AI unit activation method, terminal and network side equipment

Abstract

The invention discloses an activation method of an AI unit, a terminal and network side equipment, and belongs to the technical field of communication, the activation method of the AI unit comprises the steps that the terminal activates a target AI unit according to target information, and the target information comprises at least one of indication information and activation conditions of the AI unit, the indication information is used for activating the target AI unit.

Description

Technical Field

[0001] This application belongs to the field of communication technologies, and particularly relates to a method for activating an AI unit, a terminal, and a network-side device. Background Art

[0002] In a wireless communication network, a terminal can use an Artificial Intelligence (AI) unit for prediction and make relevant decisions according to the prediction results, such as performing channel estimation, signal processing, etc.

[0003] Currently, in order to ensure the high quality, high efficiency, and high reliability of the AI unit, so as to meet service requirements and improve work efficiency, it is necessary to manage the life cycle of the AI unit. The life cycle management of the AI unit generally refers to managing and maintaining the entire life cycle of the AI unit, including but not limited to how to activate the AI unit, how to deactivate the AI unit, etc. However, there is currently no relevant solution for how to activate the AI unit. Summary of the Invention

[0004] Embodiments of this application provide a method for activating an AI unit, a terminal, and a network-side device, which can solve the problem of how a terminal activates an AI unit.

[0005] In a first aspect, a method for activating an AI unit is provided, which is executed by a terminal. The method includes:

[0006] The terminal activates a target AI unit according to target information, where the target information includes at least one of indication information and an activation condition of the AI unit, and the indication information is used to activate the target AI unit.

[0007] In a second aspect, a method for activating an AI unit is provided, which is executed by a network-side device. The method includes:

[0008] The network-side device sends indication information, and the indication information is used to activate a target AI unit.

[0009] In a third aspect, an activation device for an AI unit is provided, including any one of the following:

[0010] An activation module, configured to activate a target AI unit according to target information, where the target information includes at least one of indication information and an activation condition of the AI unit, and the indication information is used to activate the target AI unit.

[0011] In a fourth aspect, an activation device for an AI unit is provided, including:

[0012] A sending module, configured to send indication information, and the indication information is used to activate a target AI unit.

[0013] In a fifth aspect, a terminal is provided, which includes a processor and a memory. The memory stores a program or instructions that can run on the processor. When the program or instructions are executed by the processor, the steps of the method described in the first aspect are implemented.

[0014] In a sixth aspect, a terminal is provided, including a processor and a communication interface. Among them, the processor is used to activate a target AI unit according to target information, where the target information includes at least one of indication information and an activation condition of the AI unit, and the indication information is used to activate the target AI unit.

[0015] In a seventh aspect, a network-side device is provided, which includes a processor and a memory. The memory stores a program or instructions that can run on the processor. When the program or instructions are executed by the processor, the steps of the method described in the second aspect are implemented.

[0016] In an eighth aspect, a network-side device is provided, including a processor and a communication interface. Among them, the communication interface is used to send indication information, and the indication information is used to activate a target AI unit.

[0017] In a ninth aspect, a readable storage medium is provided. A program or instructions are stored on the readable storage medium. When the program or instructions are executed by a processor, the steps of the method described in the first aspect are implemented, or the steps of the method described in the second aspect are implemented.

[0018] In a tenth aspect, a wireless communication system is provided, including: a terminal and a network-side device. The terminal can be used to execute the steps of the method described in the first aspect, and the network-side device can be used to execute the steps of the method described in the second aspect.

[0019] In an eleventh aspect, a chip is provided. The chip includes a processor and a communication interface. The communication interface is coupled to the processor, and the processor is used to run a program or instructions to implement the method described in the first aspect, or to implement the method described in the second aspect.

[0020] In a twelfth aspect, a computer program / program product is provided. The computer program / program product is stored in a storage medium. The computer program / program product is executed by at least one processor to implement the steps of the method described in the first aspect, or to implement the steps of the method described in the second aspect.

[0021] In the embodiments of the present application, the terminal can activate a target AI unit according to at least one of indication information and an activation condition of the AI unit. Thus, the terminal can determine how to activate the AI unit, thereby optimizing the process of AI lifecycle management, ensuring the high quality, high efficiency and high reliability of the AI unit, meeting business requirements and improving work efficiency. Description of the Drawings

[0022] Figure 1 is a schematic diagram of a wireless communication system according to an embodiment of the present application;

[0023] Figure 2 is a schematic flowchart of a method for activating an AI unit according to an embodiment of the present application;

[0024] Figure 3 is a schematic flowchart of a method for activating an AI unit according to an embodiment of the present application;

[0025] Figure 4 is a schematic structural diagram of an activation device for an AI unit according to an embodiment of the present application;

[0026] Figure 5 is a schematic structural diagram of an activation device for an AI unit according to an embodiment of the present application;

[0027] Figure 6 is a schematic structural diagram of a communication device according to an embodiment of the present application;

[0028] Figure 7 is a schematic structural diagram of a terminal according to an embodiment of the present application;

[0029] Figure 8 is a schematic structural diagram of a network-side device according to an embodiment of the present application. Detailed Embodiments

[0030] Next, the technical solutions in the embodiments of the present application will be clearly described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are some, but not all, of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art belong to the scope of protection of the present application.

[0031] The terms "first", "second", etc. in the present application are used to distinguish similar objects, rather than to describe a specific order or sequence. It should be understood that such terms can be interchanged under appropriate circumstances so that the embodiments of the present application can be implemented in an order other than those illustrated or described herein, and the objects distinguished by "first" and "second" are usually of the same type, and the number of objects is not limited. For example, the first object can be one or more. In addition, "or" in the present application means at least one of the connected objects. For example, "A or B" covers three scenarios, namely, Scenario 1: including A and not including B; Scenario 2: including B and not including A; Scenario 3: including both A and B. The character " / " generally indicates an "or" relationship between the associated objects before and after.

[0032] The term "indication" in this application can be either a direct indication (or an explicit indication) or an indirect indication (or an implicit indication). Among them, a direct indication can be understood as that the sender clearly informs the receiver of specific information, operations to be performed, or request results, etc. in the sent indication; an indirect indication can be understood as that the receiver determines the corresponding information according to the indication sent by the sender, or makes a judgment and determines the operations to be performed or request results, etc. according to the judgment result.

[0033] It should be noted that the technology described in the embodiments of this application is not limited to the Long Term Evolution (LTE) / LTE-Advanced (LTE-A) system, and can also be used in other wireless communication systems, such as Code Division Multiple Access (CDMA), Time Division Multiple Access (TDMA), Frequency Division Multiple Access (FDMA), Orthogonal Frequency Division Multiple Access (OFDMA), Single-carrier Frequency-Division Multiple Access (SC-FDMA), or other systems. The terms "system" and "network" in the embodiments of this application are often used interchangeably, and the described technology can be used in the above-mentioned systems and radio technologies, as well as in other systems and radio technologies. The following description describes the New Radio (NR) system for example purposes, and uses NR terms in most of the following descriptions, but these technologies can also be applied to systems other than the NR system, such as the 6th Generation (6G) communication system. th Generation, 6G) communication system.

[0034] Figure 1The block diagram of a wireless communication system to which the embodiments of the present application can be applied is shown. The wireless communication system includes a terminal 11 and a network-side device 12. Among them, the terminal 11 can be a mobile phone, a tablet personal computer, a laptop computer, a notebook computer, a personal digital assistant (PDA), a handheld computer, a netbook, an ultra-mobile personal computer (UMPC), a mobile internet device (MID), an augmented reality (AR), a virtual reality (VR) device, a robot, a wearable device, a flight vehicle, a vehicle user equipment (VUE), a shipborne device, a pedestrian user equipment (PUE), a smart home (home devices with wireless communication functions, such as refrigerators, TVs, washing machines or furniture, etc.), a game console, a personal computer (PC), a teller machine or a self-service machine, etc. Wearable devices include: smart watches, smart bracelets, smart earphones, smart glasses, smart jewelry (smart bracelets, smart bracelets, smart rings, smart necklaces, smart anklets, smart ankle chains, etc.), smart wristbands, smart clothing, etc. Among them, the vehicle user equipment can also be referred to as a vehicle terminal, a vehicle controller, a vehicle module, a vehicle component, a vehicle chip or a vehicle unit, etc. It should be noted that the specific type of the terminal 11 is not limited in the embodiments of the present application. The network-side device 12 can include an access network device or a core network device. Among them, the access network device can also be referred to as a radio access network (RAN) device, a radio access network function or a radio access network unit. The access network device can include a base station, a wireless local area network (WLAN) access point (AP) or a wireless fidelity (WiFi) node, etc.Among them, the base station can be referred to as Node B (NB), Evolved Node B (eNB), the next generation Node B (gNB), New Radio Node B (NR Node B), access point, Relay Base Station (RBS), Serving Base Station (SBS), Base Transceiver Station (BTS), radio base station, radio transceiver, Basic Service Set (BSS), Extended Service Set (ESS), home Node B (HNB), home evolved Node B, Transmission Reception Point (TRP), or some other suitable term in the art. As long as the same technical effect is achieved, the base station is not limited to specific technical terms. It should be noted that in the embodiments of this application, only the base station in the NR system is taken as an example for introduction, and the specific type of the base station is not limited.

[0035] The AI unit in the embodiments of this application can also be referred to as an AI model, a machine learning (ML) model, an ML unit, an AI structure, an AI function, an AI feature, a machine learning model, a neural network, a neural network function, a neural network function, etc. In the embodiments of this application, only the AI unit is taken as an example for illustration. Among them, the AI unit can be a processing unit that can implement specific algorithms, formulas, processing procedures, capabilities, etc. related to AI, or it can also be a processing method, algorithm, function, module, or unit for a specific data set (including at least one of the input and output of the AI unit), or it can also be a processing method, algorithm, function, module, or unit running on AI / ML-related hardware such as a Graphics Processing Unit (GPU), a Neural Network Processing Unit (NPU), a Tensor Processing Unit (TPU), an Application Specific Integrated Circuits (ASIC), etc. No specific limitation is made here.

[0036] Optionally, the identifier of the AI unit may be an AI model identifier, an AI architecture identifier, an AI algorithm identifier, or may also be the identifier of a specific dataset associated with the AI unit, or may further be the identifier of a specific scenario, environment, channel characteristic, or device related to AI / ML, or the identifier of a function, feature, capability, or module related to AI / ML. No specific limitation is made here.

[0037] Next, in conjunction with the accompanying drawings, the activation method of the AI unit, the terminal, and the network-side device provided by the embodiments of the present application will be described in detail through some embodiments and their application scenarios.

[0038] As Figure 2 shown, the embodiments of the present application provide an activation method 200 for an AI unit. This method can be executed by a terminal. In other words, the activation method of the AI unit can be executed by software or hardware installed in the terminal. The activation method of the AI unit includes the following steps.

[0039] S202: The terminal activates the target AI unit according to the target information, where the target information includes at least one of indication information and the activation condition of the AI unit, and the indication information is used to activate the target AI unit.

[0040] With or without the AI function started, the terminal can activate the target AI unit according to the target information. The target information includes at least one of the indication information and the activation condition of the AI unit. The indication information may be sent by the network-side device. The activation condition of the AI unit may be configured by the terminal, or by the network-side device, or jointly by the terminal and the network-side device. No specific limitation is made here. The target AI unit may be an unactivated AI unit in the terminal. The number of target AI units may be one or more.

[0041] In this way, since the terminal can activate the target AI unit according to at least one of the indication information and the activation condition of the AI unit, the terminal can determine how to activate the AI unit, thereby optimizing the process of AI lifecycle management, ensuring the high quality, high efficiency, and high reliability of the AI unit, meeting business requirements, and improving work efficiency.

