Intelligent agent processing method and device, equipment and storage medium

By conducting simulation and detection of the agent, the problems of service stability and user experience of the agent are solved, the automation of the intelligent inspection and timely detection and processing of problems are realized, and the stability and user experience of the agent are improved.

CN119025412APending Publication Date: 2024-11-26BAIDU INT TECH (SHENZHEN) CO LTD
View PDF 0 Cites 4 Cited by

Patent Information

Application Number
CN202411001380.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-24
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

With the increase in the number of intelligent bodies online on the intelligent body platform, the stability of intelligent body services has become more and more important to the user experience. However, the existing technology lacks an effective inspection mechanism, which makes it difficult to detect and deal with online problems in a timely manner, affecting the impact of The stable operation and user experience of the agent.

Method used

Provide an agent processing method, by determining the target agent to be processed from multiple agents, obtaining the test cases required for simulation detection, and simulate the target agent based on the test cases to realize the inspection of the agent. Automate the process, promptly discover and deal with online problems, and ensure the stable operation of the agent and the accuracy and reliability of the output results.

Benefits of technology

The automated process of intelligent body inspection has been realized, and the online problems of intelligent body are discovered and dealt with in a timely manner, the stability and user experience of intelligent body are improved, maintenance costs are reduced, and a basis for the expansion and upgrading of intelligent body are provided.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119025412A_ABST
    Figure CN119025412A_ABST
Patent Text Reader

Abstract

The invention provides an agent processing method and device, equipment and a storage medium, and relates to the technical field of data processing, in particular to the fields of artificial intelligence, intelligent assistants, virtual assistants and the like. According to the specific implementation scheme, the method comprises the steps of determining a to-be-processed target agent from a plurality of agents under the condition that the agents need to be inspected; wherein each agent in the plurality of agents is an intelligent system which is packaged with a large model and can execute a reasoning task; obtaining at least one target test case required for performing simulation detection on the target agent; and based on the at least one target test case, performing simulation detection on the target agent, and obtaining a detection result.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present disclosure relates to the field of data processing technology, and in particular to the fields of artificial intelligence, smart assistants, virtual assistants, etc. Background Art

[0002] As the number of agents online on the agent platform continues to increase, the stability of agent services is becoming more and more important to the user experience. To ensure that users can enjoy high-quality and stable services, it is necessary to establish a patrol mechanism to promptly detect and handle online problems. Summary of the invention

[0003] The present disclosure provides an intelligent agent processing method, device, equipment and storage medium.

[0004] According to one aspect of the present disclosure, there is provided an agent processing method, comprising:

[0005] When it is necessary to inspect the intelligent agent, a target intelligent agent to be processed is determined from multiple intelligent agents; wherein each of the multiple intelligent agents is an intelligent system encapsulated with a large model and capable of performing reasoning tasks;

[0006] Obtaining at least one target test case required for simulating and detecting the target intelligent agent;

[0007] Based on the at least one target test case, a simulation test is performed on the target agent to obtain a test result.

[0008] According to another aspect of the present disclosure, there is provided an intelligent agent processing device, comprising:

[0009] An object determination unit, used to determine a target intelligent agent to be processed from a plurality of intelligent agents when an intelligent agent needs to be inspected; wherein each of the plurality of intelligent agents is an intelligent system encapsulated with a large model and capable of performing reasoning tasks;

[0010] A data acquisition unit, used to acquire at least one target test case required for simulation detection of the target intelligent agent;

[0011] The detection unit is used to perform simulation detection on the target intelligent agent based on the at least one target test case and obtain a detection result.

[0012] According to another aspect of the present disclosure, there is provided an electronic device, comprising:

[0013] at least one processor; and

[0014] a memory communicatively connected to the at least one processor; wherein,

[0015] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method of any embodiment of the present disclosure.

[0016] According to another aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable the computer to execute the method according to any embodiment of the present disclosure.

[0017] According to another aspect of the present disclosure, a computer program product is provided, including a computer program. When the computer program is executed by a processor, the method according to any one of the embodiments of the present disclosure is implemented.

[0018] In this way, the disclosed scheme can utilize at least one target test case to simulate detection of the intelligent agent to be inspected (such as the target intelligent agent) to obtain the detection results. In this way, the automated process of intelligent agent inspection is realized, which provides strong support for timely discovery and handling of online problems existing in the intelligent agent to ensure the stable operation of the intelligent agent. At the same time, it effectively guarantees the accuracy and reliability of the output results of the intelligent agent, thereby improving the user experience of online users.

