Information transmission method and apparatus, and communication device

Abstract

The application discloses an information transmission method, device and communication equipment, and belongs to the technical field of communication. The information transmission method of the application comprises the following steps: a first communication equipment sends first information of an artificial intelligence (AI) model to a second communication equipment; wherein the first information comprises at least one of the following: operation time information of the AI model; whether the operation time of the AI model meets preset requirements; complexity information of the AI model; and maximum complexity of the AI model allowed on the premise of meeting time axis requirements.

Description

Technical Field

[0001] This application belongs to the field of communication technology, specifically relating to an information transmission method, apparatus, and communication equipment. Background Technology

[0002] Artificial intelligence (AI) has been widely applied in various fields. AI models can be implemented in various ways, such as neural networks, decision trees, support vector machines, and Bayesian classifiers.

[0003] Take neural networks as an example (but this doesn't limit the specific type of AI model). A neural network consists of neurons. The parameters of the neural network are optimized using optimization algorithms. An optimization algorithm is a class of algorithms that minimizes or maximizes an objective function (or loss function). The objective function is a mathematical combination of the model parameters and the data. For example, given data X and its corresponding label Y, we build a neural network model f(.). With the model, we can obtain the predicted output f(x) based on the input x, and calculate the difference between the predicted value and the true value (f(x) - Y), which is the loss function. The goal of the optimization algorithm is to find suitable parameters that minimize the value of the loss function. The smaller the loss value, the closer the model is to the reality.

[0004] Most common optimization algorithms are based on the error back propagation (BP) algorithm. The basic idea of ​​the BP algorithm is that the learning process consists of two parts: forward propagation of the signal and backward propagation of the error. During forward propagation, the input sample is introduced from the input layer, processed layer by layer by the hidden layers, and then propagated to the output layer. If the actual output of the output layer does not match the expected output, the process transitions to the error back propagation stage. Error back propagation involves propagating the output error back to the input layer layer by layer through the hidden layers, distributing the error to all units in each layer, thus obtaining the error signal of each unit. This error signal serves as the basis for adjusting the weights of each unit. This process of adjusting the weights through forward and backward propagation is repeated continuously. This continuous adjustment of weights is the learning and training process of the network. This process continues until the error of the network output is reduced to an acceptable level, or until the predetermined number of learning iterations is reached.

[0005] However, due to differences in computing power and the complexity of different AI models, the AI ​​models used by terminals or network devices may be of inappropriate complexity, thus affecting the performance of the communication system. Summary of the Invention

[0006] This application provides an information transmission method, apparatus, and communication device that can solve the problem of inappropriate complexity of AI models used in terminals or network-side devices in the prior art.

[0007] Firstly, an information transmission method is provided, including:

[0008] The first communication device sends first information about an artificial intelligence (AI) model to the second communication device; wherein the first information includes at least one of the following:

[0009] AI model computation time information;

[0010] Does the computation time of the AI ​​model meet the preset requirements?

[0011] Information on the complexity of AI models;

[0012] The maximum complexity of an AI model allowed while meeting timeline requirements.

[0013] Secondly, an information transmission device is provided, applied to a first communication device, comprising:

[0014] The sending module is configured to send first information of an artificial intelligence (AI) model to a second communication device; wherein the first information includes at least one of the following:

[0015] AI model computation time information;

[0016] Does the computation time of the AI ​​model meet the preset requirements?

[0017] Information on the complexity of AI models;

[0018] The maximum complexity of an AI model allowed while meeting timeline requirements.

[0019] Thirdly, a communication device is provided, comprising a processor and a memory, the memory storing a program or instructions executable on the processor, the program or instructions, when executed by the processor, implementing the steps of the method described in the first aspect.

[0020] Fourthly, a communication device is provided, including a processor and a communication interface, wherein the processor is used to determine first information of an AI model, and the communication interface is used to send the first information of the AI ​​model to a second communication device; wherein the first information includes at least one of the following:

[0021] AI model computation time information;

[0022] Does the computation time of the AI ​​model meet the preset requirements?

[0023] Information on the complexity of AI models;

[0024] The maximum complexity of an AI model allowed while meeting timeline requirements.

[0025] Fifthly, a readable storage medium is provided, on which a program or instructions are stored, which, when executed by a processor, implement the steps of the method described in the first aspect.

[0026] In a sixth aspect, a chip is provided, the chip including a processor and a communication interface coupled to the processor, the processor being used to run programs or instructions to implement the method as described in the first aspect.

[0027] In a seventh aspect, a computer program / program product is provided, the computer program / program product being stored in a storage medium, the computer program / program product being executed by at least one processor to perform the steps of the method as described in the first aspect.

