Answer determination method and device, computer equipment, storage medium and computer program product

Abstract

The invention relates to an answer determination method and device, computer equipment, a storage medium and a computer program product. The method comprises the following steps: decomposing problem information associated with a product networking topology design to obtain multiple pieces of sorted sub-problem information; inputting each piece of sub-question information into a pre-trained question processing model in sequence to obtain answer information corresponding to each piece of sub-question information; any one piece of sub-question information except the first sub-question information in the sub-question information carries prompt information, and the prompt information carried by the any one piece of sub-question information comprises sub-question information arranged before the any one piece of sub-question information and answer information corresponding to the sub-question information; and determining target answer information corresponding to the question information according to the answer information corresponding to each piece of sub-question information. By adopting the method, the answer determination accuracy can be improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to an answer determination method, apparatus, computer device, computer-readable storage medium, and computer program product. Background Art

[0002] At present, in the process of enterprises providing communication services or products to the outside world, how to effectively solve the problems related to the product networking topology design is crucial to improving the quality of enterprise services.

[0003] In traditional technology, in order to solve problems related to product networking topology design, the entire problem is generally input into the model and the output of the model is directly used as the answer; however, the answer generated by this method is too general, resulting in a low accuracy rate in determining the answer. Summary of the invention

[0004] Based on this, it is necessary to provide an answer determination method, device, computer equipment, computer-readable storage medium and computer program product that can improve the accuracy of answer determination in response to the above technical problems.

[0005] In a first aspect, the present application provides an answer determination method, comprising:

[0006] Decompose the problem information related to the product network topology design to obtain multiple sorted sub-problem information;

[0007] Input each of the sub-question information into a pre-trained question processing model in turn to obtain answer information corresponding to each of the sub-question information; any sub-question information except the first sub-question information in each of the sub-question information carries prompt information, and the prompt information carried by any of the sub-question information includes the sub-question information arranged before any of the sub-question information, and the answer information corresponding to the sub-question information;

[0008] According to the answer information corresponding to each of the sub-question information, target answer information corresponding to the question information is determined.

[0009] In one embodiment, the problem information associated with the product networking topology design is decomposed to obtain a plurality of sorted sub-problem information, including:

[0010] Obtain target documents associated with product network topology design;

[0011] The question information is decomposed and processed based on the target document by the question processing model to obtain a plurality of sorted sub-question information.

[0012] In one embodiment, the step of sequentially inputting each of the sub-question information into a pre-trained question processing model to obtain answer information corresponding to each of the sub-question information includes:

[0013] In the case where the current sub-question information is the first sub-question information, inputting the first sub-question information into a pre-trained question processing model to obtain answer information corresponding to the first sub-question information;

[0014] In the case where the current sub-question information is any sub-question information except the first sub-question information, determining the prompt information carried by the current sub-question information, inputting the current sub-question information and the prompt information carried by the current sub-question information into the pre-trained question processing model, and obtaining answer information corresponding to the current sub-question information;

[0015] The next sub-question information of the current sub-question information is used as the current sub-question information, and the process jumps to the step of determining the prompt information carried by the current sub-question information until the current sub-question information is the last sub-question information.

[0016] In one embodiment, the determining the prompt information carried by the current sub-question information includes:

[0017] Obtaining all sub-question information arranged before the current sub-question information, and answer information corresponding to each sub-question information in all the sub-question information;

[0018] All the sub-question information and the answer information corresponding to each sub-question information in all the sub-question information are used as the prompt information carried by the current sub-question information.

[0019] In one embodiment, the inputting the current sub-question information and the prompt information carried by the current sub-question information into the pre-trained question processing model to obtain the answer information corresponding to the current sub-question information includes:

[0020] Using a pre-trained question processing model, based on prompt information carried by the current sub-question information, identifying network component information associated with the current sub-question information;

[0021] Generating initial answer information corresponding to the current sub-question information according to the network component information;

[0022] The initial answer information is updated according to the prompt information carried by the current sub-question information to obtain the answer information corresponding to the current sub-question information.

[0023] In one embodiment, inputting the first sub-question information into a pre-trained question processing model to obtain answer information corresponding to the first sub-question information includes:

[0024] Identifying network component information associated with the first sub-question information through a pre-trained question processing model;

[0025] Answer information corresponding to the first sub-question information is generated according to the network component information.

[0026] In one embodiment, determining target answer information corresponding to the question information according to answer information corresponding to each sub-question information includes:

[0027] Filtering out the answer information corresponding to the last sub-question information from the answer information corresponding to each of the sub-question information;

[0028] The answer information corresponding to the last sub-question information is confirmed as the target answer information corresponding to the question information.

[0029] In a second aspect, the present application further provides an answer determination device, comprising:

[0030] A problem decomposition module is used to decompose the problem information associated with the product network topology design to obtain a plurality of sorted sub-problem information;

[0031] A question processing module, used for sequentially inputting each of the sub-question information into a pre-trained question processing model to obtain answer information corresponding to each of the sub-question information; any sub-question information except the first sub-question information in each of the sub-question information carries prompt information, and the prompt information carried by any of the sub-question information includes the sub-question information arranged before any of the sub-question information, and the answer information corresponding to the sub-question information;

[0032] The answer determination module is used to determine the target answer information corresponding to the question information according to the answer information corresponding to each of the sub-question information.

