Communication method and related device

Abstract

The present application provides a communication method and a related device, used for enabling computing capability of a communication node to be applied to an AI task in a learning system and improving the implementation flexibility of different AI systems of a same learning system. In the method, a first node determines first information and second information, the first information being used for indicating AI configuration information of a first AI system, and the second information being used for indicating AI configuration information of a second AI system, wherein the first AI system and the second AI system belong to a same learning system, the first AI system and the second AI system comprise at least one same node, and the learning architecture of the first AI system is different from that of the second AI system; the first node sends the first information and the second information.

Description

A communication method and related equipment Technical Field

[0001] The present application relates to the field of communications, and in particular to a communication method and related equipment. Background Art

[0002] Wireless communication can be the transmission communication between two or more communication nodes without propagating through conductors or cables. The communication nodes generally include network devices and terminal devices.

[0003] Currently, in wireless communication systems, communication nodes generally possess both signal transceiver and computing capabilities. For example, network devices with computing capabilities primarily provide computing power to support signal transceiver capabilities (for example, calculating the time and frequency domain resources required to carry signals), enabling communication between the network device and other communication nodes.

[0004] However, in communication networks, communication nodes may have excess computing power beyond just supporting the aforementioned communication tasks. Therefore, how to utilize this computing power is a pressing technical issue.

[0005] Summary of the Invention

[0006] The present application provides a communication method and related equipment for enabling the computing power of communication nodes to be applied to artificial intelligence (AI) tasks in a learning system, and for improving the implementation flexibility of different AI systems in the same learning system.

[0007] In a first aspect, the present application provides a communication method. The method is performed by a first node, or by a component of the first node (e.g., a processor, chip, or chip system), or can be implemented by a logic module or software that implements all or part of the first node's functionality. In the first aspect and its possible implementations, the method is described as being performed by the first node. In the method, the first node determines first information and second information, the first information indicating AI configuration information of a first AI system, and the second information indicating AI configuration information of a second AI system; the first AI system and the second AI system belong to the same learning system, the first AI system and the second AI system include at least one identical node, and the learning architecture of the first AI system and the learning architecture of the second AI system are different; and the first node sends the first information and the second information.

[0008] Based on the above technical solution, the first information and second information sent by the first node are used to indicate the AI ​​configuration information of the first AI system and the AI ​​configuration information of the second AI system, respectively. The first AI system and the second AI system include at least one common node, and the learning architecture of the first AI system and the learning architecture of the second AI system are different. In other words, within the same learning system where the first and second AI systems reside, the at least one common node can perform AI tasks in the first AI system based on one learning architecture, and the at least one common node can also perform AI tasks in the second AI system based on another learning architecture. Compared to the implementation method of executing AI tasks in different AI systems of the same learning system using the same learning architecture, the above technical solution allows different AI systems of the same learning system to execute AI tasks using different learning architectures. Therefore, when communication nodes in a communication system serve as nodes participating in the learning system, the computing power of the communication nodes can be applied to AI tasks in the learning system, improving the implementation flexibility of different AI systems of the same learning system.

[0009] In addition, the performance and complexity of different learning architectures are generally different, and the capabilities (such as computing power, storage capacity, etc.) and requirements of the nodes participating in the learning system may also be different. For this reason, in the above technical solution, different AI systems of the same learning system can perform AI tasks through different learning architectures respectively, which can provide more learning architectures and better match the needs and capabilities of the nodes participating in the learning system.

[0010] In addition, different learning architectures generally have different performance gains, and the same learning architecture can generally only achieve the performance gains of that one learning architecture. If the AI ​​task is complex and requires the performance gains of multiple learning architectures, the implementation method of executing the AI ​​task through the same learning architecture in different AI systems of the same learning system requires the learning processes of multiple learning architectures to be executed separately and sequentially, which is highly complex. In the above technical solution, different AI systems of the same learning system can execute AI tasks through different learning architectures, respectively, enabling the learning system to obtain the performance gains generated by multiple learning architectures, thereby simplifying the complexity of implementation.

[0011] It should be understood that the first information indicating the AI ​​configuration information of a first AI system can be understood as follows: the first information includes an index of the AI ​​configuration information of the first AI system, so that a recipient of the first information can obtain the AI ​​configuration information of the first AI system based on the index; or the first information includes the AI ​​configuration information of the first AI system, so that a recipient of the first information can obtain the AI ​​configuration information of the first AI system from the first information. Similarly, the second information indicating the AI ​​configuration information of a second AI system can be understood as follows: the second information includes an index of the AI ​​configuration information of the second AI system, so that a recipient of the second information can obtain the AI ​​configuration information of the second AI system based on the index; or the second information includes the AI ​​configuration information of the second AI system, so that a recipient of the second information can obtain the AI ​​configuration information of the second AI system from the second information.

[0012] In this application, a learning system can be understood as a system in which one or more nodes learn data through AI. The learning system can be replaced by an AI learning system, a machine learning system, etc. The first AI system and the second AI system belong to the same learning system. It can be understood that the first AI system includes some nodes in the same learning system, and the second AI system also includes some nodes in the same learning system. The same learning system may include at least two AI systems, such as the first AI system and the second AI system. Each AI system includes one or more nodes, and the first AI system and the second AI system include at least one identical node. Optionally, the at least one identical node may be a control node of the first AI system (or the second AI system), or may not be a control node of the first AI system (or the second AI system), which is not limited here.

[0013] Optionally, in the same learning system, the first AI system and the second AI system are used to perform the same AI task, or the first AI system and the second AI system are used to perform different subtasks in the same AI task.

[0014] In one implementation example, the first AI system and the second AI system are used to perform the same AI task. For example, the AI ​​task performed by the same learning system can be an autonomous driving model training task, that is, the first AI system and the second AI system are both used to perform the autonomous driving task. After obtaining the model training results corresponding to the autonomous driving task, the different AI systems can achieve the fusion of the model training results in an interactive manner. For example, the first AI system and the second AI system can respectively send the locally obtained model training results to each other, so that the two parties can further fuse the locally obtained model training results and the other party's model training results to obtain a fusion result. For another example, the first AI system and the second AI system can both send the locally obtained model training results to the control node of the learning system, so that the control node can fuse the model training results of both parties to obtain a fusion result and then send the fusion result to the first AI system and the second AI system respectively.

[0015] In another implementation example, the first AI system and the second AI system are used to perform different subtasks in the same AI task. For example, the AI ​​task performed by the same learning system can be the task of training an autonomous driving model. Generally, image recognition is a necessary link required for autonomous driving tasks. For example, image recognition tasks such as human image recognition tasks, license plate image recognition tasks, and obstacle image recognition tasks can be subtasks of the autonomous driving task. To this end, the subtask performed by the first AI system and the subtask performed by the second AI subsystem can be two different tasks among the above-mentioned human image recognition tasks, license plate image recognition tasks, and obstacle image recognition tasks. Similarly, the first AI system and the second AI system can also realize the process of model fusion by interacting with each other or by interacting with the control nodes in the learning system.

[0016] Alternatively, the first AI system and the second AI system may include identical nodes. In this case, if the first AI system and the second AI system perform different subtasks within the same AI task, the set of identical nodes used to perform one subtask may be considered the first AI system, while the set of identical nodes used to perform another subtask may be considered the second AI system.

[0017] Optionally, the first AI system includes both nodes in the second AI system and other nodes; in this case, the second AI system can be referred to as a subsystem of the first AI system, a subordinate system of the first AI system, or a subset of the first AI system. Alternatively, the second AI system includes both nodes in the first AI system and other nodes; in this case, the first AI system can be referred to as a subsystem of the second AI system, a subordinate system of the second AI system, or a subset of the second AI system.

[0018] It should be understood that the first node is located in the learning system where the first AI system and the second AI system are located. The first node is used to determine and distribute AI configuration information for the AI ​​systems included in the learning system. The first node can be a control device, control node, scheduling node, scheduling device, management and control device, or management and control node of the learning system. The learning system may include only the first AI system and the second AI system, or it may include other AI systems in addition to the first and second AI systems, without limitation. In addition, each AI system (e.g., the first AI system or the second AI system) may include one or more communication nodes, including network devices and / or terminal devices.

[0019] Optionally, the first node may be a device in a communication system, or in other words, the node in the learning system may be a communication node, for example, the first node may be a network device or a terminal device. Where the first node is a network device, the first node may be an access network device such as a base station or a macro station, or the first node may be a core network device such as a network data analytics function (NWDAF) network element.

[0020] In a possible implementation of the first aspect, the second AI system belongs to the first AI system; and the first node sending the first information and the second information includes: the first node sending the first information and the second information to a control node of the first AI system.

[0021] Based on the above technical solution, in a learning system, the second AI system belongs to the first AI system, meaning that the first AI system includes both nodes in the second AI system and other nodes. Thus, the first node can send the first and second information to the control node of the first AI system, enabling the control node of the first AI system to execute an AI task based on the AI ​​configuration information of the first AI system indicated by the first information. Furthermore, as the second AI system is a subordinate system of the first AI system, the control node of the first AI system can send part or all of the AI ​​configuration information of the second AI system indicated by the second information to the control node of the second AI system. Through a step-by-step instruction process, the control node of the second AI system can execute an AI task based on this part or all of the AI ​​configuration information.

[0022] In a possible implementation of the first aspect, the first node sending the first information and the second information includes: the first node sending the first information to a control node of the first AI system; and the first node sending the second information to a control node of the second AI system.

[0023] Based on the above technical solution, in the learning system, the first node can send the first information and the second information to the control node of the first AI system and the control node of the second AI system, respectively. Through a centralized instruction method, the control node of the first AI system and the control node of the second AI system can execute the AI ​​task based on the received information.

[0024] Optionally, in the above implementation, the relationship between the first AI system and the second AI system is not limited. For example, the first AI system may belong to the second AI system, or the second AI system may belong to the first AI system, or the first AI system and the second AI system may both include other nodes in addition to at least one common node.

[0025] In a possible implementation of the first aspect,

[0026] The AI ​​configuration information of the first AI system includes at least one of the following:

[0027] First indication information, used to indicate nodes in the AI ​​system that participate in the AI ​​task;

[0028] Second indication information, used to indicate the role of the node participating in the AI ​​task in the AI ​​system;

[0029] The third indication information is used to indicate the learning architecture of the AI ​​system;

[0030] Fourth instruction information, used to indicate an AI task of the AI ​​system;

[0031] Fifth indication information, used to indicate the communication resources of the AI ​​system;

[0032] And / or, the AI ​​configuration information of the second AI system includes at least one of the following:

[0033] First indication information, used to indicate nodes in the AI ​​system that participate in the AI ​​task;

[0034] Second indication information, used to indicate the role of the node participating in the AI ​​task in the AI ​​system;

[0035] The third indication information is used to indicate the learning architecture of the AI ​​system;

[0036] Fourth instruction information, used to indicate an AI task of the AI ​​system;

[0037] The fifth indication information is used to indicate the communication resources of the AI ​​system.

[0038] It should be understood that in a process where a node performs an AI task based on AI configuration information (for example, a first node performs an AI task based on the AI ​​configuration information of a first AI system, or a second node performs an AI task based on the AI ​​configuration information of a second AI system), the node may perform the AI ​​task based on at least one of the above items.

[0039] For example, when the AI ​​configuration information includes first indication information, the node can determine the nodes participating in the AI ​​task based on the first indication information, and the nodes participating in the AI ​​task subsequently can provide local data / local computing power, etc. for the execution of the AI ​​task.

[0040] For another example, when the AI ​​configuration information includes second indication information, the node can determine the role of the node participating in the AI ​​task based on the second indication information, and subsequently the node can schedule the corresponding local data / local computing power of each node based on the role of each node.

[0041] For another example, when the AI ​​configuration information includes third indication information, the node may determine the learning architecture of the AI ​​system based on the third indication information, and subsequently the node may schedule each node of the AI ​​system to perform corresponding AI tasks using the learning architecture.

[0042] For another example, when the AI ​​configuration information includes fourth indication information, the node may determine the AI ​​task of the AI ​​system based on the fourth indication information, and subsequently the node may schedule each node of the AI ​​system to perform the AI ​​task.

[0043] For another example, when the AI ​​configuration information includes fifth indication information, the node may determine the communication resources of the AI ​​system based on the fifth indication information, and subsequently the node may transmit the interaction data related to the AI ​​task based on the communication resources of the AI ​​system.

[0044] Based on the above technical solution, the AI ​​configuration information of the first AI system indicated by the first information or the AI ​​configuration information of the second AI system indicated by the second information may include at least one of the above items, so that the first AI system or the second AI system can perform AI tasks based on the at least one of the above items, thereby improving the flexibility of solution implementation.

[0045] In a possible implementation of the first aspect, the learning architecture includes: federated learning, federated distillation, decentralized learning, meta-learning, or segmented learning.

[0046] Based on the above technical solution, the learning architecture of the first AI system and the learning architecture of the second AI system may include any of the above implementations to improve the flexibility of the solution implementation.

[0047] In a possible implementation of the first aspect, the first AI system or the second AI system is configured to perform a second processing on the second data after performing a first processing on the first data to obtain the second data;

[0048] When the learning architecture includes federated learning (or when the learning architecture is federated learning), the first data includes global model parameters and / or gradients, the first processing includes local training processing, the second data includes local model parameters and / or gradients, and the second processing includes fusion processing; it can be understood that when the learning architecture is federated learning, the second processing may include fusion processing of local model parameters, and / or, the second processing may include fusion processing of local model gradients;

[0049] When the learning architecture includes federated distillation (or when the learning architecture is federated distillation), the first data includes global logits, the first processing includes local training processing, the second data includes local logits, and the second processing includes fusion processing; it can be understood that when the learning architecture is federated distillation, the second processing may include fusion processing of local logits;

[0050] When the learning architecture includes decentralized learning (or when the learning architecture is decentralized learning), the first data includes parameters and / or gradients of a first local model, the first processing includes a fusion processing and a local training processing, the second data includes parameters and / or gradients of a second local model, and the second processing includes a fusion processing and a local training processing;

[0051] When the learning architecture includes meta-learning (or when the learning architecture is meta-learning), the first data includes meta-model parameters and / or gradients, the first processing includes support set training processing and test set gradient calculation processing, the second data includes test set gradients and / or losses, and the second processing includes gradient fusion processing and meta-model update processing;

[0052] When the learning architecture includes segmentation learning (or, when the learning architecture is segmentation learning), the first data includes segmentation layer inference results, the first processing includes inference, gradient calculation, back transfer and parameter update processing, the second data includes segmentation layer gradient, and the second processing includes continued gradient transfer processing and parameter update processing.

[0053] Based on the above technical solution, in different learning architectures, either the first AI system or the second AI system can perform the first processing on the first data to obtain the second data, and then perform the second processing on the second data. Furthermore, the above implementation provides specific implementations of the first data, second data, first processing, and second processing in different learning architectures.

[0054] A second aspect of the present application provides a communication method, which is performed by a second node, or by a component of the second node (e.g., a processor, chip, or chip system), or can be implemented by a logic module or software that implements all or part of the second node's functionality. In the second aspect and its possible implementations, the method is described as being performed by the second node. In this method, the second node receives first information indicating AI configuration information of a first AI system; the first AI system and the second AI system belong to the same distributed learning system, the first AI system and the second AI system include at least one common node, and the learning architecture of the first AI system is different from the learning architecture of the second AI system; the second node performs an AI task based on the AI ​​configuration information of the first AI system.

