Modeling and using method and device of intelligent service model, equipment and storage medium
By inserting the ability interface function of the self-learning field into the intelligent service model and the external agent collaboration function, the problem of insufficient capabilities of the intelligent service model when handling complex tasks is solved, and the self-improvement and flexible response of the intelligent service model is realized.
Patent Information
- Application Number
- CN202310519642.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-05-08
- Publication Date
- 2024-11-08
AI Technical Summary
The existing intelligent service model is difficult to improve itself and lacks processing capabilities when facing flexible and changing user needs, resulting in a decline in service quality after the task processing volume increases.
By obtaining the agent definition file, analyzing and inserting the ability interface function of the self-learning field, external agent collaboration and ability self-learning are realized, forming an intelligent service model based on the agent.
It realizes that the intelligent service model can independently assist other agents when facing complex tasks and improve their capabilities based on the assistance results, improving the system's self-learning and processing capabilities.
Smart Images

Figure CN118916005A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of artificial intelligence, and in particular to a modeling method of an intelligent service model, a method for using an intelligent service model, a modeling device of an intelligent service model, an intelligent service device, an electronic device, and a computer-readable storage medium. Background Art
[0002] Nowadays, faced with increasingly complex user demands, the difficulty of intelligent service tasks is also increasing. Although the current intelligent service models can handle some complex tasks, they are often faced with difficult-to-solve situations in the face of flexible and changeable user demands. Moreover, the processing capacity of intelligent service models is difficult to improve as the amount of task processing increases, making it difficult for current intelligent service models to provide better services after a period of use. Summary of the invention
[0003] In view of this, the present application provides a modeling method of an intelligent service model, a method of using an intelligent service model, a modeling device of an intelligent service model, an intelligent service device, an electronic device and a computer-readable storage medium to solve or at least partially solve the above-mentioned problems.
[0004] A first aspect of an embodiment of the present application provides a modeling method for an intelligent service model, the method comprising:
[0005] Obtaining an agent definition file; the agent definition file includes a data structure of input data and output data of various capability interfaces required to complete the intelligent service task, wherein the agent definition file declares a self-learning field for the target capability interface;
[0006] Parsing the agent definition file to obtain an agent initial code including the capability interface; in the agent initial code, the interface function of the capability interface is an empty function;
[0007] Inserting a first interface function content for implementing the capability into the agent initial code for the capability interface;
[0008] In the initial code of the intelligent agent, the second interface function content for collaboration with an external intelligent agent and the self-learning code for self-learning of capabilities according to the collaboration results of the external intelligent agent are inserted for the target capability interface to obtain an intelligent service model based on the intelligent agent.
[0009] A second aspect of the embodiment of the present application provides a method for using an intelligent service model, wherein the intelligent service model is modeled using the modeling method of the intelligent service model described in the first aspect above, and the method for using the intelligent service model includes:
[0010] receiving target input data of the intelligent service task;
[0011] Calling the capability interface in the intelligent service model to implement the capability corresponding to the capability interface based on the target input data, and outputting a response action for the intelligent service task;
[0012] For the capability interface with self-learning configuration, when the capability cannot be realized through the capability interface, external agent collaboration is performed;
[0013] After the external intelligent agent collaboration is completed, the external intelligent agent collaboration result is recorded, and the capability interface is self-learned based on the recorded external intelligent agent collaboration result.
[0014] A third aspect of the embodiment of the present application further provides a modeling device for an intelligent service model, the modeling device comprising:
[0015] An acquisition module is used to acquire an agent definition file; the agent definition file includes a data structure of input data and output data of various capability interfaces required to complete the intelligent service task, wherein the agent definition file declares a self-learning field for the target capability interface;
[0016] A parsing module, used for parsing the agent definition file to obtain an agent initial code including the capability interface; in the agent initial code, the interface function of the capability interface is an empty function;
[0017] An insertion module, used for inserting a first interface function content for realizing the capability in the initial code of the agent for the capability interface;
[0018] The insertion module is also used to insert the second interface function content for external intelligent agent collaboration in the initial code of the intelligent agent for the target capability interface, and the self-learning code for self-learning capabilities according to the results of the external intelligent agent collaboration, so as to obtain an intelligent service model based on the intelligent agent.
[0019] The fourth aspect of the embodiment of the present application also provides an intelligent service device, which is configured with an intelligent service model modeled by the modeling method of the intelligent service model described in the first aspect above, and the intelligent service device is used to implement the method of using the intelligent service model described in the second aspect above through the intelligent service model.
[0020] A fifth aspect of the embodiments of the present application further provides an electronic device, comprising: a processor, a memory, and computer program instructions stored in the memory and executable on the processor;
[0021] When the processor executes the computer program instructions, the modeling method of the intelligent service model as described in the first aspect above, or the method of using the intelligent service model as described in the second aspect above is implemented.