[0033] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0034] Decompose the problem information related to the product network topology design to obtain multiple sorted sub-problem information;

[0035] Input each of the sub-question information into a pre-trained question processing model in turn to obtain answer information corresponding to each of the sub-question information; any sub-question information except the first sub-question information in each of the sub-question information carries prompt information, and the prompt information carried by any of the sub-question information includes the sub-question information arranged before any of the sub-question information, and the answer information corresponding to the sub-question information;

[0036] According to the answer information corresponding to each of the sub-question information, target answer information corresponding to the question information is determined.

[0037] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:

[0038] Decompose the problem information related to the product network topology design to obtain multiple sorted sub-problem information;

[0039] Input each of the sub-question information into a pre-trained question processing model in turn to obtain answer information corresponding to each of the sub-question information; any sub-question information except the first sub-question information in each of the sub-question information carries prompt information, and the prompt information carried by any of the sub-question information includes the sub-question information arranged before any of the sub-question information, and the answer information corresponding to the sub-question information;

[0040] According to the answer information corresponding to each of the sub-question information, target answer information corresponding to the question information is determined.

[0041] In a fifth aspect, the present application further provides a computer program product, including a computer program, which implements the following steps when executed by a processor:

[0042] Decompose the problem information related to the product network topology design to obtain multiple sorted sub-problem information;

[0043] Input each of the sub-question information into a pre-trained question processing model in turn to obtain answer information corresponding to each of the sub-question information; any sub-question information except the first sub-question information in each of the sub-question information carries prompt information, and the prompt information carried by any of the sub-question information includes the sub-question information arranged before any of the sub-question information, and the answer information corresponding to the sub-question information;

[0044] According to the answer information corresponding to each of the sub-question information, target answer information corresponding to the question information is determined.

[0045] The above-mentioned answer determination method, device, computer equipment, storage medium and computer program product first decompose and process the problem information associated with the product networking topology design to obtain a plurality of sorted sub-problem information, and then input the first sub-problem information of the sorted plurality of sub-problem information into a pre-trained problem processing model to obtain the answer information corresponding to the first sub-problem information. For any sub-problem information except the first sub-problem information, the sub-problem information arranged before the sub-problem information and the answer information corresponding to the sub-problem information are used as the prompt information carried by the sub-problem information, and the sub-problem information and the prompt information carried by the sub-problem information are input into the pre-trained problem processing model to obtain the answer information corresponding to the sub-problem information. Finally, according to the answer information corresponding to each sub-problem information, the target answer information corresponding to the problem information is determined. In this way, in the process of solving the problem associated with the product network topology design, the problem information associated with the product network topology design is first decomposed to obtain a plurality of arranged sub-problem information, and any sub-problem information except the first sub-problem information in the arranged plurality of sub-problem information carries prompt information from the least to the most in sequence, and then these sub-problem information are processed in sequence using the prompt information from the least to the most, so that the processing basis of the sub-problem information is increasing, and then the processing accuracy of the sub-problem information is increasing, which is conducive to making the final target answer information more accurate, thereby improving the accuracy of answer determination; moreover, this method avoids the defect of low answer determination accuracy caused by inputting the entire problem into the model and directly using the output content of the model as the answer, and further improves the accuracy of answer determination. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0047] Figure 1 A schematic diagram of a flow chart of a method for determining an answer in one embodiment;

[0048] Figure 2 A schematic diagram of a PSR design architecture in one embodiment;

[0049] Figure 3 A flowchart of steps for decomposing and processing problem information associated with product networking topology design in one embodiment;

[0050] Figure 4A flowchart of steps for obtaining answer information corresponding to each sub-question information in one embodiment;

[0051] Figure 5 A flowchart of steps for obtaining answer information corresponding to each sub-question information in another embodiment;

[0052] Figure 6 A schematic flow chart of a method for determining an answer in another embodiment;

[0053] Figure 7 It is a structural block diagram of an answer determination device in one embodiment;

[0054] Figure 8 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0056] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0057] In an exemplary embodiment, Figure 1 As shown, a method for determining an answer is provided. This embodiment uses the method applied to a server as an example for illustration; it can be understood that the method can also be applied to a terminal, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. The terminal can be, but is not limited to, various personal computers, laptops, smart phones, and tablet computers; the server can be implemented as an independent server or a server cluster consisting of multiple servers. In this embodiment, the method includes the following steps:

[0058] Step S101 : decompose the problem information associated with the product network topology design to obtain a plurality of sorted sub-problem information.

[0059] Among them, product network topology design is used to represent the design process of obtaining the network topology corresponding to the product business requirements based on specific product business requirements. In actual scenarios, product network topology design is a PSR model design for PSR (Problem-Solution-Roadmap) objects.

[0060] Among them, PSR objects include RES (Resource) object, RFS (Resource Facing Service) object, CFS (Customer Facing Service) object, and PRODUCT (product) object.

[0061] Among them, RES objects are used to represent entities that can be managed and operated in a specific business environment, including network resources (such as network bandwidth), computing resources (such as memory space, storage capacity), etc.

[0062] Among them, the RFS object is used to represent resource-oriented services.

[0063] Among them, the CFS object is used to represent customer-oriented services.