[0055] Based on the above technical solution, in a learning system comprising a first AI system and a second AI system, the first information received by a second node in the first AI system indicates the AI ​​configuration information of the first AI system. The second node subsequently executes an AI task based on the AI ​​configuration information of the first AI system. The first and second AI systems include at least one identical node, and the learning architecture of the first AI system and the learning architecture of the second AI system are different. In other words, within the same learning system where the first and second AI systems reside, the at least one identical node can execute AI tasks in the first AI system based on one learning architecture, and the at least one identical node can also execute AI tasks in the second AI system based on another learning architecture. Compared to the implementation method of executing AI tasks in different AI systems within the same learning system using the same learning architecture, the above technical solution allows different AI systems within the same learning system to execute AI tasks using different learning architectures. Consequently, when communication nodes in a communication system serve as participating nodes in the learning system, the computing power of the communication nodes can be applied to AI tasks within the learning system, improving the implementation flexibility of different AI systems within the same learning system.

[0056] In addition, the performance and complexity of different learning architectures are generally different, and the capabilities (such as computing power, storage capacity, etc.) and requirements of the nodes participating in the learning system may also be different. For this reason, in the above technical solution, different AI systems of the same learning system can perform AI tasks through different learning architectures respectively, which can provide more learning architectures and better match the needs and capabilities of the nodes participating in the learning system.

[0057] In addition, different learning architectures generally have different performance gains, and the same learning architecture can generally only achieve the performance gains of that one learning architecture. If the AI ​​task is complex and requires the performance gains of multiple learning architectures, the implementation method of executing the AI ​​task through the same learning architecture in different AI systems of the same learning system requires the learning processes of multiple learning architectures to be executed separately and sequentially, which is highly complex. In the above technical solution, different AI systems of the same learning system can execute AI tasks through different learning architectures, respectively, enabling the learning system to obtain the performance gains generated by multiple learning architectures, thereby simplifying the complexity of implementation.

[0058] In this application, a learning system can be understood as a system in which one or more nodes learn data using AI. The learning system can be replaced by an AI learning system, a machine learning system, etc. The first AI system and the second AI system belong to the same learning system. It can be understood that the first AI system includes some nodes in the same learning system, and the second AI system also includes some nodes in the same learning system. The same learning system can include at least two AI systems, such as the first AI system and the second AI system. Each AI system includes one or more nodes, and the first AI system and the second AI system include at least one identical node.

[0059] Optionally, the first AI system and the second AI system include identical nodes.

[0060] Optionally, the first AI system includes both nodes in the second AI system and other nodes; in this case, the second AI system can be referred to as a subsystem of the first AI system, a subordinate system of the first AI system, or a subset of the first AI system. Alternatively, the second AI system includes both nodes in the first AI system and other nodes; in this case, the first AI system can be referred to as a subsystem of the second AI system, a subordinate system of the second AI system, or a subset of the second AI system.

[0061] Optionally, in the same learning system, the first AI system and the second AI system are used to perform the same AI task, or the first AI system and the second AI system are used to perform different subtasks in the same AI task.

[0062] It should be understood that the second node is located in the first AI system and is used to receive AI configuration information from the first AI system and, based on the AI ​​configuration information from the first AI system, control nodes in the first AI system to perform AI tasks. The second node may be a control device, control node, scheduling node, scheduling device, management and control device, or management and control node of the first AI system. Furthermore, the first AI system may include one or more communication nodes (the second node being one of these nodes), including network devices and / or terminal devices.

[0063] For example, taking the second node as a control node of a first AI system, the second node executes an AI task based on the AI ​​configuration information of the first AI system. This can be understood as the second node scheduling / controlling / instructing some or all nodes in the first AI system to execute the AI ​​task based on the AI ​​configuration information of the first AI system.

[0064] Optionally, the second node may be a device in a communication system, or in other words, the node in the first AI system may be a communication node. For example, the second node may be a network device or a terminal device. In the case where the second node is a network device, the second node may be a base station, a macro station, etc.

[0065] In a possible implementation of the second aspect, the second AI system belongs to the first AI system; the method further includes: receiving, by the second node, second information indicating AI configuration information of the second AI system; and sending, by the second node, part or all of the AI ​​configuration information of the second AI system to a control node of the second AI system.

[0066] Based on the above technical solution, in the learning system, the second AI system belongs to the first AI system. That is, the first AI system includes both nodes in the second AI system and other nodes. Therefore, the second node, acting as the control node of the first AI system, can receive the second information, allowing it to send part or all of the AI ​​configuration information of the second AI system indicated by the second information to the control node of the second AI system. Through a step-by-step instruction process, the control node of the second AI system can execute an AI task based on the part or all of the AI ​​configuration information.

[0067] In a possible implementation of the second aspect, the AI ​​configuration information of the first AI system includes at least one of the following:

[0068] First indication information, used to indicate nodes in the AI ​​system that participate in the AI ​​task;

[0069] Second indication information, used to indicate the role of the node participating in the AI ​​task in the AI ​​system;

[0070] The third indication information is used to indicate the learning architecture of the AI ​​system;

[0071] Fourth instruction information, used to indicate an AI task of the AI ​​system;

[0072] Fifth indication information, used to indicate the communication resources of the AI ​​system;

[0073] And / or, the AI ​​configuration information of the second AI system includes at least one of the following:

[0074] First indication information, used to indicate nodes in the AI ​​system that participate in the AI ​​task;

[0075] Second indication information, used to indicate the role of the node participating in the AI ​​task in the AI ​​system;

[0076] The third indication information is used to indicate the learning architecture of the AI ​​system;

[0077] Fourth instruction information, used to indicate an AI task of the AI ​​system;

[0078] The fifth indication information is used to indicate the communication resources of the AI ​​system.

[0079] Based on the above technical solution, the AI ​​configuration information of the first AI system indicated by the first information (or the AI ​​configuration information of the second AI system indicated by the second information) can include at least one of the above items, so that the first AI system or the second AI system can perform AI tasks based on the at least one of the above items, thereby improving the flexibility of solution implementation.

[0080] In a possible implementation of the second aspect, the learning architecture includes: federated learning, federated distillation, decentralized learning, meta-learning, or segmented learning.

[0081] Based on the above technical solution, the learning architecture of the first AI system and the learning architecture of the second AI system may include any of the above implementations to improve the flexibility of the solution implementation.

[0082] In a possible implementation of the second aspect, the first AI system or the second AI system is configured to perform a second processing on the second data after performing a first processing on the first data to obtain the second data;

[0083] When the learning architecture includes federated learning (or when the learning architecture is federated learning), the first data includes global model parameters and / or gradients, the first processing includes local training processing, the second data includes local model parameters and / or gradients, and the second processing includes fusion processing; it can be understood that when the learning architecture is federated learning, the second processing may include fusion processing of local model parameters, and / or, the second processing may include fusion processing of local model gradients;

[0084] When the learning architecture includes federated distillation (or when the learning architecture is federated distillation), the first data includes global logits, the first processing includes local training processing, the second data includes local logits, and the second processing includes fusion processing; it can be understood that when the learning architecture is federated distillation, the second processing may include fusion processing of local logits;

[0085] When the learning architecture includes decentralized learning (or when the learning architecture is decentralized learning), the first data includes parameters and / or gradients of a first local model, the first processing includes a fusion processing and a local training processing, the second data includes parameters and / or gradients of a second local model, and the second processing includes a fusion processing and a local training processing;

[0086] When the learning architecture includes meta-learning (or when the learning architecture is meta-learning), the first data includes meta-model parameters and / or gradients, the first processing includes support set training processing and test set gradient calculation processing, the second data includes test set gradients and / or losses, and the second processing includes gradient fusion processing and meta-model update processing;

[0087] When the learning architecture includes segmentation learning (or, when the learning architecture is segmentation learning), the first data includes segmentation layer inference results, the first processing includes inference, gradient calculation, back transfer and parameter update processing, the second data includes segmentation layer gradient, and the second processing includes continued gradient transfer processing and parameter update processing.

[0088] Based on the above technical solution, in different learning architectures, either the first AI system or the second AI system can perform the first processing on the first data to obtain the second data, and then perform the second processing on the second data. Furthermore, the above implementation provides specific implementations of the first data, second data, first processing, and second processing in different learning architectures.

[0089] In a third aspect of the present application, a communication device is provided, which is a first node, or a component of the first node (such as a processor, chip, or chip system), or a logic module or software capable of implementing all or part of the functions of the first node. In the third aspect and its possible implementations, the communication device is described as a first node, and the first node can be a terminal device or a network device.

[0090] The apparatus includes a processing unit and a transceiver unit; the processing unit is configured to determine first information and second information, the first information being configured to indicate AI configuration information of a first artificial intelligence (AI) system, and the second information being configured to indicate AI configuration information of a second AI system; wherein the first AI system and the second AI system belong to the same learning system, the first AI system and the second AI system include at least one common node, and the learning architecture of the first AI system and the learning architecture of the second AI system are different; and the transceiver unit is configured to transmit the first information and the second information.

[0091] In a possible implementation of the third aspect, the second AI system belongs to the first AI system; and the transceiver unit being configured to send the first information and the second information includes: the transceiver unit being configured to send the first information and the second information to a control node of the first AI system.

[0092] In a possible implementation of the third aspect, the transceiver unit being configured to send the first information and the second information includes: the transceiver unit being configured to send the first information to a control node of the first AI system; and the transceiver unit being configured to send the second information to a control node of the second AI system.

[0093] In a possible implementation of the third aspect, the AI ​​configuration information of the first AI system includes at least one of the following:

[0094] First indication information, used to indicate nodes in the AI ​​system that participate in the AI ​​task;

[0095] Second indication information, used to indicate the role of the node participating in the AI ​​task in the AI ​​system;

[0096] The third indication information is used to indicate the learning architecture of the AI ​​system;

[0097] Fourth instruction information, used to indicate an AI task of the AI ​​system;

[0098] Fifth indication information, used to indicate the communication resources of the AI ​​system;

[0099] And / or, the AI ​​configuration information of the second AI system includes at least one of the following:

[0100] First indication information, used to indicate nodes in the AI ​​system that participate in the AI ​​task;

[0101] Second indication information, used to indicate the role of the node participating in the AI ​​task in the AI ​​system;

[0102] The third indication information is used to indicate the learning architecture of the AI ​​system;

[0103] Fourth instruction information, used to indicate an AI task of the AI ​​system;

[0104] The fifth indication information is used to indicate the communication resources of the AI ​​system.

[0105] In a possible implementation of the third aspect, the learning architecture includes: federated learning, federated distillation, decentralized learning, meta-learning, or segmented learning.

[0106] In a possible implementation of the third aspect, the first AI system or the second AI system is configured to perform a first process on the first data to obtain second data, and then perform a second process on the second data;

[0107] When the learning architecture includes federated learning (or when the learning architecture is federated learning), the first data includes global model parameters and / or gradients, the first process includes a local training process, the second data includes local model parameters and / or gradients, and the second process includes a fusion process;

[0108] When the learning architecture includes federated distillation (or when the learning architecture is federated distillation), the first data includes global logits, the first processing includes local training processing, the second data includes local logits, and the second processing includes fusion processing;

[0109] When the learning architecture includes decentralized learning (or when the learning architecture is decentralized learning), the first data includes parameters and / or gradients of a first local model, the first processing includes a fusion processing and a local training processing, the second data includes parameters and / or gradients of a second local model, and the second processing includes a fusion processing and a local training processing;

[0110] When the learning architecture includes meta-learning (or when the learning architecture is meta-learning), the first data includes meta-model parameters and / or gradients, the first processing includes support set training processing and test set gradient calculation processing, the second data includes test set gradients and / or losses, and the second processing includes gradient fusion processing and meta-model update processing;

[0111] When the learning architecture includes segmentation learning (or, when the learning architecture is segmentation learning), the first data includes segmentation layer inference results, the first processing includes inference, gradient calculation, back transfer and parameter update processing, the second data includes segmentation layer gradient, and the second processing includes continued gradient transfer processing and parameter update processing.

[0112] In a possible implementation of the third aspect, the first AI system and the second AI system are used to perform the same AI task, or the first AI system and the second AI system are used to perform different subtasks of the same AI task.

[0113] In the third aspect of the present application, the constituent modules of the communication device can also be used to execute the steps performed in each possible implementation method of the first aspect and achieve corresponding technical effects. For details, please refer to the first aspect and will not be repeated here.

[0114] In a fourth aspect, the present application provides a communication device, which is a second node, or a component of the second node (such as a processor, chip, or chip system), or a logic module or software capable of implementing all or part of the functions of the second node. In the fourth aspect and its possible implementations, the communication device is described as a second node, and the second node can be a terminal device or a network device.

[0115] The device includes a processing unit and a transceiver unit; the transceiver unit is configured to receive first information indicating AI configuration information of a first artificial intelligence (AI) system; wherein the first AI system and the second AI system belong to the same distributed learning system, the first AI system and the second AI system include at least one common node, and the learning architecture of the first AI system and the learning architecture of the second AI system are different; and the processing unit is configured to execute an AI task based on the AI ​​configuration information of the first AI system.

[0116] In a possible implementation of the fourth aspect, the second AI system belongs to the first AI system; the transceiver unit is further configured to receive second information indicating AI configuration information of the second AI system; and the transceiver unit is further configured to send part or all of the AI ​​configuration information of the second AI system to a control node of the second AI system.

[0117] In a possible implementation of the fourth aspect, the AI ​​configuration information of the first AI system includes at least one of the following:

[0118] First indication information, used to indicate nodes in the AI ​​system that participate in the AI ​​task;

[0119] Second indication information, used to indicate the role of the node participating in the AI ​​task in the AI ​​system;

[0120] The third indication information is used to indicate the learning architecture of the AI ​​system;

[0121] Fourth instruction information, used to indicate an AI task of the AI ​​system;

[0122] Fifth indication information, used to indicate the communication resources of the AI ​​system;

[0123] And / or, the AI ​​configuration information of the second AI system includes at least one of the following:

[0124] First indication information, used to indicate nodes in the AI ​​system that participate in the AI ​​task;

[0125] Second indication information, used to indicate the role of the node participating in the AI ​​task in the AI ​​system;

[0126] The third indication information is used to indicate the learning architecture of the AI ​​system;

[0127] Fourth instruction information, used to indicate an AI task of the AI ​​system;

[0128] The fifth indication information is used to indicate the communication resources of the AI ​​system.

[0129] In a possible implementation of the fourth aspect, the learning architecture includes: federated learning, federated distillation, decentralized learning, meta-learning or segmented learning.

[0130] In a possible implementation of the fourth aspect, the first AI system or the second AI system is configured to perform a first process on the first data to obtain second data, and then perform a second process on the second data;

[0131] When the learning architecture includes federated learning (or when the learning architecture is federated learning), the first data includes global model parameters and / or gradients, the first process includes a local training process, the second data includes local model parameters and / or gradients, and the second process includes a fusion process;

[0132] When the learning architecture includes federated distillation (or when the learning architecture is federated distillation), the first data includes global logits, the first processing includes local training processing, the second data includes local logits, and the second processing includes fusion processing;

[0133] When the learning architecture includes decentralized learning (or when the learning architecture is decentralized learning), the first data includes parameters and / or gradients of a first local model, the first processing includes a fusion processing and a local training processing, the second data includes parameters and / or gradients of a second local model, and the second processing includes a fusion processing and a local training processing;

[0134] When the learning architecture includes meta-learning (or when the learning architecture is meta-learning), the first data includes meta-model parameters and / or gradients, the first processing includes support set training processing and test set gradient calculation processing, the second data includes test set gradients and / or losses, and the second processing includes gradient fusion processing and meta-model update processing;

[0135] When the learning architecture includes segmentation learning (or, when the learning architecture is segmentation learning), the first data includes segmentation layer inference results, the first processing includes inference, gradient calculation, back transfer and parameter update processing, the second data includes segmentation layer gradient, and the second processing includes continued gradient transfer processing and parameter update processing.

[0136] In a possible implementation of the fourth aspect, the first AI system and the second AI system are used to perform the same AI task, or the first AI system and the second AI system are used to perform different subtasks of the same AI task.

[0137] In the fourth aspect of the present application, the constituent modules of the communication device can also be used to execute the steps performed in each possible implementation method of the second aspect and achieve corresponding technical effects. For details, please refer to the second aspect and will not be repeated here.

