Communication method, communication apparatus, and communication system

By determining the appropriate codebook based on the distribution characteristics and mapping relationship of the data and quantizing data, the problem of large-scale resource overhead is solved and efficient data transmission is achieved.

WO2024234187A1PCT designated stage expired Publication Date: 2024-11-21HUAWEI TECH CO LTD
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
PCT/CN2023/094063
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Filing Date
2023-05-12
Publication Date
2024-11-21

AI Technical Summary

Technical Problem

When transmitting large-scale data, the prior art is difficult to effectively reduce the overhead of transmission resources, resulting in the problem of resource shortage.

Method used

By determining the appropriate codebook, data quantization is performed to reduce the overhead of transmission resources. The specific method includes determining the codebook set based on the distribution characteristics and mapping relationship of the data, and selecting a suitable codebook from it, ensuring that the transmission overhead and quantization error are within the threshold, and meeting the requirements of quantization performance and modulation coding schemes.

Benefits of technology

It effectively reduces resource overhead during data transmission, improves transmission efficiency, and ensures the quantitative performance and accuracy of data.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN2023094063_21112024_PF_FP_ABST
    Figure CN2023094063_21112024_PF_FP_ABST
Patent Text Reader

Abstract

Embodiments of the present application provide a communication method, a communication apparatus, and a communication system. The method comprises: determining a first codebook, the first codebook being used for quantifying first data; and sending first indication information, the first indication information being used for indicating the first codebook. In the present solution, the first data is quantified by means of the codebook, so that transmission resource overhead can be reduced during data transmission.
Need to check novelty before this filing date? Find Prior Art

Description

Communication method, communication device and communication system Technical Field

[0001] The present application relates to the field of communication technology, and in particular to a communication method, a communication device, and a communication system. Background Art

[0002] With the development of communication technology, the demand for large-scale data transmission between devices is increasing. When large-scale data is exchanged between different devices, it will cause huge transmission resource overhead.

[0003] With transmission resources becoming increasingly scarce, how to transmit large amounts of data is an urgent issue to be solved.

[0004] Summary of the Invention

[0005] The present application provides a communication method, a communication device, and a communication system for reducing transmission resource overhead during data transmission.

[0006] In a first aspect, embodiments of the present application provide a communication method, performed by a first apparatus, which may be a terminal device, a network device, or other device, or a module within the terminal device, network device, or other device. The method includes: determining a first codebook, the first codebook being used to quantize first data; and sending first indication information, the first indication information being used to indicate the first codebook.

[0007] In the above solution, the first data is quantized by using a codebook, which can reduce transmission resource overhead during data transmission.

[0008] In a possible implementation method, determining the first codebook includes: determining a first codebook set according to distribution characteristics of the first data, the first codebook set including at least one codebook; and determining the first codebook in the first codebook set.

[0009] The above solution determines the first codebook set in combination with the distribution characteristics of the first data, and determines the first codebook from the first codebook set, which helps to determine a suitable codebook, thereby reducing the overhead of data transmission.

[0010] In a possible implementation method, determining the first codebook in the first codebook set includes: determining the first codebook according to transmission overhead corresponding to a codebook in the first codebook set.

[0011] In the above solution, the first codebook is determined from the first codebook set according to the transmission overhead corresponding to the codebooks in the first codebook set, which helps to determine a suitable codebook, thereby reducing the overhead of data transmission.

[0012] In a possible implementation method, the transmission overhead includes codebook indication overhead and / or data transmission overhead after the first data is quantized.

[0013] In a possible implementation method, determining the first codebook in the first codebook set includes: sending first information, where the first information is used to indicate some or all codebooks in the first codebook set; receiving second information, where the second information is used to indicate some or all codebooks in the codebook indicated by the first information; and determining the first codebook based on the second information.

[0014] In the above solution, the first device and the second device jointly determine the first codebook corresponding to the first data, which can ensure that the determined first codebook can meet the quantization performance and reduce the overhead of data transmission.