[0022] The sixth aspect of the embodiment of the present application also provides a computer-readable storage medium storing computer execution instructions, which, when executed by a processor, are used to implement the modeling method of the intelligent service model as described in the first aspect above, or the method of using the intelligent service model as described in the second aspect above.
[0023] In an embodiment of the present application, an agent definition file is provided, which includes a data structure of input data and output data of various capability interfaces required to complete intelligent service tasks, wherein a self-learning field is declared for the target capability interface in the agent definition file. The embodiment of the present application first obtains the agent definition file, and then parses it to obtain an agent initial code including a capability interface, wherein, in the agent initial code, the interface function of the capability interface is an empty function. Afterwards, the first interface function content for implementing the capability is inserted into the agent initial code for each capability interface. In particular, for the target capability interface declared with a self-learning field, the second interface function content for external agent collaboration is inserted into the agent initial code, as well as a self-learning code for self-learning of capabilities according to the results of the external agent collaboration, thereby obtaining an agent-based intelligent service model.
[0024] Through the modeling method of the intelligent service model provided by the embodiment of the present application, in the initial code of the intelligent agent obtained by parsing the intelligent agent definition file, the interface function content for realizing the capability, the interface function content for external intelligent agent collaboration, and the self-learning code for self-learning the capability according to the results of the external intelligent agent collaboration are inserted for the target capability interface declared with the self-learning field, thereby obtaining an intelligent service model based on the intelligent agent that can collaborate with other intelligent agents and can self-learn. The intelligent service model modeled in this way can autonomously request the assistance of other intelligent agents when facing difficult problems, and can self-improve its capabilities according to the results of the assistance of other intelligent agents. BRIEF DESCRIPTION OF THE DRAWINGS
[0025] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments of the present application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.
[0026] Figure 1 A flowchart of a modeling method for an intelligent service model provided in an embodiment of the present application;
[0027] Figure 2 A flowchart of a method for using an intelligent service model provided in an embodiment of the present application;
[0028] Figure 3 A schematic diagram of the structure of an intelligent customer service model provided in an embodiment of the present application;
[0029] Figure 4 A structural block diagram of a modeling device for an intelligent service model provided in an embodiment of the application;
[0030] Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0031] In order to enable those skilled in the art to better understand the technical solution of the present application, the present application is described clearly and completely below in conjunction with the drawings in the embodiments of the present application. However, the present application can be implemented in many other ways different from the above description. Therefore, based on the embodiments provided by the present application, all other embodiments obtained by those of ordinary skill in the art without creative work should fall within the scope of protection of the present application.
[0032] It should be noted that the terms "first", "second", "third", etc. in the claims, description and drawings of the present application are used to distinguish similar objects and are not used to describe a specific order or sequence. The data used in this way are interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including", "having" and their variants are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0033] First, the technical background of this application is described:
[0034] Intelligent service refers to the provision of various services to users through computer equipment, such as online intelligent customer service, enabled by related technologies such as artificial intelligence, big data, and cloud computing. Today, intelligent services have been applied in various fields including e-commerce, medical care, government affairs, and tourism. Judging from the current market capacity of intelligent services, the market size of basic service software is about RMB 10 billion. With the evolution and accelerated empowerment of artificial intelligence technology, the intelligent service industry is expected to exceed RMB 30 to 60 billion. It can be seen that the intelligent service industry still has a huge market growth in the future.
[0035] Existing intelligent service products are usually intelligent systems based on procedural programming (POP) or object-oriented programming (OOP). Among them, the modeling idea of the POP mode is to decompose the task in the order of solution steps, then implement each step in the form of functions, and finally call the functions sequentially to complete the entire task. Therefore, the POP-based intelligent service usually first calls the intent recognition function to analyze the language input by the user to obtain the corresponding semantics. Then call the reply function to obtain the reply corresponding to the semantics. Therefore, the entire POP-based intelligent service is composed of multiple independent function modules, and the artificial intelligence model corresponding to each function is independent.
[0036] Understandably, intelligent service is a very complex task. The process-oriented programming modeling method cannot well decompose complex tasks into individual steps. When a new function is needed, the new function needs to be broken down into steps and then integrated into the original program. This approach requires a lot of modifications to the existing function call logic, and the modification process is extremely difficult and complicated. In addition, since the artificial intelligence models contained in the POP-based intelligent service are independent, although each artificial intelligence model can continuously improve the model performance through training and learning, for the entire intelligent service system, the intelligent service system can only call the trained artificial intelligence model to complete a certain step. It itself is not learnable and cannot improve itself.
