Intent management method and communication apparatus

Abstract

Provided in the present application are an intent management method and a communication apparatus. A large model application framework may trigger a large model to acquire information related to an intent; and the large model then provides a first parameter of the intent to the large model application framework, such that the large model application framework acquires, from a production entity, information related to the first parameter of the intent, and then on the basis of the information from the production entity side, the large model application framework manages the intent, such as creating an intent or modifying the intent. The present application helps to reduce the probability of failure of intent feasibility checks, thereby improving the intent operation efficiency, and accordingly improving the intent management efficiency and the user experience.

Description

Intent management method and communication device

[0001] This application claims priority to the Chinese patent application filed with the State Intellectual Property Office of China on January 29, 2024, with application number 202410123906.6 and invention name “A Method and Communication Device for Intent Management”, the entire contents of which are incorporated by reference into this application. Technical Field

[0002] The present application relates to the field of communications, and more particularly, to an intent management method and a communication device. Background Art

[0003] With the development of network resource management, network management service consumers no longer need to directly manage network resources. Instead, they can send messages to network management service producers to indicate structured network management intents. Upon receiving these structured network management intents, network management service producers can translate these structured network management intents into network requirements and specific operations, and then perform the corresponding operations to achieve the network management intents.

[0004] After receiving a network management intent, the network management service production entity can perform a feasibility check on the network management intent to determine whether the network management intent is feasible. Currently, the probability of feasibility check failure is high, which affects the management efficiency of the network management intent.

[0005] Summary of the Invention

[0006] The embodiments of the present application provide an intent management method and a communication device to reduce the probability of failure of intent feasibility check, thereby improving the efficiency of intent operation, and further improving the efficiency of intent management and user experience.

[0007] In a first aspect, an intent management method is provided, which can be executed by a first device. Unless otherwise specified, the "first device" can refer to the first device itself or a device that can support the first device to perform its functions.

[0008] The method includes: a first device sends first information to a second device, the first information is used to trigger the acquisition of information related to the intent from a first production entity; the first device receives second information from the second device, the second information includes a first parameter of the intent; the first device sends third information to the first production entity based on the second information, the third information is used to obtain fourth information related to the first parameter; the first device receives the fourth information from the first production entity; the first device obtains fifth information for managing the intent based on the fourth information.

[0009] Exemplarily, the intent is a network management intent.

[0010] Based on the above method, the first device can trigger the second device to obtain information related to the intent, and then the second device provides the first parameter of the intent to the first device, so that the first device can obtain information related to the first parameter of the intent from the first production entity, and then obtain information for managing the intent based on the information related to the first parameter of the intent on the first production entity side. In this way, in the above method, intents can be managed based on real-time information on the first production entity side, such as creating or modifying intents, which helps reduce the probability of intent feasibility check failure, thereby improving intent operation efficiency, and further improving intent management efficiency and user experience.

[0011] In conjunction with the first aspect, in some possible implementations, the method further includes: the first device receiving sixth information input by the user, the sixth information indicating an operation related to the intent; and the first device determining the first information based on the sixth information. In other words, the first device can determine that the user is performing an operation related to the intent based on the user input information, and thereby determine the first information.

[0012] In combination with the first aspect or any implementation thereof, in some other possible implementations, the first information includes a first prompt instance, and the thought chain of the first prompt instance includes steps for triggering obtaining information related to the intention from the first production entity.

[0013] In combination with the first aspect or any implementation thereof, in some other possible implementations, the second information also includes application programming interface (API) information, and the API information is used to indicate the API called to obtain the fourth information; the third information is used to call the API to obtain the fourth information.

[0014] Exemplarily, the API information may be an index or number of an API, or an API calling instruction.

[0015] In the above implementation, the second device determines the information that needs to be obtained and the API that needs to be called to obtain the information.

[0016] In combination with the first aspect or any implementation thereof, in some other possible implementations, the first device obtaining, based on the fourth information, fifth information for managing the intent includes: the first device sending the fourth information to the second device; and the first device receiving, from the second device, fifth information, where the fifth information is used to manage the intent, the fifth information being determined based on the fourth information. The method further includes: the first device managing the intent based on the fifth information.

[0017] In the above implementation, the first device obtains information for management intent from the second device.

[0018] In combination with the first aspect or any implementation thereof, in some other possible implementations, the first device manages the intent based on the fifth information, including: the first device performs an intention operation based on the fifth information, and the intention operation includes at least one of the following: requesting the first production entity to create the intent, requesting the first production entity to modify the intent, requesting the first production entity to delete the intent, querying the first production entity about the intent, giving up creating the intent, or giving up modifying the intent.

[0019] In combination with the first aspect or any implementation manner thereof, in some other possible implementation manners, the method further includes: the first device determining a second prompt instance based on the fourth information, where context information of the second prompt instance includes the fourth information. The first device sending the fourth information to the second device includes: the first device sending the second prompt instance to the second device.

[0020] In combination with the first aspect or any implementation thereof, in some other possible implementations, the thought chain of the second prompt instance includes steps for triggering the second device to generate management suggestions for the intention based on the fourth information.

[0021] Based on the above implementation method, the second device can be triggered to generate management suggestions for the intention based on the fourth information, thereby helping the user to issue correct operation information and complete the corresponding intention operation, which can effectively avoid erroneous intention operations.

[0022] In combination with the first aspect or any implementation thereof, in some other possible implementations, the method further includes: the first device receiving seventh information from the second device and outputting the seventh information to the user, wherein the seventh information is used to indicate the management suggestion.

[0023] Based on the above implementation method, the second device can output management suggestions for the intention to the user through the first device, thereby helping the user to issue correct operation information and complete the corresponding intention operation, which can effectively avoid erroneous intention operations.

[0024] In combination with the first aspect or any implementation thereof, in some other possible implementations, the method further includes: the first device receives eighth information input by the user and sends the eighth information to the second device, where the eighth information is feedback information of the user regarding the management suggestion.

[0025] Based on the above implementation method, the second device can provide the first device with more reasonable information for managing intent based on user feedback information, thereby helping to improve intent operation efficiency, and further improve intent management efficiency and user experience.

[0026] In combination with the first aspect or any implementation thereof, in some other possible implementations, the first information is further used to trigger the second device to modify the intention, and the first information includes the first parameter.

[0027] In a scenario where the structured intent already exists on the first device, the first device can read the intent parameters from the memory and provide them to the second device via the first message, thereby simplifying the interaction process between the second device and the user.

[0028] In combination with the first aspect or any implementation thereof, in some other possible implementations, the first device and the second device are set in the same entity, such as the first device and the second device are set in a consumption entity or a production entity; or, the first device and the second device are set in different entities, such as the first device is set in a consumption entity or a second production entity, and the second device is set in other entities (such as a cloud platform).

[0029] In combination with the first aspect or any implementation thereof, in some other possible implementations, the intention is an intention related to network management, and the first parameter includes at least one of the following parameters: an expected object, an expected target, or an expected context.

[0030] On the second aspect, an intent management method is provided, which can be executed by a second device. Unless otherwise specified, the "second device" can refer to the second device itself or a device that can support the second device to perform its functions.

[0031] The terms or features in the second aspect or its implementation that are the same as those in the first aspect or its implementation can refer to the first aspect or its implementation, and the technical effects of the second aspect or its implementation can refer to the technical effects in the first aspect or its implementation, and will not be repeated in the second aspect.

[0032] The method includes: a second device receives first information from the first device, the first information is used to trigger the acquisition of information related to the intention from a first production entity; the second device sends second information to the first device based on the first information, the second information includes a first parameter of the intention.

[0033] In combination with the second aspect, in some possible implementations, the first information includes a first prompt instance, and the thought chain of the first prompt instance includes steps for triggering acquisition of information related to the intention from the first production entity.

[0034] In combination with the second aspect or any implementation thereof, in some other possible implementations, the second information further includes API information, where the API information is used to indicate an API called to obtain fourth information related to the first parameter.

[0035] In combination with the second aspect or any implementation thereof, in some other possible implementations, the method further includes: the second device receives fourth information related to the first parameter from the first device; the second device determines fifth information for managing the intention based on the fourth information and the large model; and the second device sends the fifth information to the first device.

[0036] In combination with the second aspect or any implementation thereof, in some other possible implementations, the second device receiving third information related to the first parameter from the first device includes: the second device receiving a second prompt instance from the first device, the context of the second prompt instance including the fourth information. The second device determining, based on the fourth information and a large model, fifth information for managing the intent includes: the second device determining the fifth information based on the second prompt instance and the large model.

[0037] In combination with the second aspect or any implementation thereof, in some other possible implementations, the thought chain of the second prompt instance includes steps for triggering the second device to generate management suggestions for the intention based on the fourth information.

[0038] In combination with the second aspect or any implementation thereof, in some other possible implementations, the method further includes: the second device generates the management suggestion based on the second prompt instance; the second device sends seventh information to the first device, and the seventh information is used to indicate the management suggestion.

[0039] In combination with the second aspect or any implementation thereof, in some other possible implementations, the method further includes: the second device receiving eighth information from the first device, the eighth information being user feedback regarding the management suggestion. The second device determining the fifth information based on the second prompt instance and the large model includes: the second device determining the fifth information based on the second prompt instance, the large model, and the eighth information.

[0040] In combination with the second aspect or any implementation manner thereof, in some other possible implementation manners, the first information is further used to trigger the second device to modify the intent, the first information includes the first parameter. The second device sending the second information to the first device based on the first information includes: the second device obtaining the first parameter from the first information; and the second device sending the second information to the first device based on the first parameter.

[0041] In combination with the second aspect or any implementation thereof, in some other possible implementations, the second device sends second information to the first device based on the first information, including: the second device determines the first parameter based on the first information by interacting with the user through the first device; the second device sends the second information to the first device based on the first parameter.

[0042] In combination with the second aspect or any implementation thereof, in some other possible implementations, the first device and the second device are set in the same entity, such as the first device and the second device are set in a consumption entity or a production entity; or, the first device and the second device are set in different entities, such as the first device is set in a consumption entity or a second production entity, and the second device is set in other entities (such as a cloud platform).

[0043] In combination with the second aspect or any implementation thereof, in some other possible implementations, the intention is an intention related to network management, and the first parameter includes at least one of the following parameters: an expected object, an expected target, or an expected context.

[0044] On the third aspect, an intent management method is provided, which can be executed by a first production entity. Unless otherwise specified, "first production entity" can refer to the first production entity itself or a device that can support the first production entity to realize its functions.

[0045] The method includes: a first production entity receiving third information, where the third information is used to obtain fourth information related to a first parameter of an intent; and the first production entity sending the fourth information.

[0046] The terms or features in the third aspect or its implementation that are the same as those in the first aspect or its implementation can refer to the first aspect or its implementation, and the technical effects of the third aspect or its implementation can refer to the technical effects in the first aspect or its implementation, and will not be repeated in the third aspect.

[0047] In a fourth aspect, an intent management method is provided, which is executed by a first device and a second device. Unless otherwise specified, the "first device" or "second device" may refer to the first device or the second device itself, or may refer to a device that can support the first device or the second device to perform its functions.

[0048] The method includes: the first device sends first information, the second device receives the first information, the first information is used to trigger the acquisition of information related to the intention from the first production entity; the second device sends second information based on the first information, the first device receives the second information, the second information includes a first parameter of the intention; the first device sends third information to the first production entity based on the second information, the third information is used to obtain fourth information related to the first parameter; the first device receives the fourth information from the first production entity; the first device obtains fifth information for managing the intention based on the third information.

[0049] The steps performed by the first device in the fourth aspect or its implementation method can refer to the first aspect or its implementation method, the steps performed by the second device in the fourth aspect or its implementation method can refer to the second aspect or its implementation method, the terms or features in the fourth aspect or its implementation method that are the same as those in the first aspect or its implementation method can refer to the first aspect or its implementation method, and the technical effects of the fourth aspect or its implementation method can refer to the technical effects in the first aspect or its implementation method, and will not be repeated in the fourth aspect.