[0138] A fifth aspect of an embodiment of the present application provides a communication device, comprising at least one processor, which is used to execute programs or instructions in a memory so that the device implements the method described in the first aspect or any possible implementation method of the first aspect.

[0139] A sixth aspect of an embodiment of the present application provides a communication device, comprising at least one processor, which is used to execute programs or instructions in a memory so that the device implements the method described in the second aspect or any possible implementation method of the second aspect.

[0140] In a seventh aspect, an embodiment of the present application provides a communication device, comprising at least one logic circuit and an input / output interface; the logic circuit is used to execute the method described in the first aspect or any possible implementation of the first aspect.

[0141] In an eighth aspect, an embodiment of the present application provides a communication device, comprising at least one logic circuit and an input / output interface; the logic circuit is used to execute the method described in the second aspect or any possible implementation of the second aspect.

[0142] A ninth aspect of an embodiment of the present application provides a computer-readable storage medium for storing one or more computer-executable instructions. When the computer-executable instructions are executed by a processor, the processor executes the method described in any possible implementation of any one of the first to second aspects above.

[0143] A tenth aspect of the embodiments of the present application provides a computer program product (or computer program). When the computer program product is executed by the processor, the processor executes the method described in any possible implementation of any one of the first to second aspects above.

[0144] In an eleventh aspect of an embodiment of the present application, a chip system is provided, which includes at least one processor for supporting a communication device to implement the functions involved in any possible implementation method of any one of the first to second aspects above.

[0145] In one possible design, the chip system may further include a memory for storing program instructions and data necessary for the first communication device. The chip system may be composed of a chip or may include a chip and other discrete components. Optionally, the chip system may further include an interface circuit that provides program instructions and / or data to the at least one processor.

[0146] A twelfth aspect of an embodiment of the present application provides a communication system, which includes the communication device of the third aspect and the communication device of the fourth aspect, and / or the communication system includes the communication device of the fifth aspect and the communication device of the sixth aspect, and / or the communication system includes the communication device of the seventh aspect and the communication device of the eighth aspect.

[0147] Among them, the technical effects brought about by any design method in the third aspect to the twelfth aspect can refer to the technical effects brought about by the different design methods in the above-mentioned first aspect to the second aspect, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0148] FIG1a is a schematic diagram of a communication system provided by the present application;

[0149] FIG1b is another schematic diagram of the communication system provided by the present application;

[0150] FIG1c is another schematic diagram of the communication system provided by the present application;

[0151] FIG1d is a schematic diagram of the AI ​​processing process involved in this application;

[0152] FIG1e is another schematic diagram of the AI ​​processing process involved in this application;

[0153] FIG2a is another schematic diagram of the AI ​​processing process involved in this application;

[0154] FIG2 b is another schematic diagram of the AI ​​processing process involved in this application;

[0155] FIG2c is another schematic diagram of the AI ​​processing process involved in this application;

[0156] FIG2 d is another schematic diagram of the AI ​​processing process involved in this application;

[0157] FIG2e is another schematic diagram of the AI ​​processing process involved in this application;

[0158] FIG2f is another schematic diagram of the AI ​​processing process involved in this application;

[0159] FIG2g is another schematic diagram of the AI ​​processing process involved in this application;

[0160] FIG3a is another schematic diagram of the AI ​​processing process involved in this application;

[0161] FIG3 b is another schematic diagram of the AI ​​processing process involved in this application;

[0162] FIG3c is another schematic diagram of the AI ​​processing process involved in this application;

[0163] FIG4 is an interactive schematic diagram of the communication method provided by this application;

[0164] FIG5 is another interactive schematic diagram of the communication method provided by this application;

[0165] FIG6 is another interactive schematic diagram of the communication method provided by this application;

[0166] FIG7a is another interactive schematic diagram of the communication method provided by the present application;

[0167] FIG7 b is another interactive schematic diagram of the communication method provided by this application;

[0168] FIG8 is a schematic diagram of a communication device provided by the present application;

[0169] FIG9 is another schematic diagram of a communication device provided by the present application;

[0170] FIG10 is another schematic diagram of a communication device provided by the present application;

[0171] FIG11 is another schematic diagram of a communication device provided by the present application;

[0172] FIG12 is another schematic diagram of the communication device provided in this application. DETAILED DESCRIPTION

[0173] First, some of the terms used in the embodiments of the present application are explained to facilitate understanding by those skilled in the art.

[0174] (1) Terminal device: It can be a wireless terminal device that can receive network device scheduling and instruction information. The wireless terminal device can be a device that provides voice and / or data connectivity to the user, or a handheld device with wireless connection function, or other processing device connected to a wireless modem.

[0175] Terminal devices can communicate with one or more core networks or the Internet via a radio access network (RAN). Terminal devices can be mobile terminal devices, such as mobile phones (also known as "cellular" phones, mobile phones), computers, and data cards. For example, they can be portable, pocket-sized, handheld, computer-built-in, or vehicle-mounted mobile devices that exchange voice and / or data with the radio access network. Examples include personal communication service (PCS) phones, cordless phones, Session Initiation Protocol (SIP) phones, wireless local loop (WLL) stations, personal digital assistants (PDAs), tablet computers, and computers with wireless transceiver capabilities. A wireless terminal device may also be referred to as a system, a subscriber unit, a subscriber station, a mobile station, a mobile station (MS), a remote station, an access point (AP), a remote terminal, an access terminal, a user terminal, a user agent, a subscriber station (SS), a customer premises equipment (CPE), a terminal, a user equipment (UE), a mobile terminal (MT), etc.

[0176] As an example and not a limitation, in the embodiments of the present application, the terminal device may also be a wearable device. Wearable devices may also be referred to as wearable smart devices or smart wearable devices, etc., which are a general term for wearable devices that are intelligently designed and developed using wearable technology for daily wear, such as glasses, gloves, watches, clothing, and shoes. A wearable device is a portable device that is worn directly on the body or integrated into the user's clothes or accessories. Wearable devices are not only hardware devices, but also achieve powerful functions through software support, data interaction, and cloud interaction. Broadly speaking, wearable smart devices include those that are fully functional, large in size, and can achieve complete or partial functions without relying on smartphones, such as smart watches or smart glasses, etc., as well as those that only focus on a certain type of application function and need to be used in conjunction with other devices such as smartphones, such as various smart bracelets, smart helmets, and smart jewelry for vital sign monitoring.

[0177] The terminal may also be a drone, a robot, a terminal in device-to-device (D2D) communication, a terminal in vehicle-to-everything (V2X), a virtual reality (VR) terminal device, an augmented reality (AR) terminal device, a wireless terminal in industrial control, a wireless terminal in self-driving, a wireless terminal in remote medical, a wireless terminal in smart grid, a wireless terminal in transportation safety, a wireless terminal in smart city, a wireless terminal in smart home, etc.

[0178] In addition, the terminal device may also be a terminal device in a communication system that has evolved after the fifth generation (5G) communication system (e.g., a sixth generation (6G) communication system) or a terminal device in a future public land mobile network (PLMN). For example, the 6G network can further expand the form and function of 5G communication terminals. 6G terminals include but are not limited to vehicles, cellular network terminals (with integrated satellite terminal functions), drones, and Internet of Things (IoT) devices.

[0179] In an embodiment of the present application, the terminal device may also obtain AI services provided by the network device. Optionally, the terminal device may also have AI processing capabilities.

[0180] (2) Network equipment: It can be a device in a wireless network. For example, the network equipment can be a RAN node (or device) that connects a terminal device to a wireless network, which can also be called a base station. Currently, some examples of RAN equipment include: base station, evolved NodeB (eNodeB), gNB (gNodeB) in a 5G communication system, transmission reception point (TRP), evolved Node B (eNB), radio network controller (RNC), Node B (NB), home base station (e.g., home evolved Node B, or home Node B, HNB), base band unit (BBU), or wireless fidelity (Wi-Fi) access point AP, etc. In addition, in a network structure, the network equipment can include a centralized unit (CU) node, a distributed unit (DU) node, or a RAN device including a CU node and a DU node.

[0181] Alternatively, a RAN node can be a macro base station, micro base station, indoor base station, relay node, donor node, or a wireless controller in a cloud radio access network (CRAN) scenario. A RAN node can also be a server, wearable device, vehicle, or vehicle-mounted device. For example, the access network device in vehicle-to-everything (V2X) technology can be a roadside unit (RSU).

[0182] In another possible scenario, multiple RAN nodes collaborate to assist the terminal in achieving wireless access, and different RAN nodes respectively implement part of the functions of the base station. For example, the RAN node can be a centralized unit (CU), a distributed unit (DU), a CU-control plane (CP), a CU-user plane (UP), or a radio unit (RU). The CU and DU can be set separately, or they can be included in the same network element, such as a baseband unit (BBU). The RU can be included in a radio frequency device or radio frequency unit, such as a remote radio unit (RRU), an active antenna unit (AAU), or a remote radio head (RRH).

[0183] In different systems, CU (or CU-CP and CU-UP), DU or RU may also have different names, but those skilled in the art can understand their meanings. For example, in an open access network (open RAN, O-RAN or ORAN) system, CU may also be called O-CU (open CU), DU may also be called O-DU, CU-CP may also be called O-CU-CP, CU-UP may also be called O-CU-UP, and RU may also be called O-RU. For the convenience of description, this application takes CU, CU-CP, CU-UP, DU and RU as examples for description. Any unit of CU (or CU-CP, CU-UP), DU and RU in this application can be implemented by a software module, a hardware module, or a combination of a software module and a hardware module.

[0184] The communication between the access network device and the terminal device follows a certain protocol layer structure. The protocol layer may include a control plane protocol layer and a user plane protocol layer. The control plane protocol layer may include at least one of the following: a radio resource control (RRC) layer, a packet data convergence protocol (PDCP) layer, a radio link control (RLC) layer, a media access control (MAC) layer, or a physical (PHY) layer. The user plane protocol layer may include at least one of the following: a service data adaptation protocol (SDAP) layer, a PDCP layer, an RLC layer, a MAC layer, or a physical layer.

[0185] For the correspondence between network elements in the ORAN system and their achievable protocol layer functions, please refer to Table 1 below.

[0186] Table 1

[0187] The network device may be any other device that provides wireless communication functionality to the terminal device. The embodiments of this application do not limit the specific technology and device form used by the network device. For ease of description, the embodiments of this application do not limit this.

[0188] The network equipment may also include core network equipment, which may include, for example, a mobility management entity (MME), a home subscriber server (HSS), a serving gateway (S-GW), a policy and charging rules function (PCRF), and a public data network gateway (PDN gateway, P-GW) in a fourth generation (4G) network; and network elements such as an access and mobility management function (AMF), a user plane function (UPF), or a session management function (SMF) in a 5G network. In addition, the core network equipment may also include other core network equipment in a 5G network and a next generation network of a 5G network.

[0189] In an embodiment of the present application, the above-mentioned network device may also have a network node with AI capabilities, which can provide AI services for terminals or other network devices. For example, it can be an AI node on the network side (access network or core network), a computing power node, a RAN node with AI capabilities, a core network element with AI capabilities, etc.

[0190] In the embodiments of the present application, the apparatus for implementing the function of the network device may be the network device, or may be a device capable of supporting the network device in implementing the function, such as a chip system, which may be installed in the network device. In the technical solutions provided in the embodiments of the present application, the technical solutions provided in the embodiments of the present application are described by taking the network device as an example.

[0191] (3) Configuration and pre-configuration: In this application, configuration and pre-configuration are used simultaneously. Configuration refers to the network device / server sending some parameter configuration information or parameter values ​​to the terminal through messages or signaling, so that the terminal can determine the communication parameters or resources during transmission based on these values ​​or information. Pre-configuration is similar to configuration, and can be parameter information or parameter values ​​pre-negotiated between the network device / server and the terminal device, or parameter information or parameter values ​​used by the base station / network device or terminal device as specified in the standard protocol, or parameter information or parameter values ​​pre-stored in the base station / server or terminal device. This application does not limit this.

[0192] Furthermore, these values ​​and parameters can be changed or updated.

[0193] (4) The terms "system" and "network" in the embodiments of the present application can be used interchangeably. "Multiple" refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships can 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 previous and next 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" includes A, B, C, AB, AC, BC or ABC. In addition, unless otherwise specified, the ordinal numbers such as "first" and "second" mentioned in the embodiments of the present application are used to distinguish multiple objects, and are not used to limit the order, timing, priority or importance of multiple objects.

[0194] (5) “Sending” and “receiving” in the embodiments of the present application indicate the direction of signal transmission. For example, “sending information to XX” can be understood as the destination of the information being XX, which can include direct sending through the air interface, as well as indirect sending through the air interface by other units or modules. “Receiving information from YY” can be understood as the source of the information being YY, which can include direct receiving from YY through the air interface, as well as indirect receiving from YY through the air interface from other units or modules. “Sending” can also be understood as the “output” of the chip interface, and “receiving” can also be understood as the “input” of the chip interface.

[0195] In other words, sending and receiving can be performed between devices, for example, between a network device and a terminal device, or can be performed within a device, for example, sending or receiving between components, modules, chips, software modules or hardware modules within the device through a bus, wiring or interface.

[0196] It is understandable that information may be processed between the source and destination of information transmission, such as coding, modulation, etc., but the destination can understand the valid information from the source. Similar expressions in this application can be understood similarly and will not be repeated.

[0197] (6) In the embodiments of the present application, "indication" may include direct indication and indirect indication, and may also include explicit indication and implicit indication. The information indicated by a certain information (such as the indication information described below) is called information to be indicated. In the specific implementation process, there are many ways to indicate the information to be indicated, such as but not limited to, directly indicating the information to be indicated, such as the information to be indicated itself or the index of the information to be indicated. The information to be indicated may also be indirectly indicated by indicating other information, wherein the other information is associated with the information to be indicated; or only a part of the information to be indicated may be indicated, while the other part of the information to be indicated is known or agreed in advance. For example, the indication of specific information may be achieved by means of the arrangement order of each information agreed in advance (such as predefined by the protocol), thereby reducing the indication overhead to a certain extent. The present application does not limit the specific method of indication. It is understandable that for the sender of the indication information, the indication information can be used to indicate the information to be indicated, and for the receiver of the indication information, the indication information can be used to determine the information to be indicated.

[0198] In this application, unless otherwise specified, the same or similar parts between the various embodiments can refer to each other. In the various embodiments of this application, and the various methods / designs / implementations in each embodiment, if there is no special explanation and logical conflict, the terms and / or descriptions between different embodiments and the various methods / designs / implementations in each embodiment are consistent and can be referenced to each other. The technical features in different embodiments and the various methods / designs / implementations in each embodiment can be combined to form new embodiments, methods, or implementations according to their inherent logical relationships. The following description of the implementation methods of this application does not constitute a limitation on the scope of protection of this application.

[0199] The present application can be applied to a long term evolution (LTE) system, a new radio (NR) system, or a communication system evolved after 5G (such as 6G, etc.). The communication system includes at least one network device and / or at least one terminal device.

[0200] Please refer to Figure 1a, which is a schematic diagram of a communication system in this application. Figure 1a exemplarily illustrates a network device and six terminal devices, namely terminal device 1, terminal device 2, terminal device 3, terminal device 4, terminal device 5, and terminal device 6. In the example shown in Figure 1a, terminal device 1 is a smart teacup, terminal device 2 is a smart air conditioner, terminal device 3 is a smart gas pump, terminal device 4 is a vehicle, terminal device 5 is a mobile phone, and terminal device 6 is a printer.

[0201] As shown in Figure 1a, the AI ​​configuration information sending entity can be a network device. The AI ​​configuration information receiving entity can be terminal devices 1-6. In this case, the network device and terminal devices 1-6 form a communication system. In this communication system, terminal devices 1-6 can send data to the network device, and the network device needs to receive data sent by terminal devices 1-6. At the same time, the network device can send configuration information to terminal devices 1-6.

