Method, device and system for model training

Abstract

The embodiment of the invention provides a method, device and system for model training. The method may comprise: acquiring, by a network management device, training data for training a first portion of a bilateral model, wherein the bilateral model comprises the first portion to be deployed at a network device and a second portion to be deployed at a terminal device; performing model training based on the training data to obtain a first part; and transmitting the first portion to the network device. Exemplary, the second portion may be trained by the server or may be trained by the network management device. In this manner, a training process of the bilateral model of the air interface may be assisted by the network management device of the management plane, in particular at least training the first portion and deploying it at the network device, thereby supporting use of the bilateral model in the air interface.

Description

Technical Field

[0001] Embodiments of the present disclosure relate to the field of communications, and more specifically, to a method, device, system, and computer-readable storage medium for model training. Background Art

[0002] In order to improve the intelligence and automation level of the network, artificial intelligence (AI) and machine learning (ML) technologies are being applied in more and more fields. The 3rd Generation Partnership Project (3GPP) working group has multiple related standardization topics that are studying the corresponding network intelligence. For example, the 3GPP Radio Access Network (RAN) working group is discussing topics including AI / ML algorithm research for the new radio (NR) air interface, which mainly studies the AI / ML-based algorithms to enhance the air interface to improve the air interface performance. The air interface uses AI / ML in two ways: unilateral AI and bilateral AI. The model is deployed on one side of the access network network equipment or terminal equipment, and the model reasoning is performed by the party corresponding to the deployment model. This AI model usage mode is called unilateral AI, and the corresponding model used is a unilateral AI model. The model is divided into two parts, which are deployed in the access network network equipment or terminal equipment respectively, and the two parties use their own deployed models for reasoning and collaborate to complete the reasoning task. This AI model usage mode is called bilateral AI, and the corresponding model used is a bilateral AI model.

[0003] The bilateral AI model includes a part deployed at the access network equipment (for example, called a decoder model) and a part deployed at the terminal device (for example, called an encoder model), but how to obtain the bilateral AI model through training still needs further study. Summary of the invention

[0004] The embodiments of the present disclosure provide a solution for model training, and specifically provide a training solution for a bilateral AI model.

[0005] In a first aspect of the present disclosure, a method for model training is provided. The method may include: obtaining training data for training a first part of a bilateral model, wherein the bilateral model includes a first part to be deployed at a network device and a second part to be deployed at a terminal device; performing model training based on the training data to obtain the first part; and sending the first part to the network device.

[0006] The embodiments of the present disclosure can be executed by a network management device. In this way, the network management device on the management plane can assist the training process of the bilateral model of the air interface, specifically training the first part and deploying it on the network device, thereby supporting the use of the bilateral model in the air interface.

[0007] In some implementations, the method may further include: performing lightweight processing on the first part so that the lightweight first part meets at least one of the following: performance requirements of the first part, energy consumption requirements for training the first part, energy consumption requirements for the reasoning task to which the first part will be used, or first model size requirements of the first part. In this way, lightweight processing can be performed on the premise of ensuring the requirements of the model to compress the model, which can reduce the amount of calculation and size of the model.

[0008] Exemplarily, the method may be executed by a second network management device, such as an element management system (EMS).

[0009] In some implementations, obtaining training data for training the first part of the bilateral model includes: receiving a first training request from a first network management device, the first training request indicating training the first part of the bilateral model, the first training request including training data for training the first part of the bilateral model and structural information of the first part. In this way, based on the first training request from the first network management device, model training can be performed using the training data and structural information provided by the first network management device, so that the first part matching the second part can be obtained.

[0010] In some implementations, the method may further include: sending a first model training report to the first network management device, the first model training report including at least one of the following: a model identifier of the bilateral model, a model identifier of the first part, an actual performance of the first part, an actual energy consumption of training the first part, an estimated energy consumption of the inference task for which the first part will be used, or an actual model size of the first part. In this way, by sending the first model training report to the first network management device, the first network management device can determine that the training of the first part has been completed, and can also learn about the training status of the first part, such as the actual energy consumption, the actual model size, etc.

[0011] In some implementations, obtaining training data for training the first part of the bilateral model includes: obtaining training data for training the first part from a network device. Exemplarily, the method also includes: receiving a first training request from a first network management device, the first training request being used to request training of the first part of the bilateral model, and also requesting training data for training the second part. Exemplarily, performing model training based on the training data to obtain the first part includes: obtaining the first part through training based on the first training request using the training data for training the first part; and determining structural information of the second part, and constructing training data for training the second part. Exemplarily, the method may also include: sending a first model training report to the first network management device, the first model training report including training data for training the second part and structural information of the second part. In this way, the training data for training the second part and the structural information of the second part can be subsequently provided to the server, and the server trains to obtain the second part.

[0012] In some implementations, obtaining training data for training the first part of the bilateral model includes: obtaining training data for training the first part and the second part from a network device. Exemplarily, when performing model training, the first part and the second part can be obtained based on the training data. In this way, the two parts of the bilateral model can be obtained by training the network management device, reducing signaling interaction.

[0013] In some implementations, the method may further include: receiving a model loading request from the first network management device, the model loading request indicating that the first part is to be loaded into the network device; and sending the first part to the network device based on the model loading request. In this way, the first part may be loaded into the network device based on the model loading request from the first network management device, which ensures that the network device uses the first part as part of a bilateral model, rather than as a unilateral model.

[0014] In some implementations, the model loading request may include at least one of the following: a first identifier of the network device, a model identifier of the first part, a second identifier of the terminal device, a model identifier of the second part, or a model identifier of the bilateral model.

[0015] In some implementations, the first training request also includes at least one of the following: a model identifier of the bilateral model, a model identifier of the first part, a model identifier of the second part, a first identifier of the network device, a second identifier of the terminal device, indication information for indicating the type of reasoning task for which the bilateral model will be used, performance requirements of the bilateral model, energy consumption requirements for bilateral model training, energy consumption requirements for reasoning tasks, a first model size requirement for the first part, or a lightweight indication.

[0016] In this way, by carrying the model identifier of the first part and / or the second part in the request, the matching and association between the first part and the second part can be determined. By carrying the identifier of the network device and / or the terminal device in the request, it is possible to determine which two parties are to deploy the bilateral model. By carrying the demand for the model in the first training request, etc., the actual available resources can be considered during the training process to ensure the availability of the trained model.

[0017] In some implementations, the method may further include: receiving a request message from the first network management device, the request message being used to request status information of the network device; and in response to the request message, sending a response message to the first network management device, the response message including at least one of the following: computing power status of the network device, storage status of the network device, energy consumption information of the network device, or a network energy consumption target value of the network device. In this way, the first network management device may determine the requirements for training based on the status information, thereby ensuring the availability of the trained model.

[0018] In some implementations, the method may further include: sending a second model training report to the first network management device, the second model training report including training data for training the second part and structural information of the second part. In this way, the training data for training the second part may be further provided to the server so that the server can train to obtain a matching second part.

[0019] Exemplarily, the method may be executed by a first network management device, such as a network management system (NMS).

[0020] In some implementations, obtaining training data for training the first part of the bilateral model includes: receiving a second training request from a server, the second training request indicating training the first part of the bilateral model, the second training request including training data for training the first part of the bilateral model and structural information of the first part. In this way, model training can be performed based on the second training request from the server using the training data and structural information provided by the server, so that the first part matching the second part can be obtained.

[0021] In some implementations, the method may further include: sending a fourth model training report to the server, the fourth model training report including at least one of the following: a model identifier of the bilateral model, a model identifier of the first part, or an actual performance of the first part. In this way, by sending the fourth model training report to the server, the server can determine that the training of the first part has been completed, so that the server can further decide to deploy the model.

[0022] In some implementations, the method may further include: receiving a notification message from the server, the notification message indicating that the second part trained by the server has been loaded into the terminal device; and sending the first part to the network device in response to the notification message. In this way, after the second part is deployed to the terminal device, the first part may be loaded into the network device based on the notification message from the server, which ensures that the network device uses the first part as part of a bilateral model, rather than a unilateral model.

[0023] In some implementations, the second training request also includes at least one of the following: a model identifier of the bilateral model, a model identifier of the first part, a model identifier of the second part, a first identifier of the network device, a second identifier of the terminal device, indication information for indicating the type of reasoning task for which the bilateral model will be used, or performance requirements of the bilateral model.

[0024] In some implementations, the method may further include: sending a request message to a second network management device, the request message including a first identifier of the network device; receiving a response message from the second network management device, the response message including at least one of the following: computing power status of the network device, storage status of the network device, energy consumption information of the network device, or a network energy consumption target value of the network device; and based on the response message, determining at least one of the following: energy consumption requirements for bilateral model training, first energy consumption requirements for the reasoning task to be used by the first part, second energy consumption requirements for the reasoning task to be used by the second part, first model size requirements for the first part, or second model size requirements for the second part. In this way, the first network management device can determine the requirements for training based on the status information, etc., so as to ensure the availability of the trained model.

[0025] In some implementations, the method may further include: sending a second model training report to the server, the second model training report including training data for training the second part and structural information of the second part. In this way, the training data for training the second part may be further provided to the server so that the server can train to obtain a matching second part.

[0026] In some implementations, the second model training report also includes at least one of the following: a model identification of the bilateral model, a model identification of the first part, a first identification of the network device, an actual performance of the first part, an actual energy consumption of training the first part, an estimated energy consumption of the reasoning task for which the first part will be used, an actual model size of the first part, a performance requirement of the second part, a second energy consumption requirement for the reasoning task for which the second part will be used, or a second model size requirement for the second part.

[0027] In a second aspect of the present disclosure, a method for model training is provided. The method includes: sending a first training request to a second network management device, the first training request indicating training a first part of a bilateral model, the bilateral model including a first part to be deployed at a network device and a second part to be deployed at a terminal device; and receiving a model training report from the second network management device, the model training report including at least a model identifier of the first part.

[0028] The method may be performed by a first network management device, in which way the first network management device may instruct a second network management device to train at least a first part of a bilateral model, thereby supporting the use of the bilateral model in an air interface.

[0029] In some implementations, the first training request includes: training data for training the first part, and structural information of the first part. In this way, the second network management device can perform training based on the training data and the structural information, thereby ensuring that the obtained first part matches and is associated with the trained second part.

[0030] In some implementations, the model training report includes a first model training report, which includes at least one of the following: a model identification of the bilateral model, a model identification of the first part, an actual performance of the first part, an actual energy consumption of training the first part, an estimated energy consumption of an inference task for which the first part will be used, or an actual model size of the first part.

[0031] In some implementations, the method may further include: receiving a second training request from the server, the second training request instructing to train the first part of the bilateral model, the second training request including: training data for training the first part, and structural information of the first part. In this way, it can be determined that the second part has been trained by the server, and then the second network management device can be instructed to train the second part.

[0032] In some implementations, the model training report includes a second model training report, and the second model training report includes training data for training the second part and structural information of the second part. In some implementations, the method may also include: sending a training request to the server, the training request including training data for training the second part and structural information of the second part. In this way, the server can perform training to obtain the second part. And, since the training data and structural information for training the second part come from the second network management device, the second part trained by the server will match the first part trained by the second network management device.

[0033] In some implementations, the second model training report further includes at least one of the following: a model identifier of the bilateral model, a model identifier of the first part, a first identifier of the network device, actual performance of the first part, actual energy consumption of training the first part, estimated energy consumption of the inference task to be used by the first part, and an actual model size of the first part. In this way, the first network management device can learn about the situation during the training process of the first part.

[0034] In some implementations, the method may further include: forwarding the second model training report to the server; receiving a notification message from the server, the notification message indicating that the second part has been loaded into the terminal device; and sending a model loading request to the second network management device, the model loading request indicating that the first part is loaded into the network device. In this way, after the server deploys the second part to the terminal device, based on the model loading request from the server, the second network management device can be instructed to deploy the first part to the network device, so that the association between the first part and the second part can be ensured, and thus the use of the bilateral model can be ensured.

[0035] In some implementations, the method further includes: receiving a second training request from a server, the second training request being used to request training of the first part of the bilateral model and also requesting training data for training the second part; sending a first training request to a second network management device, the first training request being used to request training of the first part of the bilateral model and also requesting training data for training the second part. Exemplarily, the method further includes: receiving a first model training report from a second network management device, the first model training report including training data for training the second part and structural information of the second part, the first model training report indicating that the first part has been trained. Exemplarily, the method further includes: sending a second model training report to the server, the second model training report including training data for training the second part and structural information of the second part, the second model training report indicating that the first part has been trained. In this way, the server can obtain the second part through training based on the second model training report, thereby being able to support the use of the bilateral model over the air interface.

[0036] In a third aspect of the present disclosure, a method for model training is provided. The method includes: obtaining training data for training a second part of a bilateral model, wherein the bilateral model includes a first part to be deployed at a network device and a second part to be deployed at a terminal device; performing model training based on the training data to obtain the second part; and sending the second part to the terminal device.

[0037] The method can be executed by a server. In this way, the server can assist in the training process of the bilateral model of the air interface, specifically training the second part and deploying it on the terminal device, thereby supporting the use of the bilateral model in the air interface.

[0038] In some implementations, the method may further include: sending a second training request to the first network management device, the second training request indicating training of the first part of the bilateral model, the second training request including training data for training the first part of the bilateral model, and structural information of the first part. In this way, training data and structural information for the first part may be provided, so that the management domain can train the first part to match the second part.

[0039] In some implementations, the method may further include: receiving a first model training report from the first network management device, the first model training report indicating that the first part has been trained. In this way, the server may determine that the first part has been trained by the management domain based on the first model training report.

[0040] In some implementations, obtaining training data for training the second part of the bilateral model includes: receiving a second model training report from the first network management device, the second model training report including training data for training the second part and structural information of the second part. In this way, the training data and structural information for the second part provided by the management domain can be obtained, so that the trained second part can match the first part trained by the management domain.

[0041] In some implementations, the second model training report also includes at least one of the following: a model identification of the bilateral model, a model identification of the first part, a first identification of the network device, an actual performance of the first part, a performance requirement of the second part, a second energy consumption requirement of the inference task for which the second part will be used, or a second model size requirement for the second part.