[0037] The modeling idea of the OOP model is to divide the task into multiple objects according to the function, then configure its own methods and properties for each object, and finally complete the task through the interaction of different objects in the method. In this way, when adding new functions, you only need to modify or add the corresponding methods of the relevant objects, and there is no need to make a lot of changes to the logic of the program. Therefore, the intelligent service based on OOP is generally a user-oriented online service system. That is, different service programs are formulated for different types of customers. When executing online service tasks, different service programs are called according to the customer type, the user's intention is analyzed, and then the corresponding response is obtained according to the intention.
[0038] Although the object-oriented programming modeling method can handle complex tasks, it focuses more on using the system's logical computing capabilities, and requires a lot of conversion work to use artificial intelligence capabilities. The use of models is inefficient. Moreover, it is essentially a customized function and cannot self-learn to improve system performance.
[0039] From the above description, it can be seen that whether it is a POP-based intelligent service or an OOP-based intelligent service, it essentially calls multiple customized functions to perform reply tasks. And each customized function is usually a trained artificial intelligence model.
[0040] This results in the AI model being unable to fully utilize its ability to seek help and improve itself like a "human" during use, which seriously affects the capabilities of the entire intelligent service system. Therefore, how to obtain an intelligent service system that can seek external help and self-learn and improve has become an urgent problem to be solved.
[0041] In view of the above technical problems, the present application provides a modeling method of an intelligent service model, a method for using an intelligent service model, a modeling device of an intelligent service model, an intelligent service device, an electronic device, and a computer-readable storage medium. The method, device, electronic device, and computer-readable storage medium described in the present application are further described in detail below in conjunction with specific embodiments and drawings.
[0042] Before introducing the embodiments of the present application, a brief introduction to the concept of agent-oriented programming (AOP) is given first:
[0043] AOP is a domain-specific language used to describe tasks to agents, with the Markov decision process (MDP) model as its theoretical basis. Among them, the Markov decision process is a mathematical model of sequential decision-making, which is used to simulate the random strategies and rewards that can be achieved by agents in an environment where the system state has Markov properties. In terms of application, MDP is used to model reinforcement learning problems in machine learning. By using methods such as dynamic programming and random sampling, MDP can solve the agent strategy that maximizes the return. Specifically, MDP includes a set of interactive objects, namely agents and environments. Among them, the agent is an agent for machine learning, which can perceive the state of the external environment to make decisions, take actions on the environment, and adjust the strategy through feedback from the environment. The environment is a collection of all things outside the agent, whose state will change due to the influence of the agent's actions, and its changes can be fully or partially perceived by the agent. And the environment will feedback the corresponding reward to the agent after each decision made by the agent. In general, the MDP theory abstracts the task completion process as "observation-recognition-decision-action". When performing a task, the intelligent agent needs to perceive certain information, which is called observation. Based on the observation, it needs to identify certain information, which is called state. Based on the state and observation, it needs to make certain decisions, which is called action.
[0044] Therefore, based on the AOP concept, we can sort out the observations required to complete the service task, the states identified during the completion process, and the actions required to define the input and output of the task.
[0045] The following introduces the relevant methods, devices, etc. of the intelligent service model based on agent-oriented programming (AOP) provided in the embodiments of the present application.
[0046] Figure 1A flow chart of a modeling method of an intelligent service model provided in an embodiment of the present application is shown as follows: Figure 1 As shown, the method includes the following steps 101-104.
[0047] 101. Obtain an agent definition file; the agent definition file includes a data structure of input data and output data of various capability interfaces required to complete the intelligent service task, wherein a self-learning field is declared for the target capability interface in the agent definition file.
[0048] First, an agent definition file is written using an interface description language (IDL), and the modeled electronic device can obtain the written agent definition file.
[0049] According to the intelligent service task process "observation-recognition-decision-action" abstracted from the perspective of the intelligent agent based on the MDP theory, the following contents may be specifically included in the intelligent agent definition file:
[0050] 1. Definition of observation data: may include the name, description, data type, etc. of the observation data.
[0051] 2. Definition of state data: may include the name, description, data type, etc. of the state data.
[0052] 3. Action definition: It may include the name, description, data type, etc. Action is the response to the task.
[0053] 4. Data definition of capability interface: may include the name and description of the capability, the name of the capability interface input data, the name of the capability interface output data, etc.
[0054] In an optional embodiment, the capability interface may include a recognition capability interface and a policy capability interface, wherein the recognition capability refers to the ability of the agent to obtain the state from the observation, and the policy capability refers to the ability of the agent to take what action to perform the task response according to the observation and / or state.
[0055] The input data of the recognition capability interface may include observation data, and the output data of the recognition capability interface may include state data. Accordingly, the recognition capability interface can be used to obtain state data based on the observation data analysis. The recognition capability may include intent recognition capability, slot value filling capability, etc.