[0042] Optionally, in some embodiments, when the target information includes indication information, the terminal may further perform the following operations:

[0043] The terminal receives the indication information.

[0044] The indication information received by the terminal can be the indication information received by the terminal from the network device. Among them, the indication information can be carried by Radio Resource Control (RRC) signaling, or by Downlink Control Information (DCI), or by Medium Access Control-Control Element (MAC-CE).

[0045] Optionally, in some embodiments, the indication information can be actively sent by the network device to the terminal. In this case, the indication information can indicate any one of the following:

[0046] Activate the target AI unit;

[0047] Activate the target AI unit and deactivate the activated AI unit.

[0048] The activated AI unit can be one or more activated AI units in the terminal. In the case where the indication information indicates to activate the target AI unit, the operation performed by the terminal is to activate the target AI unit. In the case where the indication information indicates to activate the target AI unit and deactivate the activated AI unit, the operation performed by the terminal is to deactivate the activated AI unit and activate the target AI unit.

[0049] Optionally, in some other embodiments, the indication information can also be sent by the network device to the terminal when receiving the AI unit activation request of the terminal. That is, the terminal receiving the indication information can include:

[0050] The terminal determines whether the activated AI unit meets the deactivation condition;

[0051] When the terminal determines that the activated AI unit meets the deactivation condition, it sends an AI unit activation request;

[0052] The terminal receives the indication information.

[0053] The deactivation condition can be configured by the terminal itself, or by the network device, or jointly configured by the terminal and the network device, which is not specifically limited here. Optionally, the deactivation condition can be related to at least one of the following:

[0054] The performance when applying the activated AI unit;

[0055] The computing and storage capabilities supported by the terminal;

[0056] The ID of the activated AI unit;

[0057] The scenarios applicable to the activated AI unit;

[0058] Function information supported by the activated AI unit;

[0059] Cell information supported by the activated AI unit;

[0060] Area information supported by the activated AI unit;

[0061] Input / output type information of the activated AI unit;

[0062] Inference accuracy supported by the activated AI unit;

[0063] Terminal computing and storage capabilities corresponding to the activated AI unit;

[0064] Dataset related to the activated AI unit;

[0065] Performance threshold of the activated AI unit;

[0066] Complexity threshold of the activated AI unit.

[0067] When the terminal determines whether the activated AI unit meets the deactivation condition, it can make a judgment based on at least one of the above factors related to the deactivation condition, and determine whether the activated AI unit meets the deactivation condition according to the judgment result. For example, the terminal can judge whether the performance when applying the activated AI unit is greater than or equal to a certain threshold. If so, it can indicate that the performance of the terminal when applying the activated AI unit is good and there is no need to deactivate the AI unit. At this time, it can be determined that the deactivation condition is not met. If not, it can indicate that the performance of the terminal when applying the activated AI unit is poor and the AI unit needs to be deactivated. At this time, it can be determined that the deactivation condition is met. Another example is that the terminal can judge whether the currently supported computing and storage capabilities match the computing and storage capabilities required by the activated AI unit. If so, it can indicate that the currently supported computing and storage capabilities of the terminal allow the terminal to continue using the AI unit. At this time, it can be determined that the deactivation condition is not met. If not, it can indicate that the currently supported computing and storage capabilities of the terminal do not allow the terminal to continue using the AI unit. At this time, it can be determined that the deactivation condition is met. Another example is that the terminal can judge whether the current scenario belongs to the scenario applicable to the activated AI unit. If so, it can indicate that the AI unit can be used in the current scenario. At this time, it can be determined that the deactivation condition is not met. If not, it can indicate that the current scenario does not match the scenario applicable to the AI unit and the AI unit cannot be used continuously. At this time, it can be determined that the deactivation condition is met. And so on. Here, it is no longer necessary to give examples one by one on how to judge whether the activated AI unit meets the deactivation condition based on other factors.

[0068] It should be noted that for at least one of the above factors related to the deactivation condition, when the terminal determines whether the activated AI unit meets the deactivation condition, it can determine that the activated AI unit meets the deactivation condition when one of the factors meets the deactivation condition, or it can determine that the activated AI unit meets the deactivation condition when multiple of the factors meet the deactivation condition. There is no specific limitation here.

[0069] When the terminal determines that the activated AI unit meets the deactivation condition, it can send an AI unit activation request to the network-side device. After receiving the AI unit activation request, the network-side device can send indication information to the terminal. After receiving the indication information, the terminal can activate the target AI unit according to the indication information. Among them, the indication information received by the terminal can indicate any one of the following:

[0070] Activate the target AI unit;

[0071] Activate the target AI unit and deactivate the activated AI unit;

[0072] The terminal selects the target AI unit to be activated.

[0073] When the indication information indicates to activate the target AI unit, the operation performed by the terminal is to activate the target AI unit. When the indication information indicates to activate the target AI unit and deactivate the activated AI unit, the operation performed by the terminal is to deactivate the activated AI unit and activate the target AI unit. When the indication information indicates that the terminal selects the target AI unit to be activated, the operation performed by the terminal is to select the target AI unit by itself for activation.

[0074] Optionally, in some embodiments, when the target information includes the activation condition of the AI unit and the indication information, and the indication information indicates to activate the target AI unit or to activate the target AI unit and deactivate the activated AI unit, when the terminal activates the target AI unit, it can determine whether the target AI unit meets the activation condition of the AI unit, and when it meets the activation condition, activate the target AI unit. Optionally, the indication information may include the unit information of the target AI unit. In this way, when the terminal determines whether the target AI unit meets the activation condition, it can determine whether the target AI unit meets the activation condition of the AI unit according to the unit information of the target AI unit included in the indication information, and when it is determined that the target AI unit meets the activation condition, activate the target AI unit.

[0075] The unit information of the target AI unit may include at least one of the following:

[0076] The ID of the target AI unit;

[0077] The scenario applicable to the target AI unit;

[0078] Function information supported by the target AI unit;

[0079] Cell information supported by the target AI unit;

[0080] Area information supported by the target AI unit;

[0081] Input / output type information of the target AI unit;

[0082] Inference accuracy supported by the target AI unit;

[0083] Terminal computing and storage capabilities corresponding to the target AI unit;

[0084] Dataset related to the target AI unit;

[0085] Performance threshold of the target AI unit;

[0086] Complexity threshold of the target AI unit.

[0087] Optionally, the activation condition may be related to at least one of the following:

[0088] Terminal computing and storage capabilities supported;

[0089] ID of the target AI unit;

[0090] Scenarios applicable to the target AI unit;

[0091] Function information supported by the target AI unit;

[0092] Cell information supported by the target AI unit;

[0093] Area information supported by the target AI unit;

[0094] Input / output type information of the target AI unit;

[0095] Inference accuracy supported by the target AI unit;

[0096] Terminal computing and storage capabilities corresponding to the target AI unit;

[0097] Dataset related to the target AI unit;

[0098] Performance threshold of the target AI unit;

[0099] Complexity threshold of the target AI unit.

[0100] When determining whether a target AI unit meets the activation condition based on the unit information of the target AI unit, the terminal may make a judgment based on the unit information of the target AI unit and at least one of the above factors related to the activation condition, and determine whether the target AI unit meets the activation condition according to the judgment result. For example, the terminal may determine whether the currently supported computing and storage capabilities allow the activation of the target AI unit. If so, it may indicate that the terminal supports the use of the target AI unit, and at this time, it can be determined that the activation condition is met. If not, it may indicate that the terminal does not support the use of the target AI unit, and at this time, it can be determined that the activation condition is not met. For another example, the terminal may determine whether the current scenario belongs to the scenario applicable to the target AI unit. If so, it may indicate that the target AI unit can be used in the current scenario, and at this time, it can be determined that the activation condition is met. If not, it may indicate that the current scenario does not match the scenario applicable to the target AI unit and the target AI unit cannot be used, and at this time, it can be determined that the activation condition is not met. For yet another example, the terminal may determine whether the performance threshold of the target AI unit meets the service requirements. If so, it may indicate that the target AI can be used to execute relevant services, and at this time, it can be determined that the activation condition is met. If not, it may indicate that it is not suitable to use the target AI to execute relevant services, and at this time, it can be determined that the activation condition is not met. And so on. Here, no further examples will be given on how to determine whether the target AI unit meets the activation condition based on other factors.

[0101] It should be noted that for at least one of the above factors related to the activation condition, when determining whether the target AI unit meets the activation condition, the terminal may determine that the target AI unit meets the activation condition when one of the factors meets the activation condition, or may determine that the target AI unit meets the activation condition when multiple of the factors meet the activation condition. There is no specific limitation here.

[0102] Optionally, in some embodiments, if the terminal determines that the target AI unit does not meet the activation condition, the terminal may perform at least one of the following operations:

[0103] The terminal sends a first message, where the first message represents that the target AI unit does not meet the activation condition;

[0104] The terminal receives a second message, where the second message instructs to cancel the activation of the target AI unit;

[0105] The terminal cancels the activation of the target AI unit;

[0106] The terminal sends a third message, where the third message represents that the target AI unit is not activated;

[0107] The terminal receives a fourth message, where the fourth message instructs to execute a first operation;

[0108] The terminal executes the first operation.

[0109] For example, when the terminal determines that the target AI unit does not meet the activation condition, it can send the first information to the network-side device to inform the network-side device that the target AI unit does not meet the activation condition. After receiving the first information, the network-side device can send the second information to the terminal, and the second information can indicate to cancel the activation of the target AI unit. After receiving the second information, the terminal can cancel the activation of the target AI unit, that is, not activate the target AI unit. Optionally, when the terminal cancels the activation of the target AI unit, it can send the third information to the network-side device to inform the network-side device that the target AI unit has not been activated. Optionally, after receiving the third information, the network-side device can send the fourth information to the terminal, and the fourth information indicates to perform the first operation. After receiving the fourth information, the terminal can perform the first operation.

[0110] For another example, when the terminal determines that the target AI unit does not meet the activation condition, it can cancel the activation of the target AI unit by itself. Optionally, when the terminal cancels the activation of the target AI unit by itself, it can send the third information to the network-side device to inform the network-side device that the target AI unit has not been activated. Optionally, after receiving the third information, the network-side device can send the fourth information to the terminal, and the fourth information indicates to perform the first operation. After receiving the fourth information, the terminal can perform the first operation.

[0111] For yet another example, when the terminal determines that the target AI unit does not meet the activation condition, it can cancel the activation of the target AI unit by itself. Optionally, when the terminal cancels the activation of the target AI unit by itself, it can send the third information to the network-side device to inform the network-side device that the target AI unit has not been activated. Optionally, after sending the third information to the network-side device, the terminal can perform the first operation by itself.

[0112] The first information and the third information sent by the terminal can be carried by RRC signaling or by uplink control information (UCI). The second information and the fourth information received by the terminal can be carried by RRC signaling, or by DCI, or by MAC-CE. The first operation can include at least one of fine-tuning the target AI unit, retraining the target AI unit, upgrading the target AI unit, downgrading the target AI unit, rolling back the target AI unit, data collection, and performing AI unit transfer with other devices. The other device can be a network-side device or the terminal's own server, etc. Performing AI unit transfer with other devices can be that the other device transfers a new AI unit to the terminal.

[0113] Optionally, in some embodiments, when the indication information indicates that the terminal selects a target AI unit to be activated, the terminal activates the target AI unit according to the target information, which may be that the terminal itself selects the target AI unit to be activated for activation. Optionally, when the target information includes the activation conditions of the AI unit, when the terminal itself selects the target AI unit for activation, the following steps may be included:

[0114] The terminal determines whether one or more AI units meet the activation conditions according to the unit information of the one or more AI units;

[0115] The terminal determines the AI units that meet the activation conditions among the one or more AI units as the target AI units, and activates the target AI units.