[0050] In the fifth aspect, an intent management method is provided, which can be executed by a first device, a second device and a first production entity. Unless otherwise specified, "first device", "second device" or "first production entity" may refer to the first device, the second device or the first production entity itself, or may refer to a device that can support the first device, the second device or the first production entity to perform its functions.

[0051] The method includes: the first device sends first information, the second device receives the first information, the first information is used to trigger the acquisition of information related to the intention from the first production entity; the second device sends second information based on the first information, the first device receives the second information, the second information includes a first parameter of the intention; the first device sends third information based on the second information, the first production entity receives the third information, the third information is used to obtain fourth information related to the first parameter; the first production entity sends the fourth information, the first device receives the fourth information; the first device obtains fifth information for managing the intention based on the third information.

[0052] The steps performed by the first device in the fifth aspect or its implementation method may refer to the first aspect or its implementation method, the steps performed by the second device in the fifth aspect or its implementation method may refer to the second aspect or its implementation method, the steps performed by the first production entity in the fifth aspect or its implementation method may refer to the third aspect or its implementation method, the terms or features in the fourth aspect or its implementation method that are the same as those in the first aspect or its implementation method may refer to the first aspect or its implementation method, the technical effects of the fourth aspect or its implementation method may refer to the technical effects in the first aspect or its implementation method, and will not be repeated in the fourth aspect.

[0053] In a sixth aspect, a communication device is provided, the device being configured to execute the method provided in any of the above aspects or implementations thereof. Specifically, the device may include units and / or modules, such as a processing unit and / or a transceiver unit, configured to execute the method provided in any of the above aspects or implementations thereof. The processing unit is configured to execute the processing steps in the method provided in any of the above aspects or implementations thereof. The transceiver unit is configured to execute the transceiver steps in the method provided in any of the above aspects or implementations thereof.

[0054] In one implementation, the device is a first device, a second device, or a first production entity. When the device is the first device, the second device, or the first production entity, the transceiver unit may be a transceiver, an input / output interface, or a communication interface; and the processing unit may be at least one processor. Optionally, the transceiver is a transceiver circuit. Optionally, the input / output interface is an input / output circuit.

[0055] In another implementation, the device is a chip, chip system, or circuit used in the first device, the second device, or the first production entity. When the device is a chip, chip system, or circuit used in the first device, the second device, or the first production entity, the transceiver unit may be an input / output interface, interface circuit, output circuit, input circuit, pin, or related circuit on the chip, chip system, or circuit; and the processing unit may be at least one processor, processing circuit, or logic circuit.

[0056] In a seventh aspect, a communication device is provided, which includes: a memory for storing programs; and at least one processor for executing computer programs or instructions stored in the memory to execute the method provided by any one of the above aspects or its implementation.

[0057] In one implementation, the device is a first device, a second device, or a first production entity.

[0058] In another implementation, the device is a chip, a chip system, or a circuit used in the first device, the second device, or the first production entity.

[0059] In an eighth aspect, a communication device is provided, comprising: at least one processor and a communication interface, wherein the at least one processor is configured to retrieve a computer program or instruction stored in a memory through the communication interface to execute the method provided by any one of the above aspects or implementations thereof. The communication interface may be implemented in hardware or software.

[0060] In one implementation, the device further includes the memory.

[0061] In a ninth aspect, a processor is provided for executing the methods provided in the above aspects.

[0062] For the operations such as sending and acquiring / receiving involved in the processor, unless otherwise specified, or if they do not conflict with their actual functions or internal logic in the relevant descriptions, they can be understood as operations such as processor output, reception, and input, or as sending and receiving operations performed by the radio frequency circuit and antenna. This application does not limit this.

[0063] In a tenth aspect, a computer-readable storage medium is provided, which stores program code for execution by a device, and the program code includes a method for executing any one of the above aspects or its implementation method.

[0064] In the eleventh aspect, a computer program product containing instructions is provided, including a computer program or instructions, which, when the computer program or instructions are run on a computer, implements the steps of the method provided by any of the above aspects or its implementation methods.

[0065] In a twelfth aspect, a chip is provided, comprising a processor and a communication interface, wherein the processor reads instructions stored in a memory through the communication interface and executes the method provided in any one of the above aspects or implementations thereof. The communication interface may be implemented in hardware or software.

[0066] Optionally, as an implementation method, the chip also includes a memory, in which a computer program or instruction is stored, and the processor is used to execute the computer program or instruction stored in the memory. When the computer program or instruction is executed, the processor is used to execute the method provided by any of the above aspects or its implementation methods.

[0067] When the method provided in this application is executed by a chip, this application does not limit the number of chips that implement the method. For example, the method can be executed by one chip or by two or more chips. Furthermore, when the number of chips implementing the method of this application is two or more, the chip manufacturers are not limited and can be the same manufacturer or different manufacturers.

[0068] In a thirteenth aspect, a communication system is provided, comprising at least one of the first device, the second device or the first production entity described above.

[0069] In a fourteenth aspect, a computer program is provided, which, when run on a computer, enables the method provided by any one of the above aspects or its implementation to be executed. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] FIG1 is a schematic structural diagram of an information object class (IOC).

[0071] FIG2 is an example of a large model without prompt assistance and a large model with prompt assistance.

[0072] FIG3 is a schematic structural diagram of a system architecture applicable to an embodiment of the present application.

[0073] FIG4 is another schematic structural diagram of a system architecture applicable to an embodiment of the present application.

[0074] Figure 5 is a schematic flowchart of the big model application framework processing natural language intent into structured intent based on the big model.

[0075] FIG6 is a schematic flowchart of an intent management method 600 provided in an embodiment of the present application.

[0076] FIG7 is a schematic flowchart of an intent management method 700 provided in an embodiment of the present application.

[0077] FIG8 is a schematic flowchart of an intent management method 800 provided in an embodiment of the present application.

[0078] FIG9 is a schematic structural diagram of a device provided in an embodiment of the present application.

[0079] FIG10 is another schematic structural diagram of the device provided in an embodiment of the present application.

[0080] FIG11 is a schematic diagram of a chip system provided in an embodiment of the present application. DETAILED DESCRIPTION

[0081] Before introducing the embodiments of the present application, the following explanation is made.

[0082] "For indicating" or "indicating" can include direct indication and indirect indication, or "for indicating" or "indicating" can be explicitly and / or implicitly indicated. The various numerical numbers such as first, second, etc. are only used for the convenience of description and are not used to limit the scope of the embodiments of this application, such as distinguishing different messages, different information, etc. "Pre-definition" can be implemented by pre-saving corresponding codes, tables or other methods that can be used to indicate relevant information in the device. This application does not limit its specific implementation method. The "protocol" involved may refer to a standard protocol in the field of communications, such as the Long Term Evolution (LTE) protocol, the New Radio (NR) protocol, and related protocols used in future communication systems. This application does not limit this. Words such as "exemplary", "for example", "exemplarily", "as (another) example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as an "example" in this application should not be construed as being preferred or advantageous over other embodiments or design schemes. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized. "Multiple" refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone, where A and B can be singular or plural. The character " / " generally indicates that the associated objects are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b and c can mean: a, or b, or c, or a and b, or a and c, or b and c, or a, b and c. Where a, b and c can be single or multiple. The description involving network element A sending a message, information or data to network element B, and network element B receiving a message, information or data from network element A, is intended to indicate to which network element the message, information or data is to be sent, and does not limit whether they are sent directly or indirectly via other network elements. Descriptions such as "when...", "in the case of...", "if...", and "if" all mean that the device will take corresponding actions under certain objective circumstances. They do not limit the time, nor do they require the device to make judgments when implementing them, nor do they imply the existence of other limitations.

[0083] In addition, the network architecture and business scenarios described in the embodiments of the present application are intended to more clearly illustrate the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided in the embodiments of the present application. Ordinary technicians in this field can know that with the evolution of network architecture and the emergence of new business scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0084] The technical solution of the present application can be applied to various network function virtualization (NFV) systems. The system can describe the current operator's business technology solutions, network construction solutions, and network operation and maintenance methods as patterns and strategies in a standard formal language, and implement the technical solutions and construction solutions based on these patterns and strategies. For example, the technical solution of the present application can be applied to one or more of the following systems: a wireless intent driven network (wIDN) system, an experiential networked intelligence (ENI) system, an intent driven management service (IDMS) system, or an open network automation platform (ONAP) system.

[0085] To facilitate understanding of the embodiments of the present application, some terms involved in the embodiments of the present application are first explained.

[0086] 1. Network management and network management services

[0087] 1.1 Network Management

[0088] Network management refers to the management of network resources, including but not limited to monitoring, controlling and recording the performance and usage of network resources, and issuing management action groups to network resources (such as devices in the network) based on the detected network status to enable the network to operate effectively. Exemplarily, network management can be at least one of monitoring, testing, configuring, analyzing, evaluating or controlling network resources. Network management can also be timely reporting and handling when network failures occur, and coordinating and maintaining the efficient operation of the network system. Among them, network resources are the objects of network management, which can also be called network objects or managed entities. Exemplarily, network resources can be base station equipment, routers, switches, core network equipment, etc. This is not particularly limited in the embodiments of the present application. For the convenience of description, the embodiments of the present application are described using managed entities as an example, but they can be replaced with other objects of network management.

[0089] 1.2 Network Management Services

[0090] A network management service refers to a service that provides network management functions. The service producing entity usually provides the network management functions to the service consuming entity through a network interface (eg, a service-based interface).

[0091] 2. Network management intention

[0092] The current 3rd Generation Partnership Project (3GPP) System Aspects Work Group 5 (SA5) adopts a comprehensive approach to modeling network resources and managing network objects in traditional interface definitions. The network management system directly adds, deletes, modifies, and queries all network resources (such as network devices) through operations such as configuration management, performance management, and fault management. This management approach imposes high barriers to operator management and O&M, and the differentiated implementations of equipment vendors hinder interoperability. To reduce management complexity and improve O&M efficiency in multi-vendor scenarios, standards organizations and projects such as 3GPP and the European Telecommunications Standards Institute (ETSI) have conducted research and standardization on network management intent technology. The main approach can be summarized as follows: the network management system does not directly manage network resources. Instead, it retains only the network management intent model, and interfaces only transmit vendor-agnostic structured network management intent, thereby shielding device vendor implementation differences. Structured network management intent is expressed declaratively, describing only the "what" and not the "how." After receiving the structured network management intent from the network management system, the device management system translates it into network requirements and specific operations based on the network status and executes the corresponding operations to achieve the network management intent.

[0093] A network management intent refers to a demand for network management services by a network management service consumer entity, and may also be referred to as a network management service demand. This term is not limited in the present application embodiments and is hereinafter referred to as an intent. For example, an intent may be an optical dedicated line service intent or an energy-saving intent, such as the intent defined in 3rd Generation Partnership Project (3GPP) standard specification 28.312.

[0094] Intent-based interaction scenarios primarily involve two roles: the network management service consumer and the network management service producer. The network management service consumer is the entity that invokes management services, or in other words, the intent service. The network management service producer is the entity that provides network management services. A more detailed description of the network management service consumer and producer can be found in Figure 3 below.

[0095] Intents can be applied to network systems to achieve desired performance targets for specific network devices within a specific scope. Intents can express expectations for network systems through formal specifications and descriptive information. Formal specifications can refer to the intent's specific syntax or semantics, while descriptive information can describe at least one of the intent's requirements, goals, or constraints.