[0064] It should be noted that the relationship between RES objects, RFS objects, CFS objects and PRODUCT objects is that multiple RES objects constitute RFS objects, one or more RFS objects constitute CFS objects, and one or more CFS objects constitute PRODUCT objects. Figure 2 shown.

[0065] The question information refers to questions related to the product network topology design.

[0066] The decomposition process refers to the process of obtaining sorted sub-question information according to the question information.

[0067] The sorted multiple sub-question information refers to multiple question information carrying sorting information obtained by decomposing the question information.

[0068] Exemplarily, the server responds to a problem processing instruction for problem information associated with the product networking topology design, and obtains problem information associated with the product networking topology design; then, the server decomposes and processes the problem information associated with the product networking topology design according to a preset decomposition instruction, and obtains multiple problem information corresponding to the problem information and carrying sorting information, as multiple sorted sub-problem information.

[0069] For example, see Figure 3, the server inputs the question information "generating product networking topology according to product business requirements" into the large language model according to the pre-set standardized prompt process, decomposes the question information through the large language model, and obtains the answer information corresponding to the question information, that is, in order to solve "generating product networking topology according to product business requirements", we need to solve "generating network service networking topology according to product business requirements" in the first step, solve "generating business service networking topology according to product business requirements" in the second step, and solve "generating product networking topology according to product business requirements" in the third step; then, the server uses "generating network service networking topology according to product business requirements" as the sub-question information with sorting information of 1, "generating business service networking topology according to product business requirements" as the sub-question information with sorting information of 2, and "generating product networking topology according to product business requirements" as the sub-question information with sorting information of 3; then, the server uses "generating network service networking topology according to product business requirements", "generating business service networking topology according to product business requirements" and "generating product networking topology according to product business requirements" as multiple sorted sub-question information.

[0070] Step S102, input each sub-question information into a pre-trained question processing model in turn to obtain answer information corresponding to each sub-question information; any sub-question information except the first sub-question information carries prompt information, and the prompt information carried by any sub-question information includes the sub-question information arranged before any sub-question information, and the answer information corresponding to the sub-question information.

[0071] Among them, the question processing model refers to a network model that can use each sub-question information to obtain the answer information corresponding to each sub-question information, such as a large language model.

[0072] The answer information is used to represent the answer corresponding to the sub-question information obtained through the question processing model.

[0073] The first sub-problem information refers to the sub-problem information with sorting information being 1.

[0074] The prompt information includes the sub-question information arranged before any sub-question information, and the answer information corresponding to the sub-question information.

[0075] Exemplarily, the server inputs the first sub-question information among the sorted multiple sub-question information into a pre-trained question processing model to obtain answer information corresponding to the first sub-question information, and for any sub-question information except the first sub-question information, the server uses the sub-question information arranged before the sub-question information and the answer information corresponding to the sub-question information as the prompt information carried by the sub-question information, and inputs the sub-question information and the prompt information carried by the sub-question information into the pre-trained question processing model to obtain the answer information corresponding to the sub-question information.

[0076] For example, see Figure 4 , the server takes the sub-question information "Generate network service networking topology according to product business requirements" with sorting information 1 as sub-question 1, the sub-question information "Generate business service networking topology according to product business requirements" with sorting information 2 as sub-question 2, and the sub-question information "Generate product networking topology according to product business requirements" with sorting information 3 as sub-question 3; then, the server inputs sub-question 1 "Generate network service networking topology according to product business requirements" into the large language model, and obtains the answer corresponding to sub-question 1 "Generate network service networking topology by assembling one or more resource components according to product business requirements"; then, the server takes sub-question 1 and the answer corresponding to sub-question 1 as the prompt information of sub-question 2, and inputs sub-question 2 "Generate business service networking topology according to product business requirements" and the prompt information of sub-question 2 into the large language model, and obtains sub-question 1. The answer corresponding to sub-question 2 is "According to the product business requirements, one or more resource components are used to assemble and generate the network service networking topology, so one or more network service components are used to generate the business service networking topology"; then, the server uses sub-question 2 and the answer corresponding to sub-question 2 as the prompt information of sub-question 3, and inputs sub-question 3 "Generate the product networking topology according to the product business requirements" and the prompt information of sub-question 3 into the large language model, and obtains the answer corresponding to sub-question 3 "According to the product business requirements, one or more resource components are used to assemble and generate the network service networking topology, so one or more network service components are used to generate the business service networking topology, so one or more business service components are used to generate the product networking topology"; finally, the server uses the answer corresponding to sub-question 1, the answer corresponding to sub-question 2, and the answer corresponding to sub-question 3 as the answer information corresponding to each sub-question information.

[0077] Step S103: determining target answer information corresponding to the question information according to the answer information corresponding to each sub-question information.

[0078] Among them, the target answer information is used to represent the final answer corresponding to the question information.

[0079] Exemplarily, the server determines the importance of the answer information corresponding to each sub-question information; then, the server selects the answer information with the greatest importance from the answer information corresponding to each sub-question information as the target answer information corresponding to the question information.