[0042] In some implementations, the method may further include: sending a notification message to the first network management device, the notification message indicating that the second part has been loaded into the terminal device. In this way, the management domain may further deploy the first part to the network device based on the notification message, thereby ensuring the use of the bilateral model.

[0043] In a fourth aspect of the present disclosure, a system is provided. The system includes: a first network management device and a second network management device, wherein the first network management device can be configured to perform the operations of the method in the second aspect or any implementation as described above, and wherein the second network management device can be configured to perform the operations of the method in the first aspect or any implementation as described above that can be performed by the second network management device.

[0044] In some implementations, the system may further include a server, wherein the server may be configured to perform the operations of the method in the third aspect or any implementation as described above. For example, the server may obtain the second part of the bilateral model through training.

[0045] In some implementations, the system may further include a network device configured to: receive a first portion of the bilateral model trained by the second network management device. In some implementations, the network device may further include: receive a second portion from the server; and send the second portion to the terminal device. In some implementations, the system may further include a terminal device configured to receive the second portion from the network device.

[0046] In some other implementations, the second network management device may be further configured to obtain the second part of the bilateral model through training.

[0047] It should be noted that some implementation methods and beneficial effects of the aforementioned first to third aspects are also applicable to the fourth aspect, and for the sake of brevity, they will not be repeated here.

[0048] In a fifth aspect of the present disclosure, a system is provided. The system includes: a first network management device and a server, wherein the first network management device can be configured to perform the operation of the method that can be performed by the first network management device in the first aspect or any implementation as described above, wherein the server can be configured to perform the operation of the method in the third aspect or any implementation as described above. For example, the server can obtain the second part of the bilateral model through training.

[0049] In some implementations, the system may further include a network device, the network device being configured to: receive a first part of the bilateral model trained by the first network management device. In some implementations, the network device is further configured to: receive a second part from the server; and send the second part to the terminal device. In some implementations, the system may further include a terminal device, configured to receive the second part from the network device.

[0050] In some other implementations, the first network management device may also be configured to obtain the second part of the bilateral model through training.

[0051] It should be noted that some implementation methods and beneficial effects of the aforementioned first to third aspects are also applicable to the fifth aspect, and for the sake of brevity, they will not be repeated here.

[0052] In a sixth aspect of the present disclosure, a method is provided. The method may include: a first network management device performs the operations of the method in the second aspect or any implementation as described above, and a second network management device performs the operations of the method in the first aspect or any implementation as described above that can be performed by the second network management device.

[0053] In some implementations, the method further includes: the server performs the operations of the method in the third aspect or any implementation as described above. For example, the server can obtain the second part of the bilateral model through training.

[0054] In some implementations, the method further includes: receiving, by the network device, the second part from the server; and sending the second part to the terminal device. In some implementations, the method further includes: receiving, by the terminal device, the second part from the network device.

[0055] In some other implementations, the method further includes: obtaining the second part of the bilateral model by training by the second network management device.

[0056] In a seventh aspect of the present disclosure, a method is provided, which includes: a first network management device performs the operations of the method executable by the first network management device in the first aspect or any implementation manner as described above.

[0057] In some implementations, the method further includes: the server performs the operations of the method in the third aspect or any implementation as described above. For example, the server can obtain the second part of the bilateral model through training.

[0058] In some implementations, the method further includes: receiving, by the network device, the second part from the server; and sending the second part to the terminal device. In some implementations, the method further includes: receiving, by the terminal device, the second part from the network device.

[0059] In some other implementations, the method further includes: obtaining the second part of the bilateral model by training by the first network management device.

[0060] In an eighth aspect of the present disclosure, a network management device is provided, comprising a component for performing operations according to the method in the aforementioned first aspect or any implementation thereof, or comprising a component for performing operations according to the method in the aforementioned second aspect or any implementation thereof. Optionally, the component can be implemented as a unit, a module, etc.

[0061] In a ninth aspect of the present disclosure, a server is provided, comprising a component for performing operations according to the method in the third aspect or any implementation thereof. Optionally, the component may be implemented as a unit, a module, etc.

[0062] In the tenth aspect of the present disclosure, a network management device is provided, including a transceiver, a processor and a memory, wherein the memory stores instructions executed by the processor, and when the instructions are executed by the processor, the network management device performs the operations of the method in the aforementioned first aspect, second aspect, or any implementation thereof.

[0063] In the eleventh aspect of the present disclosure, a server is provided, comprising a transceiver, a processor and a memory, wherein the memory stores instructions executed by the processor, and when the instructions are executed by the processor, the server performs the operations of the method in the aforementioned third aspect or any of its implementations.

[0064] In a twelfth aspect of the present disclosure, a computer-readable storage medium, such as a non-transitory computer-readable storage medium, is provided. The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the operations of the method according to the first to third aspects or any implementation thereof.

[0065] In a thirteenth aspect of the present disclosure, a chip or a chip system is provided, which includes a processing circuit configured to perform operations according to the method in the first to third aspects or any implementation manner thereof.

[0066] In a fourteenth aspect of the present disclosure, a computer program or a computer program product is provided. The computer program or computer program product is tangibly stored on a computer-readable medium and includes computer executable instructions, which, when executed, enable a device to implement the operation of the method according to the first to third aspects or any implementation thereof.

[0067] It should be noted that some embodiments and beneficial effects of the aforementioned method aspects are also applicable to device aspects, user equipment aspects, network equipment aspects, computer-readable storage media aspects, chips or chip systems aspects, computer programs or computer program products aspects, and will not be repeated here for the sake of brevity. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] The above and other features, advantages and aspects of various embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings.

[0069] In the accompanying drawings, the same or similar reference numerals represent the same or similar elements, wherein:

[0070] Figure 1 A schematic diagram of a bilateral AI model for channel state information (CSI) feedback is shown;

[0071] Figure 2 A schematic structural diagram showing a network system in which some embodiments of the present disclosure can be applied;

[0072] Figure 3 A schematic structural diagram showing another network system to which some embodiments of the present disclosure can be applied;

[0073] Figure 4 shows a signaling diagram according to some embodiments of the present disclosure;

[0074] Figure 5 shows another signaling diagram according to some embodiments of the present disclosure;

[0075] Figure 6 shows yet another signaling diagram according to some embodiments of the present disclosure;

[0076] Figure 7 shows yet another signaling diagram according to some embodiments of the present disclosure;

[0077] Figure 8 shows another signaling diagram according to some embodiments of the present disclosure;

[0078] Fig. 9 shows yet another signaling diagram according to some embodiments of the present disclosure;

[0079] Figure 10A-10C A schematic diagram showing a scenario of training and reasoning according to some embodiments of the present disclosure is shown;

[0080] Fig.11 A schematic block diagram of a network management device according to some embodiments of the present disclosure is shown;

[0081] Fig.12 A schematic block diagram of a network management device according to some embodiments of the present disclosure is shown;

[0082] Fig.13 A schematic block diagram showing a server according to some embodiments of the present disclosure; and

[0083] Fig.14 A schematic block diagram of an example device that can be used to implement embodiments of the present disclosure is shown. DETAILED DESCRIPTION

[0084] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein, which are instead provided for a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

[0085] In the description of the embodiments of the present disclosure, the term "including" and similar terms should be understood as open inclusion, that is, "including but not limited to". The term "based on" should be understood as "based at least in part on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The terms "first", "second", etc. may refer to different or the same objects. Other explicit and implicit definitions may also be included below.

[0086] The embodiments of the present disclosure may be implemented according to any appropriate communication protocol, including but not limited to third generation (3G), fourth generation (4G), fifth generation (5G), sixth generation (6G) and other cellular communication protocols, wireless local area network communication protocols such as Institute of Electrical and Electronics Engineers (IEEE) 802.11, and / or any other protocol currently known or developed in the future.

[0087] The technical solutions of the embodiments of the present disclosure are applied to communication systems that follow any appropriate communication protocols, such as: Universal Mobile Telecommunications Service (UMTS), Long Term Evolution (LTE) system, Wideband Code Division Multiple Access system (WCDMA), Code Division Multiple Access 2000 system (CDMA2000), Time Division-Synchronization Code Division Multiple Access system (TD-SCDMA), Frequency Division Duplex (FDD) system, Time Division Duplex (TDD), fifth generation (5G) system or new radio (NR), future evolved 6G and 7G communication systems, and the like.

[0088] It should be understood that the embodiments of the present disclosure may be applied to any communication system having similar problems, such as a wireless local area network (WLAN), a wired communication system, or other communication systems developed in the future.

[0089] The term "terminal device" used in the present disclosure refers to any terminal device that can perform wired or wireless communication with network devices or with each other. The terminal device may sometimes be referred to as a user equipment (UE). The terminal device may be any type of mobile terminal, fixed terminal or portable terminal. As an example, the terminal device may include a mobile phone, a site, a unit, a device, a mobile terminal (MT), a subscription station, a portable subscription station, an Internet node, a communicator, a desktop computer, a laptop computer, a notebook computer, a tablet computer, a personal communication system device, a personal navigation device, a personal digital assistant (PDA), a positioning device, a radio broadcast receiver, an e-book device, a gaming device, an Internet of Things (IoT) device, a vehicle-mounted device, an aircraft, a virtual reality (VR) device, an augmented reality (AR) device, a wearable device, a terminal device in a 5G network or any terminal device in an evolved public land mobile network (PLMN), other devices that can be used for communication, or any combination of the above. The embodiments of the present disclosure do not limit this.

[0090] The term "network device" used in the present disclosure refers to an entity or node that can be used to communicate with a terminal device, for example, it can be an access network network device (i.e., access network device). The access network device can be a device deployed in the RAN to provide wireless communication functions for mobile terminals, for example, it can be a RAN network device. The access network device may include various types of base stations (BS). As an example, the access network device may include various forms of macro base stations, micro base stations, pico base stations, femto base stations, relay stations, access points, satellites, remote radio units (RRU), radio heads (RH), remote radio heads (RRH), etc., and may also be devices that assume base station functions in device-to-device (D2D) communication and machine-to-machine (M2M) communication. In systems using different wireless access technologies, the names of access network devices may be different, such as evolved NodeB (eNB or eNodeB) in Long Term Evolution (LTE) networks, NodeB (NB) in 3G networks, gNB or NR NB in ​​5G networks, and so on. In some scenarios, the access network equipment may include a centralized unit (CU) and / or a distributed unit (DU). The CU and DU may be placed in different places, for example: the DU is remotely located in an area with high traffic volume, and the CU is placed in a central computer room. Alternatively, the CU and DU may be placed in the same computer room. The CU and DU may also be different components under one rack. The access network equipment may also be a device deployed in an open RAN (O-RAN) to provide wireless communication functions for mobile terminals, for example, the access network equipment may include an open CU (O-CU) and / or an open DU (O-DU). For the convenience of description, in the subsequent embodiments of the present disclosure, the above-mentioned devices providing wireless communication functions for mobile terminals are collectively referred to as network devices, and the embodiments of the present disclosure are no longer specifically limited.

[0091] As described above, the AI ​​model used in the air interface (i.e., the RAN domain AI model) may include a unilateral AI model (referred to as a unilateral model) and a bilateral AI model (referred to as a bilateral model). For example, the unilateral AI model may be used for beam management, positioning, etc., and the bilateral AI model may be used for CSI feedback, etc. For the bilateral AI model, its characteristics may include: the model consists of two parts, and the model output of the first part will be used as the model input of the second part; model training needs to be carried out in coordination so that the original input and the final output are as close as possible; and model reasoning needs to be carried out in coordination, and the models used by the two parties performing reasoning need to match.

[0092] Taking CSI feedback as an example, Figure 1 A schematic diagram of a bilateral AI model 100 for CSI feedback is shown. Figure 1 In FIG, a decoder model deployed at an access network device (e.g., BS) 110 and an encoder model deployed at a terminal device (e.g., UE) 120 are shown. In a frequency division multiplexing network, the BS needs to obtain CSI feedback for the downlink to perform precoding to achieve performance gain. However, the UE will consume more air interface resources to transmit the original CSI. Therefore, in order to reduce the overhead of CSI feedback, a bilateral AI model can be used to implement the CSI compression and reconstruction mechanism. Specifically, a bilateral AI model can be constructed by using a neural network structure of an autoencoder (AutoEncoder), which consists of an encoder model and a decoder model. The encoder model is deployed on the UE side to convert the original CSI matrix (expressed as X t ) Use neural network to reduce dimension and get compressed CSI (expressed as Y t ); The decoder model is deployed on the BS side to recover the received compressed CSI using a neural network to obtain the estimated original CSI (expressed as ).

[0093] The lifecycle management of the model involves model training, model testing, model simulation, model loading, and model reasoning. Model lifecycle management can be performed by the management domain, where specific model training and model reasoning can be performed by the management domain, RAN domain, and / or core network domain. For bilateral AI models reasoned in the RAN domain, the management domain can assist in performing model training, and then deploy the model to access network devices and terminal devices for model reasoning. However, the details of how the management domain specifically assists the bilateral AI model training in the RAN domain still need further study.

[0094] Figure 2 A schematic structural diagram of a network system 200 to which some embodiments of the present disclosure can be applied is shown. Figure 2The 3GPP network domain 250 and the server 290 outside the 3GPP network domain 250 are included. The 3GPP network domain 250 includes the NMS 210, the EMS 220, the access network device 230 and the terminal device 240, wherein the NMS 210 and / or the EMS 220 perform the model training function, and the access network device 230 and the terminal device 240 perform the model reasoning function. The server 290 may be a UE server, which may be a server provided by the manufacturer of the terminal device 240, or a server provided by the chip manufacturer in the terminal device 240. For example, the server 290 may perform the model training function, provide software download or upgrade services for the terminal device 240, etc.

