Task processing method and device, electronic equipment and storage medium

Abstract

The invention discloses a task processing method and device, electronic equipment and a storage medium. The method comprises the steps of obtaining a to-be-processed first task; determining a target agent for processing the first task from the first agent set and / or the second agent set; under the condition that the target agent is the target first agent, allocating a first task to the target first agent, so as to process the first task based on the task information through the target first agent to obtain a target task processing result; and under the condition that the target agent comprises a plurality of target second agents, allocating a first task to the plurality of target second agents, so as to process the first task based on the task information through each target second agent to obtain a task processing result corresponding to each target second agent, determining a target task processing result of the first task based on each task processing result; according to the method and the device, the processing efficiency of the intelligent agent during task processing is effectively improved.

Description

Technical Field

[0001] This disclosure relates to the field of computer technology, and more specifically to a task processing method, apparatus, electronic device, and storage medium. Background Technology

[0002] With the development of computer technology, intelligent agents have become the core execution subjects in various task execution scenarios, such as data annotation, information review, and content generation. Intelligent agents can be artificial intelligence agents or human intelligent agents.

[0003] Currently, task execution solutions mainly fall into two categories: single agent execution and multi-agent collaborative execution. In scenarios where agents perform tasks, either single agents are used to execute different tasks separately, or all agents are used to execute tasks collaboratively. This results in existing agents being unable to effectively balance the efficiency, quality, and cost of task execution. There is a lack of a task processing flow that can balance the efficiency, quality, and cost of task execution, leading to low processing efficiency of agents when processing tasks. Summary of the Invention

[0004] This application provides a task processing method, apparatus, electronic device, and storage medium. By providing a task processing flow that can balance the efficiency, quality, and cost of task execution, it can adapt to task execution scenarios of different scales and types, and effectively improve the processing efficiency of intelligent agents when processing tasks.

[0005] In a first aspect, embodiments of this application provide a task processing method applied to a task processing system. The task processing system is configured with a first set of intelligent agents and a second set of intelligent agents. The first intelligent agents in the first set of intelligent agents are configured to process tasks independently, and the second intelligent agents in the second set of intelligent agents are configured to perform task processing using a swarm of intelligent agents. The method includes: Obtain the first task to be processed, which has corresponding task information set. Determine the target agent for processing the first task from the first set of agents and / or the second set of agents; When the target agent is a first target agent, the first task is assigned to the first target agent so that the first target agent processes the first task based on the task information to obtain the target task processing result; When the target agent includes multiple target second agents, the first task is assigned to the multiple target second agents so that each target second agent processes the first task based on the task information to obtain the task processing result corresponding to each target second agent, and the target task processing result of the first task is determined based on the task processing result of each target second agent.

[0006] Secondly, embodiments of this application provide a task processing apparatus applied to a task processing system. The task processing system is configured with a first set of intelligent agents and a second set of intelligent agents. The first intelligent agents in the first set of intelligent agents are configured to process tasks independently, and the second intelligent agents in the second set of intelligent agents are configured to perform task processing using a swarm of intelligent agents, including: The acquisition unit is used to acquire the first task to be processed, and the first task has corresponding task information set. A determining unit is configured to determine a target agent for processing the first task from the first set of agents and / or the second set of agents; The first allocation unit is configured to allocate the first task to the target first intelligent agent when the target intelligent agent is the target first intelligent agent, so that the target first intelligent agent can process the first task based on the task information to obtain the target task processing result; The second allocation unit is configured to allocate the first task to the multiple target second intelligent agents when the target intelligent agent includes multiple target second intelligent agents, so that each target second intelligent agent processes the first task based on the task information to obtain the task processing result corresponding to each target second intelligent agent, and determines the target task processing result of the first task based on the task processing results.

[0007] Thirdly, embodiments of this application also provide an electronic device, including a memory storing multiple instructions; a processor loading instructions from the memory to execute the steps of any task processing method provided in embodiments of this application.

[0008] Fourthly, embodiments of this application also provide a computer-readable storage medium storing a plurality of instructions adapted for loading by a processor to perform the steps of any of the task processing methods provided in embodiments of this application.

[0009] Fifthly, embodiments of this application also provide a computer program product, including a computer program or instructions, which, when executed by a processor, implement the steps in a processing method for any task provided in embodiments of this application.

[0010] The solution adopted in this application embodiment can be used to complete the task by assigning the first task to be processed to a target first intelligent agent of a first intelligent agent set or to multiple target second intelligent agents of a second intelligent agent set. This provides a task processing flow that can take into account the efficiency, quality and cost of task execution, and can adapt to task execution scenarios of different scales and types, effectively improving the processing efficiency of intelligent agents when processing tasks. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a schematic diagram of a task processing system provided in the embodiments of this application; Figure 2 This is a schematic flowchart of one embodiment of the task processing method provided in this application. Figure 3 This is a schematic flowchart of another embodiment of the task processing method provided in this application. Figure 4 This is a schematic diagram of the structure of the task processing device provided in the embodiments of this application; Figure 5 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0013] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. At the same time, in the description of the embodiments of this application, the terms "first," "second," etc., are only used to distinguish descriptions and should not be construed as indicating or implying relative importance. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0014] This application provides a task processing method, apparatus, electronic device, and computer-readable storage medium. Specifically, this embodiment will be described from the perspective of a task processing apparatus, which can be integrated into an electronic device. That is, the task processing method of this application embodiment can be executed by an electronic device. Optionally, the electronic device may include a terminal device. The terminal device may be a mobile phone, tablet computer, smart Bluetooth device, laptop computer, or personal computer (PC), etc.