[0080] In the above-mentioned answer determination method, the question information associated with the product networking topology design is first decomposed and processed to obtain a plurality of sorted sub-question information, and then the first sub-question information among the sorted plurality of sub-question information is input into a pre-trained question processing model to obtain the answer information corresponding to the first sub-question information, and for any sub-question information except the first sub-question information, the sub-question information arranged before the sub-question information and the answer information corresponding to the sub-question information are used as the prompt information carried by the sub-question information, and the sub-question information and the prompt information carried by the sub-question information are input into the pre-trained question processing model to obtain the answer information corresponding to the sub-question information, and finally, according to the answer information corresponding to each sub-question information, the target answer information corresponding to the question information is determined. In this way, in the process of solving the problem associated with the product network topology design, the problem information associated with the product network topology design is first decomposed to obtain a plurality of arranged sub-problem information, and any sub-problem information except the first sub-problem information in the arranged plurality of sub-problem information carries prompt information from the least to the most in sequence, and then these sub-problem information are processed in sequence using the prompt information from the least to the most, so that the processing basis of the sub-problem information is increasing, and then the processing accuracy of the sub-problem information is increasing, which is conducive to making the final target answer information more accurate, thereby improving the accuracy of answer determination; moreover, this method avoids the defect of low answer determination accuracy caused by inputting the entire problem into the model and directly using the output content of the model as the answer, and further improves the accuracy of answer determination.

[0081] In an exemplary embodiment, the above step S101 decomposes the problem information associated with the product networking topology design to obtain a plurality of sorted sub-problem information, specifically including the following contents: obtaining a target document associated with the product networking topology design; decomposing the problem information based on the target document through a problem processing model to obtain a plurality of sorted sub-problem information.

[0082] The target document is used to represent the product business requirement document associated with the product networking topology design.

[0083] Exemplarily, the server responds to a problem handling instruction for problem information associated with the product networking topology design, obtains a target document associated with the product networking topology design, and the problem information associated with the product networking topology design; then, the server inputs the target document and the problem information into a problem handling model, and decomposes the problem information based on the target document through the problem handling model to obtain a plurality of sorted sub-problem information.

[0084] In this embodiment, by extracting and obtaining the target document associated with the product networking topology design, the problem processing model decomposes the problem information based on the comprehensive and accurate relevant information collected, which is beneficial to improving the accuracy of the decomposition of the problem information and facilitating the subsequent problem information processing process.

[0085] In an exemplary embodiment, Figure 5 As shown, the above step S102, sequentially inputting each sub-question information into the pre-trained question processing model to obtain the answer information corresponding to each sub-question information, specifically includes the following steps:

[0086] Step S501, when the current sub-question information is the first sub-question information, the first sub-question information is input into a pre-trained question processing model to obtain answer information corresponding to the first sub-question information.

[0087] Step S502, when the current sub-question information is any sub-question information except the first sub-question information, determine the prompt information carried by the current sub-question information, input the current sub-question information and the prompt information carried by the current sub-question information into a pre-trained question processing model, and obtain the answer information corresponding to the current sub-question information.

[0088] Step S503: taking the next sub-question information of the current sub-question information as the current sub-question information, and jumping to the step of determining the prompt information carried by the current sub-question information, until the current sub-question information is the last sub-question information.

[0089] The current sub-question information refers to any sub-question information among the sub-question information, and may be the first sub-question information, or may be any sub-question information except the first sub-question information.

[0090] Exemplarily, the server determines the current sub-problem information of each sub-problem information; when the current sub-problem information is the first sub-problem information, the server inputs the first sub-problem information into a pre-trained question processing model to obtain answer information corresponding to the first sub-problem information output by the question processing model; when the current sub-problem information is any sub-problem information except the first sub-problem information, the server determines the prompt information carried by the current sub-problem information, and inputs the current sub-problem information and the prompt information carried by the current sub-problem information into the pre-trained question processing model to obtain answer information corresponding to the current sub-problem information output by the question processing model; then, the server uses the next sub-problem information of the current sub-problem information as the current sub-problem information, and jumps to the step of determining the prompt information carried by the current sub-problem information until the current sub-problem information is the last sub-problem information; then, the server uses the answer information corresponding to all the current sub-problem information as the answer information corresponding to each sub-problem information.

[0091] In this embodiment, for the first sub-problem information, it is directly input into the pre-trained problem processing model to obtain the corresponding answer information, which is equivalent to providing a basis for solving subsequent sub-problem information; for other sub-problems except the first sub-problem, the answer information of the previous question can be used as a prompt, so that the problem processing is more coherent and logical.

[0092] In an exemplary embodiment, the above step S502, determining the prompt information carried by the current sub-question information, specifically includes the following contents: obtaining all sub-question information arranged before the current sub-question information, and the answer information corresponding to each sub-question information in all the sub-question information; using all the sub-question information and the answer information corresponding to each sub-question information in all the sub-question information as the prompt information carried by the current sub-question information.

[0093] Exemplarily, the server obtains all sub-question information arranged before the current sub-question information, and the answer information corresponding to each sub-question information in all the sub-question information; then, the server uses all the sub-question information and the answer information corresponding to each sub-question information in all the sub-question information as the prompt information carried by the current sub-question information.

[0094] For example, see Figure 4, when the current sub-question information is sub-question 2, the server obtains sub-question 1 arranged before sub-question 2, and the answer corresponding to sub-question 1, and uses sub-question 1 and the answer corresponding to sub-question 1 as the prompt information carried by sub-question 2; when the current sub-question information is sub-question 3, the server obtains sub-question 1 and sub-question 2 arranged before sub-question 3, and the answers corresponding to sub-question 1 and sub-question 2, and uses sub-question 1 and sub-question 2, and the answers corresponding to sub-question 1 and sub-question 2 as the prompt information carried by sub-question 3.