[0095] Figure 3 A schematic structural diagram of a network system 300 in which some embodiments of the present disclosure can be applied is shown. The system 300 is based on the O-RAN architecture, where O-RAN is intended to implement an intelligent, open access network. O-RAN is a concept proposed based on the interoperability and standardization of RAN, which provides the same interconnection standard for white box hardware and software from different vendors. Figure 3 As shown, the system 300 includes a service management and orchestration framework (SMO) 305, which includes a non-real time RAN intelligent controller (non-RT RIC or NRT RIC) 310. The system 300 also includes a near-realtime RAN intelligent controller (near-RT RIC or nRT RIC) 320, an access network device 330, a terminal device 340, and a server 390, wherein the access network device 330 includes an O-CU 332, an O-DU 334, and an O-RU 336. The non-RT RIC 310 and / or the near-RT RIC 320 perform a model training function, while the O-CU 332, the O-DU 334, and the terminal device 340 perform a model reasoning function. Server 390 may be a UE server, which may be a server provided by the manufacturer of terminal device 340 or a server provided by a chip manufacturer in terminal device 340. For example, server 390 may perform model training functions or provide software download or upgrade services for terminal device 340.

[0096] Embodiments of the present disclosure relate to performing model training of a bilateral AI model by a management domain, wherein the management domain may include Figure 2 The NMS 210 and EMS 220 shown, or the management domain may include Figure 3 The non-RT RIC 310 and the near-RT RIC 320 are shown. It is understandable that the management domain for performing model training may also be referred to as a management plane, and may include other devices or network elements or systems in other scenarios or architectures, etc., which are not listed one by one in this disclosure.

[0097] In the following more detailed embodiments, for the convenience of description, the following Figure 2 However, it should be understood that the following embodiments can also be applied to the following system 200. Figure 3 In the system 300 shown, for example, the NMS 210 in the following may be replaced by a non-RT RIC 310, the EMS 220 may be replaced by a near-RT RIC 320, the access network device 230 may be replaced by an access network device 330 (or replaced by an O-CU 332 and an O-DU 334), the terminal device 240 may be replaced by a terminal device 250, and the server 290 may be replaced by a server 390, so that the following embodiments may be applied to the system 300. In addition, the exemplary embodiments in the following may also be applied to other systems or architectures not shown in this document, which are not listed one by one here.

[0098] In an embodiment of the present disclosure, the bilateral AI model may include a first part deployed in an access network device and a second part deployed in a terminal device, and the first part and the second part are matched (or associated). Exemplarily, the matching first and second parts can be understood as: when the output of the second part is used as the input of the first part, the output of the first part can be approximated to the input of the second part. For the convenience of description, in the following embodiments, it is assumed that the first part is a decoder model and the second part is an encoder model. Figures 4 to 10C Elaborate in more detail.

[0099] Figure 4 A schematic diagram of a signaling interaction 400 according to some embodiments of the present disclosure is shown. Figure 4 Involving server 290, NMS 210, EMS 220, access network device 230 and terminal device 240. It can be understood that although signaling interaction 400 is combined with Figure 2 However, the process can also be applied to other systems, which is not limited by the present disclosure. Figure 4 In the process, the server 290 and the EMS 220 perform model training. Specifically, the server 290 obtains the part of the bilateral AI model to be deployed to the terminal device 240 (called the encoder model or the UE-side model), and the EMS 220 obtains the part of the bilateral AI model to be deployed to the access network device 230 (called the decoder model or the BS-side model).

[0100] At 401, NMS210 may, for example, open AI / ML management services to server 290 through a management service open interface. For example, AI / ML management services may include model training, model deployment notifications, etc. It is understandable that since server 290 does not belong to the 3GPP network domain and cannot directly call services in the 3GPP network, NMS210 may provide services to server 290 by opening network capabilities. It should be noted that step 401 is a prerequisite step of this embodiment, and the present disclosure does not limit the specific implementation of the network capability opening method. In this way, interfaces for model training services and model loading services for bilateral AI models can be opened to server 290 outside the 3GPP network domain to ensure the normal execution of model training.

[0101] At 402, the server 290 obtains training data. Specifically, the terminal device 240 may perform operations such as data collection, measurement, and calculation to obtain training data, and the terminal device 240 may report the training data to the server 290. For example, the terminal device 240 may provide the training data to the server 290 through a private interface, such as through a Hyper Text Transport Protocol (HTTP) communication protocol. Assuming that the present embodiment is intended to train a bilateral AI model for CSI feedback, the training data may include the original CSI measured by the terminal device 240.

[0102] At 403, the server 290 performs model training based on the training data to obtain an encoder model (ie, a UE-side model) and an initial decoder model. The model training process may involve multiple rounds of iterations. The details of the model training may refer to the prior art, and the present disclosure is not limited thereto.

[0103] At 404, the server 290 sends a first model training request to the NMS 210. The first model training request includes: training data for training a decoder model (i.e., a BS-side model) and structural information of the decoder model. The first model training request may also include one or more of the following: a model identifier of a bilateral AI model, an identifier of the access network device 230, an identifier of a terminal device group, a model identifier of an encoder model (i.e., a UE-side model), an indication of the type of reasoning task for which the bilateral AI model will be used, model performance requirements (such as performance requirements of a decoder model), bilateral model training instructions, etc.

[0104] Exemplarily, after completing the model training at 403, the server 290 may construct training data for training the decoder model (i.e., the BS side model) by the management domain based on the training data acquired at 402, and optionally based on the initial decoder model obtained by the model training. Exemplarily, the first model training request may directly include training data for training the decoder model (i.e., the BS side model). Exemplarily, the server 290 may store training data for training the decoder model (i.e., the BS side model), and the first model training request may include the address of the storage to indicate the training data for training the decoder model (i.e., the BS side model). Thereafter, the management domain (e.g., NMS 210 or EMS 220) may obtain training data for training the decoder model (i.e., the BS side model) based on the storage address. For example, the management domain (e.g., NMS 210 or EMS 220) may download the training data for training the decoder model (i.e., the BS side model) from the storage address via the File Transfer Protocol (FTP).

[0105] The first model training request includes the structural information of the decoder model. By indicating the structural information, it can be ensured that the decoder model trained by the management domain (such as NMS 210 or EMS220) matches (or docks) the encoder model trained by the server 290, thereby ensuring the availability of the bilateral AI model. Exemplarily, the structural information of the decoder model may include one or more of the following: the network type of the decoder model, the number of network layers, the number of nodes in the input layer, the number of nodes in the output layer, etc., where the network type may be a convolutional neural network, a fully connected neural network, a recurrent neural network, etc. Optionally, the first model training request may include an initial decoder model to implicitly indicate the structural information of the decoder model. In this way, the management domain (such as EMS220) can perform training on the basis of the initial decoder model (such as in 408), which can reduce the amount of training calculations and improve the efficiency of training.

[0106] The first model training request may also include a model identifier of a bilateral AI model and / or a model identifier of an encoder model (i.e., a UE-side model). In some examples, the model identifier of the encoder model (i.e., the UE-side model) may uniquely represent or indicate an encoder model trained by server 290. In some examples, the model identifier of the bilateral AI model may be used to uniquely identify the bilateral AI model. Optionally, the model identifier of the bilateral AI model may be the same as the model identifier of a decoder model to be deployed at the access network device 230. For example, if a decoder model has been previously trained and a model identifier of the decoder model has been assigned, the bilateral AI model may be indicated by the model identifier of the decoder model, and the previously trained decoder model may be retrained. Optionally, the model identifier of the bilateral AI model may be a unique identifier generated based on a combination of a model identifier of an encoder model and a model identifier of a decoder model.

[0107] The first model training request may also indicate the type of reasoning task that the bilateral AI model will be used for, for example, the first model training request includes indication information about the type of reasoning task that the bilateral AI model will be used for. Exemplarily, the type of reasoning task, i.e., the reasoning type, may indicate the actual use case that the bilateral AI model will be used for, for example, the reasoning type may be CSI feedback, beam management, etc. Optionally, if the bilateral AI model is being trained for the first time or the model identifier of the bilateral AI model has not yet been assigned (unknown), the model to be trained may be indicated by the reasoning type.

[0108] The first model training request may also include an identifier of the access network device 230 (i.e., a base station identifier) ​​and / or an identifier of a terminal device group (i.e., a UE group identifier). The identifier of the access network device 230 (i.e., a base station identifier) ​​may be used to indicate the access network device to which the decoder model can be applied. For example, the base station identifier may be implemented as a network unique identifier name (Distinguished Name, DN) or a geographic area indication. The identifier of the terminal device group (i.e., a UE group identifier) ​​may be used to indicate the terminal device to which the encoder model trained by the server 290 can be applied. For example, the terminal device group includes at least the terminal device 240 shown in the figure. For example, the UE group identifier may be implemented as the manufacturer name, software version, chip manufacturer name, or chip model of the UE.

[0109] The first model training request may also include a bilateral model training indication, which may indicate the need to train the bilateral AI model. In some examples, the bilateral model training indication may be represented as a Boolean value, for example, "true" indicates that the bilateral AI model needs to be trained, and "false" indicates that the bilateral AI model does not need to be trained (i.e., training a unilateral AI model). In other examples, the bilateral model training indication may be represented as an encoded value, for example, "1" indicates that the bilateral AI model needs to be trained, and "0" indicates that the bilateral AI model does not need to be trained (i.e., training a unilateral AI model). It is understandable that the bilateral model training indication may be represented in other forms, and the present disclosure is not limited to this. In some embodiments, the bilateral model training indication may be implicitly indicated. For example, the first model training request includes training data for training the decoder model (i.e., the BS side model) and the structural information of the decoder model. Then, NMS210 may know that the decoder model needs to be trained, but NMS210 may not know whether the decoder model to be trained is a unilateral AI model or a BS side model in the bilateral AI model. Furthermore, the first model training request may also include the model identifier of the encoder model that has been trained by the server 290. Then, NMS210 may determine that the decoder model to be trained is to be matched with the encoder model that has been trained by the server 290, so that NMS210 can determine that the current training is for the bilateral AI model. In other words, it can be considered that the first model training request implicitly indicates the bilateral model training indication.

[0110] The first model training request may optionally further indicate a performance requirement of the decoder model. For example, the performance requirement may indicate an accuracy requirement of the decoder model when it is subsequently used for an inference task, such as 90%, 85%, or other values.

[0111] Optionally, at 405, NMS 210 sends a device query request message to EMS 220; and at 406, EMS 220 sends a device query response message to NMS 210. Exemplarily, the device query request message may be used to request to query the status information of the access network device, for example, the device query request message may be referred to as a device status query request message, etc. Correspondingly, the device query response message is used to provide the status information of the queried access network device, for example, the device query response message may be referred to as a device status query response message.

[0112] In some embodiments, the status information of the access network device may include one or more of the following: computing power status, storage status, energy consumption status, and network energy consumption target value of the access network device. Exemplarily, the device query response message may include computing resource utilization to indicate the computing power status. Exemplarily, the device query response message may include storage resource utilization to indicate the storage status. Exemplarily, the device query response message may include the current energy consumption or energy efficiency value to indicate the energy consumption status. Exemplarily, the device query response message may include the network's target value for energy consumption to indicate the network's target value for energy consumption, for example, the network's target value for energy consumption is: the overall energy consumption within a certain geographical range is less than 500 kilojoules (kJ).

[0113] In this way, through 405 and 406, NMS210 can obtain the status information of the access network device, and then NMS210 can determine the requirements for training based on the status information, for example, the requirements can be indicated to EMS220 in the second model training request below, thereby ensuring the availability of the first part of the trained bilateral AI model (i.e., the decoder model) at the access network device, thereby supporting the use of the bilateral AI model.

[0114] At 407, NMS 210 sends a second model training request to EMS 220. The second model training request may include part or all of the first model training request as described above. The second model training request includes training data for training a decoder model (i.e., a BS-side model) and structural information of the decoder model. In this way, the training data and structural information for training the decoder model can be provided to EMS 220, so that EMS 220 obtains a decoder model based on this training, thereby ensuring matching with the trained encoder model.

[0115] The first model training request may also include one or more of the following: a model identifier of the bilateral AI model, an identifier of the access network device 230, an identifier of the terminal device group (including the terminal device 240), a model identifier of the encoder model (i.e., the UE-side model) that has been trained by the server 290, information indicating the type of reasoning task for which the bilateral AI model will be used, model performance requirements, bilateral model training instructions, etc.

[0116] In some embodiments, the second model training request may also include one or more of the following about the decoder model: training energy consumption requirement, reasoning energy consumption requirement, model size (size) requirement, or lightweight indication, etc. Specifically, NMS210 may determine one or more of the following based on the device query response message received at 406: training energy consumption requirement, reasoning energy consumption requirement, model size (size) requirement, or lightweight indication, etc. For example, the computing power state and storage state are used to refer to determine the computing and storage requirements for bilateral AI model training and reasoning, and the energy consumption state and network energy consumption target value are used to refer to determine the energy consumption requirements for bilateral AI model training and reasoning. Exemplarily, the training (or reasoning) energy consumption requirement may include the energy consumption (e.g., duration, power) during the training (reasoning) process, etc. Exemplarily, the model size requirement may represent the maximum requirement (or upper limit) of the storage space occupied by the trained model, such as 100MB. The lightweight indication is used to indicate whether the model needs to be lightweight processed, and the method of lightweight processing, such as pruning, quantization, and distillation. Model pruning can also be called model sparsification. For example, some weight parameters of the model can be directly deleted. Model quantization can compress the weight parameters of the model. Model distillation compresses the model based on knowledge distillation.

[0117] At 408, EMS 220 performs model training based on the second model training request. Specifically, by using the training data for training the decoder model (i.e., BS-side model) indicated in the second model training request and the structural information of the decoder model, EMS 220 obtains the decoder model (i.e., BS-side model) through training.

[0118] Optionally, at 409, EMS220 performs lightweight processing on the decoder model. In some examples, if the second model training request includes a lightweight indication indicating that lightweight processing is required, EMS220 performs lightweight processing. In other examples, even if the second model training request does not include a lightweight indication, or the second model training request indicates that lightweight processing is not required, EMS220 can decide to perform lightweight processing on its own. For example, EMS220 can decide to perform lightweight processing based on model performance requirements, training energy consumption requirements, reasoning energy consumption requirements, model size requirements, etc. of the decoder model. Exemplarily, EMS220 can perform quantization, pruning, and other processing on the decoder model to obtain a decoder model after lightweight processing. It can be understood that lightweight processing is to compress the model under the premise of minimizing model performance requirements (such as accuracy) to reduce the amount of calculation and size (such as the number of parameters) of the model.