[0015] The task processing method provided in this application can be applied to interactive systems, such as terminal devices and servers. The terminal can be a device that includes both receiving and transmitting hardware, i.e., a device with receiving and transmitting hardware capable of performing bidirectional communication over a bidirectional communication link. The terminal device and the server can communicate bidirectionally via a network.

[0016] Optionally, the server can be a standalone server, or a server network or server cluster, including but not limited to computers, network hosts, single network servers, multiple network server sets, or cloud servers composed of multiple servers. Cloud servers consist of a large number of computers or network servers based on cloud computing.

[0017] In one embodiment of this disclosure, the task processing method can run on a local terminal device or a server. When the game interaction method runs on a server, the method can be implemented and executed based on a cloud interaction system, wherein the cloud interaction system includes a server and a client device.

[0018] Please see Figure 1 , Figure 1This is a schematic diagram of a task processing system provided in an embodiment of this application. The system may include at least one terminal, at least one server, at least one database, and a network. A user's terminal can connect to different servers via the network. The terminal is any device with computing hardware capable of supporting and executing software products corresponding to model generation. Furthermore, when the system includes multiple terminals, multiple servers, and multiple networks, different terminals can connect to each other through different networks and servers. The network can be a wireless network or a wired network, such as a wireless local area network (WLAN), local area network (LAN), cellular network, 2G network, 3G network, 4G network, 5G network, etc. Additionally, different terminals can also connect to other terminals or servers using their own Bluetooth networks or hotspot networks. For example, multiple users can connect online through different terminals via appropriate networks and synchronize with each other to support multi-user access. Furthermore, the system may include multiple databases coupled to different servers, and can continuously store information related to the operating environment in the databases while different users are using the system online.

[0019] The following detailed description is provided in conjunction with the accompanying drawings. In this embodiment, the execution subject is a terminal device as an example. It should be noted that the order of description in the following embodiments is not intended to limit the preferred order of the embodiments. Although a logical order is shown in the flowcharts, in some cases, the steps shown or described may be performed in a different order than that shown in the accompanying drawings.

[0020] Please see Figure 2 , Figure 2 This is a flowchart illustrating a task processing method provided in an embodiment of this application. The specific flow of the task processing method can be summarized in steps 101 to 104, wherein: Step 101: Obtain the first task to be processed. The first task has corresponding task information.

[0021] The task information for the first task may include the task's deadline, the estimated time required for the task, the current answer options for the task, and the task's confidence level, etc.

[0022] Step 102: Determine the target agent for processing the first task from the first set of agents and / or the second set of agents.

[0023] Specifically, the task processing method provided in this application embodiment is applied to a task processing system. The task processing system is configured with a first set of intelligent agents and a second set of intelligent agents. The first intelligent agent in the first set of intelligent agents is configured to process tasks independently, and the second intelligent agent in the second set of intelligent agents is configured to process tasks using a group of intelligent agents.

[0024] In one embodiment, tasks can be preferentially assigned to agents with high task processing capabilities. The step "determining the target agent for processing the first task from the first set of agents and / or the second set of agents" includes: Based on the task processing capabilities of each of the first and second agents, a target agent for processing the first task is determined from the set of first agents and / or the set of second agents.

[0025] Specifically, the task processing capabilities of each first agent and each second agent can be obtained, and the first agent with the strongest task processing capability can be selected from the set of first agents as the target first agent. If all first agents in the set of first agents are assigned tasks to process, then multiple target second agents can be selected from the set of second agents to process the first task.

[0026] In this context, task processing capability is determined by the agent's credibility profile and credibility level. Credibility level is a variable / fluctuating confidence interval of the profile, linked to the number of correct / incorrect answers, the most recent verification time, and the verification ratio. Credibility level is one of the factors considered in task allocation. Agents with high credibility are allowed to complete tasks independently. As credibility decreases over time, the agent is required to collaborate with multiple agents to complete the task. In this embodiment, higher credibility level equates to stronger task processing capability.

[0027] In another embodiment, the target agent can be determined based on the quality requirements of the task. The step "the first task has corresponding task quality requirements; determining the target agent for processing the first task from the first set of agents and / or the second set of agents" includes: Based on the task quality requirements, target agents that match the task quality requirements are determined from the first set of agents and / or the second set of agents.

[0028] The task quality requirement can be a confidence level requirement. For example, a container goods checkout task has a high accuracy requirement (i.e., a confidence level requirement), specifically requiring an accuracy of 98.5%. In this case, agents with a confidence level (i.e., accuracy) of 0.985 or higher can be selected from the first agent set to process the first task. If all first agents in the first agent set are assigned tasks, then multiple target second agents with a confidence level (i.e., accuracy) of 0.9 or higher are selected from the second agent set to collaboratively process the first task.

[0029] Step 103: If the target agent is the first target agent, assign the first task to the first target agent so that the first target agent processes the first task based on the task information to obtain the target task processing result.