[0056] The input data of the decision capability interface may include observation data and / or state data, and the output data of the decision capability interface may include the response action to be taken. Accordingly, the decision capability interface can be used to determine the response action to be taken based on the observation data and / or state data. The decision capability may include the ability to output a response, the ability to perform a required operation, etc.
[0057] In an optional implementation, all the data structures involved may be first defined through a structure (Struct), and then when defining the observation data, state data, action and capability interfaces, it is sufficient to specify which data structure is required.
[0058] In addition, in the embodiment of the present application, for the target capability interface in the agent definition file, if it is desired that the corresponding capability has the ability of self-learning (i.e., the ability of self-improvement), the definition of the self-learning field can be added to the target capability interface, and the ability interface with the self-learning field can be configured with the self-learning capability accordingly, so that the ability interface can gradually improve its ability as the task processing volume increases. Among them, the target capability interface can be one or more capability interfaces defined in the agent definition file.
[0059] 102. Parse the agent definition file to obtain the agent initial code including the capability interface; in the agent initial code, the interface function of the capability interface is an empty function.
[0060] The agent definition file written in I DL can be parsed by a parser to obtain the agent initial code. Since the agent definition file does not define the implementation logic for the capability interface, the interface function of the capability interface in the parsed agent initial code is an empty function.
[0061] 103. Insert the first interface function content for implementing the capability into the capability interface in the initial code of the intelligent body.
[0062] Modelers can develop first interface function contents for realizing corresponding capabilities according to capability interfaces involved in the initial code of the agent, and the electronic device can insert the first interface function contents corresponding to each capability interface into the empty function position of the corresponding capability interface in the initial code of the agent. In this way, each capability interface can realize the corresponding capability according to the corresponding first interface function contents.
[0063] 104. Insert the second interface function content for external intelligent agent collaboration and the self-learning code for capability self-learning according to the external intelligent agent collaboration results into the initial code of the intelligent agent for the target capability interface, so as to obtain an intelligent service model based on the intelligent agent.
[0064] The self-learning of the capability interface is a further improvement on the existing capabilities, so some external help is needed. In the embodiment of the present application, the self-learning of the capability interface can be realized based on the collaboration results of the external intelligent agent. Therefore, for the target capability interface declared with the self-learning field, it is necessary to configure the ability to collaborate with other intelligent agents when it cannot process itself, and the ability to self-learn according to the collaboration results of the external intelligent agent.
[0065] Therefore, for the target capability interface declared with a self-learning field, the modeler can develop the second interface function content for collaboration with an external intelligent agent, and the electronic device can also insert the second interface function content corresponding to each capability interface into the empty function position of the corresponding capability interface in the initial code of the intelligent agent, as a capability of the capability interface to process tasks by seeking external help.
[0066] By connecting to other intelligent agents for collaboration, firstly, it can ensure that the user's needs are responded to, that is, to ensure the smooth completion of the intelligent service tasks; secondly, the collaboration results can be used as external help required for self-learning of capabilities, that is, the intelligent service model can learn how to handle tasks when encountering the same or similar problems based on the task processing results of other intelligent agents.
[0067] At this point, the target capability interface has multiple capabilities to realize the corresponding capabilities, request external help, and improve itself according to external help. Through the above steps, an intelligent service model based on intelligent agents that can collaborate with other intelligent agents and can self-learn can be obtained.
[0068] In an embodiment of the present application, an agent definition file is provided, which includes a data structure of input data and output data of various capability interfaces required to complete intelligent service tasks, wherein a self-learning field is declared for the target capability interface in the agent definition file. The embodiment of the present application first obtains the agent definition file, and then parses it to obtain an agent initial code including a capability interface, wherein, in the agent initial code, the interface function of the capability interface is an empty function. Afterwards, the first interface function content for implementing the capability is inserted into the agent initial code for each capability interface. In particular, for the target capability interface declared with a self-learning field, the second interface function content for external agent collaboration is inserted into the agent initial code, as well as a self-learning code for self-learning of capabilities according to the results of the external agent collaboration, thereby obtaining an agent-based intelligent service model.
[0069] Through the modeling method of the intelligent service model provided by the embodiment of the present application, in the initial code of the intelligent agent obtained by parsing the intelligent agent definition file, the interface function content for realizing the capability, the interface function content for external intelligent agent collaboration, and the self-learning code for self-learning the capability according to the results of the external intelligent agent collaboration are inserted for the target capability interface declared with the self-learning field, thereby obtaining an intelligent service model based on the intelligent agent that can collaborate with other intelligent agents and can self-learn. The intelligent service model modeled in this way can autonomously request the assistance of other intelligent agents when facing difficult problems, and can self-improve its capabilities according to the results of the assistance of other intelligent agents.