[0119] In this way, EMS220 can obtain a trained decoder model that matches the encoder model trained by server 290 through 408, or through 408 and 409. It should be noted that although Figure 4 409 is shown as a step independent of 408, but in other embodiments, 408 may include 409, that is, lightweight processing may be performed during model training. It is understandable that since the training at 408 and 409 is based on the training data and structural information provided by the server for training the decoder model, the trained decoder model can match the encoder model trained by the server. Since each requirement in the second model training request is considered at 408 and 409, it is ensured that the trained decoder model can adapt to the status information at the access network device, ensuring the availability of the decoder model.

[0120] At 410, EMS 220 sends a second model training report to NMS 210 to indicate that the decoder model has been trained. Exemplarily, the second model training report may include a model identifier of the trained decoder model and an identifier of the access network device 230. Exemplarily, the second model training report may also include one or more of the following: actual performance of the decoder model, actual training energy consumption of the decoder model, actual inference energy consumption of the decoder model, and actual model size of the decoder model.

[0121] At 411, NMS 210 sends a first model training report to server 290 to indicate that the decoder model has been trained. Exemplarily, the first model training report may include part or all of the second model training report. For example, the first model training report may include: a model identifier of the trained decoder model, and the actual performance of the decoder model.

[0122] At 412, the server 290 sends the encoder model to the terminal device 240. In some examples, the encoder model can be expressed in the form of model parameters or carried in the form of a model file. Specifically, the server 290 can send a model sending request to the terminal device 240, and the model sending request includes the encoder model itself, or includes model parameters or the file address of the model file. Exemplarily, the model sending request may also include a bilateral model indication, a model identifier of the encoder model, and a model identifier of the decoder model. In this way, the terminal device 240 can directly obtain the encoder model, or determine the encoder model through model parameters, or obtain the encoder model from the file address of the model file. The terminal device 240 can determine that the encoder model is a UE-side model in the bilateral AI model based on the bilateral model indication. In addition, the terminal device 240 can determine which decoder model matches the encoder model based on the model identifier of the encoder model and the model identifier of the decoder model. In this way, the terminal device 240 can use the encoder model from the server 290 as the UE-side model of the bilateral AI model instead of using it as a unilateral model, thus ensuring the association with the decoder model at the access network device 230.

[0123] At 413, server 290 sends a model delivery notification to NMS 210 to indicate that the encoder model has been deployed at terminal device 240. Exemplarily, the model delivery notification may include a bilateral model indication and a model identifier of the encoder model, indicating that the encoder model has been delivered.

[0124] At 414, NMS210 sends a model loading request to EMS220, and then at 415, EMS220 loads the decoder model to the access network device 230. Exemplarily, the model loading request may include a model identifier of the decoder model and an identifier of the access network device 230, which is used to instruct EMS220 to deploy the decoder model to the access network device 230. Exemplarily, the model loading request may also include a bilateral model indication, a model identifier of the encoder model, and a terminal device group identifier. In this way, the access network device 230 can use the decoder model from EMS220 as a BS-side model of the bilateral AI model instead of a unilateral model, so as to ensure the association with the encoder model at the terminal device 240.

[0125] In some examples, the decoder model may be expressed in the form of model parameters or carried in the form of a model file. Exemplarily, EMS220 may send another model loading request to the access network device 230, which may include the decoder model itself, or include the model parameters of the decoder model or the file address of the model file of the decoder model. In this way, the access network device 230 can directly obtain the decoder model, or determine the decoder model by the model parameters, or obtain the decoder model from the file address of the model file. Exemplarily, another model loading request may also include part or all of the model loading request as described above, for example, another model loading request includes a bilateral model indication, a model identifier of the encoder model, and a terminal device group identifier. In this way, the access network device 230 can determine that the decoder model is a BS-side model in a bilateral AI model based on the bilateral model indication. In addition, the access network device 230 can determine which encoder model matches the decoder model and which / those terminal devices the encoder model is deployed on based on the model identifier and the terminal device group identifier of the encoder model. For example, the bilateral model indication, the model identifier of the encoder model, and the terminal device group identifier can be understood as bilateral model association information to determine which encoder model matches the decoder model and which terminal device / devices the encoder model is deployed on.

[0126] In this way, through 401 to 415, the server 290 can be supported to call the model management service, and then the server 290 and the EMS 220 collaboratively complete the model training of the bilateral AI model, and are respectively deployed at the terminal device 240 and the access network device 230. In some embodiments, during the training process, the state information of the access network device 230 can be considered at the same time, so that a suitable model can be trained based on the computing power, storage, and energy consumption requirements of the access network device 230.

[0127] Further, at 416, the bilateral AI model may be used in a subsequent model reasoning process. As an example, the model reasoning process for CSI feedback may refer to the aforementioned combined Figure 1 Description.

[0128] By combining Figure 4The described signaling interaction 400 can increase the service types (i.e., model training service, model loading service) opened by the management plane (such as NMS210), and support the server 290 outside the 3GPP network domain 250 to call the model training service and model loading service of the management plane. The first / second model training request may include a bilateral model indication, a model identifier of the UE-side model, a model structure of the BS-side model, and training data of the BS-side model, thereby supporting the management plane (such as EMS220) to train the BS-side model with the training data provided by the external server 290. The model loading request / another model loading request may include a bilateral model indication and / or bilateral model association information, thereby indicating that the UE-side model has been deployed, thereby supporting the deployment and reasoning of the bilateral AI model.

[0129] Figure 5 A schematic diagram of another signaling interaction 500 according to some embodiments of the present disclosure is shown. Figure 5 Involving server 290, NMS 210, EMS 220, access network device 230 and terminal device 240. It can be understood that although signaling interaction 500 is combined with Figure 2 However, the process can also be applied to other systems, which is not limited by the present disclosure. Figure 5 In the process, the server 290 and the EMS 220 perform model training. Specifically, the server 290 obtains the part of the bilateral AI model to be deployed to the terminal device 240 (called the encoder model or the UE-side model), and the EMS 220 obtains the part of the bilateral AI model to be deployed to the access network device 230 (called the decoder model or the BS-side model).

[0130] At 501, NMS 210 may open AI / ML management services to server 290. This process is similar to the above Figure 4 The 401 in the above code is similar, so we will not go into details here to avoid repetition.

[0131] At 502, the server 290 sends a third model training request to the NMS 210. The third model training request may include an indication of training data for requesting UE-side model training. The third model training request may also include one or more of the following: a model identifier of a bilateral AI model, an identifier of the access network device 230, an identifier of a terminal device group (including the terminal device 240), an indication of the type of reasoning task for which the bilateral AI model will be used, model performance requirements (such as performance requirements of a decoder model), and a bilateral model training indication.

[0132] In some examples, the indication of training data for requesting UE-side model training can be represented as a Boolean value, for example, "true" indicates that training data for training the UE-side model is required, and "false" indicates that training data for training the UE-side model is not required. In other examples, the indication of training data for requesting UE-side model training can be represented as a coded value, for example, "1" indicates that training data for training the UE-side model is required, and "0" indicates that training data for training the UE-side model is not required.

[0133] In some embodiments, the bilateral model training indication and the indication of training data for requesting UE-side model training can be jointly indicated by a Boolean value or a code. For example, "true" indicates that the bilateral AI model needs to be trained, and training data for training the UE-side model is required. For example, "false" indicates that the bilateral AI model does not need to be trained. For example, the coded value "00" indicates that the bilateral AI model does not need to be trained; "10" indicates that only the BS-side model needs to be trained; and "11" indicates that the bilateral AI model needs to be trained, and training data for training the UE-side model is required.

[0134] For example, the model identifier of the bilateral AI model, the identifier of the access network device 230, the identifier of the terminal device group (including the terminal device 240), the indication information of the type of reasoning task for which the bilateral AI model will be used, the model performance requirements, the bilateral model training instructions, etc., can refer to the aforementioned combined Figure 4 The terms described in the embodiments are not repeated here.

[0135] Optionally, at 503, NMS 210 sends a device query request message to EMS 220; and at 504, EMS 220 sends a device query response message to NMS 210. For 503 and 504, please refer to the above combined Figure 4 The described 405 and 406 are not repeated here.

[0136] At 505, NMS210 sends a fourth model training request to EMS220. The fourth model training request may include part or all of the third model training request as described above. The fourth model training request includes an indication of training data for requesting UE-side model training. The fourth model training request may also include one or more of the following: a model identifier of a bilateral AI model, an identifier of the access network device 230, an identifier of a terminal device group (including the terminal device 240), an indication of the type of reasoning task for which the bilateral AI model will be used, model performance requirements, and a bilateral model training indication.

[0137] In some embodiments, the fourth model training request may also include one or more of the following about the decoder model: training energy consumption requirement, inference energy consumption requirement, model size requirement, or lightweight indication, etc. Figure 4 As described in detail in the embodiment of the second model training request.

[0138] At 506, EMS220 obtains training data. Specifically, EMS220 obtains training data collected by terminal device 240 through access network device 230. Specifically, terminal device 240 can perform data collection, measurement, calculation and other operations to obtain training data, and terminal device 240 can send the training data to access network device 230 via an uplink, and then access network device 230 can provide (for example, based on an instruction or request from EMS220) to EMS220. Assuming that this embodiment is intended to train a bilateral AI model for CSI feedback, the training data may include the original CSI measured by terminal device 240.

[0139] At 507 , the EMS 220 performs model training using the training data based on the fourth model training request to obtain a decoder model (ie, a BS-side model) and an initial encoder model (ie, a UE-side model).

[0140] Optionally, at 508, EMS 220 may perform lightweight processing on the decoder model. In some examples, if the fourth model training request includes a lightweight indication indicating that lightweight processing is required, EMS 220 performs lightweight processing. In other examples, even if the fourth model training request does not include a lightweight indication, or the fourth model training request indicates that lightweight processing is not required, EMS 220 may decide to perform lightweight processing at its own discretion. Figure 4 The embodiment described in 409.

[0141] Thus, EMS220 can obtain a trained decoder model through 507, or through 507 and 508. It should be noted that although Figure 5 508 is shown as a step independent of 507, but in other embodiments, 507 may include 508, that is, lightweight processing can be performed during model training.

[0142] At 509, EMS 220 sends a fourth model training report to NMS 210. Exemplarily, the fourth model training report may indicate that the decoder model has been trained, and the fourth model training report includes training data for training the encoder model (ie, UE-side model) and structural information of the encoder model.

[0143] Exemplarily, after EMS220 completes model training at 507, it can construct training data for training the encoder model (i.e., UE-side model) by server 290 based on the training data acquired at 506, and optionally based on the initial encoder model obtained by model training. Exemplarily, the fourth model training report can directly include training data for training the encoder model (i.e., UE-side model). Exemplarily, the fourth model training report can include a storage address of training data for training the encoder model (i.e., UE-side model), so that server 290 can download the training data for training the encoder model (i.e., UE-side model) from the storage address via FTP.

[0144] Exemplarily, the fourth model training report may include a model identifier of the trained decoder model and an identifier of the access network device 230. Exemplarily, the fourth model training report may also include one or more of the following: actual performance of the decoder model, actual training energy consumption of the decoder model, actual inference energy consumption of the decoder model, and actual model size of the decoder model.

[0145] At 510, NMS210 sends a third model training report to server 290. Exemplarily, the third model training report may indicate that the decoder model has been trained, and the third model training report includes training data for training the encoder model (i.e., the UE-side model) and structural information of the encoder model. In this way, the training data and structural information for training the encoder model can be provided to server 290, so that server 290 obtains the encoder model based on this training, thereby ensuring matching with the trained decoder model. Exemplarily, the third model training report may include part or all of the fourth model training report. For example, the third model training report may include: a model identifier of a trained decoder model, an identifier of the access network device 230, and the actual performance of the decoder model.

[0146] Optionally, the third model training report may also include a model performance requirement of the encoder model (ie, the UE side model). For example, the model performance requirement of the encoder model indicates the accuracy requirement of the encoder model when it is subsequently used for reasoning tasks, such as 90%, 88%, or other values.

[0147] At 511, the server 290 performs model training based on the third model training report to obtain a trained encoder model. Specifically, the server 290 can build a model based on the structural information of the encoder model, and perform model training based on the training data used to train the encoder model (i.e., the UE side model). It can be understood that since the training at 511 is based on the training data and structural information provided by EMS220 for training the encoder model, the trained encoder model can match the decoder model trained by EMS220. In some examples, during the training process of 511, the model performance requirements for the encoder model in the third model training report can also be considered to ensure that the trained encoder model can adapt to the status information at the terminal device and ensure the availability of the encoder model.

[0148] At 512, the server 290 sends the encoder model to the terminal device 240. At 513, the server 290 sends a model sending notification to the NMS 210 to indicate that the encoder model has been deployed at the terminal device 240. At 514, the NMS 210 sends a model loading request to the EMS 220, and then at 515, the EMS 220 loads the decoder model to the access network device 230. At 516, the bilateral AI model can be used in the subsequent model reasoning process. Specifically, 512 to 516 can be respectively referred to in conjunction with Figure 4 For the sake of brevity, the descriptions of 412 to 416 are not repeated here.

[0149] By combining Figure 5The described signaling interaction 500 can support the server 290 to call the model management service, and then the server 290 and the EMS 220 collaboratively complete the model training of the bilateral AI model, and are deployed at the terminal device 240 and the access network device 230, respectively. In some embodiments, during the training process, the state information of the access network device 230 can be considered at the same time, so that the appropriate model can be trained based on the computing power, storage, and energy consumption requirements of the access network device 230. The service types (i.e., model training services, model loading services) opened by the management plane (such as NMS 210) can be increased to support the server 290 outside the 3GPP network domain 250 to call the model training service and model loading service of the management plane. The third / fourth model training request can include a bilateral model indication and an indication of the training data for requesting UE-side model training, and the third / fourth model training report includes training data for training the encoder model and the structural information of the encoder model, thereby supporting the server 290 outside the 3GPP network domain 250 to train the UE-side model using the training data provided by the management plane. The model loading request / another model loading request may include a bilateral model indication and / or bilateral model association information, thereby indicating that the UE side model has been deployed, thereby being able to support the deployment and reasoning of the bilateral AI model.