[0116] The above one or more AI units may be existing and unactivated AI units in the terminal. For each AI unit, the unit information of the AI unit may include at least one of the following:

[0117] The ID of the AI unit;

[0118] The scenarios applicable to the AI unit;

[0119] The function information supported by the AI unit;

[0120] The cell information supported by the AI unit;

[0121] The area information supported by the AI unit;

[0122] The input / output type information of the AI unit;

[0123] The inference accuracy supported by the AI unit;

[0124] The terminal operation and storage capabilities corresponding to the AI unit;

[0125] The data set related to the AI unit;

[0126] The performance threshold value of the AI unit;

[0127] The complexity threshold value of the AI unit.

[0128] When the terminal determines whether one or more AI units meet the activation conditions according to the unit information of the one or more AI units, it may determine whether each AI unit meets the activation conditions separately. Among them, for an AI unit, the activation conditions may be related to at least one of the following:

[0129] The operation and storage capabilities supported by the terminal;

[0130] The ID of the AI unit;

[0131] Scenarios applicable to the AI unit;

[0132] Function information supported by the AI unit;

[0133] Cell information supported by the AI unit;

[0134] Area information supported by the AI unit;

[0135] Input and output type information of the AI unit;

[0136] Inference accuracy supported by the AI unit;

[0137] Terminal operation and storage capabilities corresponding to the AI unit;

[0138] Datasets related to the AI unit;

[0139] Performance threshold of the AI unit;

[0140] Complexity threshold of the AI unit.

[0141] When determining whether the AI unit meets the activation condition based on the unit information of the AI unit, it can be determined according to the unit information of the AI unit and at least one of the above factors related to the activation condition, and whether the AI unit meets the activation condition can be determined according to the judgment result. For the specific implementation method, reference can be made to the specific implementation method of the above terminal for determining whether the target AI unit meets the activation condition based on the unit information of the target AI unit, which will not be elaborated here.

[0142] After the terminal determines whether one or more AI units meet the activation condition, it can determine the AI units that meet the activation condition as the target AI units and activate the target AI units. Among them, the number of AI units that meet the activation condition can be one or more.

[0143] Optionally, in some embodiments, when the target information includes the activation condition of the AI unit, the terminal activates the target AI unit according to the target information, which may include:

[0144] The terminal determines whether the target AI unit meets the activation condition;

[0145] When the terminal determines that the target AI unit meets the activation condition, the terminal activates the target AI unit.

[0146] The activation condition can be configured by the terminal itself, or by the network-side device, or jointly configured by the terminal and the network-side device, which is not specifically limited here. Optionally, the activation condition is related to at least one of the following:

[0147] The operation and storage capabilities supported by the terminal;

[0148] The ID of the target AI unit;

[0149] Scenarios applicable to the target AI unit;

[0150] Function information supported by the target AI unit;

[0151] Cell information supported by the target AI unit;

[0152] Area information supported by the target AI unit;

[0153] Input / output type information of the target AI unit;

[0154] Inference accuracy supported by the target AI unit;

[0155] Terminal computing and storage capabilities corresponding to the target AI unit;

[0156] Dataset related to the target AI unit;

[0157] Performance threshold of the target AI unit;

[0158] Complexity threshold of the target AI unit.

[0159] When determining whether the target AI unit meets the activation condition, the terminal can make a judgment based on at least one of the above factors related to the activation condition, and determine whether the target AI unit meets the activation condition according to the judgment result. For example, the terminal can judge whether the currently supported computing and storage capabilities allow the activation of the target AI unit. If so, it can indicate that the terminal supports the use of the target AI unit, and at this time, it can be determined that the activation condition is met. If not, it can indicate that the terminal does not support the use of the target AI unit, and at this time, it can be determined that the activation condition is not met. For another example, the terminal can judge whether the current scenario belongs to the scenarios applicable to the target AI unit. If so, it can indicate that the target AI unit can be used in the current scenario, and at this time, it can be determined that the activation condition is met. If not, it can indicate that the current scenario does not match the scenarios applicable to the target AI unit and the target AI unit cannot be used, and at this time, it can be determined that the activation condition is not met. And so on. Here, no further examples will be given on how to judge whether the target AI unit meets the activation condition based on other factors.

[0160] It should be noted that for at least one of the above factors related to the activation condition, when determining whether the target AI unit meets the activation condition, the terminal can determine that the target AI unit meets the activation condition when one of the factors meets the activation condition, or can determine that the target AI unit meets the activation condition when multiple of the factors meet the activation condition. There is no specific limitation here.

[0161] Optionally, in some embodiments, when the terminal determines whether the target AI unit meets the activation condition, it may include:

[0162] The terminal determines whether the activated AI unit meets the deactivation condition;

[0163] When the terminal determines that the activated AI unit meets the deactivation condition, it determines whether the target AI unit meets the activation condition.

[0164] That is to say, the premise for the terminal to determine whether the target AI unit meets the activation condition is that the activated AI unit in the terminal meets the deactivation condition. When the activated AI unit meets the deactivation condition, it is necessary to activate the target AI unit. Before activating the target AI unit, it is necessary to determine whether the target AI unit meets the activation condition, and when it meets the activation condition, activate the target AI unit.

[0165] The above deactivation condition can be configured by the terminal itself, or by the network-side device, or jointly configured by the terminal and the network-side device, and no specific limitation is made here. Optionally, the deactivation condition may be related to at least one of the following:

[0166] The performance when applying the activated AI unit;

[0167] The computing and storage capabilities supported by the terminal;

[0168] The ID of the activated AI unit;

[0169] The scenarios applicable to the activated AI unit;

[0170] The function information supported by the activated AI unit;

[0171] The cell information supported by the activated AI unit;

[0172] The area information supported by the activated AI unit;

[0173] The input / output type information of the activated AI unit;

[0174] The inference accuracy supported by the activated AI unit;

[0175] The terminal computing and storage capabilities corresponding to the activated AI unit;

[0176] The data set related to the activated AI unit;

[0177] The performance threshold of the activated AI unit;

[0178] The complexity threshold of the activated AI unit.

[0179] When the terminal determines whether the activated AI unit meets the deactivation condition, it can make a judgment based on at least one of the above factors related to the deactivation condition, and determine whether the activated AI unit meets the deactivation condition according to the judgment result. The specific implementation method can be the same as the specific implementation method when the terminal determines whether the activated AI unit meets the deactivation condition, which will not be elaborated here.

[0180] Optionally, in some embodiments, when the terminal activates the target AI unit, it may include:

[0181] After the terminal performs the first processing or the second processing on the target AI unit, it applies the target AI unit; or,

[0182] After the terminal deactivates the activated AI unit and performs the first processing or the second processing on the target AI unit, it applies the target AI unit.

[0183] Specifically, if the indication information indicates to activate the target AI unit, then when the terminal activates the target AI unit, the operations it performs can be to first perform the first processing or the second processing on the target AI unit, and then apply the target AI unit. Among them, the first processing includes reading or loading at least part of the information of the target AI unit, and the second processing includes reading or loading the remaining information of the target AI unit. Applying the target AI unit can be using the target AI unit to perform relevant service processing. If the indication information indicates to activate the target AI unit and deactivate the activated AI unit, then when the terminal activates the target AI unit, the operations it performs can be to first deactivate the activated AI unit, then perform the first processing or the second processing on the target AI unit, and finally apply the target AI unit. Or, when the terminal supports parallel processing capabilities, it can also be to deactivate the activated AI unit, and perform the first processing or the second processing on the target AI unit in parallel during deactivation, and then apply the target AI unit. If the indication information indicates that the terminal executes to select the target AI unit to be activated or the terminal selects the target AI unit to be activated according to the activation condition by itself, then when the terminal activates the target AI unit, it can be to perform the first processing or the second processing on the target AI unit and then apply the target AI unit, or to deactivate the activated AI unit and perform the first processing or the second processing on the target AI unit and then apply the target AI unit.

[0184] The above-mentioned first processing or second processing of the target AI unit may be to perform the first processing on the target AI unit (in this case, the first processing may include reading or loading all information of the target AI unit), or to perform the first processing and the second processing on the target AI unit (in this case, the first processing may include reading or loading partial information of the target AI unit). For example, when the target AI unit is small, the reading or loading process of the target AI unit can be completed at one time. At this time, when the terminal reads or loads the target AI unit, it may be to perform the first processing on the target AI unit, and the first processing includes reading or loading all information of the target AI unit. When the target AI unit is large, the reading or loading process of the target AI unit may not be completed at one time. At this time, when the terminal reads or loads the target AI unit, it may be to perform the first processing and the second processing on the target AI unit. The first processing includes reading or loading partial information of the target AI unit, and the second processing includes reading or loading the remaining information of the target AI unit.

[0185] Optionally, in some embodiments, after activating the target AI unit, the terminal may further perform the following operations:

[0186] The terminal sends the fifth information, and the fifth information represents that the terminal has completed the activation of the target AI unit. The fifth information is carried by the RRC signaling or UCI.

[0187] In the embodiments of the present application, when the terminal activates the target AI unit according to the indication information, certain latency requirements need to be met. Specifically, when the number of target AI units is one, the activation latency of the one target AI unit (that is, the actual time used by the terminal when activating one target AI unit) needs to be less than or equal to the first latency. The first latency can be configured by the network-side device or agreed by the protocol. When the number of target AI units is multiple, the activation latency of the multiple target AI units (that is, the actual time used by the terminal when activating multiple target AI units) needs to be less than or equal to the second latency. The second latency is less than or equal to the sum of the activation latencies of the multiple target AI units. The second latency can be configured by the network-side device or agreed by the protocol. In this way, by restricting the activation latency of the AI unit, the process of AI lifecycle management can be optimized, and the definition of the communication system can be improved.

[0188] Optionally, for one target AI unit, the activation latency of the target AI unit may include at least one of the following:

[0189] The latency between the terminal sending the AI unit activation request and receiving the indication information;

[0190] The processing or parsing latency of the signaling carrying the indication information by the terminal;

[0191] The latency of the terminal for Hybrid Automatic Repeat reQuest (HARQ) feedback;

[0192] The latency of the terminal for deactivating an activated AI unit;

[0193] The latency of the terminal for performing a first process on a target AI unit, where the first process includes reading or loading partial information of the target AI unit;

[0194] The latency of the terminal for performing a second process and application on a target AI unit, where the second process includes reading or loading the remaining information of the target AI unit.

[0195] For example, in the case where the network - side device instructs the terminal to activate the target AI unit through RRC or DCI, the activation delay of the target AI unit may include the processing or parsing delay of the signaling (i.e., RRC or DCI) carrying the indication information by the terminal, the delay of the terminal's first - stage processing of the target AI unit, and the delay of the terminal's second - stage processing and application of the target AI unit. In the case where the network - side device instructs the terminal to activate the target AI unit and de - activate the activated AI unit through MAC - CE, the activation delay of the target AI unit may include the processing or parsing delay of the signaling (i.e., MAC - CE) carrying the indication information by the terminal, the delay of the terminal's HARQ feedback, the delay of the terminal's de - activation of the activated AI unit, the delay of the terminal's first - stage processing of the target AI unit, and the delay of the terminal's second - stage processing and application of the target AI unit. In the case where the network - side device instructs the terminal to activate the target AI unit through RRC or DCI according to the terminal's AI unit activation request, the activation delay of the target AI unit may include the delay between the terminal sending the AI unit activation request and receiving the indication information, the processing or parsing delay of the signaling carrying the indication information by the terminal, the delay of the terminal's first - stage processing of the target AI unit, and the delay of the terminal's second - stage processing and application of the target AI unit. In the case where the network - side device instructs the terminal to activate the target AI unit and de - activate the activated AI unit through MAC - CE according to the terminal's AI unit activation request, the activation delay of the target AI unit may include the delay between the terminal sending the AI unit activation request and receiving the indication information, the processing or parsing delay of the signaling carrying the indication information by the terminal, the delay of the terminal's HARQ feedback, the delay of the terminal's de - activation of the activated AI unit, the delay of the terminal's first - stage processing of the target AI unit, and the delay of the terminal's second - stage processing and application of the target AI unit. In the case where the network - side device instructs the terminal to select the target AI unit to be activated by itself through RRC or DCI, the activation delay of the target AI unit may include the processing or parsing delay of the signaling carrying the indication information by the terminal, the delay of the terminal's first - stage processing of the target AI unit, and the delay of the terminal's second - stage processing and application of the target AI unit. Here, no more examples will be given one by one. Among them, the delay of the terminal's HARQ feedback may include the delay between the terminal receiving the downlink data transmission and the terminal sending the confirmation indication.