[0095] In this embodiment, for processing the current sub-problem, all sub-problem information and answer information arranged before it are obtained as prompt information, thereby providing rich background knowledge for solving the current sub-problem, which is conducive to improving the accuracy of processing the current sub-problem.

[0096] In an exemplary embodiment, the above step S502 inputs the current sub-question information and the prompt information carried by the current sub-question information into a pre-trained question processing model to obtain answer information corresponding to the current sub-question information, which specifically includes the following contents: through the pre-trained question processing model, based on the prompt information carried by the current sub-question information, the network component information associated with the current sub-question information is identified; based on the network component information, the initial answer information corresponding to the current sub-question information is generated; based on the prompt information carried by the current sub-question information, the initial answer information is updated to obtain the answer information corresponding to the current sub-question information.

[0097] The network component information refers to the PSR object corresponding to the current sub-problem information.

[0098] The initial answer information refers to the answer corresponding to the current sub-question information obtained based on the network component information.

[0099] Exemplarily, the server inputs the current sub-question information and the prompt information carried by the current sub-question information into a pre-trained question processing model, and identifies the network component information associated with the current sub-question information based on the prompt information carried by the current sub-question information through the pre-trained question processing model; then, the server generates initial answer information corresponding to the current sub-question information based on the network component information associated with the current sub-question information; then, the server extracts the answer information corresponding to the previous question information of the current question information from the prompt information carried by the current sub-question information, and updates the initial answer information based on the answer information to obtain the updated initial answer information as the answer information corresponding to the current sub-question information.

[0100] For example, see Figure 4, when the current sub-question information is sub-question 2, through the pre-trained question processing model, the server identifies the network component information associated with the current sub-question information as "network service component" based on the prompt information carried by sub-question 2 "generating network service networking topology according to product business requirements" and "generating network service networking topology by assembling one or more resource components according to product business requirements"; then, the server generates the initial answer information corresponding to sub-question 2 "generating business service networking topology by using one or more network service components" according to the network component information "network service component"; then, the server extracts the answer information corresponding to the previous question information of sub-question 2 from the prompt information carried by sub-question 2, that is, the answer corresponding to sub-question 1 "generating network service networking topology by assembling one or more resource components according to product business requirements", and updates the initial answer information according to the answer information to obtain the answer corresponding to sub-question 2 "generating network service networking topology by assembling one or more resource components according to product business requirements, so generating business service networking topology by using one or more network service components".

[0101] In this embodiment, network component information associated with the current sub-question is identified based on the prompt information carried by the current sub-question information, and initial answer information is generated based on the identified network component information. The initial answer is then updated based on the prompt information, thereby making full use of various information sources, gradually optimizing the answer, and thus improving the accuracy of the answer.

[0102] In an exemplary embodiment, the above step S501 inputs the first sub-question information into a pre-trained question processing model to obtain answer information corresponding to the first sub-question information, which specifically includes the following contents: identifying the network component information associated with the first sub-question information through the pre-trained question processing model; and generating the answer information corresponding to the first sub-question information based on the network component information.

[0103] Exemplarily, the server inputs the first sub-question information into a pre-trained question processing model, and identifies the network component information associated with the first sub-question information through the pre-trained question processing model; then, the server generates answer information corresponding to the first sub-question information based on the network component information.

[0104] For example, see Figure 4 , the server identifies the network component information corresponding to sub-question 1 as "resource component"; then, based on the network component information "resource component", the server generates the answer corresponding to sub-question 1 "Use one or more resource components to assemble and generate the network service networking topology according to product business requirements."

[0105] In this embodiment, for the first sub-question information, the pre-trained question processing model can quickly identify the network component information associated with it. The entire process does not require human intervention, avoiding the manual processing method that is easy to consume a lot of time, which is conducive to improving the efficiency of answer determination.

[0106] In an exemplary embodiment, the above step S103 determines the target answer information corresponding to the question information based on the answer information corresponding to each sub-question information, and specifically includes the following contents: from the answer information corresponding to each sub-question information, filter out the answer information corresponding to the last sub-question information; confirm the answer information corresponding to the last sub-question information as the target answer information corresponding to the question information.

[0107] The last sub-question information refers to the corresponding sub-question information arranged at the end. In an actual scenario, the last sub-question information refers to sub-question 3.

[0108] Exemplarily, the server filters out the answer information corresponding to the last sub-question information from the answer information corresponding to each sub-question information; then, the server confirms the answer information corresponding to the last sub-question information as the target answer information corresponding to the question information associated with the product network topology design.

[0109] For example, see Figure 4 The server selects the answer corresponding to sub-question 3 from the answers corresponding to sub-question 1, the answers corresponding to sub-question 2, and the answers corresponding to sub-question 3, "One or more resource components are used to assemble and generate the network service networking topology according to the product business requirements, so one or more network service components are used to generate the business service networking topology, so one or more business service components are used to generate the product networking topology", and confirms the answer corresponding to sub-question 3 as the target answer information corresponding to the question information associated with the product networking topology design.