[0028] In this embodiment, the first communication device sends computation time-related information and / or complexity-related information of the AI ​​model to the second communication device, enabling the first and second communication devices to align computation time and complexity to determine a suitable AI model, thereby improving the performance of the AI ​​model in the communication system and further enhancing system performance. Attached Figure Description

[0029] Figure 1 A block diagram illustrating a wireless communication system to which embodiments of this application may be applied;

[0030] Figure 2 This diagram illustrates the steps of the information transmission method provided in the embodiments of this application.

[0031] Figure 3 This is a schematic diagram illustrating the structure of the information transmission device provided in the embodiments of this application;

[0032] Figure 4 This is a schematic diagram of the structure of the communication device provided in the embodiments of this application;

[0033] Figure 5 This is a schematic diagram of the structure of the terminal provided in the embodiments of this application;

[0034] Figure 6 This is a schematic diagram illustrating the structure of the network-side device provided in the embodiments of this application. Detailed Implementation

[0035] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0036] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such terms can be used interchangeably where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first" and "second" are generally of the same class, not limited in number; for example, a first object can be one or more. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0037] It is worth noting that the technologies described in this application are not limited to Long Term Evolution (LTE) / LTE-Advanced (LTE-A) systems, but 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), and other systems. The terms "system" and "network" in this application are often used interchangeably, and the described technologies can be used with the systems and radio technologies mentioned above, as well as with other systems and radio technologies. The following description describes New Radio (NR) systems for illustrative purposes, and NR terminology is used in most of the following description; however, these technologies can also be applied to applications beyond NR systems, such as 6th generation (6G) radio systems. th Generation 6G communication system.

[0038] Figure 1This diagram illustrates a block diagram of a wireless communication system applicable to embodiments of this application. The wireless communication system includes a terminal 11 and a network-side device 12. Terminal 11 can be a mobile phone, tablet computer, laptop computer, personal digital assistant (PDA), handheld computer, netbook, ultra-mobile personal computer (UMPC), mobile internet device (MID), augmented reality (AR) / virtual reality (VR) device, robot, wearable device, vehicle-mounted device (VUE), pedestrian terminal (PUE), smart home (home devices with wireless communication capabilities, such as refrigerators, televisions, washing machines, or furniture), game console, personal computer (PC), ATM, or self-service machine, etc. Wearable devices include: smartwatches, smart bracelets, smart headphones, smart glasses, smart jewelry (smart bracelets, smart chains, smart rings, smart necklaces, smart anklets, smart anklets, etc.), smart wristbands, smart clothing, etc. It should be noted that the specific type of terminal 11 is not limited in this embodiment. Network-side equipment 12 may include access network equipment or core network equipment. Access network equipment 12 may also be referred to as radio access network equipment, radio access network (RAN), radio access network function, or radio access network unit. Access network equipment 12 may include base stations, WLAN access points, or WiFi nodes, etc. Base stations may be referred to as Node B, evolved Node B (eNB), access point, base transceiver station (BTS), radio base station, radio transceiver, Basic Service Set (BSS), Extended Service Set (ESS), home B node, home evolved B node, Transmitting Receiving Point (TRP), or any other suitable term in the field, 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 this application embodiment, only a base station in an NR system is used as an example for description, and the specific type of base station is not limited.

[0039] The information transmission method, apparatus, and communication equipment provided in this application will be described in detail below with reference to the accompanying drawings and through some embodiments and application scenarios.

[0040] Please see Figure 2 , Figure 2 This is a schematic diagram illustrating the steps of an information transmission method provided in an embodiment of this application. The information transmission method includes:

[0041] Step 201: The first communication device sends first information about the artificial intelligence (AI) model to the second communication device; wherein the first information includes at least one of the following:

[0042] Information on the computation time of AI models;

[0043] Whether the computation time of the AI ​​model meets the preset requirements; Optionally, send an instruction if the computation time of the AI ​​model meets the preset requirements, or send an instruction if the computation time of the AI ​​model does not meet the preset requirements, or send an instruction regardless of whether the AI ​​model meets the preset requirements.

[0044] Complexity information of AI models;

[0045] The maximum complexity of an AI model allowed while meeting timeline requirements.

[0046] As an optional embodiment, the first communication device can be a network-side device or a terminal, and the second communication device can also be a network-side device or a terminal. For example, the first communication device is a network-side device and the second communication device is a terminal; or, for another example, the first communication device is a terminal and the second communication device is a network-side device.

[0047] Optionally, the AI ​​model mentioned above can be an AI model used by the first communication device or an AI model currently used by the first communication device.