[0096] An intent can include at least one intent expectation. Each intent expectation can represent the performance requirements of a network management service consuming entity for a specific network object. An intent expectation can consist of at least one expected target and at least one expected object. Each expected target can represent the performance requirements of the network management service consuming entity for a specific attribute of the expected object. The expected object can be the network object on which the intent acts, and the expected object can also be called the action object. For example, an intent expectation can include two expected targets, one of which is an average downlink throughput greater than 5 Mbps and the other is an end-to-end latency less than 10 milliseconds. If the average downlink throughput of the expected object is greater than 5 Mbps, then the downlink throughput target is met. If the end-to-end latency of the expected object is less than 10 milliseconds, then the latency target is met. If both expected targets of an intent expectation are met, then the intent expectation is met. If one or more intent expectations are not met, or if one or more expected targets are not met, then the intent is not met.

[0097] The intention, intention expectation or expected goal is met or not met, which can also be described as: the intention, intention expectation or expected goal is achieved or not achieved, the intention, intention expectation or expected goal is realized or not realized, or the intention, intention expectation or expected goal is achieved or not achieved, etc.

[0098] An intent may also include an intent context. The intent context may represent constraints or conditions imposed by the network management service consumer on the intent. For example, the network management service consumer may constrain the duration of the intent through the intent context.

[0099] Intents may also include an expected context. The expected context may represent the constraints or conditions that a network management service consumer expects for an intent. For example, the network management service consumer may use the expected context to constrain the expected duration of the intent.

[0100] The intent may also include a target context, which may represent the constraints or conditions imposed by the network management service consumer on a desired target. For example, the network management service consumer may constrain the duration of the desired target through the target context.

[0101] The intent may also include an object context, which may indicate constraints or conditions on the network object to be acted upon by the network management service consumer. For example, the network management service consumer may use the object context to indicate the area where the intended network object is located.

[0102] For example, the intent may be expressed as:

[0103] For example, an intent can be expressed as: "Ensure that the handover failure rate of a certain cell is less than 2% when the load is greater than 80%." "Cell" is the expected object or action object, and "Cell" can correspond to Expectation Object O; "Load > 80%" is the target context, and "Load > 80%" can correspond to Target Context (such as C_1); "Handover failure rate < 2%" is the expected target, and "Handover failure rate < 2%" can correspond to Expectation Target (such as T_1).

[0104] The intent expectation, expected target and intent context introduced above can be used as information elements in the IOC in the intent creation request. The network management service consumer entity can express its demand for network management services by sending the above information elements to the network management service producer entity.

[0105] The information elements in the intent IOC can use a specific structure to express the demand for network management services.

[0106] Exemplarily, FIG1 shows a schematic structural diagram of an intention IOC.

[0107] Referring to Figure 1 , the intent IOC includes an intent expectation and an intent context. The intent expectation includes an expected object, an expected target, and an expected context. The expected object includes an expected object context, and the expected target includes a target context. It is understood that non-contextual information elements (such as intent IOC, intent expectation, expected object, or expected target) can include contextual information elements (such as intent context, expected object context, target context, expected context, etc.), and contextual information elements can be used to constrain non-contextual information elements.

[0108] It is understood that Figure 1 is a schematic illustration of the relationship between attributes in an intent IOC. Based on network management service requirements, the intent IOC may also include other attributes, and this application does not specifically limit this. It is also understood that this application does not specifically limit the number of attributes of the same type. For example, an intent IOC may include one or more parallel intent expectation attributes, and an intent expectation may also include one or more parallel desired target attributes.

[0109] 3. Intent Template

[0110] An intent template refers to a templated description information of an intent, and may also be referred to as a network management service requirement template. This name is not limited in the embodiments of the present application. The intent template may be used to specify the syntax and semantics of the various attribute expressions introduced above. For example, the intent template may specify the field types and value types that may be used for each attribute in the IOC. As an example, the intent template may include intent expectation description information for specifying how the intent expectation is to be expressed. The intent expectation description information may specify the types of expected target attributes that the intent expectation may include, such as latency, bandwidth, maximum number of users, etc. The intent expectation description information may also specify the field type of each expected target. For example, the field type of the expected target for latency is a string, the field type for the expected target for bandwidth is a string, and the field type for the expected target for the maximum number of users is an integer. As another example, the intent template may also include intent context description information that specifies how the intent context and the expected context are expressed. The intent context description information may be used to specify the field type of the intent context, etc. This application does not specifically limit this.

[0111] 4. Intent Expression

[0112] In the embodiment of the present application, the intent expression is an informational expression that expresses the intent, and can be the instantiation result of the intent template. The consumer of the intent can instantiate the intent IOC according to the intent template to generate an intent expression, and the instantiated intent expression can be used for a specific network service. For example, the intent expression can be represented by a list of a set of attribute and value key-value pairs (for example, a set of [attribute, value]) of the intent.

[0113] Exemplarily, the intent expression includes a performance indicator information instance, a network object information instance, and a context information instance.

[0114] For example, a performance indicator information instance may include the value or value range of the target indicator requirement corresponding to the performance indicator information. For example, if the target indicator requirement corresponding to a performance indicator information template includes a bandwidth parameter, a performance indicator information instance may include a bandwidth parameter value of 20 Mbps, or a bandwidth parameter value range of 10 Mbps to 15 Mbps. For example, if the target indicator requirement corresponding to a performance indicator information template includes a latency parameter, a latency parameter instance may include a latency parameter value of 2 s, or a latency parameter value range of 1 s to 5 s.

[0115] The target network object indicated by the network object information instance is the instantiation result of the target network object indicated by the network object information template. Exemplarily, the target network object indicated by the network object information instance is one of the target network objects indicated by the network object information template, i.e., one of the physical or logical entities that implement the network management service. In this case, the network object information instance indicates a specific network object. For example, the network object information template indicates a network service, while the network object information instance indicates an energy-saving service.

[0116] The context information instance is used to refer to the constraints of the schematic expression. Exemplarily, the context information instance includes the constraint parameters of the intent expression. For example, the context information instance may include the source and sink information corresponding to the target network object indicated by the network object instance.

[0117] Intent expectation information instances include performance indicator information instances and network object information instances.

[0118] In the embodiment of the present application, the "intention" transmitted through the message refers to the intention expression.

[0119] 5. Intent Examples

[0120] An intent instance refers to an intent processing process created locally by a network management service production entity based on a received intent expression (e.g., a performance indicator information instance, a network object information instance, a context information instance).

[0121] The intent processing process is used to process intent expressions. For example, the intent processing process may include an identifier that identifies the process; computing resources that are used to perform operations such as intent translation and querying a database (e.g., a semantic knowledge base); and storage resources that are used to store intent expressions and intent translation results.

[0122] Among them, intent translation is used to convert intent expressions into corresponding management action groups, which can be instructions acting on physical entities or logical entities, such as adjusting the antenna tilt angle of a base station or turning on the energy-saving switch of a cell.

[0123] 6. Feasibility Check of Intention

[0124] The feasibility check of the intent can be performed by the production entity to check whether the intent is feasible. After receiving the intent creation request or modification request from the consumer entity, the production entity can automatically perform the feasibility check of the intent, thereby obtaining an intent feasibility check report. The feasibility check may include: checking the satisfaction of intent fulfillment, or whether there are potential conflicts (potential conflicts between one or more intent instances), etc., wherein the potential conflicts include at least one of conflicts between different intents, conflicts between different intent expectations of the same intent, or conflicts between different expected goals of the same intent.

[0125] 7. Large Model

[0126] Big models are used to provide big model services. Big models refer to neural network models with extremely large parameters. Big models can also be called foundation models. Big models play an important role in many fields and applications. Common application scenarios include natural language processing, computer vision, speech recognition and synthesis, recommendation systems, financial risk control, intelligent dialogue systems, game artificial intelligence (AI), and healthcare. In the future, big models will evolve towards multimodality. Multimodal big models can process data in multiple modalities. For example, multimodal big models can process natural language as well as text, images, or videos.

[0127] A large language model (LLM) is a deep learning model trained using large amounts of text data. LLMs can generate natural language text or understand the meaning of text. LLMs can handle a variety of natural language tasks, such as text classification, question answering, and conversation.

[0128] 8. Prompt

[0129] Prompt is an interactive method based on natural language processing. It uses machines to parse natural language, enabling communication between users and machines. Prompt primarily converts natural language into machine-readable instructions by building a corresponding corpus and semantic parsing model. Prompt can generate fine-grained tasks by adding prompt information to the original text based on context, thought processes, or prompt words, guiding large models to better complete tasks.

[0130] The design principles of prompts include writing clear and specific instructions and giving the model time to think.

[0131] The design points of prompt include: Instruct, Context, Constraint, Example and Input. Among them, Instruct is used to provide task instructions to be performed and to state specific goals. Context is used to allow artificial intelligence (AI) (or a large model) to stand in the role's perspective and provide context around the task. Constraint is used to give steps to complete the task (such as a chain of thought (COT)), clearly specify the language style, and specify the output template. Example is used to give a specific example. It should be noted that the division method of the various design points of the prompt here is only an example, and the prompt can also adopt other division methods without limitation.

[0132] 9. Prompt Example

[0133] A prompt instance is a collection of information sent to the big model that contains the user's goal and the context that describes this goal.

[0134] Prompt instances can be composed in various ways, with varying amounts of content. The more detailed the prompt instance, the higher the quality of the output guidance for the larger model. Prompt instances primarily come in the following forms: Zero-Shot, Few-Shot, COT, or Role-Play. Zero-Shot only requires simple input, making it easy to use but often failing to demonstrate the powerful capabilities of the larger model. Few-Shot, while providing the larger model with input and output examples, can effectively improve the quality of its final output. COT effectively activates the larger model's reasoning capabilities by introducing thought chaining within the prompt. Role-Play allows prompt instances to be defined in a role-playing manner, maximizing guidance for the larger model to achieve its task objectives.

[0135] Here is an example of a prompt instance that plays the role of a "book reviewer".

[0136] ##Role: Book Reviewer

[0137] ##Profile:

[0138] - Language: Chinese

[0139] -Description: I am an experienced book reviewer who is good at conveying reading notes in concise and clear language.

[0140] ##Goals:

[0141] I hope to use the prescribed framework to output the key points of the book, so as to help readers quickly understand the core ideas and conclusions of the book.

[0142] ##Principles:

[0143] -The output content must be organized according to the given format and cannot deviate from the framework requirements

[0144] -Only 3 views will be output

[0145] -Only the content in the knowledge base will be output. Books that are not in the knowledge base will be directly told to users that they do not understand

[0146] -Good at using expressions

[0147] - Proficient in using A grammar to generate structured text

[0148] ##Workflow:

[0149] 1. User provides the title of the book

[0150] 2. Based on the information provided by the user, generate reading notes in format A that conform to the following framework:

[0151] ===

[0152] -Book: <Book Name>

[0153] -Author: <Author Name>

[0154] - Time: <Publish Time>

[0155] -Question: <The core question this book attempts to answer>

[0156] -Summary: <100 words summarizing the core ideas of the book>

[0157] ###Viewpoint

[0158] <Viewpoint Description>

[0159] ###Golden Quotes

[0160] <Output three golden sentences related to the viewpoint>

[0161] ###Case

[0162] <Examples related to the viewpoint, output multiple cases, each case no less than 50 words>

[0163] ===

[0164] Among them, roles and introductions can correspond to contexts, goals can correspond to instructions, and principles and workflows can correspond to constraints. The constraints in the above example include thought chains and output templates.

[0165] Users can ask questions using natural language through the human-computer interface. The device then feeds the natural language input into a large model, which then infers or analyzes the input. The large model then displays the inference or analysis results to the user through the device's human-computer interface, effectively answering the user's question. Sometimes, large models struggle to directly provide correct answers to user questions, especially complex ones. Using prompts can unlock the potential of large models, enabling them to better complete tasks. Prompts have become a common method for large models to handle tasks.

[0166] Figure 2 shows an example of using and not using prompt to assist a large model. In Figure 2, the large model is an LLM.