[0019] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0021] Figure 1 This is a schematic flow chart of an intelligent agent processing method according to an embodiment of the present application. Figure 1 ;

[0022] Figure 2 This is a schematic flow chart of an intelligent agent processing method according to an embodiment of the present application. Figure 2 ;

[0023] Figure 3 This is a schematic flow chart of an intelligent agent processing method according to an embodiment of the present application. Figure 3 ;

[0024] Figure 4 is a flowchart of an intelligent agent processing method in a specific example according to an embodiment of the present application;

[0025] Figure 5 is a schematic diagram of the structure of an intelligent agent processing device according to an embodiment of the present application;

[0026] Figure 6 It is a block diagram of an electronic device used to implement the intelligent agent processing method of the embodiment of the present disclosure. DETAILED DESCRIPTION

[0027] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0028] The term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there may be three relationships. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. The term "at least one" in this article means any combination of at least two of any one or more of a plurality of. For example, including at least one of A, B, and C can mean including any one or more elements selected from the set consisting of A, B, and C. The terms "first" and "second" in this article refer to multiple similar technical terms and distinguish them. They do not mean to limit the order or to limit them to only two. For example, the first feature and the second feature refer to two types / two features. The first feature can be one or more, and the second feature can also be one or more.

[0029] In addition, in order to better illustrate the present disclosure, numerous specific details are given in the specific embodiments below. It should be understood by those skilled in the art that the present disclosure can also be implemented without certain specific details. In some examples, methods, means, components and circuits well known to those skilled in the art are not described in detail in order to highlight the main purpose of the present disclosure.

[0030] The following is a description of the related technologies of the embodiments of the present disclosure. The following related technologies are optional solutions that can be arbitrarily combined with the technical solutions of the embodiments of the present disclosure, and they all belong to the protection scope of the embodiments of the present disclosure.

[0031] As the number of agents (hereinafter referred to as agents) on the agent platform continues to increase, the stability of agent services is becoming more and more important to the user experience. To ensure that users can enjoy high-quality and stable services, it is necessary to establish a patrol mechanism to promptly detect and handle online problems.

[0032] Based on this, the disclosed solution proposes an intelligent agent processing method to regularly inspect the intelligent agent.

[0033] Specifically, Figure 1This is a schematic flow chart of an intelligent agent processing method according to an embodiment of the present application. Figure 1 The method may be optionally applied to electronic devices, such as personal computers, servers, server clusters, and other electronic devices.

[0034] Further, the method includes at least part of the following contents. Figure 1 As shown, including:

[0035] Step S101: when it is necessary to inspect an intelligent agent, a target intelligent agent to be processed is determined from multiple intelligent agents.

[0036] Here, in one example, each of the multiple agents can specifically refer to an intelligent system that can collect information in a specific environment, and use a built-in knowledge base and reasoning mechanism to autonomously plan and implement actions based on the collected information to achieve a specific goal. Furthermore, the agent described in the disclosed solution can also be called an intelligent assistant or a virtual assistant.

[0037] Further, in one example, each of the multiple agents may be specifically an intelligent system that encapsulates a large model and can perform reasoning tasks. For example, in one example, the agent may be specifically a chatbot, a text writing assistant, or an image generation assistant.

[0038] Furthermore, the large model encapsulated by the agent can be specifically a large language model, or a generative model such as an image generation model, and the present disclosure does not limit the specific large model encapsulated by the agent. Furthermore, the present disclosure does not limit the number of large models encapsulated by the agent, for example, one or more large models can be encapsulated based on the specific functions to be implemented.

[0039] Step S102: Obtain at least one target test case required for simulating and detecting the target intelligent agent.

[0040] It should be noted that when two or more target test cases are obtained, two or more target test cases are also used to perform simulated detection on the intelligent agent in parallel to obtain detection results. At this time, the inspection result contains all detection information obtained from the simulated detection.

[0041] Step S103: Based on the at least one target test case, simulate and test the target agent and obtain a test result.

[0042] Here, the detection result is at least used to indicate whether the target agent has an abnormality. Further, in one example, the detection result may also include abnormality information of the agent when the agent has an abnormality.

[0043] In this way, the disclosed scheme can utilize at least one target test case to simulate detection of the intelligent agent to be inspected (such as the target intelligent agent) to obtain the detection results. In this way, the automated process of intelligent agent inspection is realized, which provides strong support for timely discovery and handling of online problems existing in the intelligent agent to ensure the stable operation of the intelligent agent. At the same time, it effectively guarantees the accuracy and reliability of the output results of the intelligent agent, thereby improving the user experience of online users.