[0048] In at least one embodiment of this application, the computation time of the AI ​​model includes at least one of the following:

[0049] The initial computation time of the AI ​​model within the AI ​​computing unit can also be referred to as the computation time of the AI ​​model itself; and,

[0050] Data transmission time refers to the time it takes for the relevant data processed by the AI ​​model to be transmitted.

[0051] It should be noted that the aforementioned AI computing unit is used for the operation, reasoning, running, or calculation of the AI ​​model, and can be a unit, module, chip, or hardware, without specific limitations.

[0052] As an optional embodiment, the data transmission time includes at least one of the following:

[0053] The moment when input data arrives at the input interface of the AI ​​computing unit from the first module;

[0054] The second time the AI ​​computing unit reads data from the input interface;

[0055] The AI ​​computing unit outputs data to the third time at the output interface;

[0056] The fourth time from the output interface to the second module;

[0057] The second module reads data from the input interface of the second module at the fifth time.

[0058] The first module described above is the source module for input data, while the second module described above is the destination module for output data. The first module can be an AI computing unit or a non-AI element unit, or the second module can be an AI computing unit or a non-AI computing unit; no specific limitation is made here.

[0059] Optionally, the non-AI computing unit can be a unit, module, chip, or hardware used for non-AI functions.

[0060] In at least one optional embodiment of this application, the complexity of the AI ​​model includes any one of the following:

[0061] Total number of computations performed by the AI ​​model;

[0062] Total number of computations for the AI ​​model;

[0063] The total number of operations (OPs) for the AI ​​model.

[0064] When the complexity of the AI ​​model is determined by the first communication device, the method further includes at least one of the following:

[0065] Based on operations (OPs), all operations of the AI ​​model are converted into OP counts;

[0066] Based on multiplication or addition, all operations of the AI ​​model are converted into the number of multiplications or additions;

[0067] The nonlinear function of the AI ​​model is converted into the number of operations (OPs); for example, the common activation function ReLU is converted into 3 OPs.

[0068] The nonlinear function of the AI ​​model is converted into the number of multiplications or additions.

[0069] The phrase "based on operations (OPs), all operations of the AI ​​model are converted into OP counts" can be understood as follows: generally, one multiplication or one addition is counted as one OP. Of course, it can also be agreed that one multiplication or addition is multiple OPs. Currently, network processors (NPUs) generally use FLOPS (floating-point operations per second) to represent the computing power of network processors (NPUs) or graphics processing units (GPUs). Floating-point operations include all operations involving decimals and are more time-consuming than integer operations. TFLOPS (Tops) is commonly used as a metric to measure NPU / GPU performance / computing power; for example, the HiSilicon 3519AV100 chip has a computing power of 1.7 Tops.

[0070] The above statement, "based on multiplication or addition, all operations of the AI ​​model are converted into the number of multiplications or additions," can be understood as follows: Although current methods for measuring AI model complexity and NPU / GPU capabilities consider one multiplication and one addition to have the same complexity, in reality, due to factors such as chip implementation, the complexity of multiplication and addition may differ. For example, let one addition be denoted as 1 operation (OP), while multiplication is denoted as M OPs (M is a positive real number, such as M = 1.2).

[0071] Optionally, the number of operation operations (OPs), multiplications, or additions required for different nonlinear functions can vary. For example, ReLU requires fewer OPs than Sigmoid.

[0072] Optionally, different hardware may require different operation operations (OPs), multiplications, or additions for the same nonlinear function.

[0073] In at least one embodiment of this application, the first communication device may include one AI model or multiple AI models. If the first communication device includes one AI model, and the AI ​​model has a computation time or a complexity, then the computation time or complexity of the AI ​​model in the first information sent by the first communication device is the computation time or complexity corresponding to the AI ​​model.

[0074] In another optional embodiment of this application, the first communication device may include multiple computation times or multiple complexities. Scenarios in which the first communication device includes multiple computation times or multiple complexities include:

[0075] Scenario 1: The first communication device includes a single AI model, which has multiple computation times. For example, if the first communication device has multiple AI computing units, then the single AI model has multiple computation times. Alternatively, if the first communication device needs to simplify the AI ​​model, then the single AI model has multiple computation times. Simplifying the AI ​​model can be understood as using methods such as knowledge distillation, pruning, low-rank approximation, and quantization to simplify and reduce the size of the AI ​​model while maintaining essentially the same performance. Of course, the AI ​​model can also be made more complex to improve performance.

[0076] Scenario 2: The first communication device includes multiple AI models with the same function, and these multiple AI models have multiple computation times; for example, for a certain function or communication module, the second communication device sends out multiple AI models, or the protocol stipulates multiple AI models, or the first communication device has multiple AI models of its own, or the number of AI models sent by the second communication device, the protocol stipulates, and the first communication device has more than 1.