[0202] For example, in Figure 1a, terminal devices 4 and 6 can also form a communication system. Terminal device 5 serves as a network device, i.e., the AI ​​configuration information sending entity; terminal devices 4 and 6 serve as terminal devices, i.e., the AI ​​configuration information receiving entities. For example, in a connected vehicle system, terminal device 5 sends AI configuration information to terminal devices 4 and 6, respectively, and receives data from them. Correspondingly, terminal devices 4 and 6 receive AI configuration information from terminal device 5 and send data to terminal device 5.

[0203] Taking the communication system shown in Figure 1a as an example, in addition to performing communication-related services, different devices (including between network devices and network devices, between network devices and terminal devices, and / or between terminal devices) may also perform AI-related services. For example, as shown in Figure 1b, taking the network device as a base station as an example, the base station can perform communication-related services and AI-related services with one or more terminal devices, and different terminal devices can also perform communication-related services and AI-related services. For another example, as shown in Figure 1c, taking the terminal devices including a TV and a mobile phone as an example, communication-related services and AI-related services can also be performed between the TV and the mobile phone.

[0204] The technical solution provided in this application can be applied to a wireless communication system (e.g., the system shown in FIG. 1a , FIG. 1b , or FIG. 1c ). For example, an AI network element can be introduced into the communication system provided in this application to implement some or all AI-related operations. The AI ​​network element can also be referred to as an AI node, AI device, AI entity, AI module, AI model, or AI unit, etc. The AI ​​network element can be a network element built into the communication system. For example, the AI ​​network element can be an AI module built into: an access network device, a core network device, a cloud server, or a network management (OAM) to implement AI-related functions. The OAM can be a network management device for a core network device and / or a network management device for an access network device. Alternatively, the AI ​​network element can also be an independently set network element in the communication system. Optionally, the terminal or the chip built into the terminal can also include an AI entity to implement AI-related functions.

[0205] The following is a brief introduction to artificial intelligence (AI) that may be involved in this application.

[0206] Artificial intelligence (AI) can imbue machines with human intelligence. For example, it can enable machines to simulate certain intelligent human behaviors using computer hardware and software. Machine learning methods can be used to achieve AI. In machine learning, a machine uses training data to learn (or train) a model. This model represents the mapping from input to output. The learned model can be used for inference (or prediction), meaning that the model can be used to predict the output corresponding to a given input. This output can also be called an inference result (or prediction result).

[0207] Machine learning can include supervised learning, unsupervised learning, and reinforcement learning. Among them, unsupervised learning can also be called unsupervised learning.

[0208] Supervised learning uses machine learning algorithms to learn the mapping relationship between sample values ​​and sample labels based on collected sample values ​​and sample labels, and then expresses this learned mapping relationship using an AI model. The process of training a machine learning model is the process of learning this mapping relationship. During training, sample values ​​are input into the model to obtain the model's predicted values. The model parameters are optimized by calculating the error between the model's predicted values ​​and the sample labels (ideal values). Once the mapping relationship is learned, the learned mapping can be used to predict new sample labels. The mapping relationship learned by supervised learning can include linear mappings or nonlinear mappings. Based on the type of label, the learning task can be divided into classification tasks and regression tasks.

[0209] Unsupervised learning uses algorithms to discover inherent patterns in collected sample values. One type of unsupervised learning algorithm uses the samples themselves as supervisory signals, meaning the model learns the mapping from one sample to another. This is called self-supervised learning. During training, the model parameters are optimized by calculating the error between the model's predictions and the samples themselves. Self-supervised learning can be used in signal compression and decompression recovery applications. Common algorithms include autoencoders and generative adversarial networks.

[0210] Reinforcement learning, unlike supervised learning, is a type of algorithm that learns problem-solving strategies through interaction with the environment. Unlike supervised and unsupervised learning, reinforcement learning problems lack explicit label data for "correct" actions. Instead, the algorithm must interact with the environment to obtain reward signals from the environment, and then adjust its decision-making actions to maximize the reward signal value. For example, in downlink power control, the reinforcement learning model adjusts the downlink transmit power of each user based on the overall system throughput fed back by the wireless network, hoping to achieve higher system throughput. The goal of reinforcement learning is also to learn the mapping between environmental states and optimal decision-making actions. However, because the labels for "correct actions" are not available in advance, network optimization cannot be achieved by calculating the error between actions and "correct actions." Reinforcement learning training is achieved through iterative interaction with the environment.

[0211] A neural network (NN) is a specific model in machine learning technology. According to the universal approximation theorem, NNs can theoretically approximate any continuous function, enabling them to learn arbitrary mappings. Traditional communication systems require extensive expert knowledge to design communication modules. However, deep learning communication systems based on neural networks can automatically discover implicit patterns in massive data sets and establish mapping relationships between data, achieving performance superior to traditional modeling methods.

[0212] The idea of ​​a neural network originates from the neuronal structure of the brain. For example, each neuron performs a weighted sum operation on its input values ​​and outputs the result through an activation function. Figure 1d shows a schematic diagram of a neuron structure. Assume that the input of a neuron is x = [x0, x1, ..., x n ], and the weights corresponding to each input are w=[w,w1,…,w n ], where n is a positive integer, w i and x i It can be a decimal, an integer (such as 0, a positive integer or a negative integer, etc.), or a complex number. i As x i The weight of x i Weighted. The bias of the weighted sum of the input values ​​according to the weight is, for example, b. The activation function can take many forms. Assuming that the activation function of a neuron is: y = f(z) = max(0,z), then the output of the neuron is: For another example, if the activation function of a neuron is: y = f(z) = z, then the output of the neuron is: b can be a decimal, an integer (eg, 0, a positive integer, or a negative integer), or a complex number, etc. The activation functions of different neurons in a neural network can be the same or different.

[0213] Furthermore, neural networks generally include multiple layers, each of which may include one or more neurons. Increasing the depth and / or width of a neural network can improve its expressive power, providing more powerful information extraction and abstract modeling capabilities for complex systems. The depth of a neural network can refer to the number of layers it comprises, and the number of neurons in each layer can be referred to as the width of that layer. In one implementation, a neural network includes an input layer and an output layer. The input layer processes the input information received by the neural network through neurons, passing the processing results to the output layer, which then obtains the output of the neural network. In another implementation, a neural network includes an input layer, a hidden layer, and an output layer. The input layer processes the input information received by the neural network through neurons, passing the processing results to an intermediate hidden layer. The hidden layer performs calculations on the received processing results to obtain a calculation result, which is then passed to the output layer or the next adjacent hidden layer, which ultimately obtains the output of the neural network. A neural network can include one hidden layer or multiple hidden layers connected in sequence, without limitation.

[0214] A neural network is, for example, a deep neural network (DNN). Depending on how the network is constructed, a DNN can include a feedforward neural network (FNN), a convolutional neural network (CNN), and a recurrent neural network (RNN).

[0215] A characteristic of FNN networks is that neurons in adjacent layers are fully connected. This characteristic typically requires a large amount of storage space and results in high computational complexity. Figure 1e shows a schematic diagram of an FNN network.

[0216] CNN is a neural network specifically designed to process data with a grid-like structure. For example, time series data (discrete sampling along the time axis) and image data (discrete sampling along two dimensions) can both be considered grid-like data. CNNs do not utilize all input information at once for computation. Instead, they use a fixed-size window to intercept a portion of the information for convolution operations, significantly reducing the computational complexity of model parameters. Furthermore, depending on the type of information intercepted by the window (e.g., people and objects in an image represent different types of information), each window can use a different convolution kernel, enabling CNNs to better extract features from the input data.

[0217] RNNs are a type of DNN that utilizes feedback time series information. Their input consists of a new input value at the current moment and their own output value at the previous moment. RNNs are suitable for capturing temporally correlated sequence features and are particularly well-suited for applications such as speech recognition and channel coding.

[0218] During the machine learning model training process, a loss function can be defined. This function describes the gap or discrepancy between the model's output and the ideal target value. Loss functions can be expressed in various forms, and there are no restrictions on their specific form. The model training process can be viewed as adjusting some or all of the model's parameters to keep the loss function below a threshold or meet the target.

[0219] A model may also be referred to as an AI model, rule, or other name. An AI model can be considered a specific method for implementing an AI function. An AI model represents a mapping relationship or function between the input and output of a model. AI functions may include one or more of the following: data collection, model training (or model learning), model information release, model inference (or model reasoning, inference, or prediction, etc.), model monitoring or model verification, or inference result release, etc. AI functions may also be referred to as AI (related) operations, or AI-related functions.

[0220] The following is an exemplary description of the implementation process of the neural network with reference to the accompanying drawings.

[0221] 1. Fully connected neural network.

[0222] Also known as a multilayer perceptron (MLP), an MLP consists of an input layer (left), an output layer (right), and multiple hidden layers (center). Each layer of the MLP contains several nodes, called neurons. Neurons in adjacent layers are connected to each other.

[0223] Optionally, considering neurons in two adjacent layers, the output h of a neuron in the next layer is the weighted sum of all neurons x connected to it in the previous layer and passes through an activation function, which can be expressed as: h=f(wx+b).

[0224] Among them, w is the weight matrix, b is the bias vector, and f is the activation function.

[0225] Alternatively, the output of the neural network can be recursively expressed as: y = f n (w n f n-1 (…)+b n ).

[0226] Where n is the index of the neural network layer, 1<=n<=N, where N is the total number of neural network layers.

[0227] In other words, a neural network can be understood as a mapping from an input data set to an output data set. Neural networks are typically initialized randomly, and the process of obtaining this mapping from random w and b using existing data is called neural network training.

[0228] Optionally, the specific training method is to use a loss function to evaluate the output of the neural network. As shown in Figure 2b, the error can be backpropagated, and the neural network parameters (including w and b) can be iteratively optimized through gradient descent until the loss function reaches a minimum value, i.e., the "optimal point" in Figure 2b. It is understandable that the neural network parameters corresponding to the "optimal point" in Figure 2b can be used as the neural network parameters in the trained AI model information.

[0229] Alternatively, the gradient descent process can be expressed as:

[0230] Among them, θ is the parameter to be optimized (including w and b), L is the loss function, and η is the learning rate, which controls the step size of gradient descent. represents the derivative operation, represents the derivative of θ with respect to L.

[0231] Alternatively, the backpropagation process can utilize the chain rule for partial derivatives. As shown in Figure 2c, the gradient of the previous layer parameters can be recursively calculated from the gradient of the next layer parameters, which can be expressed as:

[0232] Among them, w ij is the weight of node j connecting to node i, s i is the weighted sum of the inputs to node i.

[0233] 2. Federated Learning (FL)

[0234] The concept of federated learning effectively addresses the current challenges facing AI development. While fully protecting user data privacy and security, it enables collaboration between edge devices and central servers to efficiently complete model learning tasks. As shown in Figure 2d, the FL architecture is the most widely used training architecture in the FL field. The FedAvg algorithm is the foundation of FL. Its algorithmic process is roughly as follows:

[0235] (1) The center initializes the model to be trained And broadcast it to all client devices.

[0236] (2) In the round t∈[1,T], client k∈[1,K] based on the local dataset For the received global model Perform E epochs of training to obtain local training results Report it to the central node.

[0237] (3) The central node aggregates and collects the local training results from all (or some) clients. Assume that the client set that uploads the local model in round t is The center will use the number of samples of the corresponding client as the weight to perform weighted averaging to obtain a new global model. The specific update rule is: The center then sends the latest version of the global model Broadcast to all client devices for a new round of training.

[0238] (4) Repeat steps (2) and (3) until the model finally converges or the number of training rounds reaches the upper limit.

[0239] In addition to reporting local models You can also use the local gradient of training After reporting, the central node averages the local gradients and updates the global model according to the direction of the average gradient.

[0240] As you can see, in the FL framework, datasets exist on distributed nodes. Distributed nodes collect local datasets, perform local training, and report the local training results (models or gradients) to the central node. The central node itself does not have a dataset; it is only responsible for fusing the training results of distributed nodes to obtain a global model and send it to the distributed nodes.

[0241] 3. Decentralized learning.

[0242] Different from federated learning, another distributed learning architecture, decentralized learning, is shown in Figure 2e. Considering a fully distributed system without a central node, the design goal f(x) of a decentralized learning system is generally the goal f of each node. i The mean of (x), that is Where n is the number of distributed nodes, x is the parameter to be optimized. In machine learning, x is the parameter of the machine learning (such as neural network) model. Each node uses local data and local target f i (x) Calculate local gradient Then it is sent to the neighboring nodes that can be communicated. After any node receives the gradient information sent by its neighbor, it can update the parameters x of the local model according to the following formula

[0243] Among them, N i is the set of neighbor nodes of node i, |N i | represents the number of elements in the neighbor node set of node i, that is, the number of neighbor nodes of node i. Through information interaction between nodes, the decentralized learning system will eventually learn a unified model.

[0244] 4. Scalable distributed learning.

[0245] In the aforementioned federated learning and decentralized learning approaches, nodes need to exchange complete neural network model parameters or gradients. If a neural network model can be split into multiple sub-models, the parameters or gradients of some of these sub-models can be exchanged.

[0246] Exemplarily, the scalable neural network model can be divided into: a width scalable neural network model shown in FIG2f and a depth scalable neural network model shown in FIG2g, which are respectively constructed by horizontal superposition or vertical cascade of sub-models. For example, in FIG2f, for one of the neural network models a in the complete neural network model, by vertically expanding the number of layers of the neural network, the neural network models b / c / d / e in FIG2f can be obtained. For another example, in FIG2g, for one of the neural network models a in the complete neural network model, by horizontally expanding the number of neurons in the same layer of the neural network, the neural network models b / c / d / e in FIG2g can be obtained.

[0247] It should be noted that in Figure 2f, although model a and model e have the same number of model layers, neural network models b / c / d / e are all expanded based on model a; generally, compared to neural network model a, neural network models b / c / d have more parameters and better performance. For this reason, in practical applications, neural network model a can be used alone as neural network model e (possibly with poor performance), or it can be expanded into neural network models b / c / d with more parameters based on performance and complexity requirements. Similarly, in Figure 2g, neural network model a can be used alone as neural network model e (possibly with poor performance), or it can be expanded into neural network models b / c / d with more parameters based on performance and complexity requirements.

[0248] Distributed learning based on scalable neural network models is called scalable distributed learning. In a scalable distributed learning system, nodes can exchange parameters or gradients of sub-models instead of the parameters or gradients of the complete model, thereby reducing communication overhead.

[0249] The technical solutions provided in this application can be applied to wireless communication systems (e.g., the systems shown in Figures 1a and 1b). In wireless communication systems, communication nodes generally have both signal transceiver capabilities and computing capabilities. For example, network devices with computing capabilities primarily provide computing power to support signal transceiver capabilities (e.g., calculating the time and frequency domain resources that carry the signal), thereby enabling communication between the network device and other communication nodes.

[0250] However, in communication networks, communication nodes may have excess computing power beyond just supporting the aforementioned communication tasks. Therefore, how to utilize this computing power is a pressing technical issue.

[0251] In one possible implementation, the computing power of the communication node can be used in the centralized scenario shown in Figure 2d or the decentralized scenario shown in Figure 2e. Generally, for a certain node, in the above scenario, the learning system where the node is located is used to perform a certain AI task, or in other words, the data received or sent by any node in the learning system belongs to the same AI task. However, a node can often communicate with multiple other nodes. For this reason, in a learning system where multiple nodes are located, a node can form an AI system with some nodes to perform an AI task, and the node can also form another AI system with other nodes to perform an AI task. It should be noted that the two AI tasks performed by the node may be the same AI task, or they may be different subtasks in the same AI task. The following will take the latter as an example and explain it in combination with the example shown in Figure 3a.