[0044] Furthermore, since the disclosed solution collects the inspection data of the intelligent body (for example, the detection results described above) through regular inspections of the intelligent body, in practical applications, it is also possible to identify the bottlenecks and optimization points of the intelligent body based on the obtained inspection data, thereby providing a strong basis for the subsequent expansion and upgrading of the intelligent body.

[0045] Here, in a specific example, the multiple agents satisfy at least one of the following:

[0046] Different agents perform different target tasks;

[0047] The big models encapsulated by different agents are different or the same.

[0048] In this way, the disclosed solution is applicable to intelligent bodies of different types and purposes, has strong compatibility, and is easy to deploy quickly; in addition, the disclosed solution also reduces the development and maintenance costs required for intelligent body inspections.

[0049] Furthermore, in one example, after the detection result is obtained and it is determined that the detection result is an abnormality in the target intelligent agent, prompt information can also be generated, for example, prompt information (such as alarm information) is generated for the abnormal information of the target intelligent agent. In this way, it is convenient to timely discover and deal with intelligent agents with problems, thereby ensuring the stable operation of the intelligent agent online.

[0050] For example, in one example, after the inspection is completed, the inspection results are sent to the inspection platform through the inspection task callback and / or message middleware, and the inspection platform parses the received data to obtain the analysis results; when it is determined that the analysis result is that there is an abnormality in the target intelligent agent, based on the basic information of the target intelligent agent and the abnormal information of the target intelligent agent, prompt information for the target intelligent agent is generated to prompt manual processing in a timely manner.

[0051] In addition, it should be noted that in actual applications, the inspection time can be pre-set based on actual needs, and then whether the intelligent agent needs to be inspected is determined based on the pre-set inspection time. Further, if it is determined that the inspection is required based on the inspection time, the target intelligent agent to be processed is determined from multiple intelligent agents, so that regular automatic inspection of the intelligent agent is realized to ensure the normal operation of the intelligent agent.

[0052] Furthermore, in one example, the inspection timing described above may be determined based on actual scenario requirements, such as setting a fixed inspection time, etc., and the disclosed solution does not impose specific restrictions on this.

[0053] Furthermore, in a specific example, the following agent screening strategy may be used to obtain the target agent; specifically, the above-mentioned determination of the target agent to be processed from the multiple agents may specifically include:

[0054] Obtain characteristic parameters of multiple agents in an online state;

[0055] From the multiple agents, at least one target agent whose characteristic parameters satisfy the preset rules is selected.

[0056] That is to say, in this example, when it is necessary to inspect the intelligent agents, an inspection of the intelligent agents in the online state is triggered, and the target intelligent agents are screened out based on the characteristic parameters of the intelligent agents in the online state.

[0057] For example, in one example, the characteristic parameters of the agent include but are not limited to at least one of the following: page views (PV), online time (which can be recorded as ΔT), and update time.

[0058] Furthermore, it should be noted that the online status in this example may specifically refer to a status in which the product (such as an intelligent agent) has been deployed in a production environment and is able to operate normally according to established functions at any time to provide continuous services or complete specific operating instructions.

[0059] In addition, it should be noted that the preset rules can be set based on the characteristic parameters of the intelligent agent; for example, in one example, the preset rules can be set based on at least one of the following characteristic parameters of the intelligent agent: page views, online time, and number of queries per second.

[0060] For example, the target agent can be obtained by screening based on the following methods:

[0061] Method 1: Filter out target agents with PV values ​​greater than or equal to 50 and PV values ​​less than 100 from multiple agents. Further, the filtered target agents can be added to queue 1.

[0062] Method 2: Filter out the agents whose update time is less than the first preset time, for example, less than one week, but whose PV value is greater than or equal to 100 from multiple agents; further, the filtered target agents can be added to queue 2.

[0063] Method three: select a target agent with a PV value greater than or equal to 100 from multiple agents, and further add the selected target agent to queue 3.

[0064] Method four, from multiple agents, select a target agent whose online time ΔT is less than or equal to a second preset time, for example, 10 minutes; further, the selected target agent can be added to queue 4.

[0065] In this way, by using the above preset rules, the target intelligent agent that needs to be simulated and tested can be screened.

[0066] Furthermore, the preset rules may also include priority rules, for example, setting the priorities of the queues obtained above, such as from high to low: priority queue 4> priority queue 3> priority queue 2> priority queue 1. In this way, the intelligent agents with high priorities can be simulated and tested first, thereby effectively optimizing the resource allocation of intelligent agent inspections, ensuring the stable operation of key intelligent agents, and thus ensuring that users can enjoy high-quality and stable services.