[0150] Figure 6 A schematic diagram of yet another signaling interaction 600 according to some embodiments of the present disclosure is shown. Figure 6 Involving NMS210, EMS220, access network equipment 230 and terminal equipment 240. It can be understood that although the signaling interaction 600 is combined Figure 2 However, the process can also be applied to other systems, which is not limited by the present disclosure. Figure 6 In the process, EMS 220 performs model training to obtain the part of the bilateral AI model to be deployed to the terminal device 240 (called the encoder model or the UE-side model), and the part to be deployed to the access network device 230 (called the decoder model or the BS-side model).

[0151] Optionally, at 601 , the NMS 210 sends a device query request message to the EMS 220 ; and at 602 , the EMS 220 sends a device query response message to the NMS 210 .

[0152] Exemplarily, the device query request message can be used to request to query the status information of the access network device and to request to query the status information of the terminal device group. For example, the device query request message includes the identifier of the access network device 230 and the identifier of the terminal device group (such as including the terminal device 240). Correspondingly, the device query response message is used to provide the status information of the queried access network device and the status information of the terminal device group. In some embodiments, the status information of the access network device may include one or more of the following: the computing power status, storage status, energy consumption status, and network energy consumption target value of the access network device. In some embodiments, the status information of the terminal device group may include one or more of the following: the computing power status, storage status, energy consumption status, and network energy consumption target value of each terminal device in the terminal device group.

[0153] At 603, NMS 210 sends a fourth model training request to EMS 220. The fourth model training request may also include one or more of the following: a model identifier of the bilateral AI model, an identifier of the access network device 230, an identifier of the terminal device group (including the terminal device 240), indication information of the type of reasoning task for which the bilateral AI model will be used, model performance requirements (including performance requirements of the decoder model and / or performance requirements of the encoder model), and bilateral model training instructions.

[0154] In some embodiments, the fourth model training request may further include one or more of the following about the decoder model and / or the encoder model: training energy consumption requirement, inference energy consumption requirement, model size requirement, or lightweight indication, etc. For example, NMS210 may determine one or more of the following based on the device query response message: training energy consumption requirement, inference energy consumption requirement, model size requirement, or lightweight indication, etc. For example, NMS210 may determine one or more of the following about the decoder model based on the status information of the access network device 230: a first training energy consumption requirement, a first inference energy consumption requirement, a first model size requirement, or a first lightweight indication, etc. For example, NMS210 may determine one or more of the following about the encoder model based on the status information of the terminal device 240: a second training energy consumption requirement, a second inference energy consumption requirement, a second model size requirement, or a second lightweight indication, etc.

[0155] Optionally, the fourth model training request may include an indication of requesting training data for UE-side model training, for example, the indication is "false", that is, no training data for the UE-side model is required.

[0156] At 604, EMS 220 obtains training data. This process is combined with the above Figure 5 The embodiment of 506 is similar.

[0157] At 605, EMS 220 performs model training using the training data based on the fourth model training request to obtain a bilateral AI model, which includes a decoder model (i.e., a BS side model) and an encoder model (i.e., a UE side model). In this way, the two parts of the bilateral AI model can be obtained by training by EMS 220, avoiding the introduction of the server to participate in the training, and no longer needing to open additional training-related interfaces to the server, thereby reducing signaling interaction.

[0158] Optionally, at 606, EMS220 may perform lightweight processing on the bilateral AI model. In some examples, if the fourth model training request includes a first lightweight indication and a second lightweight indication, indicating that lightweight processing is required, EMS220 performs lightweight processing on the bilateral AI model (including the encoder model and the decoder model). In other examples, even if the fourth model training request does not include any lightweight indication, or the fourth model training request indicates that lightweight processing is not required, EMS220 may decide at its own discretion to perform lightweight processing on the encoder model and / or decoder model in the bilateral AI model. The operation on lightweight processing is combined with the aforementioned Figure 4 The process described for 409 is similar.

[0159] In this way, EMS220 can obtain a trained bilateral AI model through 605, or through 605 and 606, including a trained encoder model and a trained decoder model. It should be noted that although Figure 6 606 is shown as a step independent of 605, but in other embodiments, 605 may include 606, that is, lightweight processing can be performed during model training.

[0160] At 607, EMS220 sends a fourth model training report to NMS210. Exemplarily, the fourth model training report includes a bilateral model indication to indicate that the bilateral AI model (including the encoder model and the decoder model) has been trained. Exemplarily, the fourth model training report may also include one or more of the following: a model identifier of the encoder model, an identifier of the terminal device group, a model identifier of the decoder model, an identifier of the access network device 230, an actual performance of the bilateral AI model, an actual training energy consumption of the bilateral AI model, an actual reasoning energy consumption of the bilateral AI model, an actual model size of the bilateral AI model, etc. For example, the actual performance of the bilateral AI model includes the actual performance of the encoder model and / or the actual performance of the decoder model, the actual training energy consumption of the bilateral AI model includes the actual training energy consumption of the encoder model and / or the actual training energy consumption of the decoder model, the actual reasoning energy consumption of the bilateral AI model includes the actual reasoning energy consumption of the encoder model and / or the actual reasoning energy consumption of the decoder model, and the actual model size of the bilateral AI model includes the actual model size of the encoder model and / or the actual model size of the decoder model.

[0161] At 608, NMS 210 sends a model loading request to EMS 220, the model loading request is used to instruct EMS 220 to deploy the bilateral AI model, specifically, to deploy the encoder model to the terminal device group (including the terminal device 240), and to deploy the decoder model to the access network device 230. For example, the model loading request includes a bilateral model indication, a model identifier of the encoder model, an identifier of the terminal device group, a model identifier of the decoder model, and an identifier of the access network device 230.

[0162] At 609 , the EMS 220 loads the bilateral AI model to the access network device 230 based on the model loading request; and at 610 , the access network device 230 transfers the encoder model to the terminal device 240 .

[0163] Exemplarily, EMS220 may send another model loading request to access network device 230 for deploying the encoder model to terminal device 240 and the decoder model to access network device 230. In some examples, the decoder model and / or the encoder model may be expressed in the form of model parameters or carried in the form of model files. For example, another model loading request includes the decoder model itself, or includes the model parameters of the decoder model or the file address of the model file of the decoder model. Another model loading request includes the encoder model itself, or includes the model parameters of the encoder model or the file address of the model file of the encoder model. Exemplarily, another model loading request may also include a bilateral model indication, a model identifier of the encoder model, an identifier of the terminal device group, a model identifier of the decoder model, and an identifier of the access network device 230.

[0164] Exemplarily, the access network device 230 may send the encoder model to each terminal device in the terminal device group, including the terminal device 240. In some examples, the access network device 230 may send one or more of the following to the terminal device 240: the encoder model, the bilateral model indication, the model identifier of the encoder model, and the model identifier of the decoder model.

[0165] At 611, the bilateral AI model may be used in subsequent model reasoning processes.

[0166] By combining Figure 6 The described signaling interaction 600 can complete the model training of the bilateral AI model by EMS220, and be deployed at the terminal device 240 and the access network device 230 respectively. In some embodiments, during the training process, the status information of the access network device and the terminal device can be considered at the same time, so that the appropriate model can be trained based on the computing power, storage, and energy consumption requirements of the access network device and the terminal device. The fourth model training request and the fourth model training report may include a bilateral model indication, and the model loading request / another model loading request may include a bilateral model indication and / or bilateral model association information, so as to support the deployment and reasoning of the bilateral AI model.

[0167] In the above combination Figures 4 to 6 In the embodiment of the present invention, the EMS 220 of the management plane performs model training, for example, obtaining a trained BS side model ( Figure 4 or Figure 5 ) or get a trained bilateral AI model ( Figure 6 ). During the training process, bilateral training instructions can be used to ensure the match between the BS side model and the UE side model, that is, it is used as a bilateral AI model. When deploying the model, bilateral instructions can be used to enable the access network device and the terminal device to determine the association between the BS side model and the UE side model, thereby avoiding considering the deployed model as a unilateral model. In this way, the match and association between the BS side model and the UE side model during the model reasoning process can be ensured, thereby ensuring the performance of the model reasoning.

[0168] Figure 7 A schematic diagram of a signaling interaction 700 according to some embodiments of the present disclosure is shown. Figure 7 Involving server 290, NMS 210, EMS 220, access network device 230 and terminal device 240. It can be understood that although signaling interaction 700 is combined with Figure 2 However, the process can also be applied to other systems, which is not limited by the present disclosure. Figure 7In the process, the server 290 and the NMS 210 perform model training. Specifically, the server 290 obtains the part of the bilateral AI model to be deployed to the terminal device 240 (called the encoder model or the UE-side model), and the NMS 210 obtains the part of the bilateral AI model to be deployed to the access network device 230 (called the decoder model or the BS-side model).

[0169] At 701, NMS 210 may open AI / ML management services to server 290. At 702, server 290 obtains training data. At 703, server 290 performs model training based on the training data to obtain an encoder model (i.e., a UE-side model) and an initial decoder model. At 704, server 290 sends a first model training request to NMS 210. Optionally, at 705, NMS 210 sends a device query request message to EMS 220; and at 706, EMS 220 sends a device query response message to NMS 210. Specifically, 701 to 706 may be respectively referred to in conjunction with Figure 4 For the sake of brevity, steps 401 to 406 described above will not be repeated here.

[0170] At 707, NMS 210 performs model training based on the first model training request. Specifically, by using the training data for training the decoder model (i.e., BS-side model) indicated in the first model training request and the structural information of the decoder model, NMS 210 obtains the decoder model (i.e., BS-side model) through training.

[0171] Optionally, at 708, NMS210 performs lightweight processing on the decoder model. Specifically, NMS210 can determine one or more of the following about the decoder model based on the device query response message received at 706: training energy consumption requirement, inference energy consumption requirement, or model size requirement, etc. For example, NMS210 can decide to perform lightweight processing based on model performance requirements, training energy consumption requirements, inference energy consumption requirements, model size requirements, etc. Exemplarily, NMS210 can perform quantization, pruning, etc. on the decoder model to obtain a decoder model after lightweight processing. It can be understood that lightweight processing is to compress the model under the premise of minimizing the model performance requirements (such as accuracy) to reduce the model's computational complexity, size (such as the number of parameters), etc.

[0172] In this way, NMS 210 can obtain a trained decoder model that matches the encoder model trained by server 290 through 707, or through 707 and 708. It should be noted that although Figure 7 708 is shown as a step independent of 707, but in other embodiments, 707 may include 708, that is, lightweight processing can be performed during model training.

[0173] At 709, the NMS 210 sends a first model training report to the server 290 to indicate that the decoder model has been trained. For example, the first model training report may include a model identifier of the trained decoder model and actual performance of the decoder model.

[0174] At 710, the server 290 sends the encoder model to the terminal device 240. At 711, the server 290 sends a model sending notification to the NMS 210 to indicate that the encoder model has been deployed at the terminal device 240. Specifically, 710 to 711 can be respectively referred to in conjunction with Figure 4 For the sake of brevity, the described steps 412 to 413 are not repeated here.

[0175] At 712, NMS210 loads the decoder model to the access network device 230. Exemplarily, NMS210 may send a model loading request to the access network device 230. Exemplarily, NMS210 may deploy the decoder model to the access network device 230 via EMS220. For example, NMS210 may send a model loading request to EMS220, and then EMS220 forwards the model loading request to the access network device 230. For example, the model loading request includes the decoder model itself, or includes model parameters of the decoder model or a file address of a model file of the decoder model. For example, the model loading request may also include a bilateral model indication, a model identifier of the encoder model, and a terminal device group identifier.

[0176] At 713, the bilateral AI model may be used in subsequent model reasoning processes.

[0177] By combining Figure 7The described signaling interaction 700 can support the server 290 to call the model management service, and then the server 290 and the NMS 210 collaboratively complete the model training of the bilateral AI model, and are deployed at the terminal device 240 and the access network device 230, respectively. In some embodiments, during the training process, the state information of the access network device 230 can be considered at the same time, so that the appropriate model can be trained based on the computing power, storage, and energy consumption requirements of the access network device 230. The service types (i.e., model training services, model loading services) opened by the management plane (such as NMS 210) can be increased to support the server 290 outside the 3GPP network domain 250 to call the model training service and model loading service of the management plane. The first model training request can include a bilateral model indication, a model identifier of the UE side model, a model structure of the BS side model, and training data of the BS side model, thereby supporting the management plane (such as NMS 210) to train the BS side model with the training data provided by the external server 290. The model loading request may include a bilateral model indication and / or bilateral model association information, thereby indicating that the UE-side model has been deployed, thereby supporting the deployment and reasoning of the bilateral AI model.

[0178] Figure 8 A schematic diagram of another signaling interaction 800 according to some embodiments of the present disclosure is shown. Figure 8 Involving server 290, NMS 210, EMS 220, access network device 230 and terminal device 240. It can be understood that although signaling interaction 800 is combined with Figure 2 However, the process can also be applied to other systems, which is not limited by the present disclosure. Figure 8 In the process, the server 290 and the NMS 210 perform model training. Specifically, the server 290 obtains the part of the bilateral AI model to be deployed to the terminal device 240 (called the encoder model or the UE-side model), and the NMS 210 obtains the part of the bilateral AI model to be deployed to the access network device 230 (called the decoder model or the BS-side model).

[0179] At 801, NMS 210 may open AI / ML management services to server 290. At 802, server 290 sends a third model training request to NMS 210. Optionally, at 803, NMS 210 sends a device query request message to EMS 220; and at 804, EMS 220 sends a device query response message to NMS 210. Specifically, 801 to 804 may be referred to in conjunction with Figure 5 For the sake of brevity, steps 501 to 504 described are not repeated here.