[0110] In this embodiment, by filtering out the answer information of the last sub-question from the answer information of each sub-question and confirming it as the target answer information corresponding to the question information, the most critical answer can be directly obtained without the need for tedious searching and judging among the answers to numerous sub-questions, which is conducive to improving the efficiency of answer determination.

[0111] Furthermore, the answer determination method provided in the present application also includes a training step for a pre-trained question processing model, which specifically includes the following contents: obtaining sample question information, actual sub-question information after decomposition of the sample question information, and actual sub-answer information corresponding to the actual sub-question information; iteratively training the question processing model to be trained based on the sample question information, the actual sub-question information and the actual sub-answer information to obtain a pre-trained question processing model.

[0112] Specifically, the server obtains sample question information, actual sub-question information after decomposition of the sample question information, and actual sub-answer information corresponding to the actual sub-question information; then, the server inputs the sample question information into the question processing model to be trained, and obtains predicted sub-question information after decomposition of the sample question information, and ranking information of each predicted sub-question information; then, the server inputs each predicted sub-question information into the question processing model to be trained in turn according to the ranking information of each predicted sub-question information, and obtains predicted sub-answer information corresponding to each predicted sub-question information; then, the server obtains a first loss value and a root loss value according to the difference between the predicted sub-question information after decomposition of the sample question information and the actual sub-question information. A second loss value is obtained according to the difference between the predicted sub-answer information corresponding to the predicted sub-question information and the actual sub-answer information corresponding to the actual sub-question information; then, the server obtains a target loss value based on the first loss value and the second loss value; for example, the server performs weighted summation processing on the first loss value and the second loss value to obtain the target loss value; then, the server adjusts the model parameters of the problem processing model to be trained according to the target loss value; then, the server re-trains the problem processing model after the model parameters are adjusted until the target loss value obtained by the trained problem processing model is less than the loss value threshold, then stops training, and uses the trained problem processing model as the pre-trained problem processing model.

[0113] In an exemplary embodiment, Figure 6 As shown, another answer determination method is provided, which is described by taking the method applied to a server as an example, and includes the following steps:

[0114] Step S601, obtaining a target document associated with a product network topology design.

[0115] Step S602 : Decomposing the problem information associated with the product network topology design based on the target document through the problem processing model to obtain a plurality of sorted sub-problem information.

[0116] Step S603, when the current sub-question information is the first sub-question information, the network component information associated with the first sub-question information is identified through a pre-trained question processing model, and answer information corresponding to the first sub-question information is generated according to the network component information.

[0117] Step S604, when the current sub-question information is any sub-question information except the first sub-question information, obtain all sub-question information arranged before the current sub-question information, and the answer information corresponding to each sub-question information in all the sub-question information, and use all the sub-question information and the answer information corresponding to each sub-question information in all the sub-question information as the prompt information carried by the current sub-question information.

[0118] Step S605, through the pre-trained question processing model, based on the prompt information carried by the current sub-question information, identify the network component information associated with the current sub-question information, generate initial answer information corresponding to the current sub-question information according to the network component information, and update the initial answer information according to the prompt information carried by the current sub-question information to obtain the answer information corresponding to the current sub-question information.

[0119] Step S606, taking the next sub-question information of the current sub-question information as the current sub-question information, and jumping to obtain all sub-question information arranged before the current sub-question information, and the answer information corresponding to each sub-question information in all the sub-question information, and taking all the sub-question information and the answer information corresponding to each sub-question information in all the sub-question information as the prompt information carried by the current sub-question information, until the current sub-question information is the last sub-question information.

[0120] Step S607: Filter out the answer information corresponding to the last sub-question information from the answer information corresponding to each sub-question information, and confirm the answer information corresponding to the last sub-question information as the target answer information corresponding to the question information.

[0121] In the above-mentioned answer determination method, in the process of solving the problem associated with the product networking topology design, the problem information associated with the product networking topology design is first decomposed to obtain a plurality of arranged sub-problem information, and any sub-problem information except the first sub-problem information in the arranged plurality of sub-problem information carries prompt information from the least to the most in sequence, and then these sub-problem information are processed in sequence using the prompt information from the least to the most, so that the processing basis of the sub-problem information is increasing, and then the processing accuracy of the sub-problem information is increasing, which is conducive to making the target answer information obtained in the end more accurate, thereby improving the accuracy of answer determination; moreover, this method avoids the defect of low answer determination accuracy caused by inputting the entire problem into the model and directly using the output content of the model as the answer, and further improves the accuracy of answer determination.

[0122] In an exemplary embodiment, in order to more clearly illustrate the answer determination method provided by the embodiment of the present application, the answer determination method is specifically described below with a specific embodiment. In one embodiment, the present application also provides ordered prompts from least to most to realize the complex reasoning of product networking topology design under the PSR mode. In the process of solving problems associated with product networking topology design, the problem information associated with product networking topology design is first decomposed and processed to obtain a plurality of sorted sub-problem information, and then the first sub-problem information in the sorted plurality of sub-problem information is input into a pre-trained problem processing model to obtain the answer information corresponding to the first sub-problem information. For any sub-problem information except the first sub-problem information, the sub-problem information arranged before the sub-problem information and the answer information corresponding to the sub-problem information are used as the prompt information carried by the sub-problem information, and the sub-problem information and the prompt information carried by the sub-problem information are input into the pre-trained problem processing model to obtain the answer information corresponding to the sub-problem information. Finally, according to the answer information corresponding to each sub-problem information, the target answer information corresponding to the problem information is determined. Specifically includes the following contents:

[0123] In the initial stage of product networking topology design, a divide-and-conquer strategy was adopted, breaking down the complex design task into several more manageable subtasks with ordered prompts from the least to the most. Specifically, the problems were divided into networking topology design problems for RFS, CFS, and PRODUCT in an orderly manner, and each subtask was carried out for a specific network component. Next, these subtasks were answered one by one in the established order. This approach not only helps the model to gradually understand the entire design process in depth, but also gradually solve the entire complex design task. In these two stages, only a small number of prompts were used to guide the model, avoiding large-scale fine-tuning of the model. This strategy not only improves the efficiency of the design, but also ensures the flexibility and accuracy of the model in dealing with complex problems; the specific contents are as follows:

[0124] (i) In the initial stage of problem decomposition, a set of standardized prompt processes is adopted, which clearly guides how to break down a complex task into multiple easy-to-operate subtasks, such as Figure 3 As shown. The prompts of this process are pre-set to ensure the consistency and systematicness of task decomposition, thus laying a solid foundation for subsequent design work. This approach ensures that each subtask is precisely defined and has a clear execution path, which promotes the efficiency and orderliness of the entire design process.

[0125] (ii) During the subtask solving phase, the prompt structure consists of three key parts to ensure the coherence and effectiveness of the problem solving process. Figure 4 shown.

[0126] (1) Example Guidance: Provide fixed examples that explain in detail how to solve specific sub-problems, providing clear solution strategies and methodologies for the model.

[0127] (2) Problem and solution record: A possibly empty list that records previously answered sub-problems and their corresponding generated solutions. This list may be empty, but it provides a reference frame for the model to help it review and learn from solved problems.

[0128] (3) Problem prompts: Clearly indicate the next sub-problem to be solved, ensuring that the model can focus and process the remaining tasks one by one in sequence. This structured prompting method not only helps the model to systematically understand and solve each sub-problem, but also promotes the orderly progress of the entire design task. In this way, it can be ensured that each sub-task receives appropriate attention and is solved in an efficient and precise manner.

[0129] In the above-mentioned embodiment, in the process of solving the problem associated with the product networking topology design, the problem information associated with the product networking topology design is first decomposed to obtain a plurality of arranged sub-problem information, and any sub-problem information except the first sub-problem information in the arranged plurality of sub-problem information carries prompt information from the least to the most in sequence, and then these sub-problem information are processed in sequence using the prompt information from the least to the most, so that the processing basis of the sub-problem information is increasing, and then the processing accuracy of the sub-problem information is increasing, which is conducive to making the target answer information finally obtained more accurate, thereby improving the accuracy of answer determination; moreover, the method avoids the defect of low answer determination accuracy caused by inputting the entire problem into the model and directly using the output content of the model as the answer, and further improves the accuracy of answer determination. At the same time, this embodiment proposes a complex reasoning scheme for realizing product networking topology design under the PSR mode by using ordered prompts from least to most. From least to most prompts, the language model is endowed with the ability to solve higher-level problems. This embodiment combines top-down problem refinement with bottom-up solution construction, which is highly consistent with the hierarchical design philosophy of the PSR mode, thereby successfully realizing in-depth and complex reasoning of product networking topology design under the PSR mode. This embodiment can be applied to the networking topology design recommendation link in the AI ​​(Artificial Intelligence) assisted design module.

[0130] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0131] Based on the same inventive concept, the embodiment of the present application also provides an answer determination device for implementing the answer determination method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more answer determination device embodiments provided below can refer to the limitations of the answer determination method above, and will not be repeated here.

[0132] In an exemplary embodiment, Figure 7 As shown, an answer determination device is provided, comprising: a question decomposition module 701, a question processing module 702 and an answer determination module 703, wherein:

[0133] The problem decomposition module 701 is used to decompose the problem information associated with the product networking topology design to obtain a plurality of sorted sub-problem information.

[0134] The question processing module 702 is used to input each sub-question information into a pre-trained question processing model in turn to obtain answer information corresponding to each sub-question information; any sub-question information except the first sub-question information carries prompt information, and the prompt information carried by any sub-question information includes the sub-question information arranged before any sub-question information, and the answer information corresponding to the sub-question information.

[0135] The answer determination module 703 is used to determine the target answer information corresponding to the question information according to the answer information corresponding to each sub-question information.

[0136] In an exemplary embodiment, the problem decomposition module 701 is also used to obtain a target document associated with the product networking topology design; and decompose the problem information based on the target document through the problem processing model to obtain a plurality of sorted sub-problem information.

[0137] In an exemplary embodiment, the question processing module 702 is also used to, when the current sub-question information is the first sub-question information, input the first sub-question information into a pre-trained question processing model to obtain answer information corresponding to the first sub-question information; when the current sub-question information is any sub-question information other than the first sub-question information, determine the prompt information carried by the current sub-question information, input the current sub-question information and the prompt information carried by the current sub-question information into the pre-trained question processing model to obtain answer information corresponding to the current sub-question information; use the next sub-question information of the current sub-question information as the current sub-question information, and jump to the step of determining the prompt information carried by the current sub-question information, until the current sub-question information is the last sub-question information.