[0167] Figure 2 (a) shows the case where prompt is not introduced, as follows:

[0168] User input: Blue sky, white clouds, and rainbows belong to Category A, computers, tables, and bridges belong to Category B, so which category does a volcano belong to?

[0169] LLM's answer: In this example the volcano is not classified, so it is impossible to determine which category the volcano belongs to.

[0170] Among them, without the introduction of prompt, LLM only simply associates the text. For unassociated content, LLM cannot find a direct correspondence.

[0171] Figure 2(b) shows the introduction of prompts. Prompts can guide LLM to complete tasks based on user input, as follows:

[0172] User input: Blue sky, white clouds, and rainbows belong to Category A, computers, tables, and bridges belong to Category B, so which category does a volcano belong to?

[0173] Prompt: 1. Analyze the common characteristics of the examples in category A; 2. Analyze the common characteristics of the examples in category B; 3. Which category does the volcano better fit? 4. Determine which category the volcano belongs to.

[0174] LLM's answer: Based on the analysis, the volcano is classified as Category A.

[0175] Among them, when prompt is introduced, LLM can analyze step by step according to the prompt to get the final answer, namely:

[0176] 1. Analyze the commonalities of Category A: Examples in Category A are all related to weather and natural phenomena;

[0177] 2. Analyze the commonalities of Category B: The examples in Category B are all man-made objects;

[0178] 3. Volcanoes are natural geological phenomena and are closer to Class A;

[0179] 4. Volcanoes belong to Class A.

[0180] As can be seen from the above examples, combining prompts with user input can give LLM stronger analytical capabilities and help LLM output results that satisfy users.

[0181] The above describes the relevant terms involved in the embodiments of the present application. The following describes the system architecture applicable to the embodiments of the present application in conjunction with Figures 3 and 4.

[0182] FIG3 is a schematic structural diagram of a system architecture applicable to an embodiment of the present application.

[0183] The system architecture shown in FIG3 includes a network management service consumption entity 210 and a network management service production entity 220 .

[0184] 1) Network management service consumption entity 210

[0185] The caller or entity that invokes the network management service is called a network management service consumer. Network management service consumer 210 can invoke management services, or intent services, to manage the network system. For example, it can create, modify, delete, and query intent expressions. For example, network management service consumer 210 can instantiate an intent template, set the intent IOC, generate an intent expression, and send the intent expression to the network management service producer 220 to implement expectations for the network system.

[0186] The network management service consuming entity 210 may also be referred to as a network management service consumer, an intent consumer, an intent owner, a management service consumer (MnS consumer), a network management service consuming network element, an intent management service consumer, or an intent service consumer. In future communication systems, the network management service consuming entity may also have other names, which are not specifically limited in this application. For ease of description, the network management service consuming entity will be referred to as a consuming entity hereinafter.

[0187] The capabilities or functions of a consumer entity can be deployed on a network element, referred to as a network management service consumer network element. The capabilities or functions of a consumer entity can also be deployed on other devices, which is not limited in this embodiment of the present application. For ease of description, this embodiment of the present application uses a consumer entity as an example, but any other device that deploys the capabilities or functions of a consumer entity can be substituted.

[0188] 2) Network Management Service Producer 220: The entity that provides network management services is called a network management service producer. Network Management Service Producer 220 can be used to provide network management services. For example, Network Management Service Producer 220 can receive intent expressions from Network Management Service Consumer 210 and perform intent translation, intent management, and intent execution and maintenance based on the intent expressions.

[0189] The network management service production entity 220 may also be referred to as a network management service providing entity, a network management provider, a management service provider (MnS provider), an intent provider, an intent handler, a network management service production network element, an intent management service provider, an intent service provider, or a network management service producer. In future communication systems, the network management service production entity may also have other names, which are not specifically limited in this application. For ease of description, the network management service production entity will be referred to as a production entity below.

[0190] The capabilities or functions of a production entity can be deployed on a network element, referred to as a network management service production network element. The capabilities or functions of a production entity can also be deployed on other devices, which is not limited in this embodiment of the present application. For ease of description, this embodiment of the present application uses a production entity as an example, but any other device that deploys the capabilities or functions of a production entity can be substituted.

[0191] The consumption entity and / or production entity can be a management function, or a management function entity, or a management entity, or a network device, or a network element, etc., and this application does not limit this.

[0192] The above-mentioned entity can be a network element in a hardware device, a software function running on dedicated hardware, or a virtualized function instantiated on a platform (for example, a cloud platform). It is understandable that the above-mentioned entity can be implemented by one device, or it can be implemented by multiple devices together. In addition, the above-mentioned entity can also be a functional module within a system, such as a network management system (NMS) or an equipment management system (EMS), or the above-mentioned entity can also be a functional module within a device, such as one or more functional modules within a network equipment (NE), and the network equipment can be an access network device or a core network device, etc.

[0193] As an example, the consumption entity may be deployed in an NMS, and the production entity may be deployed in a different EMS. The interaction between the consumption entity and the production entity may be performed through information exchange through the interface between the NMS and the EMS.

[0194] As another example, the consumption entity may be deployed in the EMS, and the production entity may be deployed in different NEs. The interaction between the consumption entity and the production entity may be carried out through information exchange through the interface between the EMS and the NE.

[0195] The solution of the present application can also be applied to other systems including corresponding entities, and the present application does not limit this.

[0196] FIG4 is another schematic structural diagram of a system architecture applicable to an embodiment of the present application.

[0197] The system architecture shown in Figure 4 includes consumer entities, production entities, and a large model.

[0198] The big model can be deployed within the consuming entity as shown in Figure 4, or it can be deployed outside the consuming entity (such as on a cloud platform, etc.), without limitation. The consuming entity can include a big model application framework, which can include units such as a prompt generator, memory, and an intent tool. The prompt generator can include a prompt scheduler and an intent prompt assembly module. The big model application framework in the consuming entity can process northbound natural language, identify intent content, guide the big model to generate appropriate structured intents and API call instructions, and send the received structured intents and API call instructions to the producing entity. Specifically, the big model application framework can be used to: determine the role type played by the big model, compose prompt instances based on built-in prompt templates, assist users in interacting with the big model, and interact with the producing entity. The producing entity can include intent processing functions that can receive and execute structured intents sent by the consuming entity.

[0199] It should be noted that the large model application framework can also be replaced with other names, such as AI module, AI agent, prompt unit or prompt engine. It should also be noted that the large model application framework can also be deployed in other locations, such as other devices or cloud platforms. It should also be noted that the division of the various functions or units of the large model application framework shown in Figure 4 is only exemplary and does not limit the embodiments of this application.

[0200] The following describes the process of the big model application framework processing natural language intent into structured intent based on the big model in conjunction with Figure 5.

[0201] Figure 5 is a schematic flowchart of the big model application framework processing natural language intent into structured intent based on the big model.

[0202] Figure 5 uses the LLM as the large model and the LLM as an example, with the LLM configured outside the consuming entity. The LLM application framework processes natural language intents into structured intents based on the LLM, primarily through two key steps: generating an intent prompt instance (as shown in steps 501 to 509 below), and generating a structured intent and API call instruction (as shown in steps 510 to 512 below).

[0203] Step 501: The LLM application framework receives natural language sent (or input) by a user.

[0204] Exemplarily, the LLM application framework receives a message sent (or input) by a user stating that "the average downlink throughput of area A is guaranteed to be no less than 20 Mbps."

[0205] In step 502 , the prompt scheduler in the LLM application framework assembles a schedule prompt instance according to the natural language input by the user and a schedule prompt template.

[0206] The scheduling prompt instance is used to assist or prompt the LLM to communicate with the user to determine the type of role the LLM needs to play.

[0207] Step 503: The LLM application framework sends the scheduling prompt instance to the LLM.

[0208] Accordingly, the LLM receives the scheduling prompt instance sent by the LLM application framework.

[0209] In step 504 , the LLM uses natural language to interact with the user based on the scheduling prompt instance, assisting the LLM application framework in determining the user's purpose, determining the user's operation intention, and clarifying the type of role that the LLM needs to play subsequently.

[0210] The role types that the LLM needs to play can include common chat roles and / or intent action roles. Intent action roles can include at least one of the following: intent_create role, intent_modify role, intent_delete role, intent_query role, or intent_report role.

[0211] The following uses the role type that the LLM needs to play as an example to illustrate the creation of a role with intent.

[0212] It should be noted that the communication between LLM and users goes through the LLM application framework because users are not aware of the existence of LLM.

[0213] In step 505 , the prompt scheduler in the LLM application framework initializes (or obtains) the intent prompt template according to the role type determined in step 504 .

[0214] For example, when the role type is an intention to create a role, the prompt scheduler in the LLM application framework matches and obtains the intention to create a prompt template. An example of an intention to create a prompt template is as follows:

[0215] Role: Intent Action

[0216] Goal: Intent Creation

[0217] Context:

[0218] 1. Status information of production entities;

[0219] 2. Intent model description;

[0220] 3. Intent API description;

[0221] 4. Description of the intention scenario;

[0222] constraint:

[0223] (COT):

[0224] Step 1: Confirm the intended scenario;

[0225] Step 2: Extract key elements of intent based on communication;

[0226] Step 3: Generate structured intent + examples;

[0227] Step 4: Generate API call instructions + examples.

[0228] Step 506: The prompt scheduler in the LLM application framework sends the intent prompt template to the intent prompt assembly module.

[0229] Accordingly, the intent prompt assembly module receives the intent prompt template from the prompt dispatcher.

[0230] Step 507: The intent prompt assembly module obtains information for filling in the intent prompt template from the memory.

[0231] Among them, the information used to fill in the intent prompt template may include at least one of the following information: user customized information, intent template, intent API information, status information of the production entity, or intent scenario information. Among them, the user customized information may include the user's historical conversations and / or intent identification list, etc. The intent template, intent API and intent scenario may be defined in the standard protocol. For example, the intent template includes at least one of the following templates: intent creation template, intent modification template, intent deletion template, intent query template, or intent report template, etc. For example, the intent API includes at least one of the following: intent creation API, intent modification API, intent deletion API, intent query API, or intent report API, etc. For example, the intent scenario includes at least one of the following: service activation scenario or service assurance scenario, etc.

[0232] In conjunction with the example in step 505 , the information used to fill in the intent prompt template may include at least one of the following information: status information of a production entity, intent template, intent API information, or intent scenario information.

[0233] Step 508: The intent prompt assembly module assembles the intent prompt instance according to the acquired information.

[0234] In conjunction with the example in step 505, an example of an intent prompt instance is as follows:

[0235] prompt={

[0236] {"role":"intent-create","context":"{role description}"}

[0237] {"role":"user","context":"Ensure the average downlink throughput of area A is no less than 20Mbps"}

[0238] }

[0239] Role Description:

[0240] #Role: Intent Creation Assistant#Instruct

[0241] ##Profile:

[0242] - Language: Chinese

[0243] -Description: I am an AI assistant role that assists in building intent creation requests

[0244] ##Goals:

[0245] By communicating with users, we extract the key elements of user creation intent, build structured intents and trigger intent creation API calls.

[0246] #Context

[0247] 1. Status information of the current production entity:

[0248] 1) Supports processing area A; 2) Supports LTR and NR formats

[0249] 2. Structured intent template: includes key elements: expectation, object, target, context, etc.<TS28312_IntentNrm.yaml>

[0250] 3. List of Intent Scenarios: ...<TS28312_IntentExpectations.yaml>

[0251] 4. Intent API List: ...<TS28532_ProvMns.yaml>

[0252] #Contraint

[0253] ##Workflow:

[0254] You will follow the steps below to extract information and generate a structured intent, and then send an intent creation request to the production entity by calling the intent creation API:

[0255] Step 1: Communicate with the user to confirm the intended scenario (such as service assurance scenario, service activation scenario, etc.)