[0067] It should be pointed out that the above is only an example. In actual applications, specific preset rules and priority division rules built into the preset rules can be set according to the specific needs of the intelligent body inspection, and the present disclosure does not impose specific restrictions on this.

[0068] In this way, the disclosed scheme provides a specific solution for screening the target intelligent body. The scheme is simple and efficient, and it also effectively avoids the full inspection of the intelligent body, saves the network bandwidth and computing resources required for the intelligent body inspection, and reduces maintenance costs, thereby improving the efficiency and quality of the intelligent body inspection.

[0069] Figure 2 This is a schematic flow chart of an intelligent agent processing method according to an embodiment of the present application. Figure 2 The method can be optionally applied to electronic devices, such as personal computers, servers, server clusters, and other electronic devices. It can be understood that the above Figure 1 The relevant contents of the method shown can also be applied to this example, and the relevant contents will not be described in detail in this example.

[0070] Further, the method includes at least part of the following contents. Figure 2 As shown, including:

[0071] Step S201: when it is necessary to inspect an intelligent agent, a target intelligent agent to be processed is determined from multiple intelligent agents.

[0072] Here, each of the multiple agents is an intelligent system that encapsulates a large model and can perform reasoning tasks.

[0073] Here, for the relevant content about the intelligent agent, please refer to the above description and will not be repeated here.

[0074] Step S202: calling the model interface of the large language model.

[0075] Step S203: using the called large language model, obtaining at least one target test case for simulating detection of the target intelligent agent.

[0076] Step S204: Based on the at least one target test case, simulate and test the target agent and obtain a test result.

[0077] In this way, the disclosed solution can use the large language model to obtain at least one target test case for the target intelligent agent, and then use the at least one target test case to perform simulated detection on the target intelligent agent. In this way, the accuracy and reliability of the inspection results of the target intelligent agent are effectively improved, thereby laying the foundation for ensuring the quality and performance of the target intelligent agent and improving the user experience.

[0078] Further, in a specific example, at least one target test case can be obtained in the following manner; specifically, the above-described use of the called large language model to obtain at least one target test case for simulating detection of the target agent (for example, the above-described step S203) specifically includes:

[0079] Step S203-1: Determine the characteristic information of the target agent.

[0080] Here, the characteristic information of the target agent includes at least one of the following: name, function description information, test template, and the number of test cases to be generated.

[0081] Step S203-2: Generate a target prompt word based on the characteristic information of the target intelligent agent.

[0082] Step S203-3: input the target prompt word into a large language model to obtain a plurality of preset test cases for the target intelligent agent.

[0083] In this way, the reasoning ability of the large language model can be effectively utilized to obtain preset test cases suitable for the target intelligent agent, providing strong support for effective simulation and detection of the target intelligent agent.

[0084] Step S203 - 4 : Based on a plurality of preset test cases for the target intelligent agent, at least one target test case for simulating detection of the target intelligent agent is obtained.

[0085] For example, in one example, each preset test case for the target intelligent agent obtained by using a large language model can be directly used as a target test case. Alternatively, in another example, at least one target test case for simulating the detection of the target intelligent agent can be randomly screened out from a plurality of preset test cases for the target intelligent agent. Alternatively, in yet another example, at least one target test case for simulating the detection of the target intelligent agent can be screened out from a plurality of preset test cases for the target intelligent agent based on a preset case screening strategy. For example, in one example, the case screening strategy can be specifically: screen out the test case with the least processing resources occupied by the simulation detection.

[0086] In this way, the disclosed scheme provides a specific scheme for generating test cases for a target intelligent agent, that is, using the characteristic information of the target intelligent agent to obtain target prompt words, and then using a large language model to automatically generate test cases for the target intelligent agent (such as target test cases) according to the target prompt words. In this way, the testing cost is effectively reduced. Moreover, the large language model used in the disclosed scheme can generate test cases according to the specific needs and environment of the intelligent agent. Therefore, the adaptability of the test is effectively improved, which lays a foundation for the subsequent accurate evaluation of the performance and stability of the target intelligent agent, and also lays a foundation for ensuring that high-quality intelligent agents can run stably in the future.

[0087] In a specific example of the disclosed solution, the above-mentioned calling of the model interface of the large language model (for example, the above-mentioned step S202) specifically includes:

[0088] Step S202-1: Check whether there is a preset test case for the target agent in the preset question library. If yes, proceed to step S202-2; otherwise, proceed to step S202-3.