[0196] For multiple target AI units, optionally, when the multiple target AI units can be activated in parallel, the second time delay can be less than or equal to the maximum value of the activation time delays of the multiple target AI units. When the differences among the multiple target AI units are within the specified differences, the activation time delays of the multiple target AI units can be less than or equal to a third time delay, and this third time delay is less than the second time delay. That is to say, when the differences among the multiple target AI units are within the specified differences, the time delay requirement for activating the multiple target AI units can be further shortened. Among them, the differences among the multiple target AI units are related to at least one of the structure, parameters, complexity, size, quantization level, function, corresponding terminal operation and storage capabilities, and applicable scenarios of the A1 unit.

[0197] In the embodiments of the present application, the terminal can activate the target AI unit according to at least one of the indication information and the activation condition of the AI unit. Thereby, the terminal can determine how to activate the AI unit, so as to optimize the process of AI life cycle management, ensure the high quality, high efficiency and high reliability of the AI unit, meet the service requirements and improve the work efficiency.

[0198] As Figure 3 shown, the embodiments of the present application provide an activation method 300 for an AI unit. This method can be executed by a network-side device. In other words, this activation method for the AI unit can be executed by software or hardware installed in the network-side device. This activation method for the AI unit includes the following steps.

[0199] S302: The network-side device sends indication information, and the indication information is used to activate the target AI unit.

[0200] When communicating with the terminal, the network-side device can send indication information to the terminal, and the indication information is used to activate the target AI unit. The target AI unit can be an AI unit that is not activated in the terminal. The number of target AI units can be one or more.

[0201] The indication information can be carried by RRC signaling, or by DCI, or by MAC-CE.

[0202] Optionally, in some embodiments, the network-side device can actively send the indication information to the terminal. In this case, the indication information can indicate any one of the following:

[0203] Activate the target AI unit;

[0204] Activate the target AI unit and deactivate the activated AI unit.

[0205] Optionally, in some other embodiments, the network device may also send indication information to the terminal when receiving an AI unit activation request from the terminal. That is, the network device sending the indication information may include:

[0206] The network device receives an AI unit activation request;

[0207] The network device sends indication information according to the AI unit activation request.

[0208] The AI unit activation request may be sent by the terminal to the network device when it determines that the activated AI unit does not meet the activation conditions. For the specific implementation of how the terminal determines that the activated AI unit does not meet the activation conditions, reference can be made to Figure 2 the corresponding steps in the illustrated embodiments, which will not be repeated here. After receiving the AI unit activation request, the network device may send indication information to the terminal according to the AI unit activation request. Among them, the indication information sent by the network device may indicate any one of the following:

[0209] Activate the target AI unit;

[0210] Activate the target AI unit and deactivate the activated AI unit;

[0211] The terminal selects the target AI unit to be activated.

[0212] The target AI unit may be an AI unit selected or determined by the network device according to actual service requirements. The method for the network device to select or determine the target AI unit is not limited here. The target AI unit may be an existing AI unit of the terminal. Optionally, when the terminal does not have a target AI unit, the network device may also indicate the target AI unit to the terminal. The number of target AI units may be one or more.

[0213] After the network device sends the indication information to the terminal, the terminal may activate the target AI unit according to the indication information. For the specific implementation of how the terminal activates the target AI unit according to the indication information, reference can be made to Figure 2 the corresponding steps in the illustrated embodiments, which will not be elaborated here.

[0214] Optionally, in some embodiments, when the indication information indicates to activate the target AI unit or activate the target AI unit and deactivate the activated AI unit, the indication information may include the unit information of the target AI unit. The unit information of the target AI unit can be used by the terminal to determine whether the target AI unit meets the activation conditions. Among them, the unit information of the target AI unit may include at least one of the following:

[0215] The ID of the target AI unit;

[0216] Scenarios applicable to the target AI unit;

[0217] Function information supported by the target AI unit;

[0218] Cell information supported by the target AI unit;

[0219] Area information supported by the target AI unit;

[0220] Input and output type information of the target AI unit;

[0221] Inference accuracy supported by the target AI unit;

[0222] Terminal computing and storage capabilities corresponding to the target AI unit;

[0223] Dataset related to the target AI unit;

[0224] Performance threshold of the target AI unit;

[0225] Complexity threshold of the target AI unit.

[0226] For the terminal, after receiving the indication information, it can determine whether the target AI unit meets the activation condition according to the unit information of the target AI unit included in the indication information. If it meets the condition, the target AI unit is activated. For the specific implementation method of how the terminal determines whether the target AI unit meets the activation condition, reference can be made to Figure 2 the specific implementation of the corresponding steps in the embodiments shown, which will not be repeated here.

[0227] Optionally, in some embodiments, after the terminal determines whether the target AI unit meets the activation condition, the determination result may be that the target AI unit does not meet the activation condition. In this case, the terminal can report relevant information to the network-side device, and the network-side device can perform at least one of the following operations:

[0228] The network-side device receives the first information, and the first information represents that the target AI unit does not meet the activation condition;

[0229] The network-side device sends the second information, and the second information instructs to cancel the activation of the target AI unit;

[0230] The network-side device receives the third information, and the third information represents that the target AI unit is not activated;

[0231] The network-side device sends the fourth information, and the fourth information instructs to perform the first operation.

[0232] For example, when the terminal determines that the target AI unit does not meet the activation condition, it may send a first message to the network-side device to inform the network-side device that the target AI unit does not meet the activation condition. The network-side device may receive the first message and then send a second message to the terminal. The second message may indicate to cancel the activation of the target AI unit. After receiving the second message, the terminal may cancel the activation of the target AI unit, that is, not activate the target AI unit. Optionally, when the terminal cancels the activation of the target AI unit, it may send a third message to the network-side device to inform the network-side device that the target AI unit is not activated. The network-side device may receive the third message. Optionally, after receiving the third message, the network-side device may send a fourth message to the terminal. The fourth message indicates to perform a first operation. For another example, when the terminal determines that the target AI unit does not meet the activation condition, it may cancel the activation of the target AI unit by itself. Optionally, when the terminal cancels the activation of the target AI unit by itself, it may send a third message to the network-side device to inform the network-side device that the target AI unit is not activated. The network-side device may receive the third message. Optionally, after receiving the third message, the network-side device may send a fourth message to the terminal. The fourth message indicates to perform a first operation.

[0233] The first message and the third message received by the network-side device may be carried by RRC signaling or UCI. The second message and the fourth message sent by the network-side device may be carried by RRC signaling, DCI, or MAC-CE. The first operation may include at least one of fine-tuning the target AI unit, retraining the target AI unit, upgrading the target AI unit, downgrading the target AI unit, rolling back the target AI unit, performing data collection, and performing AI unit transfer with other devices. The other device may be a network-side device or the terminal's own server, etc. Performing AI unit transfer with other devices may be that the other device transfers a new AI unit to the terminal.

[0234] Optionally, in some embodiments, when the terminal activates the target AI unit, it may send a fifth message to the network-side device. At this time, the network-side device may receive the fifth message. The fifth message represents that the terminal has completed the activation of the target AI unit. The fifth message is carried by RRC signaling or UCI.

[0235] In the embodiments of the present application, the network-side device may send indication information. The terminal may, when receiving the indication information, activate the target AI unit according to the indication information. Thus, the terminal can determine how to activate the AI unit, thereby optimizing the process of AI life cycle management, ensuring the high quality, high efficiency, and high reliability of the AI unit, meeting the service requirements, and improving work efficiency.

[0236] To facilitate understanding of the method for activating the AI unit provided in the embodiments of the present application, the following will be described by taking some more specific embodiments as examples.

[0237] Embodiment 1: The network-side device instructs the terminal to activate the target AI unit.

[0238] The activation process of Embodiment 1 is applicable to the following two scenarios:

[0239] Scenario 1: Before receiving the indication information, the AI function of the terminal is not enabled;

[0240] Scenario 2: Before receiving the indication information, the AI function of the terminal is enabled but the previously used AI unit has been deactivated.

[0241] The activation process may include the following steps:

[0242] Step 1: The network-side device sends indication information to the terminal, the indication information indicates to activate the target AI unit, and the indication information includes the unit information of the target AI unit.

[0243] The indication information may be carried by RRC signaling, DCI, or MAC-CE.

[0244] The unit information of the target AI unit may include at least one of the ID of the target AI unit, the scenario applicable to the target AI unit, the function information supported by the target AI unit, the cell information supported by the target AI unit, the area information supported by the target AI unit, the input / output type information of the target AI unit, the inference accuracy supported by the target AI unit, the terminal operation and storage capabilities corresponding to the target AI unit, the dataset related to the target AI unit, the performance threshold value of the target AI unit, and the complexity threshold value of the target AI unit.

[0245] Step 2: The terminal determines whether the target AI unit meets the activation conditions according to the unit information of the target AI unit.

[0246] When the terminal determines that the target AI unit meets the activation conditions, it can execute Step 4. When the terminal determines that the target AI unit does not meet the activation conditions, it can execute Step 3.

[0247] Step 3: The terminal sends the first information to the network-side device, and the first information represents that the target AI unit does not meet the activation conditions.

[0248] For the network-side device, after receiving the first information, it can send the second information to the terminal, and the second information instructs to cancel the activation of the target AI unit.

[0249] Optionally, Step 3 can also be replaced by the terminal itself canceling the activation of the target AI unit, and then sending the third information to the network-side device, and the third information represents that the target AI unit is not activated.

[0250] Optionally, when the terminal deactivates the target AI unit, it can also perform the first operation by itself, or the network-side device can send the fourth information to the terminal to instruct the terminal to perform the first operation. The first operation may include at least one of fine-tuning the target AI unit, retraining the target AI unit, upgrading the target AI unit, downgrading the target AI unit, rolling back the target AI unit, collecting data, and performing AI unit transfer with other devices.

[0251] Step 4: The terminal performs a first processing on the target AI unit, and the first processing includes reading or loading partial information of the target AI unit.

[0252] Step 5: The terminal performs a second processing and application on the target AI unit, and the second processing includes reading or loading the remaining information of the target AI unit.

[0253] Optionally, after the terminal activates the target AI unit, it can send the fifth information to the network-side device, and the fifth information indicates that the terminal has completed the activation of the target AI unit.

[0254] For the above activation process, the starting point is that the terminal receives the indication information sent by the network-side device, and the ending point can be that the terminal completes the second processing and application of the target AI unit, or that the terminal completes the activation of the target AI unit and sends the fifth information to the network-side device for reporting.

[0255] During the above activation process, if the number of target AI units is 1, the activation delay of the target AI unit at least includes:

[0256] The processing or parsing delay of the signaling carrying the indication information by the terminal;

[0257] The delay of the terminal's HARQ feedback (for the case where the indication information is carried by MAC-CE, it may include the delay between the terminal receiving the downlink data transmission and the terminal sending the confirmation indication);

[0258] The delay of the terminal's first processing of the target AI unit;

[0259] The delay of the terminal's second processing and application of the target AI unit.