[0180] At 805, NMS210 obtains training data. Specifically, NMS210 can obtain training data collected by terminal device 240 via EMS220 and access network device 230. Specifically, terminal device 240 can perform data collection, measurement, calculation and other operations to obtain training data, and terminal device 240 can send the training data to access network device 230 via an uplink, and then access network device 230 can provide (for example, based on an instruction or request from EMS220) to EMS220, so that NMS210 can obtain training data from EMS220. Assuming that this embodiment is intended to train a bilateral AI model for CSI feedback, the training data may include the original CSI measured by terminal device 240.

[0181] At 806 , the NMS 210 performs model training using the training data based on the third model training request to obtain a decoder model (ie, a BS-side model) and an initial encoder model (ie, a UE-side model).

[0182] Optionally, at 807, NMS210 may perform lightweight processing on the decoder model. Specifically, NMS210 may determine one or more of the following about the decoder model based on the device query response message received at 804: training energy consumption requirement, inference energy consumption requirement, or model size requirement, etc. For example, NMS210 may decide to perform lightweight processing based on model performance requirements, training energy consumption requirements, inference energy consumption requirements, model size requirements, etc. Exemplarily, NMS210 may perform quantization, pruning, etc. on the decoder model to obtain a decoder model after lightweight processing. In this way, NMS210 may obtain a trained decoder model through 806, or through 806 and 807. It should be noted that although in Figure 8 807 is shown as a step independent of 806, but in other embodiments, 806 may include 807, that is, lightweight processing can be performed during model training.

[0183] At 808, NMS 210 sends a third model training report to server 290. Exemplarily, the third model training report may indicate that the decoder model has been trained, and the third model training report includes training data for training the encoder model (ie, UE-side model) and structural information of the encoder model.

[0184] Exemplarily, after completing the model training at 806, NMS210 may construct training data for training the encoder model (i.e., UE-side model) by server 290 based on the training data acquired at 805, and optionally based on the initial encoder model obtained by the model training. Exemplarily, the third model training report may directly include the training data for training the encoder model (i.e., UE-side model). Exemplarily, the third model training report may include the storage address of the training data for training the encoder model (i.e., UE-side model), so that server 290 may download the training data for training the encoder model (i.e., UE-side model) from the storage address via FTP.

[0185] Exemplarily, the third model training report may include a model identifier of the trained decoder model and an identifier of the access network device 230. Exemplarily, the third model training report may also include one or more of the following: actual performance of the decoder model, actual training energy consumption of the decoder model, actual inference energy consumption of the decoder model, and actual model size of the decoder model.

[0186] At 809, the server 290 performs model training based on the third model training report to obtain a trained encoder model. Specifically, the server 290 can build a model based on the structural information of the encoder model, and perform model training based on the training data used to train the encoder model (i.e., the UE-side model). It can be understood that since the training at 809 is based on the training data and structural information provided by NMS210 for training the encoder model, the trained encoder model can match the decoder model trained by NMS210. In some examples, during the training process at 809, the model performance requirements for the encoder model in the third model training report can also be considered to ensure that the trained encoder model can adapt to the status information at the terminal device and ensure the availability of the encoder model.

[0187] At 810, the server 290 sends the encoder model to the terminal device 240. At 811, the server 290 sends a model delivery notification to the NMS 210 to indicate that the encoder model has been deployed at the terminal device 240. At 812, the NMS 210 can deploy the decoder model to the access network device 230 via the EMS 220. At 813, the bilateral AI model can be used in the subsequent model reasoning process. Specifically, 810 to 813 can be respectively referred to in conjunction with Figure 7 For the sake of brevity, the described steps 710 to 713 are not repeated here.

[0188] By combining Figure 8The described signaling interaction 800 can support the server 290 to call the model management service, and then the server 290 and the NMS 210 collaboratively complete the model training of the bilateral AI model, and are deployed at the terminal device 240 and the access network device 230, respectively. In some embodiments, during the training process, the status information of the access network device 230 can be considered at the same time, so that the appropriate model can be trained based on the computing power, storage, and energy consumption requirements of the access network device 230. The service types (i.e., model training services, model loading services) opened by the management plane (such as NMS 210) can be increased to support the server 290 outside the 3GPP network domain 250 to call the model training service and model loading service of the management plane. The third model training request can include a bilateral model indication and an indication of the training data for requesting UE-side model training, and the third model training report includes training data for training the encoder model and the structural information of the encoder model, thereby supporting the server 290 outside the 3GPP network domain 250 to train the UE-side model using the training data provided by the management plane. The model loading request may include a bilateral model indication and / or bilateral model association information, thereby indicating that the UE-side model has been deployed, thereby supporting the deployment and reasoning of the bilateral AI model.

[0189] Fig. 9 A schematic diagram of yet another signaling interaction 900 according to some embodiments of the present disclosure is shown. Fig. 9 Involving NMS210, EMS220, access network equipment 230 and terminal equipment 240. It can be understood that although the signaling interaction 900 is combined with Figure 2 However, the process can also be applied to other systems, which is not limited by the present disclosure. Fig. 9 In the process, NMS 210 performs model training to obtain the part of the bilateral AI model to be deployed to the terminal device 240 (called the encoder model or the UE-side model), and the part to be deployed to the access network device 230 (called the decoder model or the BS-side model).

[0190] Optionally, at 901, NMS 210 sends a device query request message to EMS 220; and at 902, EMS 220 sends a device query response message to NMS 210. Specifically, 901 and 902 may be referred to in conjunction with Figure 6 For the sake of brevity, the described steps 601 and 602 are not repeated here.

[0191] In 903, NMS 210 obtains training data. Specifically, NMS 210 can obtain the training data collected by terminal device 240 via EMS 220 and access network device 230. This process can refer to the above-mentioned combination Figure 8 The description of 805 will not be repeated here.

[0192] At 904, NMS 210 performs model training using the training data to obtain a bilateral AI model, which includes a decoder model (i.e., a BS side model) and an encoder model (i.e., a UE side model). In this way, the two parts of the bilateral AI model can be obtained by training by NMS 210, avoiding the introduction of the server to participate in the training, and no longer needing to open additional training-related interfaces to the server, thereby reducing signaling interactions.

[0193] Optionally, at 905, NMS210 may perform lightweight processing on the bilateral AI model. Specifically, NMS210 may determine one or more of the following about the decoder model and / or encoder model based on the device query response message received at 902: training energy consumption requirement, inference energy consumption requirement, or model size requirement, etc. For example, NMS210 may determine one or more of the following about the decoder model based on the status information of the access network device 230: a first training energy consumption requirement, a first inference energy consumption requirement, or a first model size requirement, etc. For example, NMS210 may determine one or more of the following about the encoder model based on the status information of the terminal device 240: a second training energy consumption requirement, a second inference energy consumption requirement, or a second model size requirement, etc.

[0194] For example, NMS210 may decide to perform lightweight processing on the encoder model and / or decoder model based on the first and / or second model performance requirements, the first and / or second training energy consumption requirements, the first and / or second inference energy consumption requirements, the first and / or second model size requirements, etc. Exemplarily, NMS210 may perform quantization, pruning, etc. on the encoder model and / or decoder model to obtain a bilateral AI model after lightweight processing. In this way, NMS210 may obtain a trained bilateral AI model, including an encoder model and a decoder model, through 904, or through 904 and 905. It should be noted that although in Fig. 9 905 is shown as a step independent of 904, but in other embodiments, 904 may include 905, that is, lightweight processing can be performed during model training.

[0195] At 906, NMS 210 may load the bilateral AI model to access network device 230 via EMS 220; and at 907, access network device 230 transmits the encoder model to terminal device 240. For example, NMS 210 may send a model loading request to access network device 230.

[0196] For example, NMS 210 may send a model loading request to EMS 220, and then EMS 220 forwards the model loading request to access network device 230. For example, the model loading request includes the decoder model itself, or includes the model parameters of the decoder model or the file address of the model file of the decoder model; the model loading request also includes the encoder model itself, or includes the model parameters of the encoder model or the file address of the model file of the encoder model. For example, the model loading request may also include a bilateral model indication, a model identifier of the encoder model, a model identifier of the decoder model, an identifier of the access network device 230, and a terminal device group identifier.

[0197] Exemplarily, the access network device 230 may send the encoder model to each terminal device in the terminal device group, including the terminal device 240. In some examples, the access network device 230 may send one or more of the following to the terminal device 240: the encoder model, the bilateral model indication, the model identifier of the encoder model, and the model identifier of the decoder model.

[0198] At 908, the bilateral AI model may be used in a subsequent model reasoning process.

[0199] By combining Fig. 9 The described signaling interaction 900 can complete the model training of the bilateral AI model by NMS210, and be deployed at the terminal device 240 and the access network device 230 respectively. In some embodiments, during the training process, the status information of the access network device and the terminal device can be considered at the same time, so that the appropriate model can be trained based on the computing power, storage, and energy consumption requirements of the access network device and the terminal device. The bilateral model indication and / or bilateral model association information can be included in the model loading request, so as to support the deployment and reasoning of the bilateral AI model.

[0200] In the above combination Figures 7 to 9 In the embodiment, the NMS 210 of the management plane performs model training, for example, obtaining a trained BS side model ( Figure 7 or Figure 8 ) or get a trained bilateral AI model ( Fig. 9 ). During the training process, bilateral training instructions can be used to ensure the match between the BS side model and the UE side model, that is, it is used as a bilateral AI model. When deploying the model, bilateral instructions can be used to enable the access network device and the terminal device to determine the association between the BS side model and the UE side model, thereby avoiding considering the deployed model as a unilateral model. In this way, the match and association between the BS side model and the UE side model during the model reasoning process can be ensured, thereby ensuring the performance of the model reasoning.

[0201] Fig. 10A1010 is a schematic diagram of a training scenario 1010 according to some embodiments of the present disclosure. Scenario 1010 corresponds to an embodiment in which a UE-side model is trained by a server and a BS-side model is trained by an NMS or an EMS, as described above. Figure 4 or Figure 7 As shown. Fig. 10A As shown, the server obtains training data from the UE and provides the training data for training the BS side model to the NMS or EMS. Furthermore, the server deploys the trained UE side model to the UE, while the NMS or EMS deploys the trained BS side model to the BS.

[0202] Fig. 10B 1020 is a schematic diagram of a training scenario according to some embodiments of the present disclosure. Scenario 1020 corresponds to an embodiment in which a UE-side model is trained by a server and a BS-side model is trained by an NMS or an EMS, as described above. Figure 5 or Figure 8 As shown. Fig. 10B As shown, the NMS or EMS obtains training data from the UE via the BS, and provides the training data for training the UE-side model to the server. Furthermore, the server deploys the trained UE-side model to the UE, while the NMS or EMS deploys the trained BS-side model to the BS.

[0203] Fig. 10C 1030 is a schematic diagram of a training scenario according to some embodiments of the present disclosure. Scenario 1030 corresponds to an embodiment in which a UE-side model and a BS-side model are trained by an NMS or an EMS, as described above. Figure 6 or Fig. 9 As shown. Fig. 10C As shown, the NMS or EMS obtains training data from the UE via the BS. In addition, the NMS or EMS provides both the trained BS-side model and the UE-side model to the BS, and the BS then deploys the UE-side model to the UE.

[0204] Through the example embodiments of the present disclosure as described above, the model training interface for bilateral AI model training is enhanced, such as the interface between the server 290 and the NMS 210, and the interface between the NMS 210 and the EMS 220, so that the collaboration between the server 290 and the management plane (including the NMS 210 and the EMS 220) can be enhanced; the model loading interface for bilateral AI model loading is also enhanced, such as the interface between the server 290 and the NMS 210, between the NMS 210 and the EMS 220, and between the NMS 210 / EMS 220 and the access network device 230, and between the server 290 and the terminal device 240, so that the successful deployment of the model can be determined.

[0205] In this way, the embodiments of the present disclosure implement the training process of the bilateral AI model of the air interface assisted by the management plane, thereby supporting the use (inference) of the bilateral AI model in the air interface.

[0206] It should be understood that the above Figure 4 to Figure 1 0 is combined with Figure 2 Although the system 200 shown in FIG. 1 is described, the embodiments of the present disclosure may also be applied to Figure 3 The system 300 or other systems not shown in this document. Figure 4 to Figure 1 0, involving interactions between different devices (or network elements, or systems), but the embodiments of the present disclosure do not limit the specific form of the interactive signaling, for example, it can be in various forms such as messages, notifications, commands, requests, responses, reports, information, signaling, etc.

[0207] It should be understood that although the encoder model and the decoder model are used to represent the UE side model and the BS side model respectively in the above embodiment, in actual situations, the bilateral AI model is not limited to the architecture of the encoder and the decoder. For example, the bilateral AI model may include a first part deployed at the access network device and a second part deployed at the terminal device. For example, the encoder model and the decoder model in the above embodiment may be replaced by the second part and the first part respectively.

[0208] It should also be understood that the methods, situations, categories and divisions of the embodiments in the present disclosure are only for the convenience of description and should not constitute special limitations. The features in various methods, categories, situations and embodiments can be combined with each other when it is logical.

[0209] It should also be understood that the above content is only to help those skilled in the art better understand the embodiments of the present disclosure, rather than to limit the scope of the embodiments of the present disclosure. Those skilled in the art may make various modifications, changes or combinations based on the above content. Such modifications, changes or combinations are also within the scope of the embodiments of the present disclosure.

[0210] It should also be understood that the description of the above content focuses on emphasizing the differences between the various embodiments, and the same or similar points can be referenced or borrowed from each other. For the sake of brevity, they will not be repeated here.

[0211] Fig.11 FIG. 1 is a schematic block diagram of a network management device 1100 according to some embodiments of the present disclosure. The network management device 1100 may be implemented as follows: Figure 2 The network management device 1100 may be implemented as the NMS 210 or EMS 220 shown in the figure, or may be implemented as a part of the NMS 210 or EMS 220 (such as a chip), etc., and the present disclosure is not limited to this. Figure 3The non-RT RIC 310 or near-RT RIC 320 shown in the figure may be implemented as a part of the non-RT RIC 310 or near-RT RIC 320 (such as a chip), etc., and the present disclosure is not limited to this. Fig.11 As shown, the network management device 1100 includes a training data acquisition module 1110 , a model training module 1120 , and a model sending module 1130 .