[0256] Step 2: Communicate with the user to extract the key elements of the intent in the scenario

[0257] Step 3: Compose an intent instance using format #1

[0258] Example: {“userLabel”: <content>, “intentContext”: <content>, “intentExpectations”: [“expectationVerb”: <content>, {“expectationObjuct”: <…>, “expectationTargets”: <…>}]}

[0259] Step 4: Use the intent instance in format #1 as input parameter to generate the API call instruction for intent creation and send it to the production entity in format #1

[0260] For example:

[0261] Step 509: The intent prompt assembly module sends the intent prompt instance to the LLM.

[0262] Accordingly, the LLM receives the intent prompt instance from the intent prompt assembly module.

[0263] In step 510 , LLM uses natural language to interact with the user based on the intent prompt instance and the thought chain in the intent prompt instance, to assist the LLM application framework in confirming the key element information required to constitute the structured intent.

[0264] Among them, the key element information may include at least one of the following information: user label (userLabel), intent context (intentContext) or intent expectation (intentExpectations). Among them, the intent expectation may include at least one of the following information: expectation ID (expectationID), expectation type (expectationVerb), expectation object (expectationObject), expectation target (expectationObject), or expectation context (expectationTargets). The intent target may include at least one of the following information: target type (objectType), target instance (objectInstance), or target context (objectContext).

[0265] In step 511 , the LLM determines the structured intent and API call instructions based on the key element information.

[0266] Among them, intent templates and API information can be provided through intent prompt instances.

[0267] Combining the examples in step 501 and step 505, an example of structured intent is as follows:

[0268] In combination with the examples in step 501 and step 505, an example of an intent API call instruction is as follows, where the intent instance refers back to the structured intent described above:

[0269] Step 512: LLM sends the structured intent and API call instructions to the intent tool in the LLM application framework.

[0270] Accordingly, the intent tool receives structured intent and API call instructions from the LLM.

[0271] In step 513 , the LLM application framework sends the structured intent and API call instructions to the production entity through the intent tool.

[0272] Accordingly, the production entity receives structured intent and API call instructions from the LLM application framework.

[0273] In step 514, the production entity completes intent processing according to the received structured intent and API call instruction, and feeds back the intent processing result to the LLM application framework through the intent tool in the LLM application framework.

[0274] In step 515 , the LLM application framework and the LLM communicate with the user based on natural language regarding the processing results of the intended process and archive relevant information.

[0275] In this way, based on the above process, the consumer entity completes the process of converting natural language intent into structured intent and the related API call process based on LLM and prompt.

[0276] From the above process, we can see that when the LLM application framework processes natural language intent into structured intent:

[0277] 1) The LLM application framework obtains information from the memory to fill the intent prompt template. The information in the memory is statically configured. In this case, the information in the memory may not be sufficient to reflect the current situation (such as the current status of the production entity).

[0278] 2) Although different intent prompt templates can correspond to different intent scenarios, because LLM does not have real-time information on the target side (i.e., the production entity side), when the production entity processes the intent according to the corresponding structured intent and intent API call instructions, the intent operation may fail due to failure to pass the feasibility check, affecting operational efficiency.

[0279] For example, a user sends a natural language intent "guaranteeing that the average downlink throughput in area A is no less than 20 Mbps." Based on the process shown in Figure 5, LLM generates a structured intent #1 and sends it to the production entity through the consumer entity. The expected objects of structured intent #1 include TAC 8, TAC 9, and TAC 10 in area A, and the expected target is an average downlink throughput of no less than 20 Mbps. In fact, the cell of TAC 9 in area A has been out of service, and intent #1 conflicts with the already activated intent #2, and intent #2 has a higher priority than intent #1. In this case, intent #1 will fail to be created because the feasibility check fails.

[0280] In response to the above problems, the present application provides an intent management method and a communication device, in order to reduce the probability of failure of intent feasibility check, thereby improving the efficiency of intent operation, and further improving the efficiency of intent management and user experience.

[0281] The following describes the method embodiments of the present application.

[0282] FIG6 is a schematic flowchart of an intent management method 600 provided in this application.

[0283] The method shown in Figure 6 can be performed by the first device, the second device and the first production entity. Unless otherwise specified, the "first device", "second device" or "first production entity" may refer to the first device, the second device or the first production entity itself, or may refer to a device that can support the first device, the second device or the first production entity to perform its functions.

[0284] Among them, the first device is a device deployed or provided with a large model application framework. The second device is a device deployed or provided with a large model. The first device and the second device can be set in the same entity, that is, the first device and the second device are modules or logical devices deployed in the same entity, such as the first device and the second device are different modules in the consumer entity or the second production entity respectively. The first device and the second device can also be set in different entities, such as the first device is set in the consumer entity or the second production entity, and the second device is set in other entities (such as a third-party cloud platform, etc.), without limitation.

[0285] Method 600 includes at least part of the following.

[0286] Step 601: A first device sends first information to a second device.

[0287] Accordingly, the second device receives the first information from the first device.

[0288] The first information may be used to trigger acquisition of information related to the intention from the first production entity; alternatively, the first information may be used to trigger updating of information related to the intention.

[0289] Among them, the intent can be various types of intent. The type of intent can be related to the goal contained in the intent. For example, the intent that contains goals related to image processing is an image processing intent. For example, the intent that contains goals related to network management is a network management intent. For example, the goal of the intent includes the performance indicator value reaching the corresponding performance indicator range. When the performance indicator is a throughput indicator, such as the average downlink throughput is greater than or equal to 5Mbps, the intent can be called a throughput intent; when the performance indicator is an energy-saving indicator, such as energy saving is greater than or equal to 10 kilowatts, the intent can be called an energy-saving intent.

[0290] Information related to intent can be understood as information related to intent parameters, where the intent parameters can be the various parameters included in the intent template. Taking network management intent as an example, intent parameters can include at least one of the following information: user tag, intent context, or intent expectation. The intent expectation can include at least one of the following information: expectation ID, expectation type, expectation object, expectation target, or expectation context. The intent target can include at least one of the following information: target type, target instance, or target context.

[0291] Specifically, the information related to the intended parameters may be information stored, acquired, and maintained by the first production entity. For example, if the first intended parameter is area A, the information related to the intended parameters may be the cells in service within area A stored or maintained by the first production entity. The term "intent attribute" may also be understood as "intent information element," "intent field," or the like, without limitation.

[0292] Specifically, the information related to the parameters of the intention may include: the value or state value of the parameters of the intention, and / or, information about other intentions that conflict with the value of the parameters of the intention. Taking network management intention as an example, the information related to the parameters of the intention may include: the value or state value of the parameters in the "expected object", "expected target" and "intention context" contained in the intention instance, as well as information about other intentions that may conflict with the value of the parameters in the "expected object", "expected target" and "intention context" contained in the intention instance. For example, for the network management intention of "ensuring that the switching failure rate of area A is <2% when the load is >80%, the parameters of the intention may include the expected type "performance guarantee type" and the expected object "area A", and the information related to the intention may be information related to the "performance guarantee type" and "area A". For example, the information related to "area A" may include two types of information, one type is the cells that are still in service in area A, and the second type is information about other intentions in area A (such as intentions in area A that conflict with the current intention, or the priority of intentions that already exist in area A, etc.). For another example, information related to the “performance guarantee type” may include whether there is an energy-saving type intention.

[0293] The embodiments of the present application do not limit the implementation method of the first information. Exemplarily, the first information includes a first prompt instance, and the thought chain of the first prompt instance includes steps for triggering the acquisition of information related to the intention from the first production entity. In some implementations, the thought chain of the first prompt instance may also include other steps, such as steps for triggering the second device to confirm the intention scenario, steps for triggering the second device to determine the parameters of the intention, steps for triggering the second device to generate a structured intention, or steps for triggering the second device to generate an API call instruction, etc. It should be noted that the step for triggering the acquisition of information related to the intention from the first production entity can be integrated with other steps, or it can be a separate step independent of other steps, and is not limited.

[0294] It should be noted that in step 601, the first device prompts, guides or triggers the second device through the first information to perform the operation of obtaining information related to the intention from the first production entity, and the specific information to be obtained from the first production entity is determined or decided by the second device.

[0295] Step 602: The second device sends second information to the first device based on the first information.

[0296] Accordingly, the first device receives the second information from the second device.

[0297] Specifically, in step 602, the second device may determine the second information based on the first information, and then send the second information to the first device.

[0298] The second information may include a first parameter of the intention. The first parameter may include one or more parameters, which may be parameters of the intention to be updated determined or identified by the second device. The values ​​of these parameters in the second device may be different from the values ​​stored or maintained by the first production entity. For example, the second device determines that the cells in service in area A are TAC 8, TAC 9, and TAC 10, but the information actually stored or maintained by the first production entity is: the cells in service in area A are TAC 8 and TAC 9.

[0299] Specifically, the first parameter may be all or part of the parameters of the intent. For example, when the intent is an intent related to network management, the first parameter includes at least one of the following parameters: an expected object, an expected target, or an expected context. The description of the expected object, the expected target, and the expected context can refer to the term introduction section and will not be described in detail. For example, for the network management intent of "ensuring that the handover failure rate of area A is <2% when the load is >80%", the cells in service in area A need to be updated, so the first parameter may include area A.

[0300] The embodiments of the present application do not limit the implementation manner in which the second device determines the second information based on the first information. In one possible implementation manner, the second device may determine the first parameter based on the first information, and determine the second information based on the first parameter. Exemplarily, when the first information includes a first prompt instance and the thought chain of the first prompt instance includes a step for triggering the acquisition of information related to the intention from the first production entity, the second device responds to the step in the thought chain of the first prompt instance for triggering the acquisition of information related to the intention from the first production entity, determines the first parameter of the intention, and then determines the second information based on the first parameter.

[0301] The embodiments of the present application do not limit the implementation manner in which the second device determines the first parameter of the intention.

[0302] One possible implementation method is that in the intent creation scenario, the first information is also used to trigger the second device to generate a structured intent to be created. For example, the thought chain of the first prompt instance also includes the step of triggering the second device to generate a structured intent to be created. The second device can determine the first parameter of the intent based on the first information by interacting with the user through the first device.

[0303] Another possible implementation method is that in a scenario where the structured intent already exists in the first device (such as an intent modification scenario, an intent query scenario, or an intent deletion scenario, etc.), the first device can read the parameters of the intent from the memory and provide it to the second device through the first information. The second device can obtain the parameters of the intent from the first information and determine the first parameter from the obtained parameters. For example, in the intent modification scenario, the first information includes a first prompt instance, and the thought chain of the first prompt instance also includes the step of triggering the second device to modify the structured intent. The parameters of the intent can be carried in the context of the first prompt instance, or in the thought chain of the first prompt instance. The second device can obtain the parameters of the intent from the first prompt instance and determine the first parameter from the obtained parameters.

[0304] In other embodiments, the second information may further include API information, where the API information may be used to indicate the API to be called to obtain the fourth information related to the first parameter of the intent. For example, the API information may be an API index or number, or an API call instruction. In other words, the second device determines the information to be obtained and the API to be called to obtain the information.

[0305] Step 603: The first device sends third information to the first production entity based on the second information.

[0306] Accordingly, the first production entity receives third information from the first device.

[0307] The third information may be used to obtain fourth information related to the first parameter of the intention.

[0308] Exemplarily, the third information includes the first parameter of the intent. For example, the second device may use the first parameter of the intent as an input parameter for calling an API. It should be noted that when the second information does not include API information, the second device may determine the API called to obtain the fourth information. When the second information includes API information, the second device may call the API indicated by the API information to obtain the fourth information, that is, the third information may be used to call the API indicated by the API information to obtain the fourth information. For example, the third information may include an API call instruction, the input parameter of the API call instruction is the first parameter, and the API is the API indicated by the API information. Taking the API as prompt_fetch as an example, the third information may be as follows:

[0309] Step 604: The first production entity sends fourth information to the first device.