[0260] If the number of target AI units is more than one, when the terminal activates multiple target AI units, the total activation delay of the terminal should be less than or equal to the sum of the delay requirements of each target AI unit. Optionally, when the terminal can activate more than one target AI unit in parallel, the activation delay is the activation delay of the largest single AI unit currently activated in parallel. Optionally, when there are more than one target AI units among multiple target AI units within a defined difference, the corresponding activation delay requirements can be shortened. The difference between AI units is related to at least one of the structure, parameters, complexity, size, quantization level, function, corresponding terminal operation and storage capabilities, and applicable scenarios of the A1 unit.

[0261] Embodiment 2: The terminal initiates the activation of an AI unit to the network-side device.

[0262] The activation process of Embodiment 2 can be applied to the following two scenarios:

[0263] Scenario 1: Before the terminal receives the indication information or before it sends an activation request by itself, the AI function is not enabled;

[0264] Scenario 2: Before the terminal receives the indication information or before it sends an activation request by itself, the AI function is enabled but the previously used AI unit has been deactivated.

[0265] The activation process may include the following steps:

[0266] Step 1: The terminal can use the AI units that meet the activation conditions among one or more AI units as target AI units and activate the target AI units. Alternatively, the terminal can send an AI unit activation request to the network-side device.

[0267] When the terminal sends an AI activation request to the network-side device, it can execute Step 2.

[0268] Step 2: The network-side device sends indication information to the terminal, and the indication information indicates to activate the target AI unit or indicates that the terminal selects the target AI unit to be activated.

[0269] In the case where the indication information indicates to activate the target AI unit, Step 3 can be executed; in the case where the indication information indicates that the terminal selects the target AI unit to be activated, the terminal can use the AI units that meet the activation conditions among one or more AI units as target AI units and activate the target AI units.

[0270] Step 3: The terminal performs a first process on the target AI unit, and the first process includes reading or loading partial information of the target AI unit.

[0271] Step 4: The terminal performs a second process and application on the target AI unit, and the second process includes reading or loading the remaining information of the target AI unit.

[0272] Optionally, after activating the target AI unit, the terminal may send fifth information to the network-side device, where the fifth information indicates that the terminal has completed the activation of the target AI unit.

[0273] For the above activation process, the starting point is that the terminal sends an AI unit activation request to the network-side device, and the ending point may be that the terminal completes the second processing and application of the target AI unit, or the terminal completes the activation of the target AI unit and sends the fifth information to the network-side device for reporting.

[0274] During the above activation process, if the number of target AI units is 1 and the indication information indicates the activation of the target AI unit, the activation delay of the target AI unit at least includes:

[0275] The delay between the terminal sending the AI unit activation request and receiving the indication information;

[0276] The processing or parsing delay of the signaling carrying the indication information by the terminal;

[0277] The delay of the terminal performing HARQ feedback (for the case where the indication information is carried by MAC-CE, it may include the delay between the terminal receiving the downlink data transmission and the terminal sending the confirmation indication);

[0278] The delay of the terminal performing the first processing on the target AI unit;

[0279] The delay of the terminal performing the second processing and application on the target AI unit.

[0280] If the number of target AI units is 1 and the indication information indicates that the terminal selects the target AI unit to be activated by itself, the activation delay of the target AI unit at least includes:

[0281] The delay between the terminal sending the AI unit activation request and receiving the indication information;

[0282] The processing or parsing delay of the signaling carrying the indication information by the terminal;

[0283] The delay of the terminal performing HARQ feedback (for the case where the indication information is carried by MAC-CE, it may include the delay between the terminal receiving the downlink data transmission and the terminal sending the confirmation indication);

[0284] The delay of the terminal performing the first processing on the target AI unit;

[0285] The delay of the terminal performing the second processing and application on the target AI unit.

[0286] If the number of target AI units is more than one, when the terminal activates multiple target AI units, the total activation latency of the terminal should be less than or equal to the sum of the latency requirements of each target AI unit. Optionally, when the terminal can activate more than one target AI unit in parallel, the activation latency is the activation latency of the largest single AI unit currently activated in parallel. Optionally, when more than one of the multiple target AI units are within a defined difference, the corresponding activation latency requirement can be shortened, and the difference between AI units is related to at least one of the structure, parameters, complexity, size, quantization level, function, corresponding terminal operation and storage capabilities, and applicable scenarios of the A1 unit.

[0287] Embodiment 3: The network-side device instructs the terminal to perform a handover of AI units, including deactivating the activated AI unit and activating the target AI unit.

[0288] The handover process of Embodiment 3 can be applied to the following scenarios:

[0289] Before the terminal receives the indication information, the AI function has been enabled and the previously used AI unit is still in the activated state.

[0290] The handover process may include the following steps:

[0291] Step 1: The network-side device sends indication information to the terminal. The indication information instructs to activate the target AI unit and deactivate the activated AI unit, and the indication information includes the unit information of the target AI unit.

[0292] The indication information can be carried by RRC signaling, DCI, or MAC-CE.

[0293] The unit information of the target AI unit may include at least one of the ID of the target AI unit, the applicable scenario of the target AI unit, the function information supported by the target AI unit, the cell information supported by the target AI unit, the area information supported by the target AI unit, the input / output type information of the target AI unit, the inference accuracy supported by the target AI unit, the corresponding terminal operation and storage capabilities of the target AI unit, the dataset related to the target AI unit, the performance threshold value of the target AI unit, and the complexity threshold value of the target AI unit.

[0294] Step 2: The terminal determines whether the target AI unit meets the activation conditions according to the unit information of the target AI unit.

[0295] When the terminal determines that the target AI unit meets the activation conditions, it can execute Step 4. When the terminal determines that the target AI unit does not meet the activation conditions, it can execute Step 3.

[0296] Step 3: The terminal sends first information to the network-side device, where the first information indicates that the target AI unit does not meet the activation condition.

[0297] For the network-side device, after receiving the first information, it may send second information to the terminal, where the second information instructs to cancel the activation of the target AI unit.

[0298] Optionally, Step 3 may also be replaced by the terminal itself canceling the activation of the target AI unit and then sending third information to the network-side device, where the third information indicates that the target AI unit is not activated.

[0299] Optionally, when the terminal cancels the activation of the target AI unit, it may also perform a first operation by itself, or alternatively, the network-side device may send fourth information to the terminal, where the fourth information instructs the terminal to perform the first operation. The first operation may include at least one of fine-tuning the target AI unit, retraining the target AI unit, upgrading the target AI unit, downgrading the target AI unit, rolling back the target AI unit, collecting data, and performing AI unit transfer with other devices.

[0300] Step 4: The terminal deactivates the activated AI unit.

[0301] Step 5: The terminal performs a first process on the target AI unit, where the first process includes reading or loading partial information of the target AI unit.

[0302] Optionally, when the terminal supports parallel processing capabilities, Steps 4 and 5 may be executed in parallel.

[0303] Step 6: The terminal performs a second process and application on the target AI unit, where the second process includes reading or loading the remaining information of the target AI unit.

[0304] Optionally, after activating the target AI unit, the terminal may send fifth information to the network-side device, where the fifth information indicates that the terminal has completed the activation of the target AI unit.

[0305] For the above switching process, the starting point is when the terminal receives the indication information sent by the network-side device, and the ending point may be when the terminal completes the second process and application of the target AI unit, or when the terminal completes the activation of the target AI unit and sends the fifth information to the network-side device for reporting.

[0306] During the above switching process, if the number of target AI units is 1, the activation delay (i.e., the switching delay) of the target AI unit at least includes:

[0307] The processing or parsing delay of the signaling carrying the indication information by the terminal;

[0308] The latency of HARQ feedback at the terminal (for the case where the indication information is carried by MAC-CE, it may include the latency between the terminal receiving downlink data transmission and the terminal sending an acknowledgment indication);

[0309] The latency of deactivating an activated AI unit at the terminal;

[0310] The latency of the terminal performing a first process on a target AI unit;

[0311] The latency of the terminal performing a second process and application on a target AI unit.

[0312] If the number of target AI units is multiple, when the terminal activates multiple target AI units, the total activation latency should be less than or equal to the sum of the latency requirements of each target AI unit. Optionally, when the terminal can activate more than 1 target AI unit in parallel, the activation latency is the activation latency of the largest single AI unit currently activated in parallel. Optionally, when there are more than 1 target AI units among multiple target AI units within a defined difference, the corresponding activation latency requirement can be shortened, and the difference between AI units is related to at least one of the structure, parameters, complexity, size, quantization level, function, corresponding terminal operation and storage capabilities, and applicable scenarios of the A1 unit.

[0313] Embodiment 4: The terminal initiates a handover of an AI unit to the network-side device, including activating a target AI unit and deactivating an activated AI unit.

[0314] The handover process of Embodiment 4 can be applied to the following scenarios:

[0315] Before the terminal receives the indication information or before sending an activation request by itself, the AI function is enabled and the previously used AI unit is still in an activated state.

[0316] The handover process may include the following steps:

[0317] Step 1: When the terminal determines that the activated AI unit meets the deactivation condition, the terminal can use the AI unit that meets the activation condition among one or more AI units as the target AI unit, activate the target AI unit, and deactivate the activated AI unit, or the terminal can send an AI unit activation request to the network-side device.

[0318] The deactivation condition can be configured by the terminal, or by the network-side device, or jointly configured by the terminal and the network-side device. The specific implementation of how the terminal determines whether the activated AI unit meets the deactivation condition and the specific implementation of how the terminal selects a target AI unit that meets the activation condition for activation can refer to the Figure 2 and Figure 3 specific implementation of the corresponding steps therein, which will not be repeated here.

[0319] When the terminal sends an AI activation request to the network - side device, step 2 can be executed.

[0320] Step 2: The network - side device sends indication information to the terminal. The indication information indicates the activation of the target AI unit and the de - activation of the activated AI unit, or indicates that the terminal selects the target AI unit to be activated.

[0321] When the indication information indicates the activation of the target AI unit and the de - activation of the activated AI unit, step 3 can be executed; when the indication information indicates that the terminal selects the target AI unit to be activated, the terminal can use the AI unit that meets the activation conditions among one or more AI units as the target AI unit, activate the target AI unit, and de - activate the activated AI unit.

[0322] Step 3: The terminal de - activates the activated AI unit.

[0323] Step 4: The terminal performs a first - stage processing on the target AI unit. The first - stage processing includes reading or loading partial information of the target AI unit.

[0324] Optionally, when the terminal supports parallel processing capabilities, steps 3 and 4 can be executed in parallel.

[0325] Step 5: The terminal performs a second - stage processing and application on the target AI unit. The second - stage processing includes reading or loading the remaining information of the target AI unit.

[0326] Optionally, after activating the target AI unit, the terminal can send fifth - type information to the network - side device. The fifth - type information indicates that the terminal has completed the activation of the target AI unit.

[0327] For the above - mentioned handover process, the starting point is that the terminal sends an AI unit activation request to the network - side device, and the ending point can be that the terminal completes the second - stage processing and application of the target AI unit, or the terminal completes the activation of the target AI unit and sends the fifth - type information to the network - side device for reporting.

[0328] During the above - mentioned handover process, if the number of target AI units is 1 and the indication information indicates the activation of the target AI unit and the de - activation of the activated AI unit, the activation latency of the target AI unit at least includes:

[0329] The latency between the terminal sending the AI unit activation request and receiving the indication information;

[0330] The latency for the terminal to de - activate the activated AI unit;

[0331] The processing or parsing latency of the signaling carrying the indication information by the terminal;

[0332] The latency of HARQ feedback at the terminal (for the case where the indication information is carried by MAC-CE, it may include the latency between the terminal receiving the downlink data transmission and the terminal sending the confirmation indication);

[0333] The latency of the terminal performing the first processing on the target AI unit;

[0334] The latency of the terminal performing the second processing and application on the target AI unit.