[0030] Specifically, when the target agent is the first target agent, the first target agent can process the first task independently, thereby obtaining the target task processing result of the first task.

[0031] Step 104: If the target intelligent agent includes multiple target second intelligent agents, the first task is assigned to the multiple target second intelligent agents so that each target second intelligent agent processes the first task based on the task information to obtain the task processing result corresponding to each target second intelligent agent, and the target task processing result of the first task is determined based on the task processing result of each target second intelligent agent.

[0032] Specifically, the step "determining the target task processing result of the first task based on the processing results of each of the tasks" includes: The results of each task are aggregated using preset rules to obtain the target truth value; The target task processing result of the first task is determined based on the target truth value.

[0033] The preset rules can include majority voting, Bayesian estimation (e.g., Bayesian truth inference), interval estimation (e.g., Wilson interval estimation), and so on.

[0034] For example, the first task is the smart vending machine product recognition task. This task is an objective problem with a unique and definite truth value (objective fact). Even if it were a subjective problem, the answers could be aggregated into a final result. In this case, different target second agents may give different answers, so a truth inference method is needed to aggregate these answers. This can be done using methods such as majority voting, Bayesian estimation, and interval estimation to obtain the final target truth value, which is then used as the result of the target task.

[0035] Among these, majority voting is the most basic and easiest-to-use result aggregation method. Specifically, after assigning the task to N target agents, all valid results are collected, and the result with the most occurrences is selected as the final task result. If a tie occurs, additional votes can be added, random selection can be performed, or the result can be weighted according to the agent's confidence level. This is the lowest-cost and most efficient method for implementing agent tasks, requiring no complex calculations and relying purely on frequency statistics. Specifically, it includes the following two rules: (1) Simple majority: The candidate is elected if the number of votes for the result is greater than any other single result (no majority required); (2) Absolute majority: The result must be ≥ 50% + 1 of the total valid votes to be selected as the final result (more rigorous, to avoid interference from scattered results).

[0036] For example, the first task is assigned to 5 target agents, requiring them to determine whether the user's intent is to place an order. The return result is: yes, yes, no, yes, no. There are a total of 5 valid votes. "Yes" gets 3 votes and "no" gets 2 votes. Since this satisfies an absolute majority, the final judgment result is "yes".

[0037] Bayesian estimation is a probability-based statistical inference method. Its core idea is to use new observational data (agent results) to correct the prior probability of the event's true value, ultimately obtaining a more realistic posterior probability. Bayesian truth inference addresses over-probability scenarios where agents have varying capabilities and results contain noise. Through Bayes' theorem, it combines the historical accuracy (prior) of each agent with the current returned result (observation) to calculate the posterior probability of each candidate result as the "true value," ultimately selecting the result with the highest probability as the final task result.

[0038] Interval estimation does not directly provide a unique true value, but rather calculates a "numerical interval that is highly likely to contain the true value" and provides the confidence level of that interval (e.g., at a 95% confidence level, the true value falls within [0.72, 0.88]). For proportional results in binary classification tasks (e.g., after overshooting, k agents return "correct", and there are a total of n agents), at a specified confidence level (commonly 95%), the range of true accuracy is calculated. The narrower the interval, the more reliable the result; at the same time, the interval can be used to determine whether the result is statistically significant.

[0039] It should be noted that experiments can be conducted on simulated data or historical response data of the intelligent agent platform. By selecting a truth inference method and combining it with the profile of the intelligent agent (such as historical accuracy), it is possible to statistically determine which intelligent agents provide answers with a high degree of confidence that lead to task convergence, and these answers are considered correct.

[0040] Based on the above description, the following examples will further illustrate the task processing method of this application. Specific embodiments of the task processing method are described below.

[0041] In one embodiment, after the step of "determining the target task processing result of the first task based on the processing results of each of the tasks", the method further includes: The task processing quality of each of the target second agents for the first task is evaluated based on the target task processing results. Based on the task processing quality of each target second intelligent agent, update the task processing capability of each target second intelligent agent.

[0042] For example, the answers from multiple target second agents will be inferred by the truth inference module to produce a truth value. Based on the truth value, the correctness of each target second agent's answer can be further judged (or how much it differs from the truth value), thereby updating the accuracy profile and credibility of the target second agents. Target second agents that complete the task will participate in reward decomposition. In the agent task collaboration system, reward decomposition is the process of dividing and distributing the total reward obtained after completing the task according to rules to the participating agents that meet the conditions. In this embodiment, rewards can be allocated only to target second agents that actually complete the task; agents that do not complete the task or are terminated will not participate in the allocation.

[0043] In one embodiment, the method further includes: During the processing of the first task by the target first agent, if the first task enters a first task processing state, a new target first agent with a higher task processing capability than the target first agent is determined from the set of first agents, and the first task is assigned to the new target first agent so that the target first agent can process the first task based on the task information to obtain the target task processing result.

[0044] The first task processing state can be a state where additional agents are urgently needed. The task in this state is one that urgently needs additional agents, meaning that due to an agent going offline or remaining unresponsive for an extended period, another agent is required to respond. In this case, based on the current task confidence and the current option distribution, the crowdlib SDK can be invoked to select an agent to respond. This approach has a high probability of resulting in a successful choice that leads to task convergence; hence, it's called the "replacement logic." The decision of whether to replace one agent or multiple agents depends on the specific circumstances.