[0077] Scenario 3: The first communication device includes a single AI model, which has multiple levels of complexity. For example, if the first communication device has multiple AI computing units, then the single AI model has multiple levels of complexity; or, for example, if the first communication device needs to simplify the AI ​​model, then the single AI model has multiple levels of complexity. Simplifying the AI ​​model can be understood as using methods such as knowledge distillation, pruning, low-rank approximation, and quantization to simplify and reduce the size of the AI ​​model while maintaining essentially the same performance; of course, the AI ​​model can also be made more complex to improve performance.

[0078] Scenario 4: The first communication device includes multiple AI models with the same function, and these multiple AI models have multiple levels of complexity; for example, for a certain function or communication module, the second communication device sends out multiple AI models, or the protocol stipulates multiple AI models, or the first communication device has multiple AI models of its own, or the number of AI models sent by the second communication device, the protocol stipulates, and the first communication device has more than 1.

[0079] Scenario 5: The first communication device includes multiple AI models belonging to different functions, and these multiple AI models have multiple computation times;

[0080] Scenario 6: The first communication device includes multiple AI models belonging to different functions, and these multiple AI models have multiple levels of complexity.

[0081] As an optional embodiment, for both scenario 1 and scenario 2, the computation time information of the AI ​​model in the first information includes any one of the following:

[0082] All computation time;

[0083] The K1 computation times that are the longest among multiple computation times;

[0084] The shortest K2 computation times among multiple computation times;

[0085] The longest computation time among multiple operations is the K3 computation time, and the shortest is the K4 computation time;

[0086] The average of all computation times;

[0087] The average of the K5 longest computation times among multiple computation times;

[0088] The average of the K6 shortest computation times among multiple computation times;

[0089] The average of the longest K7 computation times and the shortest K8 computation times;

[0090] The first communication device selects K9 computation time units;

[0091] Where K1, K2, K3, K4, K5, K6, K7, K8 and K9 are all integers greater than or equal to 1.

[0092] As another optional embodiment, for scenarios 3 and 4, the complexity information of the AI ​​model in the first information includes any one of the following:

[0093] All complexities;

[0094] The L1 complexity is the most complex among multiple complexity options;

[0095] The L2 complexity is the lowest among multiple complexity options;

[0096] Among multiple complexity levels, the highest complexity is L3 and the lowest complexity is L4;

[0097] The average of all complexities;

[0098] The average of the L5 complexities, which has the highest complexity among multiple complexity options;

[0099] The average of the L6 least complex complexityes among multiple complexity options;

[0100] The average of the L7 complexity (the highest complexity) and the L8 complexity (the lowest complexity) among multiple complexity options;

[0101] The first communication device is selected with L9 complexity;

[0102] Among them, L1, L2, L3, L4, L5, L6, L7, L8 and L9 are all integers greater than or equal to 1.

[0103] As another optional embodiment, for scenarios 5 and 6, where the first communication device includes multiple AI models belonging to different functions...

[0104] The computation time information of the AI ​​model in the first information includes: the joint computation time of multiple AI models, wherein the joint computation time includes: the sum of the computation times of multiple AI models; wherein, the computation time of an AI model can be understood as the individual computation time of that AI model;

[0105] The complexity information of the AI ​​model in the first information includes: the joint complexity of multiple AI models, and the joint complexity includes: the sum of the complexities of multiple AI models.

[0106] In at least one embodiment of this application, when the joint operation of the plurality of AI models includes non-AI operation units, the joint operation time further includes:

[0107] Running time of non-AI computing units.

[0108] The runtime of the non-AI computing unit includes at least one of the following:

[0109] The sixth time from the input data in the third module to the input interface of the non-AI computing unit;

[0110] The seventh time when the non-AI computing unit reads data from the input interface;

[0111] The non-AI computing unit outputs data to the output interface at the eighth time;

[0112] The output data is routed from the output interface to the ninth time of the fourth module;

[0113] The fourth module reads data from its input interface at the tenth time.

[0114] In at least one optional embodiment of this application, when the number of AI computing units corresponding to multiple AI models is less than the number of AI models, the joint computing time is: the sum of the computing times of the multiple AI models minus the data transmission time of at least one AI model.

[0115] Optionally, "the sum of the computation time of multiple AI models" can be understood as the sum of the individual computation time of each AI model.

[0116] For example, when N1 AI models correspond to N2 AI computing units, when N1>N2, the overall data transmission time can be reduced, and the joint computing time can be reduced, such as reducing the data transmission time of (N1-N2) single AI models.