[0335] If the number of target AI units is 1 and the indication information indicates that the terminal selects the target AI unit to be activated by itself, the activation latency of the target AI unit at least includes:

[0336] The latency between the terminal sending the AI unit activation request and receiving the indication information;

[0337] The processing or parsing latency of the terminal for the signaling carrying the indication information;

[0338] The latency of HARQ feedback at the terminal (for the case where the indication information is carried by MAC-CE, it may include the latency between the terminal receiving the downlink data transmission and the terminal sending the confirmation indication);

[0339] The latency of the terminal deactivating the activated AI unit;

[0340] The latency of the terminal performing the first processing on the target AI unit;

[0341] The latency of the terminal performing the second processing and application on the target AI unit.

[0342] If the number of target AI units is multiple, when the terminal activates multiple target AI units, the total activation latency of the terminal should be less than or equal to the sum of the latency requirements of each target AI unit. Optionally, when the terminal can activate more than 1 target AI unit in parallel, the activation latency is the activation latency of the largest single AI unit activated currently in parallel. Optionally, when there are more than 1 target AI units among multiple target AI units within the defined difference, the corresponding activation latency requirements can be shortened, and the difference between AI units is related to at least one of the structure, parameters, complexity, size, quantization level, function, corresponding terminal operation and storage capabilities, and applicable scenarios of the A1 unit.

[0343] Based on the above-mentioned Embodiments 1 to 4, the embodiments of the present application propose a method for activating and deactivating AI units in AI unit lifecycle management, which is applicable to each process including activation and deactivation in AI lifecycle management, and stipulates the corresponding radio resource management latency requirements, thereby optimizing the process of AI lifecycle management and enhancing the definition of the communication system.

[0344] The activation method of the AI unit provided by the embodiments of the present application may be executed by an activation device of the AI unit. In the embodiments of the present application, taking the activation device of the AI unit executing the activation method of the AI unit as an example, the activation device of the AI unit provided by the embodiments of the present application is described.

[0345] Figure 4 It is a schematic structural diagram of an activation device of an AI unit according to an embodiment of the present application, and this device can correspond to a terminal in other embodiments. As Figure 4 shown, the device 400 includes the following modules.

[0346] An activation module 401, configured to activate a target AI unit according to target information, where the target information includes at least one of indication information and an activation condition of the AI unit, and the indication information is used to activate the target AI unit.

[0347] Optionally, in some embodiments, when the target information includes the indication information, the device 400 further includes a receiving module;

[0348] The receiving module is configured to receive the indication information.

[0349] Optionally, in some embodiments, the device 400 further includes a sending module;

[0350] The activation module 401 is configured to determine whether the activated AI unit meets a deactivation condition;

[0351] The sending module is configured to send an AI unit activation request when it is determined that the activated AI unit meets the deactivation condition;

[0352] The receiving module is configured to receive the indication information.

[0353] Optionally, in some embodiments, the indication information indicates any one of the following:

[0354] Activate the target AI unit;

[0355] Activate the target AI unit and deactivate the activated AI unit;

[0356] The terminal selects the target AI unit to be activated.

[0357] Optionally, in some embodiments, when the indication information indicates to activate the target AI unit or to activate the target AI unit and deactivate the activated AI unit, the indication information includes unit information of the target AI unit; wherein, the activation module 401 is configured to:

[0358] Judge whether the target AI unit meets the activation condition according to the unit information;

[0359] When it is determined that the target AI unit meets the activation condition, activate the target AI unit.

[0360] Optionally, in some embodiments, it further includes at least one of the following:

[0361] A sending module, configured to send first information, where the first information represents that the target AI unit does not meet the activation condition;

[0362] A receiving module, configured to receive second information, where the second information indicates deactivating the target AI unit;

[0363] The activation module 401 is configured to deactivate the target AI unit;

[0364] The sending module is configured to send third information, where the third information represents that the target AI unit is not activated;

[0365] The receiving module is configured to receive fourth information, where the fourth information indicates performing a first operation;

[0366] The activation module 401 is configured to perform the first operation;

[0367] Among them, the first information and the third information are carried by RRC signaling or uplink control information UCI, the second information and the fourth information are carried by RRC signaling, DCI or MAC-CE, and the first operation includes at least one of fine-tuning the target AI unit, retraining the target AI unit, upgrading the target AI unit, downgrading the target AI unit, rolling back the target AI unit, performing data collection, and performing AI unit transfer with other devices.

[0368] Optionally, in some embodiments, the activation module 401 is further configured to:

[0369] Judge whether one or more AI units meet the activation condition according to the unit information of the one or more AI units;

[0370] Determine the AI units that meet the activation condition among the one or more AI units as target AI units, and activate the target AI units.

[0371] Optionally, in some embodiments, the unit information includes at least one of the following:

[0372] The ID of the AI unit;

[0373] Scenarios applicable to the AI unit;

[0374] Function information supported by the AI unit;

[0375] Cell information supported by the AI unit;

[0376] Area information supported by the AI unit;

[0377] Input / output type information of the AI unit;

[0378] Inference accuracy supported by the AI unit;

[0379] Terminal computing and storage capabilities corresponding to the AI unit;

[0380] Datasets related to the AI unit;

[0381] Performance threshold of the AI unit;

[0382] Complexity threshold of the AI unit.

[0383] Optionally, in some embodiments, when the target information includes the activation condition of the AI unit, the activation module 401 is configured to:

[0384] Determine whether the target AI unit meets the activation condition;

[0385] Activate the target AI unit when it is determined that the target AI unit meets the activation condition.

[0386] Optionally, in some embodiments, the activation module 401 is configured to:

[0387] Determine whether the activated AI unit meets the deactivation condition;

[0388] When it is determined that the activated AI unit meets the deactivation condition, determine whether the target AI unit meets the activation condition.

[0389] Optionally, in some embodiments, the deactivation condition is related to at least one of the following:

[0390] Performance when applying the activated AI unit;

[0391] Computing and storage capabilities supported by the terminal;

[0392] ID of the activated AI unit;

[0393] Scenarios applicable to the activated AI unit;

[0394] Function information supported by the activated AI unit;

[0395] Cell information supported by the activated AI unit

[0396] Area information supported by the activated AI unit

[0397] Input / output type information of the activated AI unit

[0398] Inference accuracy supported by the activated AI unit

[0399] Terminal computing and storage capabilities corresponding to the activated AI unit

[0400] Dataset related to the activated AI unit

[0401] Performance threshold of the activated AI unit

[0402] Complexity threshold of the activated AI unit

[0403] Optionally, in some embodiments, the activation condition is related to at least one of the following:

[0404] Terminal computing and storage capabilities supported by the terminal

[0405] ID of the target AI unit

[0406] Scenarios applicable to the target AI unit

[0407] Function information supported by the target AI unit

[0408] Cell information supported by the target AI unit

[0409] Area information supported by the target AI unit

[0410] Input / output type information of the target AI unit

[0411] Inference accuracy supported by the target AI unit

[0412] Terminal computing and storage capabilities corresponding to the target AI unit

[0413] Dataset related to the target AI unit

[0414] Performance threshold of the target AI unit

[0415] Complexity threshold of the target AI unit

[0416] Optionally, in some embodiments, the activation module 401 is configured to:

[0417] After performing a first process or a second process on the target AI unit, apply the target AI unit; or,

[0418] Deactivate the activated AI unit and after performing a first process or a second process on the target AI unit, apply the target AI unit;

[0419] Wherein, the first process includes reading or loading at least part of the information of the target AI unit, and the second process includes reading or loading the remaining information of the target AI unit.

[0420] Optionally, in some embodiments, it further includes:

[0421] A sending module, configured to send fifth information, where the fifth information represents that the terminal has completed the activation of the target AI unit, and the fifth information is carried by an RRC signaling or UCI.

[0422] Optionally, in some embodiments, when the number of target AI units is one, the activation delay of one target AI unit is less than or equal to a first delay; when the number of target AI units is multiple, the activation delay of multiple target AI units is less than or equal to a second delay, and the second delay is less than or equal to the sum of the activation delays of multiple target AI units;

[0423] Wherein, when multiple target AI units can be activated in parallel, the second delay is less than or equal to the maximum value of the activation delays of the multiple target AI units; when the difference between multiple target AI units is within a specified difference, the activation delays of the multiple target AI units are less than or equal to a third delay, the third delay is less than the second delay, and the difference is related to at least one of the structure, parameters, complexity, size, quantization level, function, corresponding terminal operation and storage capabilities, and applicable scenarios of the A1 unit.

[0424] Optionally, in some embodiments, the activation delay of one target AI unit includes at least one of the following:

[0425] The delay between the terminal sending an AI unit activation request and receiving the indication information;

[0426] The processing or parsing delay of the signaling carrying the indication information by the terminal;

[0427] The delay of the terminal performing hybrid automatic repeat request (HARQ) feedback;

[0428] The delay of the terminal deactivating the activated AI unit;

[0429] The latency of the first processing performed by the terminal on the target AI unit, where the first processing includes reading or loading partial information of the target AI unit;

[0430] The latency of the second processing and application performed by the terminal on the target AI unit, where the second processing includes reading or loading the remaining information of the target AI unit.

[0431] The device 400 according to an embodiment of the present application may refer to the process of the method 200 corresponding to the embodiment of the present application. Moreover, each unit / module in the device 400 and the above other operations and / or functions respectively are for implementing the corresponding processes in the method 200, and can achieve the same or equivalent technical effects. For the sake of brevity, they will not be elaborated here.

[0432] Figure 5 It is a schematic structural diagram of an activation device for an AI unit according to an embodiment of the present application. This device may correspond to a network-side device in other embodiments. As Figure 5 shown, the device 500 includes the following modules.

[0433] A sending module 501, configured to send indication information for activating a target AI unit.

[0434] Optionally, in some embodiments, the device further includes a receiving module;

[0435] The receiving module is configured to receive an AI unit activation request;

[0436] The sending module is configured to send the indication information according to the AI unit activation request.

[0437] Optionally, in some embodiments, the indication information is carried by an RRC signaling, a DCI, or a MAC-CE.

[0438] Optionally, in some embodiments, the indication information indicates any one of the following:

[0439] Activating the target AI unit;

[0440] Activating the target AI unit and deactivating an already activated AI unit;

[0441] The terminal selects the target AI unit to be activated.

[0442] Optionally, in some embodiments, when the indication information is used to indicate activating the target AI unit or activating the target AI unit and deactivating an already activated AI unit, the indication information includes unit information of the target AI unit, and the unit information is used for the terminal to determine whether the target AI unit meets the activation conditions.

[0443] Optionally, in some embodiments, the unit information includes at least one of the following:

[0444] The ID of the target AI unit;

[0445] The scenarios applicable to the target AI unit;

[0446] The function information supported by the target AI unit;

[0447] The cell information supported by the target AI unit;

[0448] The area information supported by the target AI unit;

[0449] The input / output type information of the target AI unit;

[0450] The inference accuracy supported by the target AI unit;

[0451] The terminal computing and storage capabilities corresponding to the target AI unit;

[0452] The dataset related to the target AI unit;

[0453] The performance threshold value of the target AI unit;

[0454] The complexity threshold value of the target AI unit.

[0455] Optionally, in some embodiments, the apparatus 500 further includes at least one of the following:

[0456] A receiving module, configured to receive first information, where the first information characterizes that the target AI unit does not meet the activation condition;

[0457] The sending module 501, configured to send second information, where the second information instructs to cancel the activation of the target AI unit;

[0458] The receiving module, configured to receive third information, where the third information characterizes that the target AI unit is not activated;

[0459] The sending module 501, configured to send fourth information, where the fourth information instructs to perform a first operation;

[0460] Among them, the first information and the third information are carried by RRC signaling or UCI, the second information and the fourth information are carried by RRC signaling, DCI, or MAC-CE, and the first operation includes at least one of fine-tuning the target AI unit, retraining the target AI unit, upgrading the target AI unit, downgrading the target AI unit, rolling back the target AI unit, performing data collection, and executing AI unit transfer with other devices.