[0045] Specifically, there are two ways to handle tasks in the first task processing state: (1) If the first task is in the first task processing state, then the agent A that is processing the first task needs to be replaced by someone to answer. Therefore, an agent B with a higher accuracy (i.e., reliability) than agent A can be found in the first agent set to replace agent A. Agent B processes the first task to obtain the target task processing result. This method will make the accuracy of the agent after replacement higher and higher.

[0046] (2) If the first task is in the first task processing state, the number of people needed and the accuracy of the agents of these answers can be calculated by a preset calculation method based on the current answer distribution of the task and the accuracy of these answers. Then, multiple target second agents can be found from the set of second agents to process the first task together with the agent A that is processing the first task.

[0047] In one embodiment, the method further includes: During the processing of the first task by the multiple target second agents, if the first task enters a second task processing state, a first agent whose task processing capability meets the preset conditions is determined from the first agent set, and the first task is assigned to the first agent, so that the first agent processes the first task based on the task information to obtain the target task processing result.

[0048] The second task processing state can be a non-converged state, meaning the task in this state is one that has not converged. A non-converged task is one where multiple agents have responded, but the response has failed to converge. If a replacement agent is added, their choices may differ, causing the task to remain non-converged and extending the timeline, potentially leading to a timeout. For example, a question might be sent to four people with a 2:2 answer ratio, failing to converge to the truth value; or a question might be sent to three people, with one agent disconnecting, resulting in a 1:1 answer ratio, also failing to converge. In short, a final consensus cannot be reached. Therefore, a single-agent response mode is used, directly selecting a high-capability agent from the first agent set to respond, and their answer becomes the final target task processing result. A preset condition is that the agent's task processing capability exceeds a specified value.

[0049] In one embodiment, before the step of "determining the target agent for processing the first task from the first set of agents and / or the second set of agents", the method further includes: A specified number of first agents are randomly selected from the first agent set, and the selected first agents are moved from the first agent set to the second agent set, so as to set the selected first agents as the second agents in the second agent set.

[0050] In this embodiment of the application, in order to avoid the credibility of the agent only passively decreasing, such as due to the lack of random inspection for a long time leading to a decrease in credibility and a delay in profile update, the matching system (i.e. the task processing system) randomly selects a portion of the first agents and adds them to the second agent set as second agents each time a task is matched, so that some of the first agents can enter the range of multiple people answering.

[0051] For details, please refer to Figure 3 This application provides a matching system. The matching system offers an algorithm service. The platform calls the matching system, inputting the task to be matched and the participating agents. The matching system then returns the matching result as output to the agents. When the platform requests the matching system, the system calls a profiling service to assign corresponding features to the task and agents. (1) Task: The deadline for completing the task, the estimated time required for the task, the current answer options and confidence level of the task, etc.; (2) Intelligent agent: accuracy, reliability, etc.

[0052] The matching system can divide the task into three parts: (1) Tasks that urgently need replacement agents: When an agent goes offline or does not respond for a long time, another agent is needed to answer the question. In this case, based on the current task confidence and the current option distribution, the crowdlib SDK can be called to select an agent to answer the question. There is a high probability that the agent will make a choice and the task will converge. Therefore, it is called the replacement agent logic. Depending on the actual situation, one or more agents need to be replaced.

[0053] (2) Tasks that have not converged: The task has been answered by multiple agents, but has not converged. If a replacement agent is used, the replacement agent may also have different choices, which will cause the task to still not converge and the timeline will be lengthened, resulting in a timeout. Therefore, in this case, the single-agent answering mode is used directly. A high-ability agent is selected directly from the first set of agents to answer, and its answer is used as the final result of the target task.

[0054] (3) Routine tasks: Tasks that do not belong to the above two categories. That is, ordinary tasks that the matching system initially receives but has not yet processed.

[0055] Furthermore, the agents can be divided into two tiers: (1) First Tier (i.e., the first set of agents): A certain proportion of agents with an accuracy higher than the task confidence level are selected to enter the first tier. The proportion is an configurable parameter. In the early stages of the project, most agents can be considered unreliable, so this proportion can be set slightly lower.

[0056] (2) Second echelon (i.e., the second set of intelligent agents): Other intelligent agents that do not belong to the first echelon and require collective intelligence to complete the task.

[0057] Specifically, the matching system can follow this task matching logic: first, handle problems urgently needing replacements, then handle non-converged problems, and finally handle regular problems. When handling regular problems, these tasks can be preferentially assigned to the first tier mentioned above; if any remainder are found, they can be assigned to the second tier. Optionally, when the matching system receives tasks, all tasks can be regular problems. During task processing, some tasks may transform into problems urgently needing replacements or non-converged problems. Alternatively, the matching system can directly obtain problems urgently needing replacements or non-converged problems.

[0058] Finally, the matching system sends the matching results back to the platform. After the agent responds, the answer is submitted to the truth inference module. The truth judgment module judges the convergence and updates the profile and credibility of each agent.