[0212] The training data acquisition module 1110 is configured to acquire training data for the first part of the bilateral model, wherein the bilateral model includes a first part to be deployed at the network device and a second part to be deployed at the terminal device. The model training module 1120 is configured to perform model training based on the training data to obtain the first part. The model deployment module 1130 is configured to send the first part to the network device. For example, the network device may be Figure 2 The access network device 230 shown or Figure 3 Access network equipment 330 is shown.

[0213] In some embodiments, the model training module 1120 may be further configured to perform lightweight processing on the first part; or, the network management device 1100 may further include a lightweight processing module configured to perform lightweight processing on the first part. Exemplarily, the lightweight first part meets at least one of the following: performance requirements of the first part, energy consumption requirements of training the first part, energy consumption requirements of the reasoning task to which the first part will be used, or first model size requirements of the first part.

[0214] In some implementations of the present disclosure, the network management device 1100 may be implemented as a second network management device, such as the EMS 220 or the near-RT RIC 320 .

[0215] In some examples, the training data acquisition module 1110 is specifically configured to receive a first training request from a first network management device, the first training request indicating training a first part of the bilateral model, the first training request including training data for training the first part of the bilateral model and structural information of the first part.

[0216] In some examples, the network management device 1100 may also include a reporting module configured to send a first model training report to the first network management device, the first model training report including at least one of the following: a model identification of the bilateral model, a model identification of the first part, an actual performance of the first part, an actual energy consumption of training the first part, an estimated energy consumption of an inference task for which the first part will be used, or an actual model size of the first part.

[0217] In some examples, the network management device 1100 may further include a receiving module configured to receive a model loading request from the first network management device, the model loading request indicating that the first part is loaded to the network device. The model sending module 1130 may be configured to send the first part to the network device based on the model loading request.

[0218] Optionally, the model loading request may include at least one of the following: a first identifier of the network device, a model identifier of the first part, a second identifier of the terminal device, a model identifier of the second part, or a model identifier of the bilateral model.

[0219] Optionally, the first training request also includes at least one of the following: a model identifier of the bilateral model, a model identifier of the first part, a model identifier of the second part, a first identifier of the network device, a second identifier of the terminal device, indication information for indicating the type of reasoning task for which the bilateral model will be used, performance requirements of the bilateral model, energy consumption requirements for bilateral model training, energy consumption requirements for reasoning tasks, first model size requirements for the first part, or a lightweight indication.

[0220] In some examples, the network management device 1100 may also include a request receiving module and a response sending module. The request receiving module is configured to receive a request message from the first network management device, and the request message is used to request the status information of the network device. The response sending module is configured to send a response message to the first network management device in response to the request message, and the response message includes at least one of the following: the computing power status of the network device, the storage status of the network device, the energy consumption information of the network device, or the network energy consumption target value of the network device.

[0221] In some examples, the network management device 1100 may also include a reporting module configured to send a second model training report to the first network management device, the second model training report including training data for training the second part and structural information of the second part.

[0222] Optionally, the second model training report also includes at least one of the following: a model identification of the bilateral model, a model identification of the first part, a first identification of the network device, an actual performance of the first part, an actual energy consumption of training the first part, an estimated energy consumption of the reasoning task for which the first part will be used, an actual model size of the first part, a performance requirement of the second part, a second energy consumption requirement for the reasoning task for which the second part will be used, or a second model size requirement for the second part.

[0223] The network management device 1100 is implemented as a second network management device, which can be used to implement the above combination Figures 4 to 6 For the sake of brevity, the various processes performed by EMS220 are not repeated here.

[0224] In some other implementations of the present disclosure, the network management device 1100 may be implemented as a first network management device, such as an NMS 210 or a non-RT RIC 310 .

[0225] In some examples, the training data acquisition module 1110 is specifically configured to receive a second training request from the server, the second training request indicating training the first part of the bilateral model, the second training request including training data for training the first part of the bilateral model and structural information of the first part.

[0226] In some examples, the network management device 1100 may also include a reporting module configured to send a fourth model training report to the server, the fourth model training report including at least one of the following: a model identification of the bilateral model, a model identification of the first part, or actual performance of the first part.

[0227] In some examples, the network management device 1100 may further include a receiving module configured to receive a notification message from the server, the notification message indicating that the second part trained by the server has been loaded into the terminal device. The model sending module 1130 may be configured to send the first part to the network device in response to the notification message.

[0228] Optionally, the second training request also includes at least one of the following: a model identifier of the bilateral model, a model identifier of the first part, a model identifier of the second part, a first identifier of the network device, a second identifier of the terminal device, indication information for indicating the type of reasoning task for which the bilateral model will be used, or performance requirements of the bilateral model.

[0229] In some examples, the network management device 1100 may further include a request sending module and a response receiving module. The request sending module is configured to send a request message to the second network management device, and the request message includes a first identifier of the network device. The response receiving module is configured to receive a response message from the second network management device, and the response message includes at least one of the following: the computing power status of the network device, the storage status of the network device, the energy consumption information of the network device, or the network energy consumption target value of the network device.

[0230] Exemplarily, the network management device 1100 may also include a determination module configured to determine at least one of the following based on the response message: the energy consumption requirement of bilateral model training, the first energy consumption requirement of the reasoning task for which the first part will be used, the second energy consumption requirement of the reasoning task for which the second part will be used, the first model size requirement of the first part, or the second model size requirement of the second part.

[0231] In some examples, the network management device 1100 may also include a reporting module configured to send a second model training report to the server, the second model training report including training data for training the second part and structural information of the second part.

[0232] Optionally, the second model training report also includes at least one of the following: a model identification of the bilateral model, a model identification of the first part, a first identification of the network device, an actual performance of the first part, an actual energy consumption of training the first part, an estimated energy consumption of the reasoning task for which the first part will be used, an actual model size of the first part, a performance requirement of the second part, a second energy consumption requirement for the reasoning task for which the second part will be used, or a second model size requirement for the second part.

[0233] The network management device 1100 is implemented as a first network management device, which can be used to implement the above combination Figures 7 to 9 For the sake of brevity, the various processes performed by NMS210 are not described here in detail.

[0234] Fig.12 FIG. 1 is a schematic block diagram of a network management device 1200 according to some embodiments of the present disclosure. The network management device 1200 may be implemented as follows: Figure 2 The network management device 1100 may be implemented as a part of the NMS 210 (such as a chip), etc., which is not limited in the present disclosure. Figure 3 The non-RT RIC 310 shown in the figure may be implemented as a part of the non-RT RIC 310 (such as a chip), etc., and the present disclosure is not limited to this. Fig.12 As shown, the network management device 1200 includes a sending module 1210 and a receiving module 1220 .

[0235] The sending module 1210 is configured to send a first training request to the second network management device, the first training request indicating training the first part of the bilateral model, the bilateral model including the first part to be deployed at the network device and the second part to be deployed at the terminal device. The receiving module 1220 is configured to receive a model training report from the second network management device, the model training report including at least the model identifier of the first part.

[0236] Optionally, the first training request includes: training data for training the first part, and structural information of the first part.

[0237] Optionally, the model training report includes a first model training report, which includes at least one of the following: a model identification of the bilateral model, a model identification of the first part, an actual performance of the first part, an actual energy consumption of training the first part, an estimated energy consumption of an inference task for which the first part will be used, or an actual model size of the first part.

[0238] In some examples, the receiving module 1220 is further configured to receive a second training request from the server, the second training request indicating training the first part of the bilateral model, the second training request including: training data for training the first part, and structural information of the first part.

[0239] Optionally, the model training report includes a second model training report, and the second model training report includes training data for training the second part and structural information of the second part.

[0240] Optionally, the second model training report also includes at least one of the following: the model identification of the bilateral model, the model identification of the first part, the first identification of the network device, the actual performance of the first part, the actual energy consumption of the first part training, the estimated energy consumption of the inference task for which the first part will be used, and the actual model size of the first part.

[0241] In some examples, the sending module of the network management device 1200 may also be configured to forward the second model training report to the server, and the receiving module may also be configured to receive a notification message from the server, the notification message indicating that the second part has been loaded into the terminal device. The sending module may also be configured to send a model loading request to the second network management device, the model loading request indicating that the first part is loaded into the network device.

[0242] The network management device 1200 is implemented as a first network management device, which can be used to implement the above combination Figures 4 to 6 For the sake of brevity, the various processes performed by NMS210 are not described here in detail.

[0243] Fig.13 A schematic block diagram of a server 1300 according to some embodiments of the present disclosure is shown. The server 1300 may be implemented as follows Figure 2 The server 290 shown in FIG. 1 may be implemented as a part of the server 290 (such as a chip), etc., and the present disclosure is not limited to this. The server 1300 may be implemented as Figure 3 The server 390 shown in the figure may be implemented as a part of the server 390 (such as a chip), etc., which is not limited in the present disclosure. Fig.13 As shown, the server 1300 includes a training data acquisition module 1310 , a model training module 1320 , and a model sending module 1330 .

[0244] The training data acquisition module 1310 is configured to acquire training data for training the second part of the bilateral model, wherein the bilateral model includes a first part to be deployed at the network device and a second part to be deployed at the terminal device. The model training module 1320 is configured to perform model training based on the training data to obtain the second part. The model sending module 1330 is configured to send the second part to the terminal device.

[0245] In some examples, server 1300 also includes a sending module configured to send a second training request to the first network management device, the second training request indicating training the first part of the bilateral model, the second training request including training data for training the first part of the bilateral model, and structural information of the first part.

[0246] In some examples, server 1300 also includes a receiving module configured to receive a first model training report from a first network management device, the first model training report indicating that the first part has been trained.

[0247] In some examples, the training data acquisition module 1310 is specifically configured to receive a second model training report from the first network management device, where the second model training report includes training data for training the second part and structural information of the second part.

[0248] Optionally, the second model training report also includes at least one of the following: a model identification of the bilateral model, a model identification of the first part, a first identification of the network device, an actual performance of the first part, a performance requirement of the second part, a second energy consumption requirement of the inference task for which the second part will be used, or a second model size requirement for the second part.

[0249] In some examples, the server 1300 further includes a sending module configured to send a notification message to the first network management device, the notification message indicating that the second part has been loaded into the terminal device.

[0250] The server 1300 can be used to implement the above combination Figures 4 to 5 and Figures 7 and 8 For the sake of brevity, the various processes executed by the server 290 are not repeated here.

[0251] The division of modules or units in the embodiments of the present disclosure is schematic and is only a logical function division. There may be other division methods in actual implementation. In addition, each functional unit in the disclosed embodiments may be integrated into one unit, or may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0252] The embodiment of the present disclosure also provides a system, which may include a first network management device, a second network management device, a server, a network device, and a terminal device. Exemplarily, the server may be a server 290 or a server 390, which may obtain the second part of the bilateral model through model training. Exemplarily, the first network management device may be an NMS 210 or a non-RT RIC 310, which may receive a second training request from the server. The first network management device may also send a first training request to the second network management device, or the first network management device may obtain the first part of the bilateral model through training. Exemplarily, the second network management device may be an EMS 220 or a near-RT RIC 320, which may receive a first training request from the first network management device; and perform model training based on the first training request to obtain the first part. Exemplarily, the network device may be an access network device 230 or an access network device 330 (such as an O-CU 332 and an O-DU 334), which may receive the first part from the first network management device or the second network management device. Optionally, the network device may also receive the second part and send the second part to the terminal device. Exemplarily, the terminal device may be the terminal device 240 or 340, which may receive the second part from the network device, and the second part may be obtained by training the server, the first network management device, or the second network management device.

[0253] Fig.14 1 shows a schematic block diagram of an example device 1400 that can be used to implement an embodiment of the present disclosure. The device 1400 can be implemented as or included in the aforementioned first network management device, second network management device, or server. For example, the device 1400 can be implemented as Figure 2 NMS 210, EMS 220, or server 290, Figure 3 The non-RT RIC 310, near-RT RIC 320, or server 390 in the system.

[0254] like Fig.14 As shown, device 1400 includes one or more processors 1410 , one or more memories 1420 coupled to the processors 1410 , and a communication module 1440 coupled to the processors 1410 .

[0255] The communication module 1440 may be used for two-way communication. The communication module 1440 may have at least one communication interface for communication. The communication interface may include any interface necessary for communication with other devices.

[0256] Processor 1410 may be any type suitable for the local technology network, and may include, but is not limited to, at least one of the following: a general purpose computer, a special purpose computer, a microcontroller, a digital signal processor (DSP), or one or more of a controller-based multi-core controller architecture. Device 1400 may have multiple processors, such as application specific integrated circuit chips, which are time-dependent and synchronized with a main processor.

[0257] The memory 1420 may include one or more non-volatile memories and one or more volatile memories. Examples of non-volatile memories include, but are not limited to, at least one of the following: read-only memory (ROM) 1424, erasable programmable read-only memory (EPROM), flash memory, hard disk, compact disc (CD), digital video disc (DVD), or other magnetic storage and / or optical storage. Examples of volatile memories include, but are not limited to, at least one of the following: random access memory (RAM) 1422, or other volatile memories that do not persist during the duration of a power outage.

[0258] Computer program 1430 includes computer executable instructions executed by associated processor 1410. Program 1430 may be stored in ROM 1424. Processor 1410 may perform any suitable actions and processes by loading program 1430 into RAM 1422.

[0259] The embodiment of the present disclosure can be implemented with the help of program 1430, so that the device 1400 can execute the Figure 4 to Figure 1 0. The embodiments of the present disclosure may also be implemented by hardware or by a combination of software and hardware.

[0260] The program 1430 may be tangibly embodied in a computer-readable medium that may be included in the device 1400 (such as in the memory 1420) or other storage device accessible by the device 1400. The program 1430 may be loaded from the computer-readable medium to the RAM 1422 for execution. The computer-readable medium may include any type of tangible non-volatile memory, such as ROM, EPROM, flash memory, hard disk, CD, DVD, etc.