[0252] For example, FIG3a shows a schematic diagram of a federated learning scenario in which a learning system includes multiple AI systems. In this diagram, a learning system including nine nodes, namely, node 1, node 2, ..., node 9, is taken as an example.

[0253] In an AI system 1 including nodes 1, 2, 3, 4, and 5, after nodes 2, 3, 4, and 5 update their local models, they upload the updated local models / gradients to node 1, and node 1 can subsequently perform model / gradient fusion.

[0254] In AI system 2 including nodes 3 and 5, after node 3 updates its local model, it uploads the updated local model / gradient to node 5, and node 5 can subsequently perform model / gradient fusion.

[0255] In AI system 3 including nodes 2, 6, and 7, after nodes 6 and 7 update their local models, they upload the updated local models / gradients to node 2, and node 2 can subsequently perform model / gradient fusion.

[0256] In the AI ​​system 3 including nodes 4, 8, and 9, after nodes 8 and 9 update their local models, they upload the updated local models / gradients to node 4, and node 4 can subsequently perform model / gradient fusion.

[0257] Alternatively, in Figure 3a, node 1 can serve as the control node for the learning system in which AI systems 1, 2, 3, and 4 reside. Alternatively, in addition to the nine nodes shown in Figure 3a, additional control nodes may exist. For example, as shown in Figure 3b, the control node can be connected to each node to control the configuration of each node in each AI system. For another example, as shown in Figure 3c, the control node can be connected to the control node of each AI system to configure each node in each AI system through the control node of each AI system.

[0258] As can be seen from the above implementation, an AI system can include a data aggregation node, which can receive AI data from other nodes and perform fusion processing. For example, this node can be node 1 in AI system 1, node 3 in AI system 2, node 2 in AI system 3, or node 4 in AI system 4. Accordingly, this node can be called a control node, central node, or aggregation node in the AI ​​system.

[0259] Optionally, in the scenarios shown in Figures 3a / 3b / 3c, the nodes in each AI system may be terminal devices. Alternatively, in the scenarios shown in Figures 3a / 3b / 3c, the control nodes in one or more AI systems may be network devices, while the other nodes in the one or more AI systems other than the control nodes may be terminal devices. The scenario shown in Figure 3c will be used as an example for illustration below.

[0260] As an implementation example, in the scenario shown in Figure 3c, the control node can be a core network element, the control node of AI system 1 (i.e., node 1), the control node of AI system 2 (i.e., node 3), the control node of AI system 3 (i.e., node 2), and the control node of AI system 4 (i.e., node 4) can be access network elements, and other nodes other than the control nodes (i.e., node 5, node 6, node 7, node 8, and node 9) can be terminal devices.

[0261] As another implementation example, in the scenario shown in Figure 3c, the control node can be a macro base station, the control node of AI system 1 (i.e., node 1), the control node of AI system 2 (i.e., node 3), the control node of AI system 3 (i.e., node 2), and the control node of AI system 4 (i.e., node 4) can be micro base stations or home base stations, etc., and other nodes other than the control nodes (i.e., node 5, node 6, node 7, node 8, and node 9) can be terminal devices.

[0262] However, in the scenarios shown in Figure 3a / Figure 3b / Figure 3c, the capabilities, data, or requirements of each node in different AI systems may be different. In the above scenarios, different AI systems all use the same learning architecture of federated learning, which lacks flexibility.

[0263] To address the above issues, this application provides a communication method and related equipment for enabling the computing power of communication nodes to be applied to AI tasks in a learning system, and for improving the flexibility of implementing different AI systems within the same learning system. This will be described in detail below with reference to the accompanying figures.

[0264] Please refer to FIG4 , which is a schematic diagram of an implementation of the communication method provided in this application. The method includes the following steps.

[0265] It should be noted that, in FIG4 , the first node and the second node are used as examples of the execution subjects of the interactive diagram to illustrate the method, but the present application does not limit the execution subjects of the interactive diagram. For example, the execution subject of S401 in FIG4 and the corresponding implementation is the first node, and the execution subject may also be a chip, chip system, or processor that supports the first node to implement the method, or a logic module or software that can implement all or part of the functions of the first node. The second node in S401-S402 in FIG4 and the corresponding implementation may also be replaced by a chip, chip system, or processor that supports the second node to implement the method, or may be replaced by a logic module or software that can implement all or part of the functions of the second node.

[0266] S401. A first node sends first information, and a second node receives the first information, wherein the first information indicates AI configuration information of a first AI system.

[0267] It should be understood that the first information used to indicate the AI ​​configuration information of a first AI system can be understood as follows: the first information includes an index of the AI ​​configuration information of the first AI system, so that the recipient of the first information can obtain the AI ​​configuration information of the first AI system based on the index; or the first information includes the AI ​​configuration information of the first AI system, so that the recipient of the first information can obtain the AI ​​configuration information of the first AI system from the first information. Similarly, the second information mentioned below used to indicate the AI ​​configuration information of a second AI system can be understood as follows: the second information includes an index of the AI ​​configuration information of the second AI system, so that the recipient of the second information can obtain the AI ​​configuration information of the second AI system based on the index; or the second information includes the AI ​​configuration information of the second AI system, so that the recipient of the second information can obtain the AI ​​configuration information of the second AI system from the second information.

[0268] Optionally, the first information (and / or second information) sent by the first node may be carried in one or more of an RRC layer message, a PDCP layer message, a MAC layer message, and a PHY layer message.

[0269] S402. The first node executes an AI task based on the AI ​​configuration information of the first AI system.

[0270] In this application, a learning system can be understood as a system in which one or more nodes learn data using AI. The learning system can be replaced by an AI learning system, a machine learning system, etc. The first AI system and the second AI system belong to the same learning system. It can be understood that the first AI system includes some nodes in the same learning system, and the second AI system also includes some nodes in the same learning system. The same learning system can include at least two AI systems, such as the first AI system and the second AI system. Each AI system includes one or more nodes, and the first AI system and the second AI system include at least one identical node.

[0271] Optionally, in the same learning system, the first AI system and the second AI system are used to perform the same AI task, or the first AI system and the second AI system are used to perform different subtasks in the same AI task.

[0272] In one implementation example, the first AI system and the second AI system are used to perform the same AI task. For example, the AI ​​task performed by the same learning system can be an autonomous driving model training task, that is, the first AI system and the second AI system are both used to perform the autonomous driving task. After obtaining the model training results corresponding to the autonomous driving task, the different AI systems can achieve the fusion of the model training results in an interactive manner. For example, the first AI system and the second AI system can respectively send the locally obtained model training results to each other, so that the two parties can further fuse the locally obtained model training results and the other party's model training results to obtain a fusion result. For another example, the first AI system and the second AI system can both send the locally obtained model training results to the control node of the learning system, so that the control node can fuse the model training results of both parties to obtain a fusion result and then send the fusion result to the first AI system and the second AI system respectively.

[0273] In another implementation example, the first AI system and the second AI system are used to perform different subtasks in the same AI task. For example, the AI ​​task performed by the same learning system can be the task of training an autonomous driving model. Generally, image recognition is a necessary link required for autonomous driving tasks. For example, image recognition tasks such as human image recognition tasks, license plate image recognition tasks, and obstacle image recognition tasks can be subtasks of the autonomous driving task. To this end, the subtask performed by the first AI system and the subtask performed by the second AI subsystem can be two different tasks among the above-mentioned human image recognition tasks, license plate image recognition tasks, and obstacle image recognition tasks. Similarly, the first AI system and the second AI system can also realize the process of model fusion by interacting with each other or by interacting with the control nodes in the learning system.

[0274] Alternatively, the first AI system and the second AI system may include identical nodes. In this case, if the first AI system and the second AI system perform different subtasks within the same AI task, the set of identical nodes used to perform one subtask may be considered the first AI system, while the set of identical nodes used to perform another subtask may be considered the second AI system.

[0275] Optionally, the first AI system includes both nodes in the second AI system and other nodes; in this case, the second AI system can be referred to as a subsystem of the first AI system, a subordinate system of the first AI system, or a subset of the first AI system. Alternatively, the second AI system includes both nodes in the first AI system and other nodes; in this case, the first AI system can be referred to as a subsystem of the second AI system, a subordinate system of the second AI system, or a subset of the second AI system.

[0276] It should be understood that the first node is located in the learning system where the first AI system and the second AI system are located. The first node is used to determine and distribute AI configuration information for the AI ​​systems included in the learning system. The first node can be a control device, control node, scheduling node, scheduling device, management and control device, or management and control node of the learning system. The learning system may include only the first AI system and the second AI system, or it may include other AI systems in addition to the first and second AI systems, without limitation. In addition, each AI system (e.g., the first AI system or the second AI system) may include one or more communication nodes, including network devices and / or terminal devices.

[0277] Optionally, the first node may be a device in a communication system, or in other words, the node in the learning system may be a communication node, for example, the first node may be a network device or a terminal device. Wherein, in the case where the first node is a network device, the first node may be a base station, a combination of one or more of CU-CP, CU-UP, DU, and RU in ORAN, a macro station or other access network device, or the first node may be a core network device such as a network data analytics function (NWDAF) network element. Similarly, in the first AI system, the second node may also be a network device or a terminal device.

[0278] Based on the technical solution shown in Figure 4, the first information received by the second node in step S401 indicates the AI ​​configuration information of the first AI system. Subsequently, in step S402, the second node can execute an AI task based on the AI ​​configuration information of the first AI system. The first and second AI systems include at least one identical node, and the learning architecture of the first AI system and the learning architecture of the second AI system are different. In other words, within the same learning system where the first and second AI systems reside, the at least one identical node can execute AI tasks based on one learning architecture in the first AI system, and the at least one identical node can also execute AI tasks based on another learning architecture in the second AI system. Compared to the implementation method of executing AI tasks using the same learning architecture in different AI systems within the same learning system, in the above technical solution, different AI systems within the same learning system can execute AI tasks using different learning architectures. Therefore, when communication nodes in a communication system serve as participating nodes in the learning system, the computing power of the communication nodes can be applied to AI tasks within the learning system, improving the implementation flexibility of different AI systems within the same learning system.

[0279] In addition, the performance and complexity of different learning architectures are generally different, and the capabilities (such as computing power, storage capacity, etc.) and requirements of the nodes participating in the learning system may also be different. For this reason, in the above technical solution, different AI systems of the same learning system can perform AI tasks through different learning architectures respectively, which can provide more learning architectures and better match the needs and capabilities of the nodes participating in the learning system.

[0280] In addition, different learning architectures generally have different performance gains, and the same learning architecture can generally only achieve the performance gains of that one learning architecture. If the AI ​​task is complex and requires the performance gains of multiple learning architectures, the implementation method of executing the AI ​​task through the same learning architecture in different AI systems of the same learning system requires the learning processes of multiple learning architectures to be executed separately and sequentially, which is highly complex. In the above technical solution, different AI systems of the same learning system can execute AI tasks through different learning architectures, respectively, enabling the learning system to obtain the performance gains generated by multiple learning architectures, thereby simplifying the complexity of implementation.

[0281] In one possible implementation, the method shown in FIG4 can be applied not only to the configuration process of AI configuration information for a particular AI system (i.e., the first AI system) in a learning system, but also to the configuration process of AI configuration information for other AI systems. The following description will be based on the scenario shown in FIG3c , assuming that the first AI system is AI system 1 in FIG3c , and the other AI systems include a second AI system (i.e., AI system 2 in FIG3c ).

[0282] It should be noted that in Figure 3c, the control node can serve as the first node, AI system 1 can serve as the first AI system, and the control node of AI system 1 (i.e., node 1) can serve as the second node. AI system 2 can serve as the second AI system, and the control node of AI system 2 (i.e., node 3) can serve as the third node. AI system 1 can be understood as the superior AI system of AI system 2, or, in other words, AI system 2 can be understood as the subordinate AI system of AI system 1. The following describes the configuration process of the AI ​​configuration information of the second AI system, using the processes shown in Figures 5 and 6, respectively.

[0283] Referring to Figure 5 , the first AI system may include the second AI system. The first node may serve as a control node (e.g., the control node in Figure 3c ), the second node may serve as a control node for the first AI system (e.g., node 1 in AI system 1 in Figure 3c ), and the third node may serve as a control node for the second AI system (e.g., node 3 in AI system 2 in Figure 3c ). As shown in Figure 5 , the interaction process between the nodes includes the following steps.

[0284] S501. A first node sends first information, and correspondingly, a second node receives the first information, where the first information is used to indicate AI configuration information of a first AI system.

[0285] S502. The first node sends second information, and correspondingly, the second node receives the second information, where the second information is used to indicate AI configuration information of the second AI system.

[0286] S503. The second node sends part or all of the AI ​​configuration information of the second AI system. Correspondingly, the third node receives part or all of the AI ​​configuration information of the second AI system.

[0287] S504. The second node executes the AI ​​task based on the AI ​​configuration information of the first AI system.

[0288] S505. The third node executes the AI ​​task based on part or all of the AI ​​configuration information of the second AI system.

[0289] It should be noted that the implementation process of sending and receiving the first information in FIG. 5 can refer to the implementation process of sending and receiving the first information in FIG. 4 . Similarly, the configuration process of the AI ​​configuration information of the second AI system can refer to the configuration process of the AI ​​configuration information of the first AI system. That is, the implementation process of sending and receiving the second information in FIG. 5 can also refer to the implementation process of sending and receiving the first information in FIG. 4 .

[0290] Based on the implementation process shown in Figure 5, in the learning system, the second AI system belongs to the first AI system. That is, the first AI system includes both nodes in the second AI system and other nodes. Therefore, the first node can send the first information and the second information to the control node of the first AI system, enabling the control node of the first AI system to execute the AI ​​task based on the AI ​​configuration information of the first AI system indicated by the first information. Furthermore, as the second AI system is a subordinate system of the first AI system, the control node of the first AI system can send part or all of the AI ​​configuration information of the second AI system indicated by the second information to the control node of the second AI system. Through a step-by-step instruction process, the control node of the second AI system can execute the AI ​​task based on the part or all of the AI ​​configuration information of the second AI system indicated by the second information.

[0291] Referring to Figure 6 , the first AI system may include the second AI system. The first node may serve as a control node (e.g., the control node in Figure 3c ), the second node may serve as a control node for the first AI system (e.g., node 1 in AI system 1 in Figure 3c ), and the third node may serve as a control node for the second AI system (e.g., node 3 in AI system 2 in Figure 3c ). As shown in Figure 6 , the interaction process between the nodes includes the following steps.

[0292] S601. A first node sends first information, and correspondingly, a second node receives the first information, where the first information is used to indicate AI configuration information of a first AI system.

[0293] S602. The first node sends second information, and correspondingly, the third node receives the second information, where the second information is used to indicate AI configuration information of the second AI system.

[0294] S603. The second node executes the AI ​​task based on the AI ​​configuration information of the first AI system.

[0295] S604. The third node executes the AI ​​task based on the AI ​​configuration information of the second AI system.

[0296] It should be noted that the implementation process of sending and receiving the first information in FIG. 6 can refer to the implementation process of sending and receiving the first information in FIG. 4 . Similarly, the configuration process of the AI ​​configuration information of the second AI system can refer to the configuration process of the AI ​​configuration information of the first AI system. That is, the implementation process of sending and receiving the second information in FIG. 6 can also refer to the implementation process of sending and receiving the first information in FIG. 4 .

[0297] Based on the implementation process shown in FIG6 , the first node can respectively send the first information and the second information to the control node of the first AI system (i.e., the second node) and the control node of the second AI system (i.e., the third node). Through a centralized instruction method, the control node of the first AI system and the control node of the second AI system can execute the AI ​​task based on the received information.

[0298] Optionally, in the implementation shown in FIG. 6 , the relationship between the first AI system and the second AI system is not limited. For example, the first AI system may belong to the second AI system, or the second AI system may belong to the first AI system, or the first AI system and the second AI system may both include other nodes in addition to at least one common node.