[0059] In this embodiment, the sampling inspection task and the formal task are isomorphic, meaning that a portion of the task is selected for quality inspection simultaneously. Specifically, the same task may be answered by a single agent or by multiple agents. Multiple agents' responses aggregate the truth values, thus determining whether each agent's answer is correct. Since there are many container settlement tasks, sampling inspection can be achieved by allowing a certain proportion of tasks to be answered by multiple agents (i.e., using multiple target second agents to answer). Specifically, if multiple agents answering constitute quality inspection, and quality inspection constitutes multiple agents answering, this is one scenario. Another scenario involves low-confidence agents or low-level agents where multiple agents answer each question, and each question is subject to quality inspection.

[0060] In this embodiment of the application, the quality inspection scheme is as follows: If the aggregated answers from multiple agents (i.e., multiple target second agents) converge, the aggregated result is considered the truth. A single answer is compared to the truth; if they match, it is considered correct; otherwise, it is considered incorrect. For example, the first task is smart vending machine product recognition. This is an objective problem with a uniquely determined truth (objective fact). Even in a subjective problem, answers can be aggregated into a final result. In this case, different target second agents may provide different answers, thus requiring a truth inference method to aggregate these answers. Methods such as majority voting, Bayesian estimation, and interval estimation can be used to obtain the final target truth, which is then used as the result of the target task.

[0061] It should be noted that experiments can be conducted on simulated data or historical response data of the intelligent agent platform. By selecting a truth inference method and combining it with the profile of the intelligent agent (such as historical accuracy), it is possible to statistically determine which intelligent agents provide answers with a high degree of confidence that lead to task convergence, and these answers are considered correct.

[0062] In this embodiment of the application, the timing of quality checks is as follows: A credibility profile of the agent is constructed. Credibility is a variable / fluctuating confidence interval, linked to the number of correct / incorrect answers, the most recent verification time, and the verification ratio. Initially, modeling is based on operational quality inspection rules, which can be iterated into a more general model later. Credibility is used as one of the factors for task allocation. Agents with high credibility are allowed to complete tasks independently. As credibility decreases over time, the agent is required to collaborate with other agents to complete tasks. To avoid passively decreasing credibility (e.g., due to prolonged periods without verification leading to delayed profile updates), the matching system randomly selects a portion of the first-tier agents and adds them to the second-tier agent set during each task matching process. This allows some first-tier agents to participate in multi-agent responses.

[0063] In one embodiment, the answers given by multiple target second agents are inferred by the truth inference module to obtain a truth value. Based on the truth value, the correctness of each target second agent's answer (or how much it differs from the truth value) can be further judged, thereby updating the accuracy profile and credibility of the target second agents. Furthermore, the target second agents that complete the task will participate in the reward decomposition.

[0064] In one embodiment, because the multi-agent answer aggregation scheme has high fault tolerance, it can lower the threshold of the task and allow more agents that do not meet the original standards to participate.

[0065] In one embodiment, for time-sensitive tasks, multiple agents completing the same task will inevitably lead to a longer completion time than a single agent. An over-issuance method is used to distribute tasks to multiple agents, with those completing first being retrieved first for truth inference, ensuring timeliness. Over-issuance is analogous to this: if it's initially estimated that three low-to-medium skill level second agents can complete the task, but due to time constraints and urgency, the task can be distributed to two more second agents, resulting in a total of five agents working together. This prevents slow responses from causing timeouts.

[0066] In one embodiment, the allocation algorithm takes into account factors such as the agent's credibility, whether it is idle, and the estimated response time of the task, and will give priority to highly trustworthy and idle agents.

[0067] In this embodiment, a hybrid approach of single-agent and multi-agent methods is used to complete the task, fully utilizing the agent's productivity. Agents with sufficient credibility are allowed to complete the task individually, while those without sufficient credibility can participate through swarm intelligence. This allows individual agents to participate in the multi-agent task completion process, enabling monitoring of their performance quality and ensuring the accuracy of their profiles. Since the original scheme required agents to pass an exam before participating in the annotation task, this allocation method lowers the task threshold or even eliminates the exam, allowing agents to be evaluated during task completion. Online data shows that the accuracy rate of multi-agent task completion is higher than that of single-agent completion. Because the task threshold has been relaxed, more agents can participate, resulting in a significant increase in productivity compared to the original scheme.

[0068] In summary, this application provides a task processing method. It involves obtaining a first task to be processed, which has corresponding task information; then, determining a target intelligent agent from the first intelligent agent set and / or the second intelligent agent set to process the first task; if the target intelligent agent is a target first intelligent agent, assigning the first task to the target first intelligent agent, so that the target first intelligent agent processes the first task based on the task information to obtain a target task processing result; if the target intelligent agent includes multiple target second intelligent agents, assigning the first task to the multiple target second intelligent agents, so that each target second intelligent agent processes the first task based on the task information to obtain a task processing result corresponding to each target second intelligent agent, and determining the target task processing result of the first task based on each task processing result. The method provided in this application embodiment can be used to complete tasks by assigning the first task to be processed to a target first agent of a first agent set or to multiple target second agents of a second agent set. This allows for the use of a single agent and a mixed allocation of multiple agents, providing a task processing flow that can balance the efficiency, quality, and cost of task execution. It can adapt to task execution scenarios of different scales and types, and effectively improve the processing efficiency of agents when processing tasks.