[0070] In one embodiment of the present application, the target capability interface is specifically used to implement the capability through the first interface function content based on the target input data when called, and when the capability cannot be implemented, external intelligent agent collaboration is performed through the second interface function content.
[0071] When the intelligent service model is actually used, when it is necessary to realize the capability through the target capability interface, but the target capability interface cannot realize the capability through itself, it can cooperate with an external intelligent agent. Among them, the external intelligent agent may include a person, other more advanced intelligent models, a proprietary program / device that realizes the capability, etc., and the embodiments of this application do not specifically limit this.
[0072] In one embodiment, the intelligent service model may seek cooperation from an intelligent agent "human". Accordingly, an optional method for external intelligent agent cooperation may include the following steps 201-204:
[0073] 201. Convert the target input data from the agent data format to the manual annotation format.
[0074] This step specifically includes: converting the target input data into manual annotation information, wherein the manual annotation information includes annotation task description information and multiple annotation options.
[0075] Since the data format required by human intelligence during labeling is inconsistent with the data format of the intelligent agent, a mapping function is required to convert the data format when requesting human-computer collaboration.
[0076] Description, options=mapping(observation, state)
[0077] For example:
[0078] Label task description information (description):
[0079] User: "Hello, there is sound only on one side of the headset."
[0080] Based on the above conversation, mark the user intention
[0081] Multiple annotation options:
[0082] Headphone pairing;
[0083] The connection is often disconnected;
[0084] Bluetooth cannot connect;
[0085] Only one earphone has sound;
[0086] Switch to single ear mode.
[0087] 202. For the target input data after the conversion, obtain the external intelligent agent collaboration result in the manually annotated format.
[0088] In one implementation, the target input data (i.e., Description, options) after format conversion will be sent to the crowdsourcing platform and annotated by annotators (human intelligence) on the crowdsourcing platform.
[0089] The electronic device can obtain a target annotation option selected from multiple annotation options from a crowdsourcing platform, and use the target annotation option as an external intelligent agent collaboration result in a manual annotation format.
[0090] 203. Convert the external agent collaboration results from the manual annotation format to the agent data format.
[0091] After the human intelligent labeling is completed, another Mapping function is required to convert the manual labeling format into the intelligent agent data format for use by the intelligent service model.
[0092] 204. Perform response actions for the intelligent service task according to the external intelligent agent collaboration results after the conversion.
[0093] The external agent collaboration result converted into the agent data format is the external assistance result sought by the intelligent service model for the target capability interface. The intelligent service model can then determine which action needs to be taken to respond to the task based on the result.
[0094] In another embodiment of the present application, the intelligent service model may seek cooperation from an intelligent agent "intelligent model". Accordingly, an optional method for external intelligent agent cooperation may include the following steps 301:
[0095] 301. Call an external intelligent model, and input target input data into the external intelligent model to obtain the external intelligent agent collaboration result output by the external intelligent model.
[0096] In this embodiment, the intelligent service model can call a more advanced intelligent model (such as an upgraded version of the intelligent service model), or a proprietary model for implementing the capabilities corresponding to the target capability interface, thereby collaborating through other non-human intelligent agents.
[0097] In addition, in an embodiment of the present application, the target capability interface can also be used to record the results of the external intelligent agent collaboration after the external intelligent agent collaboration is completed, and perform capability self-learning through the self-learning code based on the recorded external intelligent agent collaboration results.
[0098] Among them, the intelligent service model can control the target capability interface to perform capability self-learning at a preset time according to the recorded external intelligent agent collaboration results. The preset self-learning time can be after completing the current intelligent service task using the external intelligent agent collaboration results, or after accumulating a certain time or a certain number of external intelligent agent collaboration results, which is not specifically limited in the embodiments of the present application.
[0099] Figure 2 A flowchart of a method for using an intelligent service model provided in an embodiment of the present application, such as Figure 2 As shown, the method includes the following steps 401-404:
[0100] 401. Receive target input data of the intelligent service task.
[0101] 402. Call the capability interface in the intelligent service model to implement the capability corresponding to the capability interface based on the target input data, and output a response action for the intelligent service task.
[0102] The intelligent service model first receives the target input data of the intelligent service task, that is, the observation data, which can include user input, order information, product information, user portraits, etc. Then, the recognition capability interface is called to obtain the understanding of the target input data (state data) from the information observed in the environment (observation data), including intention, slot and other information. After that, the decision-making capability interface is called to determine what action to take (including what kind of reply content to output, what kind of instruction operation to make, etc.) to reply to the user in order to complete the response to this intelligent task. For example, if the user's intention is clear, the intelligent service model can reply directly; if the intention has a slot, but the slot is not recognized, the intelligent service model can ask the slot back or seek help from an external intelligent agent for collaboration.