[0117] In an optional embodiment of this application, the method further includes:

[0118] The first communication device determines the AI ​​model based on the AI ​​model instruction information sent by the second communication device;

[0119] or,

[0120] The first communication device determines the AI ​​model according to the agreement;

[0121] or,

[0122] The first communication device determines the AI ​​model based on its own customization.

[0123] In another optional embodiment of this application, the method further includes:

[0124] The first communication device determines the timeline based on the timeline indication information sent by the second communication device;

[0125] or,

[0126] The first communication device determines the time axis according to the protocol agreement;

[0127] or,

[0128] The first communication device determines the time axis according to its own settings.

[0129] In other words, the AI ​​model or timeline on the first communication device side comes from: the second communication device sending it, the protocol default, or the first communication device's own or custom data.

[0130] In summary, in this embodiment, the first communication device sends computation time-related information and / or complexity-related information of the AI ​​model to the second communication device, enabling the first and second communication devices to align computation time and complexity to determine a suitable AI model, thereby improving the performance of the AI ​​model in the communication system and further enhancing system performance.

[0131] The information transmission method provided in this application can be executed by an information transmission device. This application uses an information transmission device executing the information transmission method as an example to illustrate the information transmission device provided in this application.

[0132] Please see Figure 3 , Figure 3 This is a schematic diagram of the structure of the information transmission device 300 provided in the embodiments of this application. The information transmission device 300 is applied to a first communication device and includes:

[0133] The sending module 301 is configured to send first information of an artificial intelligence (AI) model to a second communication device; wherein the first information includes at least one of the following:

[0134] Information on the computation time of AI models;

[0135] Does the computation time of the AI ​​model meet the preset requirements?

[0136] Complexity information of AI models;

[0137] The maximum complexity of an AI model allowed while meeting timeline requirements.

[0138] As an optional embodiment, the computation time of the AI ​​model includes at least one of the following:

[0139] The first computation time of the AI ​​model in the AI ​​computing unit; and

[0140] Data transmission time.

[0141] As an optional embodiment, the data transmission time includes at least one of the following:

[0142] The moment when input data arrives at the input interface of the AI ​​computing unit from the first module;

[0143] The second time the AI ​​computing unit reads data from the input interface;

[0144] The AI ​​computing unit outputs data to the third time at the output interface;

[0145] The fourth time from the output interface to the second module;

[0146] The second module reads data from the input interface of the second module at the fifth time.

[0147] As an optional embodiment, the complexity of the AI ​​model includes any of the following:

[0148] Total number of computations performed by the AI ​​model;

[0149] Total number of computations for the AI ​​model;

[0150] Total number of operations (OPs) for the AI ​​model.

[0151] As an optional embodiment, the apparatus further includes:

[0152] The conversion module is used to perform at least one of the following:

[0153] Based on operations (OPs), all operations of the AI ​​model are converted into OP counts;

[0154] Based on multiplication or addition, all operations of the AI ​​model are converted into the number of multiplications or additions;

[0155] Convert the nonlinear function of the AI ​​model into the number of operations (OPs);

[0156] The nonlinear function of the AI ​​model is converted into the number of multiplications or additions.

[0157] As an optional embodiment, when the first communication device includes a single AI model with multiple computation times, or when the first communication device includes multiple AI models with the same function and multiple computation times, the computation time information of the AI ​​model in the first information includes any one of the following:

[0158] All computation time;

[0159] The K1 computation times that are the longest among multiple computation times;

[0160] The shortest K2 computation times among multiple computation times;

[0161] The longest computation time among multiple operations is the K3 computation time, and the shortest is the K4 computation time;

[0162] The average of all computation times;

[0163] The average of the K5 longest computation times among multiple computation times;

[0164] The average of the K6 shortest computation times among multiple computation times;

[0165] The average of the longest K7 computation times and the shortest K8 computation times;

[0166] The first communication device selects K9 computation time units;

[0167] Where K1, K2, K3, K4, K5, K6, K7, K8 and K9 are all integers greater than or equal to 1.

[0168] As an optional embodiment, when the first communication device includes a single AI model with multiple levels of complexity, or when the first communication device includes multiple AI models with the same function and multiple levels of complexity, the complexity information of the AI ​​model in the first information includes any one of the following:

[0169] All complexities;

[0170] The L1 complexity is the most complex among multiple complexity options;

[0171] The L2 complexity is the lowest among multiple complexity options;

[0172] Among multiple complexity levels, the highest complexity is L3 and the lowest complexity is L4;

[0173] The average of all complexities;

[0174] The average of the L5 complexities, which has the highest complexity among multiple complexity options;

[0175] The average of the L6 least complex complexityes among multiple complexity options;

[0176] The average of the L7 complexity (the highest complexity) and the L8 complexity (the lowest complexity) among multiple complexity options;

[0177] The first communication device is selected with L9 complexity;

[0178] Among them, L1, L2, L3, L4, L5, L6, L7, L8 and L9 are all integers greater than or equal to 1.