[0069] This embodiment also provides a task processing device, which can be specifically integrated into a terminal device. For example, such as Figure 4 As shown, the processing apparatus for this task may include: The acquisition unit 201 is used to acquire a first task to be processed, and the first task is configured with task information. Determining unit 202 is configured to determine the target agent for processing the first task from the first set of agents and / or the second set of agents; The first allocation unit 203 is configured to allocate the first task to the target first intelligent agent when the target intelligent agent is the target first intelligent agent, so as to obtain the target task processing result by the target first intelligent agent processing the first task based on the task information; The second allocation unit 204 is configured to allocate the first task to the multiple target second intelligent agents when the target intelligent agent includes multiple target second intelligent agents, so that each target second intelligent agent processes the first task based on the task information to obtain the task processing result corresponding to each target second intelligent agent, and determines the target task processing result of the first task based on the task processing results.

[0070] In some embodiments, the processing apparatus for the task includes a processing subunit for: The task processing quality of each of the target second agents for the first task is evaluated based on the target task processing results. Based on the task processing quality of each target second intelligent agent, update the task processing capability of each target second intelligent agent.

[0071] In some embodiments, the processing apparatus for the task includes a processing subunit for: Based on the task processing capabilities of each of the first and second agents, a target agent for processing the first task is determined from the set of first agents and / or the set of second agents.

[0072] In some embodiments, the processing apparatus for the task includes a processing subunit for: Based on the task quality requirements, target agents that match the task quality requirements are determined from the first set of agents and / or the second set of agents.

[0073] In some embodiments, the processing apparatus for the task includes a processing subunit for: During the processing of the first task by the target first agent, if the first task enters a first task processing state, a new target first agent with a higher task processing capability than the target first agent is determined from the set of first agents, and the first task is assigned to the new target first agent so that the target first agent can process the first task based on the task information to obtain the target task processing result.

[0074] In some embodiments, the processing apparatus for the task includes a processing subunit for: During the processing of the first task by the multiple target second agents, if the first task enters a second task processing state, a first agent whose task processing capability meets the preset conditions is determined from the first agent set, and the first task is assigned to the first agent, so that the first agent processes the first task based on the task information to obtain the target task processing result.

[0075] In some embodiments, the processing apparatus for the task includes a processing subunit for: A specified number of first agents are randomly selected from the first agent set, and the selected first agents are moved from the first agent set to the second agent set, so as to set the selected first agents as the second agents in the second agent set.

[0076] In some embodiments, the processing apparatus for the task includes a processing subunit for: The results of each task are aggregated using preset rules to obtain the target truth value; The target task processing result of the first task is determined based on the target truth value.

[0077] This application discloses a task processing apparatus. An acquisition unit 201 acquires a first task to be processed, the first task having corresponding task information. A determination unit 202 determines a target agent for processing the first task from a first set of agents and / or a second set of agents. A first allocation unit 203, when the target agent is a target first agent, allocates the first task to the target first agent, so that the target first agent processes the first task based on the task information to obtain a target task processing result. A second allocation unit 204, when the target agent includes multiple target second agents, allocates the first task to multiple target second agents, so that each target second agent processes the first task based on the task information to obtain a task processing result corresponding to each target second agent, and determines the target task processing result of the first task based on each task processing result. This application embodiment can assign the first task to be processed to a target first agent of a first agent set, or to multiple target second agents of a second agent set, thereby enabling the task to be completed using a single agent and a mixed allocation of multiple agents. This provides a task processing flow that can balance the efficiency, quality and cost of task execution, and can adapt to task execution scenarios of different scales and types, effectively improving the processing efficiency of agents when processing tasks.

[0078] Accordingly, this application also provides an electronic device, which can be a terminal, such as a smartphone, tablet computer, laptop computer, touch screen, game console, personal computer (PC), personal digital assistant (PDA), or other terminal device. Alternatively, the electronic device can be a server.

[0079] like Figure 5 As shown, Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device 300 includes a processor 301 with one or more processing cores, a memory 302 with one or more computer-readable storage media, and a computer program stored in the memory 302 and executable on the processor. The processor 301 and the memory 302 are electrically connected. Those skilled in the art will understand that the electronic device structure shown in the figure does not constitute a limitation on the electronic device, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0080] The processor 301 is the control center of the electronic device 300. It connects various parts of the electronic device 300 via various interfaces and lines. By running or loading software programs and / or units stored in the memory 302, and by calling data stored in the memory 302, it executes various functions and processes data of the electronic device 300, thereby providing overall monitoring of the electronic device 300. The processor 301 can be a central processing unit (CPU), a graphics processing unit (GPU), a network processor (NP), etc., and can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application.

[0081] In this embodiment, the processor 301 in the electronic device 300 loads the instructions corresponding to the processes of one or more applications into the memory 302 according to the following steps, and the processor 301 runs the applications stored in the memory 302 to realize various functions, such as: Obtain the first task to be processed, which has corresponding task information set. Determine the target agent for processing the first task from the first set of agents and / or the second set of agents; When the target agent is a first target agent, the first task is assigned to the first target agent so that the first target agent processes the first task based on the task information to obtain the target task processing result; When the target agent includes multiple target second agents, the first task is assigned to the multiple target second agents so that each target second agent processes the first task based on the task information to obtain the task processing result corresponding to each target second agent, and the target task processing result of the first task is determined based on the task processing result of each target second agent.