[0089] Step S202-2: When it is determined that there is a preset test case for the target intelligent body, determine whether the preset test case for the target intelligent body meets the detection requirements of the target intelligent body. If so, proceed to step S202-4; otherwise, proceed to step S202-3.

[0090] Step S202 - 3 : calling the model interface of the large language model to generate target test cases using the large language model.

[0091] Step S202 - 4 : using the preset test case for the target intelligent agent as the target test case to perform simulation detection on the target intelligent agent.

[0092] It should be noted that, in one example, the preset question library includes a plurality of preset test cases prepared in advance, for example, test cases used by the intelligent agent in the historical inspection process, or test cases designed based on human experience. It is understandable that the disclosed solution does not impose any specific restrictions on the construction process of the preset question library, as long as the preset question library can be constructed.

[0093] That is to say, in this example, when it is determined that there is no preset test case for the target intelligent agent in the preset question library, or when it is determined that there is a preset test case for the target intelligent agent but the preset test case for the target intelligent agent does not meet the detection requirements of the target intelligent agent, the model interface of the large language model is called to obtain the test case for the target intelligent agent (such as the target test case) using the large language model.

[0094] Furthermore, in one example, after obtaining test cases (e.g., preset test cases) for the target intelligent agent using a large language model, the current preset question library can also be updated based on the obtained test cases for the target intelligent agent, thereby achieving an optimized update of the preset question library, making the preset question library more comprehensive and practical, thereby facilitating faster identification and response to risks that may arise in the intelligent agent.

[0095] In this way, the disclosed solution can provide effective test cases for intelligent body inspection, improve the effectiveness of intelligent body inspection, and further provide data support for improving the accuracy and reliability of intelligent body inspection results, and also lay the foundation for subsequently ensuring the performance and stability of the intelligent body.

[0096] Figure 3 This is a schematic flow chart of an intelligent agent processing method according to an embodiment of the present application. Figure 3 The method can be optionally applied to electronic devices, such as personal computers, servers, server clusters, and other electronic devices. It can be understood that the above Figure 1 and Figure 2 The relevant contents of the method shown can also be applied to this example, and the relevant contents will not be described in detail in this example.

[0097] Further, the method includes at least part of the following contents. Figure 3 As shown, including:

[0098] Step S301: when it is necessary to inspect an intelligent agent, a target intelligent agent to be processed is determined from multiple intelligent agents.

[0099] Here, each of the multiple agents is an intelligent system that encapsulates a large model and can perform reasoning tasks.

[0100] Here, for the relevant content about the intelligent agent, please refer to the above description and will not be repeated here.

[0101] Step S302: Obtain at least one target test case required for simulating and detecting the target intelligent agent.

[0102] Here, regarding the method of obtaining the target test case, please refer to the above description and will not be repeated here.

[0103] Step S303: Determine whether the current number of available queries for the target agent has reached a preset maximum number of available queries; if so, proceed to step S304; otherwise, proceed to step S305.

[0104] It should be noted that, in one example, the preset maximum number of available queries can be obtained based on the preset number of queries per second (QPS) of the target agent. For example, the preset number of queries per second of the target agent is 10. In order to ensure that online requests can be responded to normally and do not affect online use, a preset number (for example, 8) of queries need to be reserved per second. At this time, the preset maximum number of available queries can be specifically equal to the preset number of queries per second of the target agent - the preset number, for example, 10-8=2.

[0105] Step S304: When it is determined that the preset maximum number of available queries has been reached, after waiting for a preset time, return to step S303.

[0106] Step S305: When it is determined that the preset maximum number of available queries has not been reached, a simulation test is performed on the target agent based on the at least one target test case, and a test result is obtained.

[0107] In this way, the disclosed solution provides a specific inspection solution, in which, while simulating the detection of the target intelligent agent, it effectively avoids the intelligent agent from occupying too many computing resources due to inspection, thereby ensuring that the user can access the target intelligent agent normally, that is, ensuring the user's usage experience.

[0108] Furthermore, in a specific example, when the target intelligent agent is simulated for detection, an online call request for the target intelligent agent is responded to to obtain response information for the online call request. That is to say, when the target intelligent agent is inspected, the user can still access and use the target intelligent agent normally, so that the service will not be interrupted due to the intelligent agent inspection, thereby effectively ensuring the continuity of the intelligent agent business and the continuity and satisfaction of the user experience.