[0179] As an optional embodiment, if the first communication device includes multiple AI models belonging to different functions,

[0180] The computation time information of the AI ​​model in the first information includes: the joint computation time of multiple AI models, and the joint computation time includes: the sum of the computation times of multiple AI models;

[0181] The complexity information of the AI ​​model in the first information includes: the joint complexity of multiple AI models, and the joint complexity includes: the sum of the complexities of multiple AI models.

[0182] As an optional embodiment, when the joint operation of the multiple AI models includes non-AI operation units, the joint operation time further includes:

[0183] Running time of non-AI computing units.

[0184] As an optional embodiment, the runtime of the non-AI computing unit includes at least one of the following:

[0185] The sixth time from the input data in the third module to the input interface of the non-AI computing unit;

[0186] The seventh time when the non-AI computing unit reads data from the input interface;

[0187] The non-AI computing unit outputs data to the output interface at the eighth time;

[0188] The output data is routed from the output interface to the ninth time of the fourth module;

[0189] The fourth module reads data from its input interface at the tenth time.

[0190] As an optional embodiment, when the number of AI computing units corresponding to multiple AI models is less than the number of AI models, the joint computing time is: the sum of the computing times of the multiple AI models minus the data transmission time of at least one AI model.

[0191] As an optional embodiment, the apparatus further includes:

[0192] The first determining module is used to determine the AI ​​model based on the AI ​​model indication information sent by the second communication device;

[0193] Alternatively, it can be used to determine the AI ​​model according to the agreement.

[0194] Alternatively, it can be used to determine the AI ​​model based on the customization of the first communication device.

[0195] As an optional embodiment, the apparatus further includes:

[0196] The second determining module is used to determine the time axis based on the time axis indication information sent by the second communication device;

[0197] Alternatively, it can be used to determine the timeline according to the agreement.

[0198] Alternatively, it can be used to determine the time axis based on the customization of the first communication device.

[0199] In this embodiment, the first communication device sends computation time-related information and / or complexity-related information of the AI ​​model to the second communication device, enabling the first and second communication devices to align computation time and complexity to determine a suitable AI model, thereby improving the performance of the AI ​​model in the communication system and further enhancing system performance.

[0200] It should be noted that the information transmission device provided in this application embodiment is a device capable of executing the above information transmission method. Therefore, all embodiments of the above information transmission method are applicable to this device and can achieve the same or similar beneficial effects.

[0201] The information transmission device in this application embodiment can 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 can be a terminal, or other devices besides a terminal. For example, the terminal can include, but is not limited to, the type of terminal 11 listed above; other devices can be servers, network attached storage (NAS), etc., and this application embodiment does not specifically limit the type.

[0202] The information transmission device provided in this application embodiment can achieve... Figures 1 to 2 The various processes implemented in the method embodiments achieve the same technical effect, and will not be described again here to avoid repetition.

[0203] Optional, such as Figure 4 As shown in the illustration, this application also provides a communication device 400, which is a first communication device including a processor 401 and a memory 402. The memory 402 stores programs or instructions that can run on the processor 401. For example, when the communication device 400 is a terminal, the program or instructions executed by the processor 401 implement the various steps of the above-described information transmission method embodiments and achieve the same technical effect. When the communication device 400 is a network-side device, the program or instructions executed by the processor 401 implement the various steps of the above-described information transmission method embodiments and achieve the same technical effect. To avoid repetition, further details are omitted here.

[0204] This application embodiment also provides a terminal, including a processor and a communication interface. The communication interface is used to send first information of an artificial intelligence (AI) model to a second communication device. The first information includes at least one of the following: computation time information of the AI ​​model; whether the computation time of the AI ​​model meets preset requirements; complexity information of the AI ​​model; and the maximum allowable complexity of the AI ​​model while meeting the timeline requirements. This terminal embodiment corresponds to the above-described terminal-side method embodiment. All implementation processes and methods of the above method embodiments can be applied to this terminal embodiment and achieve the same technical effect. Specifically, Figure 5 A schematic diagram of the hardware structure of a terminal to implement an embodiment of this application.

[0205] The terminal 500 includes, but is not limited to, at least some of the following components: radio frequency unit 501, network module 502, audio output unit 503, input unit 504, sensor 505, display unit 506, user input unit 507, interface unit 508, memory 509, and processor 510.