[0103] 403. For a capability interface with self-learning configuration, when the capability cannot be realized through the capability interface, external intelligent agent collaboration is performed.
[0104] 404. After the external intelligent agent collaboration is completed, the external intelligent agent collaboration result is recorded, and the capability interface is self-learned based on the recorded external intelligent agent collaboration result.
[0105] When the called capability interface cannot realize the capability by itself, it can collaborate with an external intelligent agent, so that the input data corresponding to the capability interface is handed over to the external intelligent agent for identification or decision-making. The external intelligent agent feeds back the identification result or decision result to the intelligent service model, which then records the identification result or decision result and uses the recorded identification result or decision result at a preset time to iteratively learn the capability interface, thereby achieving capability improvement.
[0106] In an optional example, the intelligent service model modeled by the modeling method provided in the embodiment of the present application can specifically provide online intelligent customer service, that is, provide an intelligent online question and answer system.
[0107] Figure 3 This is a schematic diagram of the structure of an intelligent customer service model provided in an embodiment of the present application. Figure 3 As shown, the intelligent customer service model includes: an input layer 501, an execution layer 502 and an output layer 503. Among them, the execution layer 502 can realize identification and decision-making through various capability interfaces.
[0108] Based on the above description, the following is an introduction to the process of the intelligent customer service model completing the online question-and-answer task: First, the intelligent customer service model needs to observe the environmental feature information, which includes, for example, the user's attribute information, the information of the products that provide intelligent customer service services, and the product order information selected by the user. Then obtain the specific corpus input by the user, that is, the specific inquiry question. Then the intelligent customer service model needs to call its own recognition ability to identify the user's intention (that is, the understanding of the input corpus) based on the observed environmental feature information. Exemplarily, the intelligent customer service model can extract keywords for intent recognition, or it can recognize intent based on language matching methods, without specific limitations. At the same time, the intelligent customer service model can also perform slot filling processing on the corpus input by the user based on the environmental feature information to supplement the guess of the user's intention. Then, the intelligent customer service model calls its own decision-making ability to decide what words to use to reply to the user based on the user's intention, that is, to generate reply information based on the user's intention.
[0109] Figure 4 A structural block diagram of a modeling device for an intelligent service model provided in an embodiment of the present application, such as Figure 4 As shown, the modeling device 600 includes:
[0110] The acquisition module 601 is used to acquire an agent definition file; the agent definition file includes a data structure of input data and output data of various capability interfaces required to complete the intelligent service task, wherein the agent definition file declares a self-learning field for the target capability interface;
[0111] A parsing module 602 is used to parse the agent definition file to obtain an agent initial code including the capability interface; in the agent initial code, the interface function of the capability interface is an empty function;
[0112] Insertion module 603, used for inserting the first interface function content for realizing the capability in the initial code of the agent for the capability interface;
[0113] The insertion module 603 is also used to insert the second interface function content for external intelligent agent collaboration in the initial code of the intelligent agent for the target capability interface, and the self-learning code for self-learning capabilities according to the results of the external intelligent agent collaboration, so as to obtain an intelligent service model based on the intelligent agent.
[0114] Optionally, the target capability interface is used to implement the capability through the first interface function content based on the target input data when called, and when the capability cannot be implemented, external intelligent agent collaboration is performed through the second interface function content.
[0115] Optionally, the collaboration module for realizing the collaboration of the external agent includes:
[0116] A conversion unit, configured to convert the target input data from an agent data format into a manual annotation format;
[0117] An acquisition unit, configured to acquire the external agent collaboration result in the manually annotated format for the target input data after the format conversion;
[0118] The conversion unit is used to convert the external agent collaboration result from the manual annotation format to the agent data format;
[0119] A response unit is used to perform a response action for the intelligent service task according to the external intelligent agent collaboration result after the conversion.
[0120] Optionally, the conversion unit is specifically used for:
[0121] Converting the target input data into manual annotation information; the manual annotation information includes annotation task description information and multiple annotation options;
[0122] The acquisition unit is specifically used for:
[0123] For the target input data after the conversion, a target annotation option selected from the multiple annotation options is obtained, and the target annotation option is used as the external intelligent agent collaboration result in the manual annotation format.
[0124] Optionally, the collaboration module for realizing the collaboration of the external agent includes:
[0125] The calling unit includes calling an external intelligent model, and inputting the target input data into the external intelligent model to obtain the external intelligent agent collaboration result output by the external intelligent model.
[0126] Optionally, the target capability interface is also used to record the external intelligent agent collaboration result after completing the external intelligent agent collaboration, and perform capability self-learning through the self-learning code according to the recorded external intelligent agent collaboration result.