[0082] The electronic device provided in this application embodiment can assign the first task to be processed to a target first intelligent agent of a first intelligent agent set, or to multiple target second intelligent agents of a second intelligent agent set, thereby completing the task using a single intelligent agent and a mixed allocation of multiple intelligent agents. This provides a task processing flow that can take into account the efficiency, quality and cost of task execution, and can adapt to task execution scenarios of different scales and types, effectively improving the processing efficiency of intelligent agents when processing tasks.

[0083] For details on the implementation of each of the above operations, please refer to the previous examples, which will not be repeated here.

[0084] Optional, such as Figure 5 As shown, the electronic device 300 also includes: a touch display screen 303, a radio frequency circuit 304, an audio circuit 305, an input unit 306, and a power supply 307. The processor 301 is electrically connected to the touch display screen 303, the radio frequency circuit 304, the audio circuit 305, the input unit 306, and the power supply 307. Those skilled in the art will understand that... Figure 5 The electronic device structure shown does not constitute a limitation on the electronic device and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0085] The touch display screen 303 can be used to display a graphical user interface (GUI) and receive operation commands generated by the user interacting with the GUI. The touch display screen 303 may include a display panel and a touch panel. The display panel can be used to display information input by the user or information provided to the user, as well as various graphical user interfaces of the electronic device. These graphical user interfaces can be composed of graphics, text, icons, video, and any combination thereof. Optionally, the display panel can be configured using a liquid crystal display (LCD), organic light-emitting diode (OLED), or other similar technologies. The touch panel can be used to collect touch operations performed by the user on or near it (such as operations performed by the user using a finger, stylus, or any suitable object or accessory on or near the touch panel), generate corresponding operation commands, and execute the corresponding program according to the operation commands. Optionally, the touch panel may include two parts: a touch detection device and a touch controller. The touch detection device detects the user's touch location and the signal generated by the touch operation, transmitting the signal to the touch controller. The touch controller receives touch information from the touch detection device, converts it into touch point coordinates, and sends it to the processor 301. It can also receive and execute commands from the processor 301. The touch panel can cover the display panel. When the touch panel detects a touch operation on or near it, it transmits the information to the processor 301 to determine the type of touch event. Subsequently, the processor 301 provides corresponding visual output on the display panel based on the type of touch event. In this embodiment, the touch panel and the display panel can be integrated into the touch display screen 303 to achieve input and output functions. However, in some embodiments, the touch panel and the touch display screen 303 can be implemented as two independent components to achieve input and output functions. That is, the touch display screen 303 can also be used as part of the input unit 306 to achieve input functions.

[0086] The radio frequency circuit 304 can be used to transmit and receive radio frequency signals to establish wireless communication with network devices or other electronic devices, and to transmit and receive signals with network devices or other electronic devices.

[0087] Audio circuitry 305 can be used to provide an audio interface between a user and an electronic device via a speaker and a microphone. Audio circuitry 305 converts received audio data into electrical signals, transmits them to the speaker, and the speaker converts them into sound signals for output. Conversely, the microphone converts collected sound signals into electrical signals, which are then received by audio circuitry 305, converted back into audio data, and then processed by processor 301 before being transmitted via radio frequency circuitry 304 to, for example, another electronic device, or output to memory 302 for further processing. Audio circuitry 305 may also include an earphone jack to facilitate communication between peripheral headphones and electronic devices.

[0088] The input unit 306 can be used to receive input numbers, characters, or user characteristic information (such as fingerprints, iris, facial information, etc.), and to generate keyboard, mouse, joystick, optical, or trackball signal inputs related to user settings and function control.

[0089] Power supply 307 is used to supply power to various components of electronic device 300. Optionally, power supply 307 can be logically connected to processor 301 through a power management system, thereby enabling functions such as charging, discharging, and power consumption management through the power management system. Power supply 307 may also include one or more DC or AC power supplies, recharging systems, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components.

[0090] although Figure 5 As not shown in the diagram, the electronic device 300 may also include a camera, sensor, wireless fidelity module, Bluetooth module, etc., which will not be described in detail here.

[0091] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0092] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be performed by instructions, or by instructions controlling related hardware. These instructions can be stored in a computer-readable storage medium and loaded and executed by a processor.

[0093] Therefore, embodiments of this application provide a computer-readable storage medium storing a plurality of computer programs, which can be loaded by a processor to execute any of the task processing methods provided in embodiments of this application. The computer program can execute the steps of the task processing method as follows: Obtain the first task to be processed, which has corresponding task information set. Determine the target agent for processing the first task from the first set of agents and / or the second set of agents; When the target agent is a first target agent, the first task is assigned to the first target agent so that the first target agent processes the first task based on the task information to obtain the target task processing result; When the target agent includes multiple target second agents, the first task is assigned to the multiple target second agents so that each target second agent processes the first task based on the task information to obtain the task processing result corresponding to each target second agent, and the target task processing result of the first task is determined based on the task processing result of each target second agent.

[0094] Because the computer program stored in the storage medium can assign the first task to be processed to a target first agent of a first agent set, or to multiple target second agents of a second agent set, the task can be completed using a single agent or a mixed allocation of multiple agents. This provides a task processing flow that can balance the efficiency, quality and cost of task execution, and can adapt to task execution scenarios of different scales and types, effectively improving the processing efficiency of agents when processing tasks.