[0109] The following is a further detailed description of the disclosed solution with reference to specific examples; the disclosed solution proposes an intelligent agent processing solution. Specifically, Figure 4 As shown, specifically including:

[0110] First of all, Figure 4 The relevant nouns are explained as follows:

[0111] (1) Large language model generation: Based on the large language model capabilities, test cases for the agent can be generated according to the agent, such as the agent's name and description information, to facilitate the construction of a preset question library.

[0112] (2) Inspection and Scheduling Center: It is mainly responsible for the construction (or update) of the preset question library and responding to the scheduled inspection and scheduling. For example, the inspection and scheduling center can be scheduled regularly based on the inspection tasks determined by the task scheduling center. It can realize priority scheduling, task management and other functions.

[0113] (3) Priority scheduling: Based on the task scheduling strategy, the agents are inspected. For example, agents with high page views (PV) are inspected and the scheduling frequency is controlled to ensure the stable operation of high-quality agents.

[0114] (4) Real machine testing: By simulating browser operations, the Agent sends test cases in the preset question library to the Agent through the Agent web address and obtains the response results.

[0115] (5) Message middleware: It is mainly used to receive log information from upstream systems. Downstream systems can subscribe to the message middleware to process and respond to related businesses.

[0116] (6) Communication software alarm: The alarm information can be pushed to the communication software and the relevant personnel can be notified in time.

[0117] (7) Intelligent agent platform: A platform that provides developers with low-cost and efficient construction of intelligent agent applications. The intelligent agent platform supports developers to create native applications based on their own needs.

[0118] Furthermore, the core process of the disclosed solution specifically includes:

[0119] Step 1: Start the initialization task of the preset question library. For example, it can be executed regularly every morning to initialize the preset question library of the agent in an incremental agent manner.

[0120] Here, the construction method of the preset question library includes but is not limited to at least one of the following: manual entry, C-end access log extraction, recommended questions, and large language model generation.

[0121] Step 2: Start the inspection task, push the intelligent agents to be inspected (corresponding to the multiple intelligent agents mentioned above) according to the task scheduling strategy, and determine the target intelligent agent to be processed from the multiple intelligent agents to be inspected.

[0122] Step 3: Use the test cases for the target intelligent agent in the preset question library to perform simulated detection on the target intelligent agent to be processed, such as interface detection, real machine detection and C-end abnormal access log capture to obtain detection results (for example, including interface detection results, real machine inspection results and message logs).

[0123] Here, the interface detection may mainly include whether the interface is unobstructed, the matching degree of the answer result, the result verification, etc. The real machine detection mainly simulates the browser operation to detect whether the intelligent agent has the following abnormal problems: page crash, network abnormality, wrong answer to the question, etc. The C-end abnormal access log capture may mainly capture page crash points, network abnormality points, backend service abnormality points, etc.

[0124] Here, all the inspection results mentioned above can be written into the quality library through an API (application programming interface) interface or a message middleware.

[0125] Step 4: The inspection results are notified to the inspection platform through inspection task callback (for example, task callback through an API interface) or message middleware.

[0126] Step 5: The inspection platform parses the data callback by the inspection task and the data in the message middleware. If it is determined that there is a problem with the target intelligent agent, the inspection problem, error information, etc. of the target intelligent agent are stored in the quality library.

[0127] Here, the quality library can receive low-quality signals and store low-quality data to provide data support for the optimization and upgrade of the target intelligent agent, so as to enhance the user's intelligent agent usage experience.

[0128] Step 6: Generate prompt information based on the basic information and error information of the target intelligent agent, and push the prompt information to the alarm platform for alarm through communication software; the prompt information can also be pushed to the manual review platform for abnormal review and / or offline operation.

[0129] Step 7: Send the manual review results to the intelligent agent platform to notify the developer, and then make repairs and re-release it online.

[0130] In summary, the disclosed solution has several advantages, including:

[0131] First, automated inspection: The disclosed solution uses automated means to regularly inspect the intelligent agent to promptly identify potential problems.

[0132] Second, quick response: After discovering the problems of the intelligent agent, the disclosed solution can respond quickly and provide corresponding treatment measures.

[0133] Third, repair suggestions: The disclosed solution can provide repair suggestions and guidance for intelligent agents with problems to help developers quickly fix the problems.

[0134] Fourth, intelligent monitoring: After the intelligent agent comes back online, the disclosed solution can continue to perform intelligent monitoring to ensure the service stability of the intelligent agent.