[0127] Optionally, the capability interface includes an identification capability interface and a decision capability interface;
[0128] The input data of the recognition capability interface includes observation data, the output data of the recognition capability interface includes status data, and the recognition capability interface is used to obtain the status data according to the analysis of the observation data;
[0129] The input data of the decision-making capability interface includes the observation data and / or the status data, the output data of the decision-making capability interface includes the response action to be taken, and the decision-making capability interface is used to determine the response action to be taken based on the observation data and / or the status data.
[0130] An embodiment of the present application also provides an intelligent service device, which is configured with an intelligent service model modeled using the modeling method of the intelligent service model as described above, and the intelligent service device is used to implement the method of using the intelligent service model as described above through the intelligent service model.
[0131] In an optional implementation, the intelligent service device may be an intelligent robot.
[0132] Next, an electronic device provided by an embodiment of the present application is introduced. Figure 5 , Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device 800 may be deployed with the intelligent service model described in the above embodiment for the functions in the above embodiment. Specifically, the electronic device 800 includes: a receiver 801, a transmitter 802, a processor 803 and a memory 804 (the number of processors 803 in the execution device 800 may be one or more, Figure 5 In the example of FIG. 801 , the processor 803 may include an application processor 8031 and a communication processor 8032. In some embodiments of the present application, the receiver 801, the transmitter 802, the processor 803 and the memory 804 may be connected via a bus or other means.
[0133] The memory 804 may include a read-only memory and a random access memory, and provides instructions and data to the processor 803. A portion of the memory 804 may also include a non-volatile random access memory (NVRAM). The memory 804 stores processor and operation instructions, executable modules or data structures, or subsets thereof, or extended sets thereof, wherein the operation instructions may include various operation instructions for implementing various operations.
[0134] The processor 803 controls the operation of the execution device. In a specific application, the various components of the execution device are coupled together through a bus system, wherein the bus system includes not only a data bus but also a power bus, a control bus, and a status signal bus, etc. However, for the sake of clarity, various buses are referred to as bus systems in the figure.
[0135] The method disclosed in the above embodiment of the present application can be applied to the processor 803, or implemented by the processor 803. The processor 803 can be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the processor 803. The above processor 803 can be a general processor, a digital signal processor (digital signal processing, DSP), a microprocessor or a microcontroller, and can further include an application specific integrated circuit (application specific integrated circuit, ASIC), a field programmable gate array (field-programmable gate array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The processor 803 can implement or execute the methods, steps and logic block diagrams disclosed in the embodiments of the present application. The general processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiment of the present application can be directly embodied as a hardware decoding processor to execute, or the hardware and software modules in the decoding processor can be combined and executed. The software module may be located in a storage medium mature in the art, such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 804, and the processor 803 reads the information in the memory 804 and completes the steps of the above method in combination with its hardware.
[0136] The receiver 801 can be used to receive input digital or character information and generate signal input related to the relevant settings and function control of the execution device. The transmitter 802 can be used to output digital or character information through the first interface; the transmitter 802 can also be used to send instructions to the disk group through the first interface to modify the data in the disk group; the transmitter 802 can also include a display device such as a display screen.
[0137] In the embodiment of the present application, the application processor 8031 in the processor 803 is used to execute the modeling method or the use method of the intelligent service model in the above embodiment. It should be noted that the specific manner in which the application processor 8031 executes each step is based on the same concept as the various method embodiments in the present application, and the technical effects it brings are the same as the various method embodiments in the present application. For specific contents, please refer to the description in the method embodiments shown in the previous embodiment of the present application, and no further description will be given here.
[0138] An embodiment of the present application also provides a chip for running instructions, which is used to execute the technical solution of the modeling method or use method of the intelligent service model in the above embodiment.
[0139] An embodiment of the present application also provides a computer-readable storage medium, which stores computer instructions. When the computer instructions are executed on a server, the server executes the technical solution of the modeling method or the use method of the intelligent service model in the above-mentioned embodiment.
[0140] An embodiment of the present application also provides a computer program product, including a computer program, which, when executed by a processor, is used to execute the technical solution of the modeling method or the use method of the intelligent service model in the above-mentioned embodiment.
[0141] The computer readable storage medium mentioned above can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The readable storage medium can be any available medium that can be accessed by a general or dedicated server.
[0142] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.
[0143] Although the present application is disclosed as above in the form of a preferred embodiment, it is not intended to limit the present application. Any technical personnel in this field may make possible changes and modifications without departing from the spirit and scope of the present application. Therefore, the scope of protection of the present application shall be based on the scope defined by the claims of the present application.