[0261] In some embodiments, the communication module 1440 in the device 1400 may be implemented as a transmitter and a receiver (or a transceiver), which may be configured to send / receive such as training requests, training reports, notifications, loading requests, etc. In addition, the device 1400 may further include one or more of a scheduler, a controller, and a radio frequency / antenna, which will not be elaborated in detail in this disclosure.

[0262] For example, Fig.14 The device 1400 may be implemented as an electronic device, or may be implemented as a chip or a chip system in an electronic device, which is not limited in the embodiments of the present disclosure.

[0263] The embodiment of the present disclosure further provides a chip, which may include an input interface, an output interface and a processing circuit. In the embodiment of the present disclosure, the input interface and the output interface may complete the interaction of signaling or data, and the processing circuit may complete the generation and processing of signaling or data information.

[0264] The embodiments of the present disclosure also provide a chip system, including a processor, for supporting a computing device to implement the functions involved in any of the above embodiments. In one possible design, the chip system may also include a memory for storing necessary program instructions and data, and when the processor runs the program instructions, the device on which the chip system is installed implements the method involved in any of the above embodiments. Exemplarily, the chip system may be composed of one or more chips, and may also include chips and other discrete devices.

[0265] An embodiment of the present disclosure further provides a processor for coupling with a memory, wherein the memory stores instructions. When the processor executes the instructions, the processor executes the methods and functions involved in any of the above embodiments.

[0266] The embodiments of the present disclosure also provide a computer program or a computer program product including instructions, which, when executed on a computer, enables the computer to execute the methods and functions involved in any of the above embodiments.

[0267] An embodiment of the present disclosure further provides a non-transitory computer-readable storage medium having computer instructions stored thereon. When a processor executes the instructions, the processor executes the methods and functions involved in any of the above embodiments.

[0268] In general, various embodiments of the present application may be implemented in hardware or dedicated circuits, software, logic, or any combination thereof. Some aspects may be implemented in hardware, while other aspects may be implemented in firmware or software, which may be performed by a controller, microprocessor, or other computing device. Although various aspects of the embodiments of the present disclosure are shown and described as block diagrams, flow charts, or using some other graphical representations, it should be understood that the boxes, devices, systems, techniques, or methods described herein may be implemented as, for example, non-limiting examples, hardware, software, firmware, dedicated circuits or logic, general hardware or controllers or other computing devices, or some combination thereof.

[0269] The present disclosure also provides at least one computer program product tangibly stored on a non-transient computer-readable storage medium. The computer program product includes computer executable instructions, such as instructions included in a program module, which are executed in a device on a real or virtual processor of the target to perform the process / method as described above with reference to the accompanying drawings. Typically, program modules include routines, programs, libraries, objects, classes, components, data structures, etc. that perform specific tasks or implement specific abstract data types. In various embodiments, the functions of program modules can be combined or divided between program modules as needed. Machine executable instructions for program modules can be executed in local or distributed devices. In distributed devices, program modules can be located in local and remote storage media.

[0270] The computer program code for realizing the method of the present disclosure can be written in one or more programming languages. These computer program codes can be provided to the processor of general-purpose computer, special-purpose computer or other programmable data processing device, so that the program code, when being executed by computer or other programmable data processing device, causes the function / operation specified in flow chart and / or block diagram to be implemented. The program code can be executed completely on computer, partly on computer, as independent software package, partly on computer and partly on remote computer or completely on remote computer or server.

[0271] In the context of the present disclosure, computer program codes or related data may be carried by any appropriate carrier to enable a device, apparatus or processor to perform the various processes and operations described above. Examples of carriers include signals, computer readable media, and the like. Examples of signals may include electrical, optical, radio, acoustic or other forms of propagation signals, such as carrier waves, infrared signals, and the like.

[0272] A computer readable medium may be any tangible medium that contains or stores a program for or related to an instruction execution system, apparatus, or device. A computer readable medium may be a computer readable signal medium or a computer readable storage medium. A computer readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination thereof. More detailed examples of computer readable storage media include an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0273] In addition, although the operation of the method of the present disclosure is described in a particular order in the accompanying drawings, this does not require or imply that these operations must be performed in this particular order, or that all the operations shown must be performed to achieve the desired result. On the contrary, the steps depicted in the flow chart can change the order of execution. Additionally or alternatively, some steps can be omitted, multiple steps can be combined into one step for execution, and / or one step can be decomposed into multiple steps for execution. It should also be noted that the features and functions of two or more devices according to the present disclosure can be embodied in one device. Conversely, the features and functions of a device described above can be further divided into being embodied by multiple devices.

[0274] The above descriptions of various implementations of the present disclosure are exemplary, non-exhaustive, and not limited to the disclosed implementations. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope of the described implementations. The selection of terms used herein is intended to explain the principles of the implementations, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the various implementations disclosed herein.

Claims

1. A method for model training, comprising: Acquire training data for training a first part of a bilateral model, wherein the bilateral model includes the first part to be deployed at a network device and a second part to be deployed at a terminal device; Performing model training based on the training data to obtain the first part; as well as The first portion is sent to the network device.

2. The method according to claim 1, further comprising: The first part is lightweighted so that the lightweighted first part meets at least one of the following: a performance requirement of the first part, an energy consumption requirement of training the first part, an energy consumption requirement of an inference task for which the first part will be used, or a first model size requirement of the first part.

3. The method according to claim 1 or 2, wherein: The obtaining of the first part of training data for training the bilateral model includes: A first training request is received from a first network management device, wherein the first training request indicates training the first part of the bilateral model, and the first training request includes training data for training the first part of the bilateral model and structural information of the first part.

4. The method according to claim 3, further comprising: Send a first model training report to the first network management device, wherein the first model training report includes at least one of the following: a model identification of the bilateral model, a model identification of the first part, an actual performance of the first part, an actual energy consumption of training the first part, an estimated energy consumption of an inference task for which the first part will be used, or an actual model size of the first part.

5. The method according to claim 3 or 4, further comprising: receiving a model loading request from the first network management device, the model loading request instructing to load the first part into the network device; as well as Based on the model loading request, the first portion is sent to the network device.

6. The method according to claim 5, wherein the model loading request includes at least one of the following: a first identifier of the network device, a model identifier of the first part, a second identifier of the terminal device, a model identifier of the second part, or a model identifier of the bilateral model.

7. The method according to any one of claims 3 to 6, wherein the first training request further comprises at least one of the following: the model identifier of the bilateral model, The model identification of the first part, The model identification of the second part, a first identifier of the network device, a second identifier of the terminal device, Indicative information for indicating the type of reasoning task for which the bilateral model is to be used, The performance requirements of the bilateral model, The energy consumption requirement of the bilateral model training, The energy consumption requirement of the inference task, a first model size requirement of the first part, or Lightweight instructions.

8. The method according to any one of claims 3 to 7, further comprising: receiving a request message from the first network management device, wherein the request message is used to request status information of the network device; as well as In response to the request message, a response message is sent to the first network management device, the response message including at least one of the following: the computing power status of the network device, the storage status of the network device, the energy consumption information of the network device, or the network energy consumption target value of the network device.

9. The method according to claim 1 or 2, wherein: The obtaining of the first part of training data for training the bilateral model includes: A second training request is received from a server, wherein the second training request indicates training the first part of the bilateral model, and the second training request includes the training data for training the first part of the bilateral model and structural information of the first part.

10. The method according to claim 9, further comprising: A fourth model training report is sent to the server, where the fourth model training report includes at least one of the following: a model identifier of the bilateral model, a model identifier of the first part, or an actual performance of the first part.

11. The method according to claim 9 or 10, further comprising: receiving a notification message from the server, wherein the notification message indicates that the second part obtained by training the server has been loaded into the terminal device; as well as In response to the notification message, the first portion is sent to the network device.

12. The method according to any one of claims 9 to 11, wherein the second training request further comprises at least one of the following: the model identifier of the bilateral model, The model identification of the first part, The model identification of the second part, a first identifier of the network device, a second identifier of the terminal device, Indicative information for indicating the type of reasoning task for which the bilateral model is to be used, or Performance requirements of the bilateral model.

13. The method according to any one of claims 9 to 12, further comprising: Sending a request message to the second network management device, wherein the request message includes a first identifier of the network device; receiving a response message from the second network management device, the response message including at least one of the following: a computing power status of the network device, a storage status of the network device, energy consumption information of the network device, or a network energy consumption target value of the network device; and Based on the response message, determine at least one of the following: the energy consumption requirement of the bilateral model training, the first energy consumption requirement of the reasoning task for which the first part will be used, the second energy consumption requirement of the reasoning task for which the second part will be used, the first model size requirement of the first part, or the second model size requirement of the second part.

14. The method according to claim 1 or 2, further comprising: A second model training report is sent to the first network management device or server, wherein the second model training report includes training data for training the second part and structural information of the second part.

15. The method according to claim 14, wherein the second model training report further includes at least one of the following: a model identification of the bilateral model, a model identification of the first part, a first identification of the network device, an actual performance of the first part, an actual energy consumption of training the first part, an estimated energy consumption of the reasoning task for which the first part will be used, an actual model size of the first part, a performance requirement of the second part, a second energy consumption requirement of the reasoning task for which the second part will be used, or a second model size requirement of the second part.

16. A method for model training, comprising: Sending a first training request to a second network management device, the first training request indicating training a first part of a bilateral model, the bilateral model comprising the first part to be deployed at a network device and a second part to be deployed at a terminal device; as well as A model training report is received from the second network management device, wherein the model training report includes at least a model identifier of the first part.

17. The method of claim 16, wherein the first training request comprises: Training data for training the first part, and structural information of the first part.

18. A method according to claim 16 or 17, wherein the model training report includes a first model training report, and the first model training report includes at least one of the following: a model identification of the bilateral model, a model identification of the first part, an actual performance of the first part, an actual energy consumption of training the first part, an estimated energy consumption of an inference task for which the first part will be used, or an actual model size of the first part.

19. The method according to claim 17 or 18, further comprising: A second training request is received from a server, where the second training request indicates training the first part of the bilateral model, and the second training request includes: training data for training the first part, and structural information of the first part.

20. The method according to claim 16, wherein the model training report comprises a second model training report, and the second model training report comprises training data for training the second part and structural information of the second part.

21. The method according to claim 20, wherein the second model training report further includes at least one of the following: a model identification of the bilateral model, a model identification of the first part, a first identification of the network device, an actual performance of the first part, an actual energy consumption of training the first part, an estimated energy consumption of an inference task for which the first part will be used, and an actual model size of the first part.

22. The method according to claim 20 or 21, further comprising: Forwarding the second model training report to the server; receiving a notification message from the server, the notification message indicating that the second part has been loaded into the terminal device; as well as A model loading request is sent to the second network management device, where the model loading request indicates that the first part is loaded into the network device.

23. A method for model training, comprising: Acquire training data for training a second part of the bilateral model, wherein the bilateral model includes a first part to be deployed at a network device and the second part to be deployed at a terminal device; Performing model training based on the training data to obtain the second part; as well as The second part is sent to the terminal device.

24. The method according to claim 23, further comprising: A second training request is sent to the first network management device, wherein the second training request indicates training the first part of the bilateral model, and the second training request includes the training data for training the first part of the bilateral model and structural information of the first part.

25. The method according to claim 24, further comprising: A first model training report is received from a first network management device, wherein the first model training report indicates that the first part has been trained.

26. The method of claim 23, wherein obtaining training data for the second part of the bilateral model comprises: A second model training report is received from the first network management device, wherein the second model training report includes training data for training the second part and structural information of the second part.

27. The method according to claim 26, wherein the second model training report further includes at least one of the following: a model identification of the bilateral model, a model identification of the first part, a first identification of the network device, an actual performance of the first part, a performance requirement of the second part, a second energy consumption requirement of the inference task for which the second part will be used, or a second model size requirement of the second part.

28. The method according to any one of claims 23 to 27, further comprising: A notification message is sent to the first network management device, the notification message indicating that the second part has been loaded into the terminal device.

29. A method for model training, comprising: The method according to any one of claims 16 to 22 is performed by a first network management device, and The method according to any one of claims 1 to 8 is performed by a second network management device.

30. The method of claim 29, further comprising: The method according to any one of claims 23 to 28 is performed by a server.

31. The method according to claim 29 or 30, further comprising: The network device receives a first part of the bilateral model trained by the second network management device; The network device receives a second portion of the bilateral model from a server; The network device sends the second part to the terminal device; as well as The terminal device receives the second part from the network device.

32. A method for model training, comprising: The method according to any one of claims 1 to 2 and 9 to 15 is performed by a first network management device; as well as The method according to any one of claims 23 to 28 is performed by a server.

33. The method of claim 32, further comprising: The network device receives a first part of the bilateral model trained by the first network management device; The network device receives a second portion of the bilateral model from the server; The network device sends the second part to the terminal device; as well as The terminal device receives the second part from the network device.

34. A communication device, comprising a module or component for executing the method according to any one of claims 1 to 28.

35. A system comprising: A first network management device, configured to perform the method according to any one of claims 16 to 22; as well as The second network management device is configured to execute the method according to any one of claims 1-8.

36. The system of claim 35, further comprising: The server is configured to execute the method according to any one of claims 23 to 28.

37. The system of claim 35 or 36, further comprising: Network devices are configured as: Receiving a first part of the bilateral model trained by the second network management device; receiving a second portion of the bilateral model from a server; sending the second part to the terminal device; as well as The terminal device is configured to receive the second part from the network device.

38. A system comprising: A first network management device, configured to perform the method as claimed in any one of claims 1 to 2 and 9 to 15; as well as The server is configured to execute the method according to any one of claims 23 to 28.

39. The system of claim 38, further comprising: Network devices are configured as: Receiving a first part of the bilateral model trained by the first network management device; receiving a second portion of the bilateral model from a server; sending the second part to the terminal device; as well as The terminal device is configured to receive the second part from the network device.

Images