance model and an action model; Based on the target question text and the target answer text, generate a dialogue summary text corresponding to the dialogue text.

2. The method as described in claim 1, wherein determining the target question text corresponding to the dialogue text includes: Based on the dialogue text, select an initial question text from the question text database; The original question text is determined from the dialogue text based on the initial question text; The initial question text and the original question text are merged, and the target question text is obtained based on the merging result.

3. The method as described in claim 2, wherein selecting an initial question text from the question text database based on the dialogue text includes: Obtain the participation status information corresponding to the dialogue text; Based on the participation status information, the problem text database is filtered, and the candidate problem texts are determined according to the filtering results. Calculate the matching degree between each candidate question text and the dialogue text, and select the initial question text based on each matching degree.

4. The method according to any one of claims 1-3, wherein determining the text to be processed based on the dialogue text and the auxiliary text corresponding to the dialogue text includes: The auxiliary text is identified, and if the identification result determines that the auxiliary text contains answer text, the auxiliary text is used as the text to be processed. Accordingly, determining the target answer text associated with the target question text based on the text to be processed includes: The answer text in the auxiliary text is used as the target answer text associated with the target question text.

5. The method according to any one of claims 1-3, wherein determining the text to be processed based on the dialogue text and the auxiliary text corresponding to the dialogue text includes: If the auxiliary text is identified and the identification result determines that there is no answer text in the auxiliary text, the dialogue text is taken as the text to be processed. Accordingly, determining the target answer text associated with the target question text based on the text to be processed includes: The dialogue text is processed for recognition, and the answer text in the dialogue text is determined based on the recognition result. The answer text in the dialogue text is then used as the target answer text associated with the target question text.

6. The method of claim 5, wherein the dialogue text is subjected to recognition processing, the answer text in the dialogue text is determined according to the recognition result, and the answer text in the dialogue text is used as the target answer text associated with the target question text, comprising: The dialogue text is input into the guidance model for guidance recognition processing to obtain the guidance answer text, and the dialogue text is input into the action model for action recognition processing to obtain the action answer text; The answer text in the dialogue text is determined based on the guided answer text and the action answer text, and the answer text in the dialogue text is used as the target answer text associated with the target question text.

7. The method as described in claim 1, comprising generating a dialogue summary text corresponding to the dialogue text based on the target question text and the target answer text, including: Determine the participation progress information of the project participation behavior, determine the dialogue summary template corresponding to the dialogue text based on the participation progress information, write the target question text and the target answer text into the dialogue summary template, and generate the dialogue summary text corresponding to the dialogue text; or, The target question text and the target answer text are fused together, and a dialogue summary text corresponding to the dialogue text is generated based on the fusion result.

8. The method of claim 1, further comprising: Input the dialogue text related to project participation behavior into the text generation model; The text generation model determines the target question text corresponding to the dialogue text, determines the text to be processed based on the dialogue text and the auxiliary text corresponding to the dialogue text, wherein the auxiliary text is associated with the project participation behavior, determines the target answer text associated with the target question text based on the text to be processed, generates the dialogue summary text corresponding to the dialogue text based on the target question text and the target answer text, and outputs the text generation model.

9. The method of claim 1, wherein determining the auxiliary text comprises: Determine the dialogue identification information of the dialogue text; The auxiliary text corresponding to the dialogue text is determined based on the dialogue identifier information.

10. A text generation method, comprising: Receive a text generation instruction submitted via a summary page for the dialogue text, wherein the summary page is associated with project participation behavior; In response to the text generation instruction, the target question text corresponding to the dialogue text is determined, and the text to be processed is determined based on the dialogue text and the auxiliary text corresponding to the dialogue text, wherein the auxiliary text is associated with the project participation behavior, and the target question text consists of a preset initial question text and the original question text associated with the user; Based on the text to be processed, a target answer text associated with the target question text is determined, wherein, when the text to be processed is the dialogue text, the target answer text is obtained by processing the dialogue text through a guidance model and an action model; Based on the target question text and the target answer text, a dialogue summary text corresponding to the dialogue text is generated, and the dialogue summary text is displayed through the summary page.

11. A text generation method, comprising: Obtain the communication dialogue text between the user and customer service, and determine the target question text corresponding to the communication dialogue text, wherein the target question text consists of a preset initial question text and the original question text associated with the user; The text to be processed is determined based on the communication dialogue text and the corresponding auxiliary text, wherein the auxiliary text is the communication template text used by the customer service representative; Based on the text to be processed, a target answer text associated with the target question text is determined, wherein, when the text to be processed is the dialogue text, the target answer text is obtained by processing the dialogue text through a guidance model and an action model; Based on the target question text and the target answer text, a customer service communication summary text corresponding to the communication dialogue text is generated.

12. A text generation system, the system comprising a server and a client; The client is used to store executable instructions for text display, and the server is used to store executable instructions for text generation; when the executable instructions for text display are executed by the client and when the executable instructions for text generation are executed by the server, the steps of the method according to any one of claims 1 to 9, 10 or 11 are implemented.

13. A computing device comprising a memory, a processor, and computer instructions stored in the memory and executable on the processor, wherein the processor, when executing the computer instructions, performs the steps of the method according to any one of claims 1 to 9, 10, or 11.

14. A computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the steps of the method according to any one of claims 1 to 9, 10, or 11.

Images