[0299] In one possible implementation, the learning architecture of the first AI system or the second AI system involved in the above embodiments includes any architecture such as federated learning, federated distillation, decentralized learning, meta-learning or segmented learning to enhance the flexibility of solution implementation.

[0300] In one possible implementation, the AI ​​configuration information involved in the above embodiments (e.g., the AI ​​configuration information of the first AI system or the AI ​​configuration information of the second AI system) includes at least one of the following:

[0301] First indication information, used to indicate nodes in the AI ​​system that participate in the AI ​​task;

[0302] Second indication information, used to indicate the role of the node participating in the AI ​​task in the AI ​​system;

[0303] The third indication information is used to indicate the learning architecture of the AI ​​system;

[0304] Fourth instruction information, used to indicate an AI task of the AI ​​system;

[0305] The fifth indication information is used to indicate the communication resources of the AI ​​system.

[0306] It should be understood that in a process where a node performs an AI task based on AI configuration information (for example, a first node performs an AI task based on the AI ​​configuration information of a first AI system, or a second node performs an AI task based on the AI ​​configuration information of a second AI system), the node may perform the AI ​​task based on at least one of the above items.

[0307] In one implementation example, when the AI ​​configuration information includes first indication information, the node may determine the nodes participating in the AI ​​task based on the first indication information. The nodes subsequently participating in the AI ​​task may provide local data, local computing power, etc. for executing the AI ​​task. For example, the nodes participating in the AI ​​task in the AI ​​system indicated by the first indication information may be determined based on node information such as the data distribution, feature distribution, requirements, user capabilities, and communication link quality of each node.

[0308] In another implementation example, when the AI ​​configuration information includes second indication information, the node may determine the roles of the nodes participating in the AI ​​task based on the second indication information. Subsequently, the node may schedule the corresponding local data / local computing power of each node based on the role of each node. Exemplarily, the node roles indicated by the second indication information may include: distributed node, aggregation node, meta-learning center node, etc.; wherein, the node role may be determined based on the capabilities of each node (e.g., computing power, communication capabilities, storage capacity, etc.).

[0309] In another implementation example, when the AI ​​configuration information includes third indication information, the node may determine the AI ​​system's learning architecture based on the third indication information. Subsequently, the node may schedule each node of the AI ​​system to execute corresponding AI tasks using the learning architecture. Exemplarily, the determination of the AI ​​system's learning architecture as indicated by the third indication information may be based on information such as the AI ​​system's dataset and the presence of a central node.

[0310] For example, in an AI system where each node has no common dataset, has a central node, and trains on the same model, the AI ​​system can be determined to employ a federated learning architecture. Alternatively, in an AI system where each node has a common dataset, a central node, and trains on a unified model, the AI ​​system can be determined to employ a federated distillation architecture. Alternatively, in an AI system where each node has no common dataset, no central node, and trains on the same model, the AI ​​system can be determined to employ a decentralized learning architecture. Alternatively, in an AI system where different parts of the same model are deployed on different nodes, the AI ​​system can be determined to employ a partitioned learning architecture.

[0311] In another implementation example, if the AI ​​configuration information includes the fourth indication information, the node may determine the AI ​​task of the AI ​​system based on the fourth indication information and subsequently schedule the various nodes of the AI ​​system to perform the AI ​​task. For example, the task targeted by the AI ​​system may be a global, unified task or a single-level task within a multi-level task, determined by the first node, which serves as the control node of the learning model.

[0312] In another implementation example, when the AI ​​configuration information includes the fifth indication information, the node may determine the communication resources of the AI ​​system based on the fifth indication information, and subsequently the node may transmit the interactive data related to the AI ​​task based on the communication resources of the AI ​​system. For example, the communication resources of an AI system may include a larger time-frequency resource, and the control node within the AI ​​system may subsequently allocate a smaller time-frequency resource for the interactive data based on the larger time-frequency resource.

[0313] Therefore, the AI ​​configuration information of the first AI system indicated by the first information or the AI ​​configuration information of the second AI system indicated by the second information may include at least one of the aforementioned items, so that the first AI system or the second AI system can perform AI tasks based on the at least one aforementioned item, thereby improving the flexibility of solution implementation.

[0314] Optionally, when the AI ​​configuration information of the second AI system shown in FIG. 5 includes five items of indication information, namely the first to fifth indication information, in step S503, the second node may send these five items of indication information to the third node. The second node may also send these five items of indication information, for example, at least one of the second and fifth indication information, to the third node.

[0315] In one possible implementation, the AI ​​system involved in the above embodiments (e.g., the first AI system or the second AI system) may be used to perform a first process on first data to obtain second data, and then perform a second process on the second data;

[0316] When the learning architecture includes federated learning (or when the learning architecture is federated learning), the first data includes global model parameters and / or gradients, the first processing includes local training processing, the second data includes local model parameters and / or gradients, and the second processing includes fusion processing; it can be understood that when the learning architecture is federated learning, the second processing may include fusion processing of local model parameters, and / or, the second processing may include fusion processing of local model gradients;

[0317] When the learning architecture includes federated distillation (or when the learning architecture is federated distillation), the first data includes global logits, the first processing includes local training processing, the second data includes local logits, and the second processing includes fusion processing; it can be understood that when the learning architecture is federated distillation, the second processing may include fusion processing of local logits;

[0318] When the learning architecture includes decentralized learning (or when the learning architecture is decentralized learning), the first data includes parameters and / or gradients of a first local model, the first processing includes a fusion processing and a local training processing, the second data includes parameters and / or gradients of a second local model, and the second processing includes a fusion processing and a local training processing;

[0319] When the learning architecture includes meta-learning (or when the learning architecture is meta-learning), the first data includes meta-model parameters and / or gradients, the first processing includes support set training processing and test set gradient calculation processing, the second data includes test set gradients and / or losses, and the second processing includes gradient fusion processing and meta-model update processing;

[0320] When the learning architecture includes segmentation learning (or, when the learning architecture is segmentation learning), the first data includes segmentation layer inference results, the first processing includes inference, gradient calculation, back transfer and parameter update processing, the second data includes segmentation layer gradient, and the second processing includes continued gradient transfer processing and parameter update processing.

[0321] Alternatively, the implementation of the above-mentioned first data, first processing, second data and second processing can also be achieved through the contents of Table 2 below.

[0322] Table 2

[0323] Thus, in different learning architectures, either the first AI system or the second AI system can perform the first processing on the first data to obtain the second data, and then perform the second processing on the second data. Furthermore, the above implementation provides specific implementations of the first data, second data, first processing, and second processing in different learning architectures.

[0324] As an implementation example, the implementation of the first data, the first processing, the second data, and the second processing will be exemplarily described below in combination with more implementation examples shown in Figures 7a and 7b.

[0325] Here we take the scenario shown in Figure 3a as an example. The learning system includes AI system 1, AI system 2, AI system 3, and AI system 4 in Figure 3a. Node 1 serves as both the control node of the learning system (i.e., the first node) and the control node of AI system 1; node 3 serves as the control node of AI system 2 (i.e., the second node), and node 5 can serve as one of the non-control nodes of AI system 2 (i.e., the third node). The following describes the interaction process between the first node, the second node, and the third node for the first data and the second data. AI system 1 can be understood as the upper-level AI system of AI system 2, or in other words, AI system 2 can be understood as the lower-level AI system of AI system 1.

[0326] As shown in FIG7 a , the data interaction process between the nodes in the first AI system and the nodes in the second AI system includes the following steps.

[0327] S701. The third node sends first data, and correspondingly, the second node receives the first data. The first data in this step is the first data in the second AI system.

[0328] S702. The second node performs a first process based on the received first data. The first process in this step is the first process in the second AI system.

[0329] S703. The second node sends the first data. Correspondingly, the first node receives the first data. The first data in this step is the first data in the first AI system.

[0330] S704. The first node performs a first process based on the received first data. The first process in this step is the first process in the first AI system.

[0331] S705. The first node sends the second data. Correspondingly, the second node receives the second data. The second data in this step is the first data in the first AI system.

[0332] S706. The second node performs a second process based on the received second data. The second process in this step is the second process in the first AI system.

[0333] S707. The second node sends the second data. Correspondingly, the third node receives the second data. The second data in this step is the first data in the second AI system.

[0334] S708. The third node performs a second process based on the received second data. The second process in this step is the second process in the second AI system.

[0335] It can be understood that, as nodes in the first AI system, the first node and the second node can implement data processing and interaction through the process of steps S703 / S704 / S705 / S706 above; as nodes in the second AI system, the second node and the third node can implement data processing and interaction through the process of steps S701 / S702 / S707 / S708 above.

[0336] Furthermore, the first and second AI systems can also perform AI tasks for scalable distributed learning. For example, these AI tasks may include any of: scalable distributed learning federated learning, scalable distributed learning federated distillation, scalable distributed learning decentralized learning, scalable distributed learning meta-learning, or scalable distributed learning segmented learning. Taking the example of the second AI system performing the AI ​​task of scalable distributed learning, different nodes in the second AI system may focus on different subtasks of the AI ​​task performed by the second AI system. Alternatively, different nodes in the second AI system may focus on different submodels within the AI ​​model used in the AI ​​task performed by the second AI system. In the following example, the second node in the second AI system focuses on the first submodel, and the third node in the second AI system focuses on the second submodel.

[0337] For example, the AI ​​task performed by the second AI system could be an autonomous driving task. Generally, image recognition is a necessary component of autonomous driving tasks. For example, image recognition tasks such as human image recognition, license plate image recognition, and obstacle image recognition can be subtasks of this autonomous driving task. Accordingly, the AI ​​task performed by the first sub-model focused on by the second node or the second sub-model focused on by the third node can be any of these image recognition tasks.

[0338] For another example, the AI ​​task performed by the second AI system may be an image classification task (or N-classification task). Generally, image classification may include image classification between different animals (e.g., classifying images of cats and dogs), image classification between different plants (e.g., classifying images of herbaceous plants and woody plants), etc. In other words, tasks such as image classification between different animals and image classification between different plants may be subtasks of the image classification task. Accordingly, the AI ​​task performed by the first sub-model focused on by the second node or the second sub-model focused on by the third node may be any of the image classification tasks between different animals and between different plants.

[0339] For example, the AI ​​task performed by the second AI system may be a machine translation task. Generally, word segmentation tasks, grammatical analysis tasks, sentence rewriting tasks, etc. may be subtasks of this machine translation task. Accordingly, the AI ​​task performed by the first sub-model focused on by the second node or the second sub-model focused on by the third node may be any of these tasks.

[0340] The following example illustrates the implementation process of the second AI system performing an AI task of scalable distributed learning using the implementation example shown in Figure 7b. As shown in Figure 7b, the process includes the following steps.

[0341] S801. The third node sends first data, and correspondingly, the second node receives the first data. The first data in this step is the first data related to the first sub-model.

[0342] S802. The second node performs a first process based on the received first data. The first process in this step is the first process related to the first sub-model.

[0343] S803. The second node sends the second data, and correspondingly, the third node receives the second data. The second data in this step is the first data related to the first sub-model.

[0344] S804. The third node performs a second process based on the received second data. The second process in this step is a second process related to the first sub-model.

[0345] S805. The third node sends the first data, and correspondingly, the second node receives the first data. The first data in this step is the first data related to the second sub-model.

[0346] S806. The second node performs a first process based on the received first data. The first process in this step is the first process related to the second sub-model.

[0347] S807. The second node sends the second data, and correspondingly, the third node receives the second data. The second data in this step is the first data related to the second sub-model.

[0348] S808. The third node performs a second process based on the received second data. The second process in this step is a second process related to the second sub-model.

[0349] It is understood that, as nodes in the second AI system, the second and third nodes can interact with data related to the first sub-model through the processes of steps S801 to S804 above, thereby implementing model updates, training, processing, iteration, etc. for the first sub-model. As nodes in the second AI system, the second and third nodes can interact with data related to the second sub-model through the processes of steps S805 to S808 above, thereby implementing model updates, training, processing, iteration, etc. for the second sub-model.

[0350] In the implementations shown in Figures 7a and 7b, the learning architecture used by the first AI system differs from the learning architecture used by the second AI system. This allows the computing power of communication nodes to be applied to AI tasks within the learning system and improves the implementation flexibility of different AI systems within the same learning system. Furthermore, different AI systems within the same learning system can execute AI tasks using different learning architectures, providing a wider range of learning architectures to choose from that better match the needs and capabilities of the nodes participating in the learning system. Furthermore, the learning system can also gain the benefits of multiple learning architectures, simplifying implementation complexity.

[0351] Optionally, FIG3c illustrates an example in which a control node sends AI configuration information to the control nodes of each AI system. In actual application of the solution, the control node may send AI configuration information to some or all nodes of each AI system (e.g., the scenario shown in FIG3b ). The implementation process of the latter may refer to the above description.

[0352] Referring to Figure 8, an embodiment of the present application provides a communication device 800. The communication device 800 can implement the functions of the first node (the first node is a terminal device or a network device) in the above method embodiment, and thus can also achieve the beneficial effects of the above method embodiment. In the embodiment of the present application, the communication device 800 can be the first node, or it can be an integrated circuit or component within the first node, such as a chip. The following embodiments are described using the communication device 800 as the first node as an example.

[0353] In one possible implementation, when the apparatus 800 is configured to execute the method executed by the first node in any of the aforementioned embodiments, the apparatus 800 includes a processing unit 801 and a transceiver unit 802; the processing unit 801 is configured to determine first information and second information, the first information being configured to indicate AI configuration information of a first artificial intelligence (AI) system, and the second information being configured to indicate AI configuration information of a second AI system; wherein the first AI system and the second AI system belong to the same learning system, the first AI system and the second AI system include at least one identical node, and the learning architecture of the first AI system and the learning architecture of the second AI system are different; and the transceiver unit 802 is configured to send the first information and the second information.

[0354] In a possible implementation, the second AI system belongs to the first AI system; and the transceiver unit 802 is configured to send the first information and the second information, including: the transceiver unit 802 is configured to send the first information and the second information to a control node of the first AI system.

[0355] In one possible implementation, the transceiver unit 802 being configured to send the first information and the second information includes: the transceiver unit 802 being configured to send the first information to a control node of the first AI system; and the transceiver unit 802 being configured to send the second information to a control node of the second AI system.

[0356] In one possible implementation, the AI ​​configuration information of the first AI system includes at least one of the following:

[0357] First indication information, used to indicate nodes in the AI ​​system that participate in the AI ​​task;

[0358] Second indication information, used to indicate the role of the node participating in the AI ​​task in the AI ​​system;

[0359] The third indication information is used to indicate the learning architecture of the AI ​​system;

[0360] Fourth instruction information, used to indicate an AI task of the AI ​​system;

[0361] Fifth indication information, used to indicate the communication resources of the AI ​​system;

[0362] And / or, the AI ​​configuration information of the second AI system includes at least one of the following:

[0363] First indication information, used to indicate nodes in the AI ​​system that participate in the AI ​​task;

[0364] Second indication information, used to indicate the role of the node participating in the AI ​​task in the AI ​​system;

[0365] The third indication information is used to indicate the learning architecture of the AI ​​system;

[0366] Fourth instruction information, used to indicate an AI task of the AI ​​system;

[0367] The fifth indication information is used to indicate the communication resources of the AI ​​system.

[0368] In one possible implementation, the learning architecture includes: federated learning, federated distillation, decentralized learning, meta-learning, or segmented learning.