[0461] Optionally, in some embodiments, it further includes:

[0462] A receiving module, configured to receive fifth information, where the fifth information indicates that the terminal has completed the activation of the target AI unit, and the fifth information is carried by RRC signaling or UCI.

[0463] The apparatus 500 according to an embodiment of the present application may refer to the process of the method 300 corresponding to the embodiment of the present application. Moreover, each unit / module in the apparatus 500 and the above other operations and / or functions respectively implement the corresponding processes in the method 300 and can achieve the same or equivalent technical effects. For the sake of brevity, details are not described herein again.

[0464] The activation apparatus of the AI unit in the embodiment of the present application may be an electronic device, such as an electronic device with an operating system, or a component in an electronic device, such as an integrated circuit or a chip. The electronic device may be a terminal or other devices other than the terminal. Exemplarily, the terminal may include, but is not limited to, the types of the terminal 11 listed above, and other devices may be a server, a Network Attached Storage (NAS), etc. The embodiment of the present application does not make specific limitations.

[0465] The activation apparatus of the AI unit provided in the embodiment of the present application can implement Figure 2 and Figure 3 each process implemented by the method embodiment of

[0466] Such as Figure 6As shown in the figure, an embodiment of the present application further provides a communication device 600, including a processor 601 and a memory 602. A program or instruction that can run on the processor 601 is stored on the memory 602. For example, when the communication device 600 is a terminal, when the program or instruction is executed by the processor 601, each step of the above-mentioned method embodiment for activating the AI unit is implemented, and the same technical effect can be achieved. When the communication device 600 is a network-side device, when the program or instruction is executed by the processor 601, each step of the above-mentioned method embodiment for activating the AI unit is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be elaborated here.

[0467] An embodiment of the present application further provides a terminal, including a processor and a communication interface. The communication interface is coupled to the processor, and the processor is used to run a program or instruction to implement the steps in the method embodiment as Figure 2 shown. This terminal embodiment corresponds to the above-mentioned terminal-side method embodiment. Each implementation process and implementation manner of the above-mentioned method embodiment can be applied to this terminal embodiment, and the same technical effect can be achieved. Specifically, Figure 7 FIG. is a schematic diagram of the hardware structure of a terminal according to an embodiment of the present application.

[0468] The terminal 700 includes, but is not limited to, at least some components such as a radio frequency unit 701, a network module 702, an audio output unit 703, an input unit 704, a sensor 705, a display unit 706, a user input unit 707, an interface unit 708, a memory 709, and a processor 710.

[0469] Those skilled in the art can understand that the terminal 700 may further include a power source (such as a battery) for supplying power to each component. The power source can be logically connected to the processor 710 through a power management system, so as to implement functions such as management of charging, discharging, and power consumption management through the power management system. Figure 7 The terminal structure shown in does not constitute a limitation on the terminal. The terminal may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements, which will not be elaborated here.

[0470] It should be understood that in the embodiments of the present application, the input unit 704 may include a Graphics Processing Unit (GPU) 7041 and a microphone 7042. The graphics processor 7041 processes the image data of static pictures or videos obtained by an image capturing device (such as a camera) in a video capture mode or an image capture mode. The display unit 706 may include a display panel 7061, and the display panel 7061 may be configured in the form of, for example, a liquid crystal display, an organic light emitting diode, etc. The user input unit 707 includes at least one of a touch panel 7071 and other input devices 7072. The touch panel 7071 is also referred to as a touch screen. The touch panel 7071 may include two parts: a touch detection device and a touch controller. The other input devices 7072 may include, but are not limited to, a physical keyboard, function keys (such as volume control keys, switch keys, etc.), a trackball, a mouse, a joystick, which will not be elaborated herein.

[0471] In the embodiments of the present application, after receiving downlink data from a network side device, the radio frequency unit 701 may transmit it to the processor 710 for processing; in addition, the radio frequency unit 701 may send uplink data to the network side device. Generally, the radio frequency unit 701 includes, but is not limited to, an antenna, an amplifier, a transceiver, a coupler, a low noise amplifier, a duplexer, etc.

[0472] The memory 709 can be used to store software programs or instructions as well as various data. The memory 709 may mainly include a first storage area for storing programs or instructions and a second storage area for storing data. Among them, the first storage area may store an operating system, application programs or instructions required for at least one function (such as a sound playback function, an image playback function, etc.). In addition, the memory 709 may include volatile memory or non-volatile memory. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), a static random access memory (SRAM), a dynamic random access memory (DRAM), a synchronous dynamic random access memory (SDRAM), a double data rate synchronous dynamic random access memory (DDR SDRAM), an enhanced synchronous dynamic random access memory (ESDRAM), a synch link dynamic random access memory (SLDRAM), and a direct rambus random access memory (DRRAM). The memory 709 in the embodiments of the present application includes but is not limited to these and any other suitable types of memory.

[0473] The processor 710 may include one or more processing units; optionally, the processor 710 integrates an application processor and a modem processor. Among them, the application processor mainly processes operations related to the operating system, user interface, and application programs, etc., and the modem processor mainly processes wireless communication signals, such as a baseband processor. It can be understood that the above modem processor may not be integrated into the processor 710 either.

[0474] Among them, the processor 710 is used to activate a target AI unit according to target information, where the target information includes at least one of indication information and an activation condition of the AI unit, and the indication information is used to activate the target AI unit.

[0475] In an embodiment of the present application, the terminal may activate a target AI unit according to at least one of the indication information and the activation condition of the AI unit. Thus, the terminal can determine how to activate the AI unit, thereby optimizing the process of AI life cycle management, ensuring the high quality, high efficiency, and high reliability of the AI unit, meeting business requirements, and improving work efficiency.

[0476] It can be understood that the implementation processes of the implementation manners mentioned in this embodiment may refer to the relevant descriptions of Method Embodiment 200 and achieve the same or corresponding technical effects. To avoid repetition, they will not be elaborated here.

[0477] An embodiment of the present application further provides a network-side device, including a processor and a communication interface. The communication interface is coupled to the processor. The processor is used to run programs or instructions to implement the steps of the method embodiment as Figure 3 shown. This network-side device embodiment corresponds to the above network-side device method embodiment. The implementation processes and implementation manners of the above method embodiment can all be applied to this network-side device embodiment and can achieve the same technical effects.

[0478] Specifically, an embodiment of the present application further provides a network-side device. As Figure 8 shown, the network-side device 800 includes: an antenna 81, a radio frequency device 82, a baseband device 83, a processor 84, and a memory 85. The antenna 81 is connected to the radio frequency device 82. In the uplink direction, the radio frequency device 82 receives information through the antenna 81 and sends the received information to the baseband device 83 for processing. In the downlink direction, the baseband device 83 processes the information to be sent and sends it to the radio frequency device 82. The radio frequency device 82 processes the received information and then sends it out through the antenna 81.

[0479] The method executed by the network-side device in the above embodiments can be implemented in the baseband device 83, and the baseband device 83 includes a baseband processor.

[0480] The baseband device 83 may include, for example, at least one baseband board, and a plurality of chips are arranged on the baseband board. As Figure 8 shown, one of the chips is, for example, a baseband processor, which is connected to the memory 85 through a bus interface to call the program in the memory 85 and execute the operations of the network device shown in the above method embodiments.

[0481] The network-side device may further include a network interface 86, and this interface is, for example, a Common Public Radio Interface (CPRI).

[0482] Specifically, the network-side device 800 in the embodiments of the present application further includes: instructions or programs stored in the memory 85 and executable on the processor 84. The processor 84 calls the instructions or programs in the memory 85 to execute Figure 5 the methods executed by the modules shown, and achieves the same technical effects. To avoid repetition, it will not be elaborated here.

[0483] The embodiments of the present application further provide a readable storage medium. Programs or instructions are stored on the readable storage medium. When the programs or instructions are executed by a processor, the various processes of the above-mentioned embodiment of the activation method of the AI unit are implemented, and the same technical effects can be achieved. To avoid repetition, it will not be elaborated here.

[0484] Wherein, the processor is the processor in the terminal described in the above embodiment. The readable storage medium includes computer-readable storage media, such as computer read-only memory ROM, random access memory RAM, magnetic disks or optical discs, etc. In some examples, the readable storage medium may be a non-transitory readable storage medium.

[0485] The embodiments of the present application further provide a chip. The chip includes a processor and a communication interface. The communication interface is coupled to the processor. The processor is used to run programs or instructions to implement the various processes of the above-mentioned embodiment of the activation method of the AI unit, and the same technical effects can be achieved. To avoid repetition, it will not be elaborated here.

[0486] It should be understood that the chip mentioned in the embodiments of the present application may also be referred to as a system-on-chip, system chip, chip system or system-on-chip, etc.

[0487] The embodiments of the present application further provide a computer program / program product. The computer program / program product is stored in a storage medium. The computer program / program product is executed by at least one processor to implement the various processes of the above-mentioned embodiment of the activation method of the AI unit, and the same technical effects can be achieved. To avoid repetition, it will not be elaborated here.

[0488] The embodiments of the present application further provide a wireless communication system, including: a terminal and a network-side device. The terminal can be used to execute the steps of the above-mentioned activation method of the AI unit, and the network-side device can be used to execute the steps of the above-mentioned activation method of the AI unit.

[0489] It should be noted that in this text, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or device comprising that element. In addition, it should be pointed out that the scope of the methods and devices in the embodiments of the present application is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in the reverse order according to the functions involved. For example, the described methods may be performed in an order different from that described, and various steps may be added, omitted or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0490] From the description of the above embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of computer software products plus the necessary general hardware platforms, and of course, can also be implemented by hardware. The computer software products are stored in storage media (such as ROM, RAM, magnetic disks, optical disks, etc.) and include several instructions for causing a terminal or a network-side device to execute the methods described in the various embodiments of the present application.

[0491] The embodiments of the present application have been described above in conjunction with the accompanying drawings, but the present application is not limited to the above specific embodiments. The above specific embodiments are merely illustrative and not restrictive. Under the inspiration of the present application, those of ordinary skill in the art can also make many forms of embodiments without departing from the purpose of the present application and the scope protected by the claims. These embodiments are all within the protection scope of the present application.

Claims

1. A method for activating an AI unit, characterized in that, including the following: The terminal activates a target AI unit according to target information, where the target information includes at least one of indication information and an activation condition of the AI unit, and the indication information is used to activate the target AI unit.

2. The method according to claim 1, wherein When the target information includes the indication information, the method further includes: The terminal receives the indication information.

3. The method according to claim 2, wherein The terminal receiving the indication information includes: The terminal determines whether the activated AI unit satisfies a deactivation condition; When the terminal determines that the activated AI unit satisfies the deactivation condition, the terminal sends an AI unit activation request; The terminal receives the indication information.

4. The method according to any one of claims 1 to 3, characterized in that, The indication information indicates any one of the following: Activating the target AI unit; Activating the target AI unit and deactivating the activated AI unit; The terminal selects a target AI unit to be activated.

5. The method according to claim 4, wherein When the indication information indicates activating the target AI unit or activating the target AI unit and deactivating the activated AI unit, the indication information includes unit information of the target AI unit; Wherein, the terminal activating the target AI unit according to the target information includes: The terminal determines whether the target AI unit satisfies the activation condition according to the unit information; When the terminal determines that the target AI unit satisfies the activation condition, the terminal activates the target AI unit.