[0206] Those skilled in the art will understand that the terminal 500 may also include a power supply (such as a battery) for supplying power to various components. The power supply may be logically connected to the processor 510 through a power management system, thereby enabling functions such as managing charging, discharging, and power consumption through the power management system. Figure 5 The terminal structure shown does not constitute a limitation on the terminal. The terminal may include more or fewer components than shown, or combine certain components, or have different component arrangements, which will not be elaborated here.

[0207] It should be understood that, in this embodiment, the input unit 504 may include a graphics processing unit (GPU) 5041 and a microphone 5042. The GPU 5041 processes image data of still images or videos obtained by an image capture device (such as a camera) in video capture mode or image capture mode. The display unit 506 may include a display panel 5061, which may be configured in the form of a liquid crystal display, an organic light-emitting diode, or the like. The user input unit 507 includes at least one of a touch panel 5071 and other input devices 5072. The touch panel 5071 is also called a touch screen. The touch panel 5071 may include a touch detection device and a touch controller. Other input devices 5072 may include, but are not limited to, physical keyboards, function keys (such as volume control buttons, power buttons, etc.), trackballs, mice, and joysticks, which will not be described in detail here.

[0208] In this embodiment, after receiving downlink data from the network-side device, the radio frequency unit 501 can transmit it to the processor 510 for processing; in addition, the radio frequency unit 501 can send uplink data to the network-side device. Typically, the radio frequency unit 501 includes, but is not limited to, antennas, amplifiers, transceivers, couplers, low-noise amplifiers, duplexers, etc.

[0209] The memory 509 can be used to store software programs or instructions, as well as various data. The memory 509 may primarily include a first storage area for storing programs or instructions and a second storage area for storing data. The first storage area may store the operating system, application programs or instructions required for at least one function (such as sound playback, image playback, etc.). Furthermore, the memory 509 may include volatile memory or non-volatile memory, or both. The non-volatile memory may be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory can be random access memory (RAM), static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct memory bus RAM (DRRAM). The memory 509 in this embodiment includes, but is not limited to, these and any other suitable types of memory.

[0210] Processor 510 may include one or more processing units; optionally, processor 510 integrates an application processor and a modem processor, wherein the application processor mainly handles operations involving the operating system, user interface, and applications, and the modem processor mainly handles wireless communication signals, such as a baseband processor. It is understood that the aforementioned modem processor may also not be integrated into processor 510.

[0211] The radio frequency unit 501 is used to send first information of an artificial intelligence (AI) model to the second communication device; wherein the first information includes at least one of the following:

[0212] Information on the computation time of AI models;

[0213] Does the computation time of the AI ​​model meet the preset requirements?

[0214] Information on the complexity of AI models;

[0215] The maximum complexity of an AI model allowed while meeting timeline requirements.

[0216] In this embodiment, the terminal sends computation time-related information and / or complexity-related information of the AI ​​model to the second communication device, enabling the terminal and the second communication device to align computation time and complexity to determine a suitable AI model, thereby improving the performance of the AI ​​model in the communication system and further enhancing system performance.

[0217] It should be noted that the information transmission device provided in this application embodiment is a device capable of executing the above information transmission method. Therefore, all embodiments of the above information transmission method are applicable to this device and can achieve the same or similar beneficial effects.

[0218] This application also provides a network-side device, including a processor and a communication interface. The communication interface is used to send first information of an artificial intelligence (AI) model to a second communication device. The first information includes at least one of the following: computation time information of the AI ​​model; whether the computation time of the AI ​​model meets preset requirements; complexity information of the AI ​​model; and the maximum allowable complexity of the AI ​​model while meeting the timeline requirements. This network-side device embodiment corresponds to the above-described network-side device method embodiment. All implementation processes and methods of the above method embodiments can be applied to this network-side device embodiment and achieve the same technical effects.

[0219] Specifically, embodiments of this application also provide a network-side device. For example... Figure 6 As shown, the network-side device 600 includes: an antenna 61, a radio frequency (RF) device 62, a baseband device 63, a processor 64, and a memory 65. The antenna 61 is connected to the RF device 62. In the uplink direction, the RF device 62 receives information through the antenna 61 and transmits the received information to the baseband device 63 for processing. In the downlink direction, the baseband device 63 processes the information to be transmitted and sends it to the RF device 62. The RF device 62 processes the received information and transmits it through the antenna 61.

[0220] The method executed by the network-side device in the above embodiments can be implemented in the baseband device 63, which includes a baseband processor.