[0369] In one possible implementation, the first AI system or the second AI system is configured to perform a first process on the first data to obtain second data, and then perform a second process on the second data;

[0370] When the learning architecture includes federated learning (or when the learning architecture is federated learning), the first data includes global model parameters and / or gradients, the first process includes a local training process, the second data includes local model parameters and / or gradients, and the second process includes a fusion process;

[0371] When the learning architecture includes federated distillation (or when the learning architecture is federated distillation), the first data includes global logits, the first processing includes local training processing, the second data includes local logits, and the second processing includes fusion processing;

[0372] When the learning architecture includes decentralized learning (or when the learning architecture is decentralized learning), the first data includes parameters and / or gradients of a first local model, the first processing includes a fusion processing and a local training processing, the second data includes parameters and / or gradients of a second local model, and the second processing includes a fusion processing and a local training processing;

[0373] When the learning architecture includes meta-learning (or when the learning architecture is meta-learning), the first data includes meta-model parameters and / or gradients, the first processing includes support set training processing and test set gradient calculation processing, the second data includes test set gradients and / or losses, and the second processing includes gradient fusion processing and meta-model update processing;

[0374] When the learning architecture includes segmentation learning (or, when the learning architecture is segmentation learning), the first data includes segmentation layer inference results, the first processing includes inference, gradient calculation, back transfer and parameter update processing, the second data includes segmentation layer gradient, and the second processing includes continued gradient transfer processing and parameter update processing.

[0375] In one possible implementation, the first AI system and the second AI system are used to perform the same AI task, or the first AI system and the second AI system are used to perform different subtasks of the same AI task.

[0376] In one possible implementation, when the apparatus 800 is configured to execute the method executed by the second node in any of the aforementioned embodiments, the apparatus 800 includes a processing unit 801 and a transceiver unit 802; the transceiver unit 802 is configured to receive first information indicating AI configuration information of a first artificial intelligence (AI) system; wherein the first AI system and the second AI system belong to the same distributed learning system, the first AI system and the second AI system include at least one common node, and the learning architecture of the first AI system and the learning architecture of the second AI system are different; and the processing unit 801 is configured to execute the AI ​​task based on the AI ​​configuration information of the first AI system.

[0377] In one possible implementation, the second AI system belongs to the first AI system; the transceiver unit 802 is further configured to receive second information indicating AI configuration information of the second AI system; and the transceiver unit 802 is further configured to send part or all of the AI ​​configuration information of the second AI system to a control node of the second AI system.

[0378] In one possible implementation, the AI ​​configuration information of the first AI system includes at least one of the following:

[0379] First indication information, used to indicate nodes in the AI ​​system that participate in the AI ​​task;

[0380] Second indication information, used to indicate the role of the node participating in the AI ​​task in the AI ​​system;

[0381] The third indication information is used to indicate the learning architecture of the AI ​​system;

[0382] Fourth instruction information, used to indicate an AI task of the AI ​​system;

[0383] Fifth indication information, used to indicate the communication resources of the AI ​​system;

[0384] And / or, the AI ​​configuration information of the second AI system includes at least one of the following:

[0385] First indication information, used to indicate nodes in the AI ​​system that participate in the AI ​​task;

[0386] Second indication information, used to indicate the role of the node participating in the AI ​​task in the AI ​​system;

[0387] The third indication information is used to indicate the learning architecture of the AI ​​system;

[0388] Fourth instruction information, used to indicate an AI task of the AI ​​system;

[0389] The fifth indication information is used to indicate the communication resources of the AI ​​system.

[0390] In one possible implementation, the learning architecture includes: federated learning, federated distillation, decentralized learning, meta-learning, or segmented learning.

[0391] In one possible implementation, the first AI system or the second AI system is configured to perform a first process on the first data to obtain second data, and then perform a second process on the second data;

[0392] When the learning architecture includes federated learning (or when the learning architecture is federated learning), the first data includes global model parameters and / or gradients, the first process includes a local training process, the second data includes local model parameters and / or gradients, and the second process includes a fusion process;

[0393] When the learning architecture includes federated distillation (or when the learning architecture is federated distillation), the first data includes global logits, the first processing includes local training processing, the second data includes local logits, and the second processing includes fusion processing;

[0394] When the learning architecture includes decentralized learning (or when the learning architecture is decentralized learning), the first data includes parameters and / or gradients of a first local model, the first processing includes a fusion processing and a local training processing, the second data includes parameters and / or gradients of a second local model, and the second processing includes a fusion processing and a local training processing;

[0395] When the learning architecture includes meta-learning (or when the learning architecture is meta-learning), the first data includes meta-model parameters and / or gradients, the first processing includes support set training processing and test set gradient calculation processing, the second data includes test set gradients and / or losses, and the second processing includes gradient fusion processing and meta-model update processing;

[0396] When the learning architecture includes segmentation learning (or, when the learning architecture is segmentation learning), the first data includes segmentation layer inference results, the first processing includes inference, gradient calculation, back transfer and parameter update processing, the second data includes segmentation layer gradient, and the second processing includes continued gradient transfer processing and parameter update processing.

[0397] In one possible implementation, the first AI system and the second AI system are used to perform the same AI task, or the first AI system and the second AI system are used to perform different subtasks of the same AI task.

[0398] It should be noted that, for details on the information execution process of the units of the above-mentioned communication device 800, please refer to the description in the method embodiment shown above in this application, and no further details will be given here.

[0399] Please refer to Fig. 9, which is another schematic structural diagram of a communication device 900 provided in this application. The communication device 900 at least includes an input and output interface 902. The communication device 900 may be a chip or an integrated circuit.

[0400] Optionally, the communication device further includes a logic circuit 901 .

[0401] The transceiver unit 802 shown in FIG8 may be a communication interface, which may be the input / output interface 902 in FIG9 , which may include an input interface and an output interface. Alternatively, the communication interface may be a transceiver circuit, which may include an input interface circuit and an output interface circuit.

[0402] Optionally, logic circuit 901 is configured to determine first information and second information, where the first information is used to indicate AI configuration information of a first artificial intelligence (AI) system, and the second information is used to indicate AI configuration information of a second AI system; wherein the first AI system and the second AI system belong to the same learning system, the first AI system and the second AI system include at least one identical node, and the learning architecture of the first AI system and the learning architecture of the second AI system are different; and input / output interface 902 is configured to transmit the first information and the second information.

[0403] Optionally, the input / output interface 902 is configured to receive first information indicating AI configuration information of a first artificial intelligence (AI) system; wherein the first AI system and the second AI system belong to the same learning system, the first AI system and the second AI system include at least one identical node, and the learning architecture of the first AI system is different from the learning architecture of the second AI system; and the logic circuit 901 is configured to execute the AI ​​task based on the AI ​​configuration information of the first AI system.

[0404] The logic circuit 901 and the input / output interface 902 may also execute other steps executed by the first node or the second node in any embodiment and achieve corresponding beneficial effects, which will not be described in detail here.

[0405] In a possible implementation, the processing unit 801 shown in FIG. 8 may be the logic circuit 901 in FIG. 9 .

[0406] Optionally, the logic circuit 901 may be a processing device, and the functions of the processing device may be partially or entirely implemented by software. The functions of the processing device may be partially or entirely implemented by software.

[0407] Optionally, the processing device may include a memory and a processor, wherein the memory is used to store a computer program, and the processor reads and executes the computer program stored in the memory to perform corresponding processing and / or steps in any one of the method embodiments.

[0408] Alternatively, the processing device may include only a processor. A memory for storing the computer program is located outside the processing device, and the processor is connected to the memory via circuits / wires to read and execute the computer program stored in the memory. The memory and processor may be integrated or physically separate.

[0409] Optionally, the processing device may be one or more chips, or one or more integrated circuits. For example, the processing device may be one or more field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), system-on-chips (SoCs), central processor units (CPUs), network processors (NPs), digital signal processors (DSPs), microcontroller units (MCUs), programmable logic devices (PLDs), or other integrated chips, or any combination of the above chips or processors.

[0410] Please refer to Figure 10, which shows the communication device 1000 involved in the above-mentioned embodiments provided in an embodiment of the present application. The communication device 1000 can specifically be a communication device serving as a terminal device in the above-mentioned embodiments. The example shown in Figure 10 is that the terminal device is implemented through the terminal device (or a component in the terminal device).

[0411] Herein, a possible logical structure diagram of the communication device 1000 is shown. The communication device 1000 may include but is not limited to at least one processor 1001 and a communication port 1002 .

[0412] The transceiver unit 802 shown in FIG8 may be a communication interface, which may be the communication port 1002 in FIG10 , which may include an input interface and an output interface. Alternatively, the communication port 1002 may be a transceiver circuit, which may include an input interface circuit and an output interface circuit.

[0413] Further optionally, the device may also include at least one of a memory 1003 and a bus 1004. In an embodiment of the present application, the at least one processor 1001 is used to control and process the actions of the communication device 1000.

[0414] In addition, the processor 1001 can be a central processing unit, a general-purpose processor, a digital signal processor, an application-specific integrated circuit, a field programmable gate array or other programmable logic device, a transistor logic device, a hardware component, or any combination thereof. It can implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. The processor can also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a digital signal processor and a microprocessor, and so on. Those skilled in the art will clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0415] It should be noted that the communication device 1000 shown in Figure 10 can be specifically used to implement the steps implemented by the terminal device in the aforementioned method embodiment and achieve the corresponding technical effects of the terminal device. The specific implementation methods of the communication device shown in Figure 10 can refer to the description in the aforementioned method embodiment and will not be repeated here.

[0416] Please refer to Figure 11, which is a structural diagram of the communication device 1100 involved in the above-mentioned embodiments provided in an embodiment of the present application. The communication device 1100 can specifically be a communication device as a network device in the above-mentioned embodiments. The example shown in Figure 11 is that the network device is implemented through the network device (or a component in the network device), wherein the structure of the communication device can refer to the structure shown in Figure 11.

[0417] The communication device 1100 includes at least one processor 1111 and at least one network interface 1114. Further optionally, the communication device also includes at least one memory 1112, at least one transceiver 1113 and one or more antennas 1115. The processor 1111, the memory 1112, the transceiver 1113 and the network interface 1114 are connected, for example, via a bus. In an embodiment of the present application, the connection may include various interfaces, transmission lines or buses, etc., which are not limited in this embodiment. The antenna 1115 is connected to the transceiver 1113. The network interface 1114 is used to enable the communication device to communicate with other communication devices through a communication link. For example, the network interface 1114 may include a network interface between the communication device and the core network device, such as an S1 interface, and the network interface may include a network interface between the communication device and other communication devices (such as other network devices or core network devices), such as an X2 or Xn interface.

[0418] The transceiver unit 802 shown in FIG8 may be a communication interface, which may be the network interface 1114 in FIG11 , which may include an input interface and an output interface. Alternatively, the network interface 1114 may be a transceiver circuit, which may include an input interface circuit and an output interface circuit.

[0419] Processor 1111 is primarily used to process communication protocols and communication data, control the entire communication device, execute software programs, and process software program data, for example, to support the communication device in performing the actions described in the embodiments. The communication device may include a baseband processor and a central processing unit. The baseband processor is primarily used to process communication protocols and communication data, while the central processing unit is primarily used to control the entire terminal device, execute software programs, and process software program data. Processor 1111 in Figure 11 may integrate the functions of both a baseband processor and a central processing unit. Those skilled in the art will appreciate that the baseband processor and the central processing unit may also be independent processors interconnected via a bus or other technology. Those skilled in the art will appreciate that a terminal device may include multiple baseband processors to accommodate different network standards, multiple central processing units to enhance its processing capabilities, and various components of the terminal device may be connected via various buses. The baseband processor may also be referred to as a baseband processing circuit or a baseband processing chip. The central processing unit may also be referred to as a central processing circuit or a central processing chip. The functionality for processing communication protocols and communication data may be built into the processor or stored in memory as a software program, which is executed by the processor to implement the baseband processing functionality.

[0420] The memory is primarily used to store software programs and data. Memory 1112 can exist independently and be connected to processor 1111. Alternatively, memory 1112 can be integrated with processor 1111, for example, within a single chip. Memory 1112 can store program code for executing the technical solutions of the embodiments of the present application, and execution is controlled by processor 1111. The various computer program codes executed can also be considered drivers for processor 1111.

[0421] Figure 11 shows only one memory and one processor. In an actual terminal device, there may be multiple processors and multiple memories. The memory may also be referred to as a storage medium or a storage device. The memory may be a storage element on the same chip as the processor, i.e., an on-chip storage element, or an independent storage element, which is not limited in the present embodiment.

[0422] The transceiver 1113 can be used to support the reception or transmission of radio frequency signals between the communication device and the terminal. The transceiver 1113 can be connected to the antenna 1115. The transceiver 1113 includes a transmitter Tx and a receiver Rx. Specifically, one or more antennas 1115 can receive radio frequency signals. The receiver Rx of the transceiver 1113 is used to receive the radio frequency signal from the antenna, convert the radio frequency signal into a digital baseband signal or a digital intermediate frequency signal, and provide the digital baseband signal or digital intermediate frequency signal to the processor 1111 so that the processor 1111 can further process the digital baseband signal or digital intermediate frequency signal, such as demodulation and decoding. In addition, the transmitter Tx in the transceiver 1113 is also used to receive a modulated digital baseband signal or digital intermediate frequency signal from the processor 1111, convert the modulated digital baseband signal or digital intermediate frequency signal into a radio frequency signal, and transmit the radio frequency signal through one or more antennas 1115. Specifically, the receiver Rx can selectively perform one or more stages of down-mixing and analog-to-digital conversion on the RF signal to obtain a digital baseband signal or a digital intermediate frequency signal. The order of the down-mixing and analog-to-digital conversion processes is adjustable. The transmitter Tx can selectively perform one or more stages of up-mixing and digital-to-analog conversion on the modulated digital baseband signal or digital intermediate frequency signal to obtain a RF signal. The order of the up-mixing and digital-to-analog conversion processes is adjustable. The digital baseband signal and the digital intermediate frequency signal may be collectively referred to as digital signals.

[0423] The transceiver 1113 may also be referred to as a transceiver unit, a transceiver, a transceiver device, etc. Optionally, a device in the transceiver unit that implements a receiving function may be referred to as a receiving unit, and a device in the transceiver unit that implements a transmitting function may be referred to as a transmitting unit. That is, the transceiver unit includes a receiving unit and a transmitting unit. The receiving unit may also be referred to as a receiver, an input port, a receiving circuit, etc., and the transmitting unit may be referred to as a transmitter, a transmitter, or a transmitting circuit, etc.

[0424] It should be noted that the communication device 1100 shown in Figure 11 can be specifically used to implement the steps implemented by the network device in the aforementioned method embodiment, and to achieve the corresponding technical effects of the network device. The specific implementation methods of the communication device 1100 shown in Figure 11 can refer to the description in the aforementioned method embodiment, and will not be repeated here one by one.

[0425] Please refer to FIG12 , which is a schematic structural diagram of the communication device involved in the above-mentioned embodiment provided in an embodiment of the present application.

[0426] It can be understood that the communication device 120 includes, for example, modules, units, elements, circuits, or interfaces, which are appropriately configured together to implement the technical solutions provided in this application. The communication device 120 can be the RAN node, terminal, core network device, or other network device described above, or a component (e.g., a chip) in these devices, used to implement the method described in the following method embodiment. The communication device 120 includes one or more processors 121. The processor 121 can be a general-purpose processor or a dedicated processor. For example, it can be a baseband processor or a central processing unit. The baseband processor can be used to process communication protocols and communication data, and the central processing unit can be used to control the communication device (e.g., RAN node, terminal, or chip), execute software programs, and process data of software programs.

[0427] Optionally, in one design, the processor 121 may include a program 123 (sometimes also referred to as code or instructions), which may be executed on the processor 121 to cause the communication device 120 to perform the methods described in the following embodiments. In yet another possible design, the communication device 120 includes a circuit (not shown in FIG. 12 ), which is configured to implement the functions of determining the first information and / or determining the second information in the above embodiments.

[0428] Optionally, the communication device 120 may include one or more memories 122 on which a program 124 (sometimes also referred to as code or instructions) is stored. The program 124 can be run on the processor 121, so that the communication device 120 executes the method described in the above method embodiment.