[0095] For details on the implementation of each of the above operations, please refer to the previous examples, which will not be repeated here.

[0096] The computer-readable storage medium may include: read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0097] Since the computer program stored in the computer-readable storage medium can execute any of the task processing methods provided in the embodiments of this application, the beneficial effects that any of the task processing methods provided in the embodiments of this application can achieve can be realized. For details, please refer to the previous embodiments, which will not be repeated here.

[0098] According to one aspect of this application, a computer program product or computer program is also provided, comprising computer instructions stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the electronic device to perform the methods provided in the various optional implementations of the above embodiments.

[0099] In the embodiments of the processing apparatus, computer-readable storage medium, electronic device, and computer program product described above, the descriptions of each embodiment have different focuses. Parts not described in detail in a particular embodiment can be referred to in the relevant descriptions of other embodiments. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process and beneficial effects of the processing apparatus, computer-readable storage medium, computer program product, electronic device, and their corresponding units described above can be referred to the description of the task processing method in the above embodiments, and will not be repeated here.

[0100] The foregoing has provided a detailed description of a task processing method, apparatus, electronic device, computer-readable storage medium, and computer program product provided by the embodiments of this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for processing a task, characterized in that, The method is applied to a task processing system, which is configured with a first set of intelligent agents and a second set of intelligent agents. The first agents in the first set of intelligent agents are configured to process tasks independently, and the second agents in the second set of intelligent agents are configured to perform task processing using a swarm of intelligent agents. The method includes: Obtain the first task to be processed, which has corresponding task information set. Determine the target agent for processing the first task from the first set of agents and / or the second set of agents; When the target agent is a first target agent, the first task is assigned to the first target agent so that the first target agent processes the first task based on the task information to obtain the target task processing result; When the target agent includes multiple target second agents, the first task is assigned to the multiple target second agents so that each target second agent processes the first task based on the task information to obtain the task processing result corresponding to each target second agent, and the target task processing result of the first task is determined based on the task processing result of each target second agent.

2. The method according to claim 1, characterized in that, After determining the target task processing result of the first task based on the processing results of each of the tasks, the method further includes: The task processing quality of each of the target second agents for the first task is evaluated based on the target task processing results. Based on the task processing quality of each target second intelligent agent, update the task processing capability of each target second intelligent agent.

3. The method according to claim 1, characterized in that, Determining the target agent for processing the first task from the first set of agents and / or the second set of agents includes: Based on the task processing capabilities of each of the first and second agents, a target agent for processing the first task is determined from the set of first agents and / or the set of second agents.

4. The method according to claim 1, characterized in that, The first task has corresponding task quality requirements; Determining the target agent for processing the first task from the first set of agents and / or the second set of agents includes: Based on the task quality requirements, target agents that match the task quality requirements are determined from the first set of agents and / or the second set of agents.

5. The method according to claim 1, characterized in that, The method further includes: During the processing of the first task by the target first agent, if the first task enters a first task processing state, a new target first agent with a higher task processing capability than the target first agent is determined from the set of first agents, and the first task is assigned to the new target first agent so that the target first agent can process the first task based on the task information to obtain the target task processing result.

6. The method according to claim 1, characterized in that, The method further includes: During the processing of the first task by the multiple target second agents, if the first task enters a second task processing state, a first agent whose task processing capability meets the preset conditions is determined from the first agent set, and the first task is assigned to the first agent, so that the first agent processes the first task based on the task information to obtain the target task processing result.

7. The method according to claim 1, characterized in that, Before determining the target agent for processing the first task from the first set of agents and / or the second set of agents, the method further includes: A specified number of first agents are randomly selected from the first agent set, and the selected first agents are moved from the first agent set to the second agent set, so as to set the selected first agents as the second agents in the second agent set.

8. The method according to claim 1, characterized in that, Determining the target task processing result of the first task based on the processing results of each of the tasks includes: The results of each task are aggregated using preset rules to obtain the target truth value; The target task processing result of the first task is determined based on the target truth value.

9. A task processing apparatus, characterized in that, This is applied to a task processing system, which is configured with a first set of intelligent agents and a second set of intelligent agents. The first agents in the first set of intelligent agents are configured to handle tasks independently, and the second agents in the second set of intelligent agents are configured to perform task processing using a swarm of intelligent agents, including: The acquisition unit is used to acquire the first task to be processed, and the first task has corresponding task information set. A determining unit is configured to determine a target agent for processing the first task from the first set of agents and / or the second set of agents; The first allocation unit is configured to allocate the first task to the target first intelligent agent when the target intelligent agent is the target first intelligent agent, so that the target first intelligent agent can process the first task based on the task information to obtain the target task processing result; The second allocation unit is configured to allocate the first task to the multiple target second intelligent agents when the target intelligent agent includes multiple target second intelligent agents, so that each target second intelligent agent processes the first task based on the task information to obtain the task processing result corresponding to each target second intelligent agent, and determines the target task processing result of the first task based on the task processing results.

10. An electronic device, characterized in that, The device includes a processor and a memory, the memory storing multiple instructions; the processor loads instructions from the memory to perform the steps of the processing method for the task as described in any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a plurality of instructions adapted for loading by a processor to perform the steps of the processing method of the task as described in any one of claims 1 to 8.

Images