[0310] Accordingly, the first device receives fourth information from the first production entity.

[0311] Among them, the fourth information can be used to indicate information related to the first parameter of the intention. The description of the information related to the first parameter of the intention can refer to the information related to the parameter of the intention in step 601, which will not be described in detail.

[0312] Specifically, after receiving the third information, the first production entity queries and obtains information related to the first parameter of the intent based on the third information, and indicates it to the first device through the fourth information. For example, the first production entity queries and obtains information related to the first parameter of the intent based on the third information based on a specific policy or relying on an internal decision, and indicates it to the first device through the fourth information.

[0313] Exemplarily, the information related to the first parameter of the intention provided by the first production entity to the first device may be real-time information related to the first parameter of the intention in the first production entity. For example, the first parameter of the intention is area A, and the information related to area A may include: the cell with TAC 9 in area A has stopped service, and / or, information on other intentions activated in area A, where the information on other intentions activated in area A may include the priority of the intention.

[0314] Step 605: The first device obtains fifth information for management intention based on the fourth information.

[0315] Exemplarily, the fifth information may be an expression carrying the user's intention. For example, the fifth information may be a structured intention defined in a standard protocol.

[0316] The embodiments of this application do not limit the implementation method by which the first device obtains the fifth information for the management intent based on the fourth information. In one possible implementation method, the first device obtains the fifth information from the second device. For example, the first device sends the fourth information to the second device, and the second device receives the fourth information from the first device in response. The second device determines the fifth information based on the fourth information and the large model. The second device sends the fifth information to the first device, and the first device receives the fifth information from the second device in response.

[0317] Based on method 600, a first device can trigger a second device to obtain information related to an intent. The second device then provides the first parameter of the intent to the first device, allowing the first device to obtain information related to the first parameter of the intent from the first production entity. Furthermore, based on the information related to the first parameter of the intent from the first production entity, the first device can obtain information for managing the intent. In this way, method 600 can manage intents, such as creating or modifying intents, based on real-time information from the first production entity. This helps reduce the probability of intent feasibility check failures, thereby improving intent operation efficiency, and ultimately improving intent management efficiency and user experience.

[0318] In other embodiments of the present application, method 600 further includes: the first device receives sixth information input by the user, wherein the sixth information can be used to indicate an operation related to the intent; and the first device determines the first information based on the sixth information. Exemplarily, the sixth information can be natural language information, and the user can input the sixth information to the first device through the human-computer interaction interface of the first device. In other words, the first device can determine that the user is performing an operation related to the intent based on the user's input information, and then determine the first information. When the first information includes a first prompt instance, the first device can determine the type of role played by the large model in the second device based on the natural language input by the user, and then obtain the intent prompt template based on the role type, and then assemble the first intent prompt instance.

[0319] In other embodiments of the present application, method 600 further includes: the first device manages the intent according to the fifth information for managing the intent. Exemplarily, the first device performs an intent operation according to the fifth information, wherein the intent operation includes at least one of the following: requesting the first production entity to create an intent, requesting the first production entity to modify the intent, requesting the first production entity to delete the intent, querying the first production entity for the intent, abandoning the creation of the intent, or abandoning the modification of the intent. In the case where an intent that conflicts with the current intent has been activated in the first production entity, if the first device still requests the first production entity to create the current intent or modify the current intent according to the fifth information, the first device may also request to delete or suspend the intent that conflicts with the current intent.

[0320] In other embodiments of the present application, method 600 further includes: the first device determines the second prompt instance based on the fourth information, wherein the context information of the second prompt instance includes the fourth information. Step 604 specifically includes: the first device sends the second prompt instance to the second device. In other words, the first device can send the fourth information to the second device by sending the second prompt instance to the second device. It should be noted that the first device can update the context of the first prompt instance based on the fourth information to obtain the second prompt instance, or assemble or generate a new second prompt instance based on the fourth information, without limitation. In this embodiment, the second device can determine the fifth information for management intent based on the second prompt instance and the large model.

[0321] In other embodiments of the present application, the thought chain of the second prompt instance may include steps for triggering the second device to generate management suggestions for the intent based on the fourth information, that is, the first device may prompt, guide or trigger the second device to generate management suggestions for the intent based on the fourth information. Subsequently, the second device may return management suggestions for the intent. For example, when the second device determines that the value of the first parameter of the intent determined in step 602 is inappropriate based on the fourth information, the second device may provide management suggestions for the intent. The second device may also not return management suggestions for the intent. For example, when the second device determines that the value of the first parameter of the intent determined in step 602 is appropriate based on the fourth information, the second device may directly return the fifth information without returning management suggestions for the intent.

[0322] For the case where the second device returns a management suggestion for the intent, method 600 further includes: the second device generates a management suggestion for the intent based on the second prompt instance; the second device sends the seventh information to the first device, and accordingly, the first device receives the seventh information from the second device, wherein the seventh information can be used to indicate the management suggestion for the intent; and the first device outputs the seventh information to the user. Exemplarily, the first device outputting the seventh information to the user can be the first device displaying the seventh information to the user through a human-computer interaction interface. In this case, the second device can output the management suggestion for the intent to the user through the first device, thereby helping the user to issue the correct operation information and complete the corresponding intended operation, which can effectively avoid incorrect intended operations.

[0323] Exemplarily, the above-mentioned management suggestions may include the results and reasons for the failure of the intent operation, and the reasons for the failure of the intent operation may indirectly indicate the management suggestions. For example, the user sends a natural language intent "to ensure that the average downlink throughput of area A is not less than 20Mbps", and LLM determines that the expected objects of intent #1 include TAC 8, TAC 9 and TAC 10 in area A. LLM obtains information related to area A on the production entity side through the LLM application framework, and learns that the cell with TAC 9 in area A has stopped service, intent #2 has been activated in area A, intent #2 conflicts with intent #1, and intent #2 has a higher priority than intent #1. In this case, the management suggestions provided by LLM to the user through the LLM application framework may be: intent #1 failed to be generated because the cell with TAC 9 in area A has stopped service, intent #2 has been activated in area A and conflicts with intent #1, and intent #2 has a higher priority.

[0324] Exemplarily, the above-mentioned management suggestions may include the results, reasons and suggestions for the failure of the intent operation. For example, the user sends a natural language intent "to ensure that the average downlink throughput of area A is not less than 20Mbps", and LLM determines that the expected objects of intent #1 include TAC 8, TAC 9 and TAC 10 in area A. LLM obtains information related to area A on the production entity side through the LLM application framework, and learns that the cell with TAC 9 in area A has stopped service, intent #2 has been activated in area A, intent #2 conflicts with intent #1, and the priority of intent #2 is higher than that of intent #1. In this case, the management suggestions provided by LLM to the user through the LLM application framework may be: intent #1 failed to be generated because the cell with TAC 9 in area A has stopped service, intent #2 has been activated in area A and conflicts with intent #1 and intent #2 has a higher priority, and it is recommended to re-specify the target area of ​​intent #1 and / or increase the priority of intent #1.

[0325] In other embodiments of the present application, method 600 further includes: the first device receives the eighth information input by the user and sends the eighth information to the second device, and accordingly, the second device receives the eighth information from the first device, wherein the eighth information is the user's feedback information regarding the management suggestion. Exemplarily, the feedback information may indicate whether the user receives the management suggestion of the second device or indicate an intention operation that is re-issued based on the management suggestion. Exemplarily, the user may input the eighth information to the first device through the human-computer interaction interface of the first device. In this case, the second device may determine the fifth information for the management intention based on the second prompt instance, the large model and the eighth information. Based on this embodiment, the second device may provide the first device with more reasonable information for the management intention, thereby helping to improve the efficiency of the intention operation, and thereby improve the management efficiency of the intention and the user experience.

[0326] The method 600 is described in detail below by taking the large model as LLM and the applicable scenario as the intention creation scenario as an example.

[0327] Figure 7 is a schematic flow chart of an intent management method 700 provided in this application. The MnS consumer and MnS producer in method 700 correspond to the consumer entity and the first production entity mentioned above, respectively. The LLM application framework and LLM correspond to the first device and the second device mentioned above, respectively. The intent prompt instance can correspond to the first information or the first prompt instance mentioned above. The key element information can correspond to the first parameter mentioned above. The intent prompt instance supplemented with real-time context information can correspond to the fourth information or the second prompt instance mentioned above. Figure 7 takes the LLM application framework and LLM settings at the consumer entity as an example.

[0328] Step 701: The LLM application framework in the MnS consumer receives natural language input from a user.

[0329] For example, the natural language input by the user is “guarantee that the average downlink throughput of area A is not less than 20 Mbps”.

[0330] Step 702: The MnS consumer entity determines the role type of the LLM.

[0331] In method 700 , the role type of the LLM is an intent-to-create role.

[0332] The implementation of step 702 may refer to steps 502 to 504 in FIG. 5 , and will not be described in detail.

[0333] Step 703: Assemble the intent prompt instance according to the role type of the LLM.

[0334] The implementation of step 703 may refer to steps 505 to 508 in Figure 5. Different from Figure 5, the thought chain of the intention prompt instance in step 703 includes the step of updating the prompt context.

[0335] For example, an example of the intent prompt instance in step 703 is as follows:

[0336] prompt={

[0337] {"role":"intent-create","context":"{role description}"}

[0338] {"role":"user","context":"Ensure the average downlink throughput of area A is no less than 20Mbps"}

[0339] }

[0340] Role Description:

[0341] #Role: Intent Creation Assistant#Instruct

[0342] ##Profile:

[0343] - Language: Chinese

[0344] -Description: I am an AI assistant role that assists in building intent creation requests

[0345] ##Goals:

[0346] By communicating with users, we extract the key elements of user creation intent, build structured intents and trigger intent creation API calls.

[0347] #Context

[0348] 1. Current MnS producer status information:

[0349] 1) Supports processing area A; 2) Supports LTR and NR formats

[0350] 2. Structured intent template: includes key elements: expectation, object, target, context, etc.<TS28312_IntentNrm.yaml>

[0351] 3. List of Intent Scenarios: ...<TS28312_IntentExpectations.yaml>

[0352] 4. Intent API List: ...<TS28532_ProvMns.yaml>

[0353] 5. Dynamic prompt information API list: ...<Prompt_fetch.yaml>

[0354] 6. Real-time context information (realtime_contextInfo): <content>

[0355] #Contraint

[0356] ##Workflow:

[0357] You will follow the steps below to extract information and generate a structured intent, and then send an intent creation request to the MnS producer by calling the intent creation API:

[0358] Step 1: Communicate with the user to confirm the intended scenario (such as service assurance scenario, service activation scenario, etc.)

[0359] Step 2: Communicate with the user to extract the key elements of the intent in the scenario. In this process, call the dynamic prompt information API to obtain the real-time context information on the MnS producer side, add the real-time context information on the MnS producer side to the intent prompt instance, and provide timely feedback to the user.

[0360] Step 3: Compose an intent instance using format #1

[0361] Example: {“userLabel”: <content>, “intentContext”: <content>, “intentExpectations”: [“expectationVerb”: <content>, {“expectationObjuct”: <…>, “expectationTargets”: <…>}]}

[0362] Step 4: Generate an API call instruction for intent creation using the intent instance in format #1 as input parameter and send it to the MnS producer in format #1

[0363] For example:

[0364] In the above example, the context of the intent prompt instance includes a list of dynamic prompt information APIs and real-time context information. The APIs in the dynamic prompt information API list are used to obtain real-time context information from the MnS producer. This real-time context information can be context information related to key elements on the MnS producer side. The definition of the steps for updating the prompt context is shown in step 2. That is, during the process of extracting the key elements of the intent, the dynamic prompt information API is called to obtain real-time context information from the MnS producer side. This real-time context information from the MnS producer side is then added to the intent prompt instance and promptly fed back to the user.