Claims
1. A modeling method for an intelligent service model, characterized in that: The method comprises: Obtaining an agent definition file; the agent definition file includes a data structure of input data and output data of various capability interfaces required to complete the intelligent service task, wherein the agent definition file declares a self-learning field for the target capability interface; Parsing the agent definition file to obtain an agent initial code including the capability interface; in the agent initial code, the interface function of the capability interface is an empty function; Inserting a first interface function content for implementing the capability into the agent initial code for the capability interface; In the initial code of the intelligent agent, the second interface function content for collaboration with an external intelligent agent and the self-learning code for self-learning of capabilities according to the collaboration results of the external intelligent agent are inserted for the target capability interface to obtain an intelligent service model based on the intelligent agent.
2. The method according to claim 1, characterized in that The target capability interface is used to realize the capability through the first interface function content based on the target input data when called. When the capability cannot be realized, external intelligent agent collaboration is carried out through the second interface function content.
3. The method according to claim 2, characterized in that The method for external agent collaboration includes: Converting the target input data from an agent data format to a manually annotated format; For the target input data after the conversion, obtaining the external agent collaboration result in the manually annotated format; Converting the external agent collaboration result from the manual annotation format to the agent data format; A response action is performed for the intelligent service task according to the external intelligent agent collaboration result after the conversion.
4. The method according to claim 3, characterized in that The step of converting the target input data from an agent data format to a manual annotation format comprises: Converting the target input data into manual annotation information; the manual annotation information includes annotation task description information and multiple annotation options; The step of obtaining the external agent collaboration result in the manually annotated format for the target input data after the format conversion includes: For the target input data after the conversion, a target annotation option selected from the multiple annotation options is obtained, and the target annotation option is used as the external intelligent agent collaboration result in the manual annotation format.
5. The method according to claim 2, characterized in that: The method for external agent collaboration includes: An external intelligent model is called, and the target input data is input into the external intelligent model to obtain an external intelligent agent collaboration result output by the external intelligent model.
6. The method according to claim 1, characterized in that The target capability interface is also used to record the external intelligent agent collaboration result after completing the external intelligent agent collaboration, and perform capability self-learning through the self-learning code according to the recorded external intelligent agent collaboration result.
7. The method according to claim 1, characterized in that The capability interface includes an identification capability interface and a decision-making capability interface; The input data of the recognition capability interface includes observation data, the output data of the recognition capability interface includes status data, and the recognition capability interface is used to obtain the status data according to the analysis of the observation data; The input data of the decision-making capability interface includes the observation data and / or the status data, the output data of the decision-making capability interface includes the response action to be taken, and the decision-making capability interface is used to determine the response action to be taken based on the observation data and / or the status data.
8. A method for using an intelligent service model, characterized in that: The intelligent service model is modeled by using the modeling method of the intelligent service model as described in any one of claims 1 to 7 above, and the use method includes: receiving target input data of the intelligent service task; Calling the capability interface in the intelligent service model to implement the capability corresponding to the capability interface based on the target input data, and outputting a response action for the intelligent service task; For the capability interface with self-learning configuration, when the capability cannot be realized through the capability interface, external agent collaboration is performed; After the external intelligent agent collaboration is completed, the external intelligent agent collaboration result is recorded, and the capability interface is self-learned based on the recorded external intelligent agent collaboration result.
9. A modeling device for an intelligent service model, characterized in that: The modeling device comprises: An acquisition module is used to acquire an agent definition file; the agent definition file includes a data structure of input data and output data of various capability interfaces required to complete the intelligent service task, wherein the agent definition file declares a self-learning field for the target capability interface; A parsing module, used for parsing the agent definition file to obtain an agent initial code including the capability interface; in the agent initial code, the interface function of the capability interface is an empty function; An insertion module, used for inserting a first interface function content for realizing the capability in the initial code of the agent for the capability interface; The insertion module is also used to insert the second interface function content for external intelligent agent collaboration in the initial code of the intelligent agent for the target capability interface, and the self-learning code for self-learning capabilities according to the results of the external intelligent agent collaboration, so as to obtain an intelligent service model based on the intelligent agent.
10. An intelligent service device, characterized in that: The intelligent service device is configured with an intelligent service model modeled by the modeling method of the intelligent service model as described in any one of claims 1 to 7 above, and the intelligent service device is used to implement the method for using the intelligent service model as described in claim 8 above through the intelligent service model.
11. The intelligent service device according to claim 10, characterized in that: The intelligent service device is an intelligent robot.
12. An electronic device, characterized in that: include: A processor, a memory, and computer program instructions stored on the memory and executable on the processor; When the processor executes the computer program instructions, the processor implements the modeling method of the intelligent service model as described in any one of claims 1 to 7, or the method for using the intelligent service model as described in claim 8.
13. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement a modeling method for an intelligent service model as described in any one of claims 1 to 7 above, or a method for using an intelligent service model as described in claim 8 above.
Citation Information
Cited By
Capability permission management method and related apparatus
WO2026137277A1