[0138] In an exemplary embodiment, the question processing module 702 is also used to obtain all sub-question information arranged before the current sub-question information, and the answer information corresponding to each sub-question information in all the sub-question information; all the sub-question information and the answer information corresponding to each sub-question information in all the sub-question information are used as the prompt information carried by the current sub-question information.

[0139] In an exemplary embodiment, the question processing module 702 is also used to identify network component information associated with the current sub-question information based on the prompt information carried by the current sub-question information through a pre-trained question processing model; generate initial answer information corresponding to the current sub-question information based on the network component information; and update the initial answer information based on the prompt information carried by the current sub-question information to obtain answer information corresponding to the current sub-question information.

[0140] In an exemplary embodiment, the question processing module 702 is further used to identify the network component information associated with the first sub-question information through a pre-trained question processing model; and generate answer information corresponding to the first sub-question information based on the network component information.

[0141] In an exemplary embodiment, the answer determination module 703 is also used to filter out the answer information corresponding to the last sub-question information from the answer information corresponding to each sub-question information; and confirm the answer information corresponding to the last sub-question information as the target answer information corresponding to the question information.

[0142] Each module in the above answer determination device can be implemented in whole or in part by software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each module.

[0143] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Figure 8 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data such as question information and answer information. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, an answer determination method is implemented.

[0144] Those skilled in the art will understand that Figure 8 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0145] In an exemplary embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0146] In an exemplary embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0147] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0148] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0149] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0150] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for determining an answer, characterized in that: The method comprises: Decompose the problem information related to the product network topology design to obtain multiple sorted sub-problem information; Input each of the sub-question information into a pre-trained question processing model in turn to obtain answer information corresponding to each of the sub-question information; any sub-question information except the first sub-question information in each of the sub-question information carries prompt information, and the prompt information carried by any of the sub-question information includes the sub-question information arranged before any of the sub-question information, and the answer information corresponding to the sub-question information; According to the answer information corresponding to each of the sub-question information, target answer information corresponding to the question information is determined.

2. The method according to claim 1, characterized in that The problem information associated with the product network topology design is decomposed to obtain a plurality of sorted sub-problem information, including: Obtain target documents associated with product network topology design; The question information is decomposed and processed based on the target document by the question processing model to obtain a plurality of sorted sub-question information.

3. The method according to claim 1, characterized in that: The step of sequentially inputting each of the sub-question information into a pre-trained question processing model to obtain answer information corresponding to each of the sub-question information includes: In the case where the current sub-question information is the first sub-question information, inputting the first sub-question information into a pre-trained question processing model to obtain answer information corresponding to the first sub-question information; In the case where the current sub-question information is any sub-question information except the first sub-question information, determining the prompt information carried by the current sub-question information, inputting the current sub-question information and the prompt information carried by the current sub-question information into the pre-trained question processing model, and obtaining answer information corresponding to the current sub-question information; The next sub-question information of the current sub-question information is used as the current sub-question information, and the process jumps to the step of determining the prompt information carried by the current sub-question information until the current sub-question information is the last sub-question information.

4. The method according to claim 3, characterized in that The determining the prompt information carried by the current sub-question information includes: Obtaining all sub-question information arranged before the current sub-question information, and answer information corresponding to each sub-question information in all the sub-question information; All the sub-question information and the answer information corresponding to each sub-question information in all the sub-question information are used as the prompt information carried by the current sub-question information.

5. The method according to claim 3, characterized in that: The step of inputting the current sub-question information and the prompt information carried by the current sub-question information into the pre-trained question processing model to obtain the answer information corresponding to the current sub-question information includes: Using a pre-trained question processing model, based on prompt information carried by the current sub-question information, identifying network component information associated with the current sub-question information; Generating initial answer information corresponding to the current sub-question information according to the network component information; The initial answer information is updated according to the prompt information carried by the current sub-question information to obtain the answer information corresponding to the current sub-question information.

6. The method according to claim 3, characterized in that: The step of inputting the first sub-question information into a pre-trained question processing model to obtain answer information corresponding to the first sub-question information includes: Identifying network component information associated with the first sub-question information through a pre-trained question processing model; Answer information corresponding to the first sub-question information is generated according to the network component information.

7. The method according to any one of claims 1 to 6, characterized in that: The step of determining target answer information corresponding to the question information according to answer information corresponding to each of the sub-question information includes: Filtering out the answer information corresponding to the last sub-question information from the answer information corresponding to each of the sub-question information; The answer information corresponding to the last sub-question information is confirmed as the target answer information corresponding to the question information.

8. An answer determination device, characterized in that: The device comprises: A problem decomposition module is used to decompose the problem information associated with the product network topology design to obtain a plurality of sorted sub-problem information; A question processing module is used to input each of the sub-question information into a pre-trained question processing model in turn to obtain answer information corresponding to each of the sub-question information; any sub-question information except the first sub-question information in each of the sub-question information carries prompt information, and the prompt information carried by any sub-question information includes the sub-question information arranged before any of the sub-question information, and the answer information corresponding to the sub-question information; The answer determination module is used to determine the target answer information corresponding to the question information according to the answer information corresponding to each of the sub-question information.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

11. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images