[0365] Step 704: The LLM application framework sends an intent prompt instance to the LLM.

[0366] Accordingly, the LLM receives intent prompt instances from the LLM application framework.

[0367] In step 705 , LLM uses natural language to interact with the user based on the intent prompt instance and the thought chain in the intent prompt instance, to assist the LLM application framework in confirming the key element information required to constitute the structured intent.

[0368] The implementation of step 705 may refer to step 510 in FIG. 5 .

[0369] In step 706 , the LLM obtains real-time context information related to the key element information from the MnS producer through the LLM application framework.

[0370] Specifically, after confirming the key element information required to constitute the structured intent, LLM can call the API in the prompt information API list (that is, the dynamic prompt information API of the MnS producer) through the LLM application framework according to the intent prompt instance to obtain real-time context information related to the key element information.

[0371] For example, the LLM application framework can call the APIs in the prompt information API list to obtain real-time context information related to the key element information during the process of determining the key element information. For example, during the process of determining the "expectation object", the MnS consumer calls the APIs in the prompt information API list to obtain the real-time context information from the MnS producer, including:

[0372] 1) The cell with TAC 9 in area A managed by the MnS producer has been taken out of service; and / or,

[0373] 2) Information about other intents that have been activated in area A.

[0374] The information of other activated intents in region A includes the priority of the intents. For example, the energy-saving intent of intent #2 in the MnS producer is activated: ... "Desired object": <region A>, "priority": <2> ….

[0375] Exemplarily, the real-time context information related to the key element information may be carried in a Realtime_ContextInfo information element.

[0376] In step 707 , the LLM application framework adds the acquired real-time context information to the intent prompt instance.

[0377] For example, the LLM application framework may add the following to the “realtime context information (realtime_contextInfo): <content>” in the intent prompt instance:

[0378] 1) The cell with TAC 9 in area A managed by the MnS producer has been taken out of service; and / or,

[0379] 2) Information about other intents activated in area A managed by the MnS producer.

[0380] In step 708 , the LLM application framework sends the intent prompt instance supplemented with real-time context information to the LLM.

[0381] In step 709, based on the intent prompt instance supplemented with real-time context information, the LLM uses natural language to interact with the user and gives the user reasonable intent instance generation suggestions to avoid unreasonable parameter settings that can only be discovered later.

[0382] Exemplarily, the intent instance generation suggestion may include: the cell with TAC 9 in area A cannot respond to the intent, and the target area information of the intent needs to be re-specified, and / or, the energy-saving intention of intent #2 has been activated, intent #2 conflicts with intent #1 to be created, and intent #2 has a higher priority, and the priority of intent #1 needs to be increased.

[0383] The above steps 705 to 709 may be executed in a loop.

[0384] Step 710: After completing the interaction with the user, the LLM determines the structured intent and API call instructions.

[0385] Step 711 : LLM sends structured intent and API call instructions to the LLM application framework.

[0386] Accordingly, the LLM application framework receives structured intent and API call instructions from the LLM.

[0387] Step 712: The LLM application framework sends a structured intent and an API call instruction to the MnS producer.

[0388] Accordingly, the MnS producer receives structured intent and API call instructions from the LLM application framework.

[0389] The implementation of steps 710 to 712 may refer to steps 511 to 513 in FIG. 5 .

[0390] In step 713, the MnS producer completes intent processing according to the received structured intent and API call instruction.

[0391] Intent processing may include creating and configuring an intent managed object instance (MOI), and executing the intent.

[0392] In method 700, by adding a prompt context update step to the thought chain of the intent prompt instance, the assembled intent prompt instance can guide the LLM to call the MnS producer's API to obtain real-time context information from the MnS producer. This dynamically updates the context information in the intent prompt instance, thereby helping the user issue the correct intent creation information and complete the corresponding intent creation operation. This effectively avoids incorrect intent creation operations and increases the probability of successful intent creation.

[0393] The method 600 is described in detail below by taking the large model as LLM and the applicable scenario as the intention modification scenario as an example.

[0394] Figure 8 is a schematic flow chart of an intent management method 800 provided in this application. The MnS consumer and MnS producer in method 800 correspond to the consumer entity and the first production entity mentioned above, respectively. The LLM application framework and LLM correspond to the first device and the second device mentioned above, respectively. The intent prompt instance can correspond to the first information or the first prompt instance mentioned above. The key element information can correspond to the first parameter mentioned above. The intent prompt instance supplemented with real-time context information can correspond to the fourth information or the second prompt instance mentioned above. Figure 8 takes the LLM application framework and LLM settings at the consumer entity as an example.

[0395] Step 801: The LLM application framework in the MnS consumer receives natural language input from a user.

[0396] For example, the natural language input by the user is "the goal of modification intention #1 is to have a throughput of no less than 30 Mbps."

[0397] Step 802: The MnS consumer entity determines the role type of the LLM.

[0398] In method 800 , the role type of the LLM is an intent-to-modify role.

[0399] The implementation of step 802 may refer to steps 502 to 504 in FIG. 5 , and will not be described in detail.

[0400] Step 803: Assemble the intent prompt instance according to the role type of the LLM.

[0401] The implementation of step 803 can refer to steps 505 to 508 in Figure 5. Different from Figure 5, the thought chain of the intent prompt instance in step 803 includes the step of updating the prompt context, and since the intent instance has been created, the key element information of the intent can be directly read (such as read from the memory of the LLM application framework) and written into the intent prompt instance. Among them, the definition of the step of updating the prompt context can be: using the key element information of the intent as an input parameter, calling the dynamic prompt information API to obtain the real-time context information on the MnS producer side, supplementing the real-time context information on the MnS producer side to the intent prompt instance, and promptly feeding back to the user. The key element information read can be written into the context of the intent prompt instance, or written into the step description of the thought chain of the intent prompt instance, without limitation.

[0402] Step 804: The LLM application framework sends an intent prompt instance to the LLM.

[0403] Accordingly, the LLM receives intent prompt instances from the LLM application framework.

[0404] In step 805, based on the intent prompt instance and following the thought chain in the intent prompt instance, LLM uses the key element information of the intent as input parameters and calls the dynamic prompt information API through the LLM application framework to obtain real-time context information on the MnS producer side.

[0405] Exemplarily, the real-time context information related to the key element information may be carried in a Realtime_ContextInfo information element.

[0406] In step 806 , the LLM application framework adds the acquired real-time context information to the intent prompt instance.

[0407] Step 807 : The LLM application framework sends the intent prompt instance supplemented with real-time context information to the LLM.

[0408] In step 808, based on the intent prompt instance supplemented with real-time context information, the LLM uses natural language to interact with the user and gives the user reasonable intent instance modification suggestions to avoid unreasonable parameter settings that can only be discovered later.

[0409] Step 809: After completing the interaction with the user, the LLM determines the modified structured intent and API call instructions.

[0410] Step 810: The LLM sends the modified structured intent and API call instructions to the LLM application framework.

[0411] Accordingly, the LLM application framework receives the modified structured intent and API call instructions from the LLM.

[0412] Step 811: The LLM application framework sends the modified structured intent and API call instruction to the MnS producer.

[0413] Accordingly, the MnS producer receives the modified structured intent and API call instructions from the LLM application framework.

[0414] The implementation of steps 809 to 811 may refer to steps 511 to 513 in FIG. 5 .

[0415] In step 812, the MnS producer completes intent processing according to the modified structured intent and API call instruction.

[0416] Intent processing may include modifying and configuring the intent MOI, and executing the modified intent.

[0417] In method 800, by adding a prompt context update step to the thought chain of the intent prompt instance, the assembled intent prompt instance can guide the LLM to call the MnS producer's API to obtain real-time context information from the MnS producer. This dynamically updates the context information in the intent prompt instance, helping the user issue the correct intent modification information and complete the corresponding intent modification operation. This effectively avoids incorrect intent modification operations and increases the probability of successful intent modification.

[0418] The above describes in detail the method embodiment provided by the present application in conjunction with Figures 1 to 8 , and the following describes the device embodiment of the present application in conjunction with Figures 9 to 11 .

[0419] It is understood that, in order to implement the functions in the above embodiments, the apparatuses in Figures 9 to 11 include hardware structures and / or software modules corresponding to the functions. Those skilled in the art should readily appreciate that, in conjunction with the various exemplary units and method steps described in the embodiments disclosed herein, the present application can be implemented in the form of hardware or a combination of hardware and computer software.

[0420] Figures 9 and 10 are schematic diagrams of possible apparatuses provided in embodiments of the present application. These apparatuses can be used to implement the functions of the first apparatus, the second apparatus, or the first production entity in the above method embodiments, thereby also achieving the beneficial effects of the above method embodiments.

[0421] As shown in FIG9 , the device 10 includes a transceiver unit 11 and a processing unit 12 .

[0422] When the apparatus 10 is used to implement the functions of the first apparatus in the above method embodiments, the transceiver unit 11 is used to execute the transceiver steps of the first apparatus, such as steps 601, 602, 603, 604, 701, 704, 706, 708, 711, 712, 801, 804, 807, 810, or 811, and the processing unit 12 is used to execute the processing steps of the first apparatus, such as steps 605, 702, 703, 707, 802, 803, or 806. When the apparatus 10 is used to implement the functions of the second apparatus in the above method embodiments, the transceiver unit 11 is used to execute the transceiver steps of the second apparatus, such as steps 601, 602, 704, 705, 708, 709, 711, 804, 807, 808, or 810, and the processing unit 12 is used to execute the processing steps of the second apparatus, such as steps 702, 710, 802, or 809. When the device 10 is used to implement the functions of the first production entity in the above-mentioned method embodiments, the transceiver unit 11 is used to execute the transceiver steps of the first production entity, such as steps 603, 604, 706, 712, 805 or 811, and the processing unit 12 is used to execute the processing steps of the first production entity, such as step 713 or 812.

[0423] For a more detailed description of the transceiver unit 11 and the processing unit 12 , please refer to the relevant description in the above method embodiment, which will not be described again here.

[0424] As shown in FIG10 , the apparatus 20 includes a processor 21. The processor 21 is coupled to a memory 23, which is used to store instructions. When the apparatus 20 is used to implement the method described above, the processor 21 is used to execute the instructions in the memory 23 to implement the functions of the processing unit 12 described above.

[0425] Optionally, the device 20 further includes a memory 23 .

[0426] Optionally, the apparatus 20 further includes an interface circuit 22. The processor 21 and the interface circuit 22 are coupled to each other. It will be appreciated that the interface circuit 22 may be a transceiver or an input / output interface. When the apparatus 20 is used to implement the method described above, the processor 21 is configured to execute instructions to implement the functions of the processing unit 12, and the interface circuit 22 is configured to implement the functions of the transceiver unit 11.

[0427] Exemplarily, when device 20 is a chip applied to a first device, a second device, or a first production entity, the chip implements the functions of the first device, the second device, or the first production entity in the above-mentioned method embodiment. The chip receives information from other modules (such as a radio frequency module or antenna) in the first device, the second device, or the first production entity, where the information is sent by the other device to the first device, the second device, or the first production entity; or the chip sends information to other modules (such as a radio frequency module or antenna) in the first device, the second device, or the first production entity, where the information is sent by the first device, the second device, or the first production entity to the other device.

[0428] 11 is a schematic diagram of a chip system 30 according to an embodiment of the present application. The chip system 30 (or also referred to as a processing system) includes a logic circuit 31 and an input / output interface 32.

[0429] The logic circuit 31 may be a processing circuit in the chip system 30. The logic circuit 31 may be coupled to a storage unit and call instructions in the storage unit so that the chip system 30 can implement the methods and functions of the various embodiments of the present application. The input / output interface 32 may be an input / output circuit in the chip system 30, outputting information processed by the chip system 30 or inputting data or signaling information to be processed into the chip system 30 for processing.