6. The method according to claim 5, wherein When the terminal determines that the target AI unit does not satisfy the activation condition, the method further includes at least one of the following: The terminal sends first information, and the first information represents that the target AI unit does not satisfy the activation condition; The terminal receives second information, and the second information indicates canceling the activation of the target AI unit; The terminal cancels the activation of the target AI unit; The terminal sends third information, and the third information represents that the target AI unit is not activated; The terminal receives fourth information, and the fourth information indicates performing a first operation; The terminal performs the first operation; Wherein, the first information and the third information are carried by RRC signaling or uplink control information UCI, the second information and the fourth information are carried by RRC signaling, DCI or MAC-CE, and the first operation includes at least one of fine-tuning the target AI unit, retraining the target AI unit, upgrading the target AI unit, downgrading the target AI unit, rolling back the target AI unit, performing data collection, and performing AI unit transfer with other devices.

7. The method according to claim 4, wherein When the indication information indicates that the terminal selects a target AI unit to be activated, the terminal activating the target AI unit according to the target information includes: The terminal determines whether one or more AI units satisfy the activation condition according to the unit information of the one or more AI units; The terminal determines the AI unit that satisfies the activation condition among the one or more AI units as the target AI unit and activates the target AI unit.

8. The method according to claim 5 or 7, characterized in that, The unit information includes at least one of the following: The ID of the AI unit; The scenarios applicable to the AI unit; Function information supported by the AI unit; Cell information supported by the AI unit; Area information supported by the AI unit; Input / output type information of the AI unit; Inference accuracy supported by the AI unit; Terminal computing and storage capabilities corresponding to the AI unit; Dataset related to the AI unit; Performance threshold of the AI unit; Complexity threshold of the AI unit.

9. The method according to claim 1, characterized in that, When the target information includes the activation condition of the AI unit, the terminal activates the target AI unit according to the target information, including: The terminal determines whether the target AI unit meets the activation condition; When the terminal determines that the target AI unit meets the activation condition, the terminal activates the target AI unit.

10. The method according to claim 9, wherein The terminal determines whether the target AI unit meets the activation condition, including: The terminal determines whether the activated AI unit meets the deactivation condition; When the terminal determines that the activated AI unit meets the deactivation condition, the terminal determines whether the target AI unit meets the activation condition.

11. The method according to claim 3 or 10, characterized in that, The deactivation condition is related to at least one of the following: Performance when applying the activated AI unit; Computing and storage capabilities supported by the terminal; ID of the activated AI unit; Scenarios applicable to the activated AI unit; Function information supported by the activated AI unit; Cell information supported by the activated AI unit; Area information supported by the activated AI unit; Input / output type information of the activated AI unit; Inference accuracy supported by the activated AI unit; Terminal computing and storage capabilities corresponding to the activated AI unit; Dataset related to the activated AI unit; Performance threshold of the activated AI unit; Complexity threshold of the activated AI unit.

12. The method according to any one of claims 1, 5, 6, 7, 9 and 10, characterized in that, The activation condition is related to at least one of the following: Computing and storage capabilities supported by the terminal; ID of the target AI unit; Scenarios applicable to the target AI unit; Function information supported by the target AI unit; Cell information supported by the target AI unit; Area information supported by the target AI unit; Input / output type information of the target AI unit; Inference accuracy supported by the target AI unit; Terminal computing and storage capabilities corresponding to the target AI unit; Dataset related to the target AI unit; Performance threshold of the target AI unit; Complexity threshold of the target AI unit.

13. The method according to claim 1, wherein The terminal activates the target AI unit according to the target information, including: After the terminal performs the first processing or the second processing on the target AI unit, the terminal applies the target AI unit; or, The terminal deactivates the activated AI unit and after performing the first processing or the second processing on the target AI unit, the terminal applies the target AI unit; Wherein, the first processing includes reading or loading at least part of the information of the target AI unit, and the second processing includes reading or loading the remaining information of the target AI unit.

14. The method according to claim 1 or 13, characterized in that The method further includes: The terminal sends fifth information, where the fifth information indicates that the terminal has completed the activation of the target AI unit, and the fifth information is carried by RRC signaling or UCI.

15. The method according to claim 1, characterized in that, When the number of target AI units is one, the activation delay of one target AI unit is less than or equal to the first delay; when the number of target AI units is multiple, the activation delay of multiple target AI units is less than or equal to the second delay, and the second delay is less than or equal to the sum of the activation delays of multiple target AI units; Among them, when multiple target AI units can be activated in parallel, the second delay is less than or equal to the maximum value of the activation delays of the multiple target AI units; when the difference between multiple target AI units is within the specified difference, the activation delays of the multiple target AI units are less than or equal to the third delay, the third delay is less than the second delay, and the difference is related to at least one of the structure, parameters, complexity, size, quantization level, function, corresponding terminal operation and storage capabilities, and applicable scenarios of the A1 unit.

16. The method according to claim 15, wherein The activation delay of one target AI unit includes at least one of the following: The delay between the terminal sending an AI unit activation request and receiving the indication information; The processing or parsing delay of the signaling carrying the indication information by the terminal; The delay of the terminal performing hybrid automatic repeat request (HARQ) feedback; The delay of the terminal deactivating the activated AI unit; The delay of the terminal performing a first process on the target AI unit, where the first process includes reading or loading partial information of the target AI unit; The delay of the terminal performing a second process and application on the target AI unit, where the second process includes reading or loading the remaining information of the target AI unit.

17. A method for activating an AI unit, characterized in that, Includes: The network-side device sends indication information, where the indication information is used to activate the target AI unit.

18. The method according to claim 17, wherein The network-side device sending the indication information includes: The network-side device receives an AI unit activation request; The network-side device sends the indication information according to the AI unit activation request.

19. The method according to claim 17 or 18, characterized in that, The indication information indicates any one of the following: Activating the target AI unit; Activating the target AI unit and deactivating the activated AI unit; The terminal selects the target AI unit to be activated.

20. The method according to claim 19, wherein When the indication information is used to indicate activating the target AI unit or activating the target AI unit and deactivating the activated AI unit, the indication information includes the unit information of the target AI unit, and the unit information is used for the terminal to determine whether the target AI unit meets the activation conditions.

21. The method according to claim 20, wherein The unit information includes at least one of the following: The ID of the target AI unit; The applicable scenario of the target AI unit; The function information supported by the target AI unit; The cell information supported by the target AI unit; The area information supported by the target AI unit; The input-output type information of the target AI unit; The inference accuracy supported by the target AI unit; The corresponding terminal operation and storage capabilities of the target AI unit; The dataset related to the target AI unit; The performance threshold of the target AI unit; The complexity threshold of the target AI unit.

22. The method according to claim 19, wherein The method further includes at least one of the following: The network-side device receives first information, which characterizes that the target AI unit does not meet the activation condition; The network-side device sends second information, which indicates deactivating the target AI unit; The network-side device receives third information, which characterizes that the target AI unit is not activated; The network-side device sends fourth information, which indicates performing a first operation; Wherein, the first information and the third information are carried by RRC signaling or UCI, the second information and the fourth information are carried by RRC signaling, DCI or MAC-CE, and the first operation includes at least one of fine-tuning the target AI unit, retraining the target AI unit, upgrading the target AI unit, downgrading the target AI unit, rolling back the target AI unit, performing data collection, and performing AI unit transfer with other devices.

23. The method according to claim 17, wherein, The method further includes: The network-side device receives fifth information, which characterizes that the terminal has completed the activation of the target AI unit, and the fifth information is carried by RRC signaling or UCI.

24. An activation device for an AI unit, characterized in that, Including: An activation module, configured to activate a target AI unit according to target information, where the target information includes at least one of indication information and the activation condition of the AI unit, and the indication information is used to activate the target AI unit.

25. The device according to claim 24, characterized in that, The apparatus further includes: A receiving module, configured to receive the indication information.

26. The device according to claim 25, characterized in that, The apparatus further includes: A judging module, configured to judge whether the activated AI unit meets the deactivation condition; A sending module, configured to send an AI unit activation request when it is determined that the activated AI unit meets the deactivation condition; The receiving module, configured to receive the indication information.

27. The device according to any one of claims 24 to 26, characterized in that The indication information indicates any one of the following: Activating the target AI unit; Activating the target AI unit and deactivating the activated AI unit; The terminal selects the target AI unit to be activated.

28. The device according to claim 27, wherein When the indication information indicates activating the target AI unit or activating the target AI unit and deactivating the activated AI unit, the unit information of the target AI unit is included in the indication information; Wherein, the activation module is configured to: Judge whether the target AI unit meets the activation condition according to the unit information; Activate the target AI unit when it is determined that the target AI unit meets the activation condition.

29. The device according to claim 28, wherein, Further includes at least one of the following: A sending module, configured to send first information, which characterizes that the target AI unit does not meet the activation condition; A receiving module, configured to receive second information, which indicates deactivating the target AI unit; The activation module, configured to deactivate the target AI unit; The sending module, configured to send third information, which characterizes that the target AI unit is not activated; The receiving module is configured to receive fourth information, where the fourth information indicates to perform a first operation; The activation module is configured to perform the first operation; Wherein, the first information and the third information are carried by RRC signaling or uplink control information UCI, the second information and the fourth information are carried by RRC signaling, DCI or MAC-CE, and the first operation includes at least one of fine-tuning the target AI unit, retraining the target AI unit, upgrading the target AI unit, downgrading the target AI unit, rolling back the target AI unit, performing data collection, and performing AI unit transfer with other devices.

30. The device according to claim 27, characterized in that, The activation module is configured to: Judge whether one or more AI units meet the activation conditions according to the unit information of the one or more AI units; Determine the AI units that meet the activation conditions among the one or more AI units as target AI units, and activate the target AI units.

31. The device according to claim 24, characterized in that, When the target information includes the activation conditions of the AI unit, the activation module is configured to: Judge whether the target AI unit meets the activation conditions; When it is determined that the target AI unit meets the activation conditions, activate the target AI unit.

32. The device according to claim 31, characterized in that, The activation module is configured to: Judge whether the activated AI unit meets the deactivation conditions; When it is determined that the activated AI unit meets the deactivation conditions, judge whether the target AI unit meets the activation conditions.

33. The device according to claim 24, characterized in that, The activation module is configured to: After performing a first process or a second process on the target AI unit, apply the target AI unit; or, Deactivate the activated AI unit and, after performing a first process or a second process on the target AI unit, apply the target AI unit; Wherein, the first process includes reading or loading at least part of the information of the target AI unit, and the second process includes reading or loading the remaining information of the target AI unit.

34. The device according to claim 24 or 33, characterized in that, It further includes: A sending module, configured to send fifth information, where the fifth information represents that the terminal has completed the activation of the target AI unit, and the fifth information is carried by RRC signaling or UCI.

35. An activation device for an AI unit, characterized in that, It includes: A sending module, configured to send indication information for activating a target AI unit.

36. The device according to claim 35, characterized in that, The device further includes a receiving module; The receiving module is configured to receive an AI unit activation request; The sending module is configured to send the indication information according to the AI unit activation request.

37. The device according to claim 35 or 36, characterized in that The indication information indicates any one of the following: Activate the target AI unit; Activate the target AI unit and deactivate the activated AI unit; The terminal selects the target AI unit to be activated.

38. A terminal, characterized in that, It includes a processor and a memory, where the memory stores a program or instruction that can run on the processor, and when the program or instruction is executed by the processor, it implements the steps of the method for activating an AI unit according to any one of claims 1 to 16.

39. A network-side device, characterized in that, It includes a processor and a memory, and the memory stores programs or instructions that can run on the processor. When the programs or instructions are executed by the processor, the steps of the activation method of the AI unit according to any one of claims 17 to 23 are implemented.

40. A readable storage medium, characterized in that, Programs or instructions are stored on the readable storage medium. When the programs or instructions are executed by a processor, the steps of the activation method of the AI unit according to any one of claims 1 to 16 are implemented, or the steps of the activation method of the AI unit according to any one of claims 17 to 23 are implemented.

Images