[0135] The disclosed solution also provides an intelligent agent processing device, such as Figure 5 As shown, including:

[0136] The object determination unit 501 is used to determine a target intelligent agent to be processed from multiple intelligent agents when it is necessary to inspect the intelligent agent; wherein each intelligent agent in the multiple intelligent agents is an intelligent system encapsulated with a large model and capable of performing reasoning tasks;

[0137] A data acquisition unit 502 is used to acquire at least one target test case required for simulation detection of the target agent;

[0138] The detection unit 503 is used to perform simulation detection on the target agent based on the at least one target test case and obtain a detection result.

[0139] In a specific example of the disclosed solution, the multiple agents satisfy at least one of the following:

[0140] Different agents perform different target tasks;

[0141] The big models encapsulated by different agents are different or the same.

[0142] In a specific example of the disclosed solution, the object determination unit is specifically used to:

[0143] Obtain characteristic parameters of multiple agents in an online state;

[0144] From the multiple agents, at least one target agent whose characteristic parameters satisfy the preset rules is selected.

[0145] In a specific example of the disclosed solution, the data acquisition unit is specifically used to call the model interface of the large language model; using the called large language model, at least one target test case for simulating detection of the target intelligent agent is obtained.

[0146] In a specific example of the disclosed solution, the data acquisition unit is specifically used to:

[0147] Determine characteristic information of the target agent, wherein the characteristic information includes at least one of the following: name, function description information, test template, and the number of test cases to be generated;

[0148] Based on the characteristic information of the target intelligent agent, generating a target prompt word;

[0149] Inputting the target prompt word into a large language model to obtain a plurality of preset test cases for the target intelligent agent;

[0150] Based on a plurality of preset test cases for the target intelligent agent, the at least one target test case for simulating detection of the target intelligent agent is obtained.

[0151] In a specific example of the disclosed solution, the data acquisition unit is specifically used to:

[0152] Detecting whether there is a preset test case for the target intelligent agent in the preset question library;

[0153] When it is determined that there is no preset test case for the target agent, a model interface of the large language model is called.

[0154] In a specific example of the disclosed solution, the detection unit is specifically used to:

[0155] Determining whether the current number of available queries for the target agent reaches a preset maximum number of available queries;

[0156] When it is determined that the preset maximum number of available queries has not been reached, a simulation test is performed on the target agent based on the at least one target test case.

[0157] In a specific example of the disclosed scheme, it also includes: a request response unit; wherein the request response unit is used to: respond to an online call request for the target intelligent agent in the case of a simulated detection of the target intelligent agent; and obtain response information for the online call request.

[0158] For the description of the specific functions and examples of each unit of the device in the embodiment of the present disclosure, reference can be made to the relevant description of the corresponding steps in the above method embodiment, which will not be repeated here.

[0159] In the technical solution disclosed herein, the acquisition, storage and application of user personal information involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0160] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0161] Figure 6 A schematic block diagram of an example electronic device 600 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0162] like Figure 6 As shown, the device 600 includes a computing unit 601, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 602 or a computer program loaded from a storage unit 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the device 600 can also be stored. The computing unit 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0163] A number of components in the device 600 are connected to the I / O interface 605, including: an input unit 606, such as a keyboard, a mouse, etc.; an output unit 607, such as various types of displays, speakers, etc.; a storage unit 608, such as a disk, an optical disk, etc.; and a communication unit 609, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 609 allows the device 600 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0164] The computing unit 601 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 601 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 601 performs the various methods and processes described above, such as the agent processing method. For example, in some embodiments, the agent processing method may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 608. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 600 via the ROM 602 and / or the communication unit 609. When the computer program is loaded into the RAM 603 and executed by the computing unit 601, one or more steps of the agent processing method described above may be performed. Alternatively, in other embodiments, the computing unit 601 may be configured to perform the agent processing method in any other appropriate manner (e.g., by means of firmware).

[0165] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0166] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0167] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0168] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0169] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0170] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0171] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0172] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the principles of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. An agent processing method, comprising: When it is necessary to inspect the intelligent agent, a target intelligent agent to be processed is determined from multiple intelligent agents; wherein each of the multiple intelligent agents is an intelligent system encapsulated with a large model and capable of performing reasoning tasks; Obtaining at least one target test case required for simulating and detecting the target intelligent agent; Based on the at least one target test case, a simulation test is performed on the target agent to obtain a test result.

2. The method according to claim 1, wherein: The multiple agents satisfy at least one of the following: Different agents perform different target tasks; The big models encapsulated by different agents are different or the same.