[0430] As a solution, the chip system 30 is used to implement the operations performed by the first device, the second device or the first production entity in the above various method embodiments.

[0431] For example, the logic circuit 31 is used to implement the processing-related operations performed by the first device, the second device or the first production entity in the above method embodiments; the input / output interface 32 is used to implement the sending and / or receiving-related operations performed by the first device, the second device or the first production entity in the above method embodiments.

[0432] The present application also provides a communication device, comprising a processor coupled to a memory, the memory being configured to store computer programs or instructions and / or data, the processor being configured to execute the computer programs or instructions stored in the memory, or to read data stored in the memory, to perform the methods described in the above method embodiments. Optionally, there are one or more processors. Optionally, the communication device includes a memory. Optionally, there are one or more memories. Optionally, the memory is integrated with the processor or provided separately.

[0433] The present application also provides a chip, including a processor, which is coupled to a memory, the memory is used to store computer programs or instructions, and the processor is used to execute the computer programs or instructions stored in the memory to implement the methods performed by the first device, the second device or the first production entity in the above-mentioned method embodiments.

[0434] The present application also provides a computer-readable storage medium storing computer instructions for implementing the methods executed by the first device, the second device or the first production entity in the above-mentioned method embodiments.

[0435] The present application also provides a computer program product, comprising a computer program or instructions, which, when executed by a computer, implement the methods performed by the first device, the second device or the first production entity in the above-mentioned method embodiments.

[0436] The present application also provides a communication system, which includes at least one of the first device, the second device, or the first production entity in the above embodiments.

[0437] The explanation of the relevant contents and beneficial effects of any of the above-mentioned devices can be referred to the corresponding method embodiments provided above, which will not be repeated here.

[0438] It is understood that the processor in the embodiments of the present application may be a central processing unit (CPU), or may be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field programmable gate arrays (FPGA), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. The general-purpose processor may be a microprocessor or any conventional processor.

[0439] The method steps in the embodiments of the present application can be implemented by hardware or by a processor executing software instructions. The software instructions can be composed of corresponding software modules, which can be stored in random access memory, flash memory, read-only memory, programmable read-only memory, erasable programmable read-only memory, electrically erasable programmable read-only memory, registers, hard disks, mobile hard disks, compact disc read-only memory (CD-ROM) or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor so that the processor can read information from the storage medium and write information to the storage medium. Of course, the storage medium can also be an integral part of the processor. The processor and the storage medium can be located in an ASIC. In addition, the ASIC can be located in the first device, the second device or the first production entity. Of course, the processor and the storage medium can also be present in the first device, the second device or the first production entity as discrete components.

[0440] In the above embodiments, all or part of the embodiments may be implemented using software, hardware, firmware, or any combination thereof. When implemented using software, all or part of the embodiments may be implemented in the form of a computer program product. The computer program product includes one or more computer programs or instructions. When the computer program or instructions are loaded and executed on a computer, the processes or functions described in the embodiments of the present application are performed in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, a network device, a user device, or other programmable device. The computer program or instructions may be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another. For example, the computer program or instructions may be transferred from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless means. The computer-readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server or data center that integrates one or more available media. The available medium may be a magnetic medium, such as a floppy disk, hard disk, or magnetic tape; an optical medium, such as a digital video disk; or a semiconductor medium, such as a solid-state drive.

[0441] In the various embodiments of the present application, unless otherwise specified or there is a logical conflict, the terms and / or descriptions between different embodiments are consistent and can be referenced by each other. The technical features in different embodiments can be combined to form new embodiments according to their inherent logical relationships.

[0442] Unless otherwise indicated, all technical and scientific terms used in the embodiments of the present application have the same meaning as those generally understood by those skilled in the art of the technical field of the application. The terms used in this application are only for the purpose of describing specific embodiments and are not intended to limit the scope of the application. It should be understood that the above are for illustration, and the examples above are only for helping those skilled in the art to understand the embodiments of the present application, rather than limiting the application embodiments to the specific numerical values ​​or specific scenarios illustrated. Those skilled in the art can obviously carry out various equivalent modifications or changes based on the examples given above, and such modifications and changes also fall within the scope of the embodiments of the present application.

Claims

1. An intention management method, characterized in that: The method comprises: The first device sends first information to the second device, where the first information is used to trigger obtaining information related to the intention from the first production entity; The first device receives second information from the second device, where the second information includes a first parameter of the intent; The first device sends third information to the first production entity according to the second information, where the third information is used to obtain fourth information related to the first parameter; The first device receives the fourth information from the first production entity; The first device obtains fifth information for managing the intention based on the fourth information.

2. The method according to claim 1, characterized in that The method further comprises: The first device receives sixth information input by a user, where the sixth information is used to indicate an operation related to the intention; The first device determines the first information according to the sixth information.

3. The method according to claim 1 or 2, characterized in that The first information includes a first prompt instance, and the thought chain of the first prompt instance includes steps for triggering obtaining the intention-related information from the first production entity.

4. The method according to any one of claims 1 to 3, characterized in that The second information further includes application program interface (API) information, where the API information is used to indicate an API called to obtain the fourth information; The third information is used to call the API to obtain the fourth information.

5. The method according to any one of claims 1 to 4, characterized in that The first device acquiring, based on the fourth information, fifth information for managing the intent, comprising: the first device sending the fourth information to the second device; and the first device receiving fifth information from the second device, the fifth information being used to manage the intent, the fifth information being determined based on the fourth information; The method further includes: the first device managing the intent according to the fifth information.

6. The method according to claim 5, characterized in that The first device manages the intent according to the fifth information, including: The first device performs an intention operation based on the fifth information, and the intention operation includes at least one of the following: requesting the first production entity to create the intention, requesting the first production entity to modify the intention, requesting the first production entity to delete the intention, querying the first production entity about the intention, giving up creating the intention, or giving up modifying the intention.

7. The method according to claim 5 or 6, characterized in that The method further includes: the first device determining a second prompt instance based on the fourth information, wherein context information of the second prompt instance includes the fourth information; The first device sending the fourth information to the second device includes: the first device sending the second prompt instance to the second device.

8. The method according to claim 7, characterized in that The thought chain of the second prompt instance includes a step for triggering the second device to generate a management suggestion for the intention according to the fourth information.

9. The method according to claim 8, characterized in that The method further comprises: The first device receives seventh information from the second device and outputs the seventh information to a user, where the seventh information is used to indicate the management suggestion.

10. The method according to claim 9, characterized in that The method further comprises: The first device receives eighth information input by the user and sends the eighth information to the second device, where the eighth information is feedback information of the user regarding the management suggestion.

11. The method according to any one of claims 1 to 10, characterized in that The first information is also used to trigger the second device to modify the intent, and the first information includes the first parameter.

12. The method according to any one of claims 1 to 11, characterized in that The first device is set in the consumer entity or the second production entity, and the second device is set in other entities; or, The first device and the second device are provided at a consumer entity.

13. The method according to any one of claims 1 to 12, characterized in that The intent is an intent related to network management, and the first parameter includes at least one of the following parameters: an expected object, an expected target, or an expected context.

14. The method according to any one of claims 1 to 13, characterized in that The method further comprises: The second device receives the first information from the first device; The second device sends the second information to the first device based on the first information.

15. An intention management method, characterized in that: The method comprises: A second device receives first information from the first device, where the first information is used to trigger obtaining information related to the intention from the first production entity; The second device sends second information to the first device based on the first information, where the second information includes a first parameter of the intent.

16. The method according to claim 15, characterized in that The first information includes a first prompt instance, and the thought chain of the first prompt instance includes steps for triggering obtaining the intention-related information from the first production entity.

17. The method according to claim 15 or 16, characterized in that The second information also includes application program interface (API) information, where the API information is used to indicate an API called to obtain fourth information related to the first parameter.

18. The method according to any one of claims 15 to 17, characterized in that The method further comprises: The second device receives fourth information related to the first parameter from the first device; The second device determines fifth information for managing the intention based on the fourth information and the macro model; The second device sends the fifth information to the first device.

19. The method according to claim 18, characterized in that The second device receiving third information related to the first parameter from the first device includes: the second device receiving a second prompt instance from the first device, the context of the second prompt instance including the fourth information; The second device determines fifth information for managing the intent based on the fourth information and the large model, including: the second device determines the fifth information based on the second prompt instance and the large model.

20. The method according to claim 19, characterized in that The thought chain of the second prompt instance includes a step for triggering the second device to generate a management suggestion for the intention according to the fourth information.

21. The method according to claim 20, characterized in that The method further comprises: The second device generates the management suggestion according to the second prompt instance; The second device sends seventh information to the first device, where the seventh information is used to indicate the management suggestion.

22. The method according to claim 21, characterized in that The method further includes: the second device receiving eighth information from the first device, the eighth information being user feedback information regarding the management suggestion; The second device determines the fifth information based on the second prompt instance and the large model, including: the second device determines the fifth information based on the second prompt instance, the large model and the eighth information.

23. The method according to any one of claims 15 to 22, characterized in that The first information is further used to trigger the second device to modify the intention, and the first information includes the first parameter; The second device sending second information to the first device according to the first information includes: The second device obtains the first parameter from the first information; The second device sends the second information to the first device according to the first parameter.

24. The method according to any one of claims 15 to 22, characterized in that The second device sending second information to the first device according to the first information includes: The second device determines the first parameter by interacting with the user through the first device according to the first information; The second device sends the second information to the first device according to the first parameter.

25. The method according to any one of claims 15 to 24, characterized in that The first device is set in the consumer entity or the second production entity, and the second device is set in other entities; or, The first device and the second device are provided at a consumer entity.

26. The method according to any one of claims 15 to 24, characterized in that The intent is an intent related to network management, and the first parameter includes at least one of the following parameters: an expected object, an expected target, or an expected context.

27. An intention management method, characterized in that: The method comprises: The first production entity receives third information, wherein the third information is used to obtain fourth information related to the first parameter of the intention; The first production entity sends the fourth information.

28. The method according to claim 27, characterized in that The intent is an intent related to network management, and the first parameter includes at least one of the following parameters: an expected object, an expected target, or an expected context.

29. An intention management method, characterized in that The method comprises: The first device sends first information, and the second device receives the first information, where the first information is used to trigger acquisition of information related to the intention from the first production entity; The second device sends second information according to the first information, and the first device receives the second information, where the second information includes a first parameter of the intent; The first device sends third information according to the second information, the first production entity receives the third information, and the third information is used to obtain fourth information related to the first parameter; The first production entity sends the fourth information, and the first device receives the fourth information; The first device obtains fifth information for managing the intention based on the third information.

30. A communication device, characterized in that: The apparatus includes modules or units for implementing the method according to any one of claims 1 to 29.

31. A communication device, characterized in that: It includes a processor and an interface circuit, wherein the interface circuit is used to receive signals from other devices outside the communication device and transmit them to the processor or send signals from the processor to other devices outside the communication device, and the processor is used to implement the method as described in any one of claims 1 to 29 through logic circuits or executing code instructions.

32. The communication device according to claim 31, wherein: The communication device is a chip or a chip system.

33. A computer-readable storage medium, characterized in that The storage medium stores a computer program or instruction, and when the computer program or instruction is executed by the communication device, the method according to any one of claims 1 to 29 is implemented.

34. A communication system, characterized in that The method comprises at least one of the following devices: a communication device for executing the method according to any one of claims 1 to 14, a communication device for executing the method according to any one of claims 15 to 26, or a communication device for executing the method according to claim 27 or 28.

35. A computer program product, characterized in that The method comprises a computer program or instructions, which implement the steps of the method according to any one of claims 1 to 29 when the computer program or the instructions are executed in a computer.

Images