[0429] Optionally, the processor 121 and / or the memory 122 may include AI modules 127 and 128, which are used to implement AI-related functions. The AI ​​module can be implemented through software, hardware, or a combination of software and hardware. For example, the AI ​​module may include a wireless intelligent control (RIC) module. For example, the AI ​​module may be a near real-time RIC or a non-real-time RIC.

[0430] Optionally, data may be stored in the processor 121 and / or the memory 122. The processor and the memory may be provided separately or integrated together.

[0431] Optionally, the communication device 120 may further include a transceiver 125 and / or an antenna 126. The processor 121 may also be sometimes referred to as a processing unit, and controls the communication device (e.g., a RAN node or terminal). The transceiver 125 may also be sometimes referred to as a transceiver unit, a transceiver, a transceiver circuit, or a transceiver, and is configured to implement the transceiver functions of the communication device through the antenna 126.

[0432] The transceiver unit 802 shown in FIG8 may be a communication interface, which may be the transceiver 125 in FIG12 . The transceiver 125 may include an input interface and an output interface. Alternatively, the transceiver 125 may be a transceiver circuit, which may include an input interface circuit and an output interface circuit.

[0433] An embodiment of the present application also provides a computer-readable storage medium, which is used to store one or more computer-executable instructions. When the computer-executable instructions are executed by a processor, the processor executes the method described in the possible implementation of the first node or the second node in the above embodiment.

[0434] An embodiment of the present application also provides a computer program product (or computer program). When the computer program product is executed by the processor, the processor executes the method of the possible implementation of the above-mentioned first node or second node.

[0435] An embodiment of the present application also provides a chip system, which includes at least one processor for supporting a communication device to implement the functions involved in the possible implementation methods of the above-mentioned communication device. Optionally, the chip system also includes an interface circuit, which provides program instructions and / or data to the at least one processor. In one possible design, the chip system may also include a memory for storing program instructions and data necessary for the communication device. The chip system can be composed of a chip, or it can include a chip and other discrete devices, wherein the communication device can specifically be the first node or the second node in the aforementioned method embodiment.

[0436] An embodiment of the present application also provides a communication system, wherein the network system architecture includes a first node and a second node in any of the above embodiments, wherein the first node may be a terminal device or a network device, and the second node may also be a terminal device or a network device.

[0437] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0438] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0439] In addition, the functional units in the various embodiments of the present application can be integrated into a processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of a software functional unit. If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the contributing part or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the various embodiments of the present application. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

Claims

1. A communication method, characterized in that: include: Determine first information and second information, wherein the first information is used to indicate AI configuration information of a first artificial intelligence (AI) system, and the second information is used to indicate AI configuration information of a second AI system; wherein the first AI system and the second AI system belong to the same learning system, the first AI system and the second AI system include at least one common node, and the learning architecture of the first AI system is different from the learning architecture of the second AI system; The first information and the second information are sent.

2. The method according to claim 1, characterized in that The second AI system belongs to the first AI system; and the sending the first information and the second information includes: The first information and the second information are sent to a control node of the first AI system.

3. The method according to claim 1, characterized in that The sending the first information and the second information comprises: Sending the first information to a control node of the first AI system; Send the second information to a control node of the second AI system.

4. The method according to any one of claims 1 to 3, characterized in that: The AI ​​configuration information of the first AI system includes at least one of the following: First indication information, used to indicate nodes participating in the AI ​​task in the AI ​​system; The second indication information is used to indicate the role of the node participating in the AI ​​task in the AI ​​system; The third indication information is used to indicate the learning architecture of the AI ​​system; Fourth indication information, used to indicate an AI task of the AI ​​system; Fifth indication information, used to indicate the communication resources of the AI ​​system; and / or, The AI ​​configuration information of the second AI system includes at least one of the following: First indication information, used to indicate nodes participating in the AI ​​task in the AI ​​system; The second indication information is used to indicate the role of the node participating in the AI ​​task in the AI ​​system; The third indication information is used to indicate the learning architecture of the AI ​​system; Fourth indication information, used to indicate an AI task of the AI ​​system; The fifth indication information is used to indicate the communication resources of the AI ​​system.

5. The method according to any one of claims 1 to 4, characterized in that: The learning framework includes: Federated learning, federated distillation, decentralized learning, meta-learning, or segmented learning.

6. The method according to any one of claims 1 to 5, characterized in that: The first AI system or the second AI system is configured to perform a second processing on the second data after performing a first processing on the first data to obtain the second data; When the learning architecture includes federated learning, the first data includes global model parameters and / or gradients, the first processing includes local training processing, the second data includes local model parameters and / or gradients, and the second processing includes fusion processing; When the learning architecture includes federated distillation, the first data includes global pairwise logits, the first processing includes local training processing, the second data includes local logits, and the second processing includes fusion processing; When the learning architecture includes decentralized learning, the first data includes parameters and / or gradients of a first local model, the first processing includes fusion processing and local training processing, the second data includes parameters and / or gradients of a second local model, and the second processing includes fusion processing and local training processing; When the learning architecture includes meta-learning, the first data includes meta-model parameters and / or gradients, the first processing includes support set training processing and test set gradient calculation processing, the second data includes test set gradients and / or losses, and the second processing includes gradient fusion processing and meta-model update processing; When the learning architecture includes segmentation learning, the first data includes segmentation layer inference results, the first processing includes inference, gradient calculation, back transfer and parameter update processing, the second data includes segmentation layer gradient, and the second processing includes continued gradient transfer processing and parameter update processing.

7. The method according to any one of claims 1 to 6, characterized in that: The first AI system and the second AI system are used to perform the same AI task, or the first AI system and the second AI system are used to perform different subtasks in the same AI task.

8. A communication method, characterized in that: include: Receive first information, where the first information is used to indicate AI configuration information of a first artificial intelligence (AI) system; wherein the first AI system and the second AI system belong to the same distributed learning system, the first AI system and the second AI system include at least one common node, and the learning architecture of the first AI system is different from the learning architecture of the second AI system; The AI ​​task is executed based on the AI ​​configuration information of the first AI system.

9. The method according to claim 8, characterized in that The second AI system belongs to the first AI system; the method further includes: receiving second information, where the second information is used to indicate AI configuration information of a second AI system; Send part or all of the AI ​​configuration information of the second AI system to a control node of the second AI system.

10. The method according to claim 9, characterized in that The AI ​​configuration information of the second AI system includes at least one of the following: First indication information, used to indicate nodes participating in the AI ​​task in the AI ​​system; The second indication information is used to indicate the role of the node participating in the AI ​​task in the AI ​​system; The third indication information is used to indicate the learning architecture of the AI ​​system; Fourth indication information, used to indicate an AI task of the AI ​​system; The fifth indication information is used to indicate the communication resources of the AI ​​system.

11. The method according to any one of claims 8 to 10, characterized in that: The AI ​​configuration information of the first AI system includes at least one of the following: First indication information, used to indicate nodes participating in the AI ​​task in the AI ​​system; The second indication information is used to indicate the role of the node participating in the AI ​​task in the AI ​​system; The third indication information is used to indicate the learning architecture of the AI ​​system; Fourth indication information, used to indicate an AI task of the AI ​​system; The fifth indication information is used to indicate the communication resources of the AI ​​system.

12. The method according to any one of claims 8 to 11, characterized in that The learning framework includes: Federated learning, federated distillation, decentralized learning, meta-learning, or segmented learning.

13. The method according to any one of claims 8 to 12, characterized in that: The first AI system or the second AI system is configured to perform a second processing on the second data after performing a first processing on the first data to obtain the second data; When the learning architecture includes federated learning, the first data includes global model parameters and / or gradients, the first processing includes local training processing, the second data includes local model parameters and / or gradients, and the second processing includes fusion processing; When the learning architecture includes federated distillation, the first data includes global pairwise logits, the first processing includes local training processing, the second data includes local logits, and the second processing includes fusion processing; When the learning architecture includes decentralized learning, the first data includes parameters and / or gradients of a first local model, the first processing includes fusion processing and local training processing, the second data includes parameters and / or gradients of a second local model, and the second processing includes fusion processing and local training processing; When the learning framework includes meta-learning, the first data includes meta-model parameters and / or gradients, and the first processing includes support set Training processing and test set gradient calculation processing, the second data includes test set gradient and / or loss, the second processing includes gradient fusion processing and meta-model update processing; When the learning architecture includes segmentation learning, the first data includes segmentation layer inference results, the first processing includes inference, gradient calculation, back transfer and parameter update processing, the second data includes segmentation layer gradient, and the second processing includes continued gradient transfer processing and parameter update processing.

14. The method according to any one of claims 8 to 13, characterized in that The first AI system and the second AI system are used to perform the same AI task, or the first AI system and the second AI system are used to perform different subtasks in the same AI task.

15. A communication device, characterized in that: including a processing unit and a transceiver unit; The processing unit is used to determine first information and second information, the first information is used to indicate AI configuration information of a first artificial intelligence AI system, and the second information is used to indicate AI configuration information of a second AI system; wherein the first AI system and the second AI system belong to the same learning system, the first AI system and the second AI system include at least one common node, and the learning architecture of the first AI system is different from the learning architecture of the second AI system; The transceiver unit is used to send the first information and the second information.

16. The device according to claim 15, characterized in that The second AI system belongs to the first AI system; the transceiver unit is used to send the first information and the second information, including: The transceiver unit is used to send the first information and the second information to a control node of the first AI system.

17. The device according to claim 15, characterized in that The transceiver unit is used to send the first information and the second information, including: The transceiver unit is used to send the first information to the control node of the first AI system; The transceiver unit is used to send the second information to the control node of the second AI system.

18. The device according to any one of claims 15 to 17, characterized in that The AI ​​configuration information of the first AI system includes at least one of the following: First indication information, used to indicate nodes participating in the AI ​​task in the AI ​​system; The second indication information is used to indicate the role of the node participating in the AI ​​task in the AI ​​system; The third indication information is used to indicate the learning architecture of the AI ​​system; Fourth indication information, used to indicate an AI task of the AI ​​system; Fifth indication information, used to indicate the communication resources of the AI ​​system; The AI ​​configuration information of the second AI system includes at least one of the following: First indication information, used to indicate nodes participating in the AI ​​task in the AI ​​system; The second indication information is used to indicate the role of the node participating in the AI ​​task in the AI ​​system; The third indication information is used to indicate the learning architecture of the AI ​​system; Fourth indication information, used to indicate an AI task of the AI ​​system; The fifth indication information is used to indicate the communication resources of the AI ​​system.

19. The device according to any one of claims 15 to 18, characterized in that The learning framework includes: Federated learning, federated distillation, decentralized learning, meta-learning, or segmented learning.

20. The device according to any one of claims 15 to 19, characterized in that The first AI system or the second AI system is configured to perform a second processing on the second data after performing a first processing on the first data to obtain the second data; When the learning architecture includes federated learning, the first data includes global model parameters and / or gradients, the first processing includes local training processing, the second data includes local model parameters and / or gradients, and the second processing includes fusion processing; When the learning architecture includes federated distillation, the first data includes global pairwise logits, the first processing includes local training processing, the second data includes local logits, and the second processing includes fusion processing; When the learning architecture includes decentralized learning, the first data includes parameters and / or gradients of a first local model, the first processing includes fusion processing and local training processing, the second data includes parameters and / or gradients of a second local model, and the second processing includes fusion processing and local training processing; When the learning architecture includes meta-learning, the first data includes meta-model parameters and / or gradients, the first processing includes support set training processing and test set gradient calculation processing, the second data includes test set gradients and / or losses, and the second processing includes gradient fusion processing and meta-model update processing; When the learning architecture includes segmentation learning, the first data includes segmentation layer inference results, the first processing includes inference, gradient calculation, back transfer and parameter update processing, the second data includes segmentation layer gradient, and the second processing includes continued gradient transfer processing and parameter update processing.

21. The device according to any one of claims 15 to 20, characterized in that The first AI system and the second AI system are used to perform the same AI task, or the first AI system and the second AI system are used to perform different subtasks in the same AI task.

22. A communication device, characterized in that: including a processing unit and a transceiver unit; The transceiver unit is used to receive first information, where the first information is used to indicate AI configuration information of a first artificial intelligence AI system; wherein the first AI system and the second AI system belong to the same distributed learning system, the first AI system and the second AI system include at least one common node, and the learning architecture of the first AI system is different from the learning architecture of the second AI system; The processing unit is configured to execute an AI task based on the AI ​​configuration information of the first AI system.

23. The device according to claim 22, characterized in that The second AI system belongs to the first AI system; The transceiver unit is further used to receive second information, where the second information is used to indicate AI configuration information of the second AI system; The transceiver unit is further configured to send part or all of the AI ​​configuration information of the second AI system to the control node of the second AI system.

24. The device according to claim 23, characterized in that The AI ​​configuration information of the second AI system includes at least one of the following: First indication information, used to indicate nodes participating in the AI ​​task in the AI ​​system; The second indication information is used to indicate the role of the node participating in the AI ​​task in the AI ​​system; The third indication information is used to indicate the learning architecture of the AI ​​system; Fourth indication information, used to indicate an AI task of the AI ​​system; The fifth indication information is used to indicate the communication resources of the AI ​​system.

25. The device according to any one of claims 22 to 24, characterized in that The AI ​​configuration information of the first AI system includes at least one of the following: First indication information, used to indicate nodes participating in the AI ​​task in the AI ​​system; The second indication information is used to indicate the role of the node participating in the AI ​​task in the AI ​​system; The third indication information is used to indicate the learning architecture of the AI ​​system; Fourth indication information, used to indicate an AI task of the AI ​​system; The fifth indication information is used to indicate the communication resources of the AI ​​system.

26. The device according to any one of claims 22 to 25, characterized in that The learning framework includes: Federated learning, federated distillation, decentralized learning, meta-learning, or segmented learning.

27. The device according to any one of claims 22 to 26, characterized in that The first AI system or the second AI system is configured to perform a second processing on the second data after performing a first processing on the first data to obtain the second data; When the learning architecture includes federated learning, the first data includes global model parameters and / or gradients, the first processing includes local training processing, the second data includes local model parameters and / or gradients, and the second processing includes fusion processing; When the learning architecture includes federated distillation, the first data includes global pairwise logits, the first processing includes local training processing, the second data includes local logits, and the second processing includes fusion processing; When the learning architecture includes decentralized learning, the first data includes parameters and / or gradients of a first local model, the first processing includes fusion processing and local training processing, the second data includes parameters and / or gradients of a second local model, and the second processing includes fusion processing and local training processing; When the learning architecture includes meta-learning, the first data includes meta-model parameters and / or gradients, the first processing includes support set training processing and test set gradient calculation processing, the second data includes test set gradients and / or losses, and the second processing includes gradient fusion processing and meta-model update processing; When the learning architecture includes segmentation learning, the first data includes segmentation layer inference results, the first processing includes inference, gradient calculation, back transfer and parameter update processing, the second data includes segmentation layer gradient, and the second processing includes continued gradient transfer processing and parameter update processing.

28. The device according to any one of claims 22 to 27, characterized in that The first AI system and the second AI system are used to perform the same AI task, or the first AI system and the second AI system are used to perform different subtasks in the same AI task.

29. A communication device, characterized in that: The device comprises at least one processor configured to execute computer programs or instructions in the memory to implement the method according to any one of claims 1 to 7.

30. A communication device, characterized in that: The device comprises at least one processor configured to execute computer programs or instructions in the memory to implement the method according to any one of claims 8 to 14.

31. A communication system, characterized in that: The system comprises the communication device according to any one of claims 15 to 21, and is used to execute the communication device according to any one of claims 22 to 28; or The system comprises the communication device according to claim 29, and the communication device according to claim 30.

32. A computer-readable storage medium, characterized in that: The medium stores instructions, and when the instructions are executed by a computer, the method according to any one of claims 1 to 14 is implemented.

33. A computer program product, characterized in that The method comprises instructions which, when executed on a computer, cause the computer to execute the method according to any one of claims 1 to 14.

Images