[0221] Baseband device 63 may include, for example, at least one baseband board on which multiple chips are disposed, such as Figure 6 As shown, one of the chips is, for example, a baseband processor, which is connected to the memory 65 via a bus interface to call the program in the memory 65 and execute the network device operation shown in the above method embodiment.

[0222] The network-side device may also include a network interface 66, such as a common public radio interface (CPRI).

[0223] Specifically, the network-side device 600 of this embodiment further includes: instructions or programs stored in a memory 65 and executable on a processor 64, wherein the processor 64 calls the instructions or programs in the memory 65 to execute. Figure 3 The methods executed by each module shown achieve the same technical effect, and to avoid repetition, they will not be described in detail here.

[0224] This application also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various processes of the above-described information transmission method embodiments and achieve the same technical effects. To avoid repetition, they will not be described again here.

[0225] The processor is the processor in the terminal described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0226] This application embodiment also provides a chip, which 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-described information transmission method embodiments and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0227] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.

[0228] This application also provides a computer program / program product, which is stored in a storage medium and executed by at least one processor to implement the various processes of the above-described information transmission method embodiments, and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0229] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0230] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a computer software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0231] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. An information transmission method, characterized in that, include: A first communication device sends first information about an artificial intelligence (AI) model to a second communication device; wherein the first information includes: the maximum complexity of the AI ​​model allowed under the premise of meeting timeline requirements; the first information is used by the first communication device and the second communication device to align the complexity of the AI ​​model; the AI ​​model is the AI ​​model used by the first communication device. Wherein, when the first communication device includes multiple AI models belonging to different functions, the complexity of the AI ​​model includes: the sum of the complexities of the multiple AI models; The time axis is determined by at least one of the following methods: The first communication device determines the time axis based on the time axis indication information sent by the second communication device; or, The first communication device determines the time axis according to the protocol agreement; The first communication device is a terminal, and the second communication device is a network-side device; The complexity of the AI ​​model includes any one of the following: Total number of computations performed by the AI ​​model; Total number of operations (OPs) for the AI ​​model; The method further includes at least one of the following: Based on operations (OPs), all operations of the AI ​​model are converted into OP counts; Based on multiplication or addition, all operations of the AI ​​model are converted into the number of multiplications or additions; Convert the nonlinear function of the AI ​​model into the number of operations (OPs); Convert the nonlinear function of the AI ​​model into the number of multiplications or additions; One multiplication or one addition is recorded as one operation; or, one multiplication or one addition is recorded as multiple operations.

2. The method according to claim 1, characterized in that, The method further includes: The first communication device determines the AI ​​model based on the AI ​​model instruction information sent by the second communication device; or, The first communication device determines the AI ​​model according to the agreement.

3. An information transmission device, applied to a first communication device, characterized in that, include: A sending module is used to send first information of an artificial intelligence (AI) model to a second communication device; wherein, the first information includes: the maximum complexity of the AI ​​model allowed under the premise of meeting timeline requirements; the first information is used by the first communication device and the second communication device to align the complexity of the AI ​​model; the AI ​​model is the AI ​​model used by the first communication device; Wherein, when the first communication device includes multiple AI models belonging to different functions, the complexity of the AI ​​model includes: the sum of the complexities of the multiple AI models; The device further includes: The second determining module is used to determine the time axis based on the time axis indication information sent by the second communication device; Alternatively, it can be used to determine the timeline according to the agreement. The first communication device is a terminal, and the second communication device is a network-side device; The complexity of the AI ​​model includes any one of the following: Total number of computations performed by the AI ​​model; Total number of operations (OPs) for the AI ​​model; The device further includes: The conversion module is used to perform at least one of the following: Based on operations (OPs), all operations of the AI ​​model are converted into OP counts; Based on multiplication or addition, all operations of the AI ​​model are converted into the number of multiplications or additions; Convert the nonlinear function of the AI ​​model into the number of operations (OPs); Convert the nonlinear function of the AI ​​model into the number of multiplications or additions; One multiplication or one addition is recorded as one operation; or, one multiplication or one addition is recorded as multiple operations.

4. The apparatus according to claim 3, characterized in that, The device further includes: The first determining module is used to determine the AI ​​model based on the AI ​​model indication information sent by the second communication device; Alternatively, it can be used to determine the AI ​​model according to the agreement.

5. A communication device, characterized in that, It includes a processor and a memory, the memory storing programs or instructions that can run on the processor, the programs or instructions being executed by the processor to implement the steps of the information transmission method as described in claim 1 or 2.

6. A readable storage medium, characterized in that, The readable storage medium stores a program or instructions, which, when executed by a processor, implement the steps of the information transmission method as described in claim 1 or 2.

Images