3. The method according to claim 1, wherein: The step of determining a target intelligent agent to be processed from a plurality of intelligent agents comprises: Obtain characteristic parameters of multiple agents in an online state; From the multiple agents, at least one target agent whose characteristic parameters satisfy the preset rules is selected.

4. The method according to any one of claims 1 to 3, wherein: The obtaining of at least one target test case required for simulating the detection of the target intelligent agent comprises: Call the model interface of the large language model; Using the called large language model, at least one target test case for simulating detection of the target intelligent agent is obtained.

5. The method according to claim 4, wherein: The method of using the called large language model to obtain at least one target test case for simulating detection of the target intelligent agent includes: Determine characteristic information of the target agent, wherein the characteristic information includes at least one of the following: name, function description information, test template, and the number of test cases to be generated; Based on the characteristic information of the target intelligent agent, generating a target prompt word; Inputting the target prompt word into a large language model to obtain a plurality of preset test cases for the target intelligent agent; Based on a plurality of preset test cases for the target intelligent agent, the at least one target test case for simulating detection of the target intelligent agent is obtained.

6. The method according to claim 4, wherein: The model interface for calling the large language model includes: Detecting whether there is a preset test case for the target intelligent agent in the preset question library; When it is determined that there is no preset test case for the target agent, a model interface of the large language model is called.

7. The method according to any one of claims 1 to 6, wherein: The performing simulation detection on the target intelligent agent based on the at least one target test case includes: Determining whether the current number of available queries for the target agent reaches a preset maximum number of available queries; When it is determined that the preset maximum number of available queries has not been reached, a simulation test is performed on the target agent based on the at least one target test case.

8. The method according to claim 7, further comprising: In the case of performing simulation detection on the target intelligent agent, responding to an online call request for the target intelligent agent; Response information for the online call request is obtained.

9. An intelligent agent processing device, comprising: An object determination unit, used to determine a target intelligent agent to be processed from a plurality of intelligent agents when an intelligent agent needs to be inspected; wherein each of the plurality of intelligent agents is an intelligent system encapsulated with a large model and capable of performing reasoning tasks; A data acquisition unit, used to acquire at least one target test case required for simulation detection of the target intelligent agent; The detection unit is used to perform simulation detection on the target intelligent agent based on the at least one target test case and obtain a detection result.

10. The device according to claim 9, wherein: The multiple agents satisfy at least one of the following: Different agents perform different target tasks; The big models encapsulated by different agents are different or the same.

11. The device according to claim 9, wherein: The object determination unit is specifically used to: Obtain characteristic parameters of multiple agents in an online state; From the multiple agents, at least one target agent whose characteristic parameters satisfy the preset rules is selected.

12. The device according to any one of claims 9 to 11, wherein: The data acquisition unit is specifically used to call the model interface of the large language model; using the called large language model, at least one target test case for simulating and detecting the target intelligent agent is obtained.

13. The device according to claim 12, wherein: The data acquisition unit is specifically used for: Determine characteristic information of the target agent, wherein the characteristic information includes at least one of the following: name, function description information, test template, and the number of test cases to be generated; Based on the characteristic information of the target intelligent agent, generating a target prompt word; Inputting the target prompt word into a large language model to obtain a plurality of preset test cases for the target intelligent agent; Based on a plurality of preset test cases for the target intelligent agent, the at least one target test case for simulating detection of the target intelligent agent is obtained.

14. The device according to claim 12, wherein: The data acquisition unit is specifically used for: Detecting whether there is a preset test case for the target intelligent agent in the preset question library; When it is determined that there is no preset test case for the target agent, a model interface of the large language model is called.

15. The device according to any one of claims 9 to 14, wherein: The detection unit is specifically used for: Determining whether the current number of available queries for the target agent reaches a preset maximum number of available queries; When it is determined that the preset maximum number of available queries has not been reached, a simulation test is performed on the target agent based on the at least one target test case.

16. The apparatus according to claim 15, further comprising: Request response unit; wherein the request response unit is used to respond to the online call request for the target intelligent agent in the case of simulated detection of the target intelligent agent; and obtain response information for the online call request.

17. An electronic device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 8.

18. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-8.

19. A computer program product, comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 8.

Citation Information

Cited By

  • Equipment control method and device, computer equipment and storage medium

    CN119717563A

  • Continuous casting quality control method based on meta-cognitive coordination architecture agent cluster

    CN120655248A

  • A continuous casting quality management and control method based on a meta-cognition coordination architecture intelligent agent cluster

    CN120655248B

  • Intelligent agent capability description method, task execution method and device and intelligent agent

    CN120688542A