[0015] In a possible implementation method, a transmission overhead corresponding to the codebook indicated by the first information is less than a transmission overhead threshold.

[0016] In a possible implementation method, determining the first codebook in the first codebook set includes: receiving third information, where the third information is used to indicate the first codebook.

[0017] In the above solution, the second device determines the first codebook corresponding to the first data, which can ensure that the determined first codebook can meet the quantization performance and reduce the overhead of data transmission.

[0018] In a possible implementation method, a quantization error of a codebook in the first codebook set is less than or equal to a quantization error threshold.

[0019] In the above solution, the quantization error of the codebooks in the first codebook set is less than or equal to the quantization error threshold. Therefore, the first codebook in the determined first codebook set is also less than or equal to the quantization error threshold, which helps to determine a suitable codebook and reduce the overhead of data transmission.

[0020] In a possible implementation method, the quantization error threshold is preconfigured or predefined.

[0021] In a possible implementation method, the quantization error threshold is related to the importance of the first data.

[0022] The above solution determines the quantization error threshold according to the importance of the first data, which helps to determine a suitable codebook for the first data.

[0023] In one possible implementation method, determining the first codebook set based on the distribution characteristics of the first data includes: determining, based on the distribution characteristics of the first data and a mapping relationship between the distribution characteristics and the codebook, the first codebook set corresponding to the distribution characteristics of the first data in the mapping relationship.

[0024] In the above solution, a mapping relationship between distribution characteristics and codebooks is pre-configured. Subsequently, the first codebook corresponding to the first data can be determined based on the mapping relationship and the distribution characteristics of the first data. This can quickly determine the codebook corresponding to the first data and reduce data transmission delay.

[0025] In a possible implementation method, the distribution characteristics of the first data include data distribution slope and / or data expectation.

[0026] In a possible implementation method, the first codebook meets a preset modulation and coding scheme MCS requirement.

[0027] In a possible implementation method, the first data is associated with a first service, a first data radio bearer or a first logical channel.

[0028] In a possible implementation method, the method further includes: quantizing the first data to obtain quantized data; determining a codeword corresponding to the quantized data according to the first codebook; and sending the codeword corresponding to the quantized data.

[0029] In the above solution, the first data is quantized by using a codebook, which can reduce transmission resource overhead during data transmission.

[0030] In a possible implementation method, the method further includes: determining a second codebook, where the second codebook is used to quantize the second data; and sending second indication information, where the second indication information is used to indicate the second codebook.

[0031] In the above solution, the second data is quantized by using a codebook, which can reduce transmission resource overhead during data transmission.

[0032] In one possible implementation method, the first data and the second data correspond to the same artificial intelligence model.

[0033] In a second aspect, embodiments of the present application provide a communication method, performed by a second apparatus or a module within the second apparatus. The second apparatus may be a terminal device, a network device, or other device, or a module within the terminal device, network device, or other device. The method includes: receiving first indication information, the first indication information being used to indicate a first codebook, the first codebook being used to quantize first data; and determining the first codebook based on the first indication information.

[0034] In the above solution, the first data is quantized by using a codebook, which can reduce transmission resource overhead during data transmission.

[0035] In a possible implementation method, the method further includes: receiving a codeword corresponding to quantized data, where the quantized data is obtained by quantizing the first data.

[0036] In the above solution, the first data is quantized by using a codebook, which can reduce transmission resource overhead during data transmission.

[0037] In a possible implementation method, the method further includes: determining the quantized data according to the first codebook and a codeword corresponding to the quantized data.

[0038] In a possible implementation method, the method further includes: receiving first information, where the first information is used to indicate some or all codebooks in a first codebook set; and sending second information, where the second information is used to indicate some or all codebooks in the codebook indicated by the first information, where the codebooks indicated by the second information include the first codebook.

[0039] In the above solution, the first device and the second device jointly determine the first codebook corresponding to the first data, which can ensure that the determined first codebook can meet the quantization performance and reduce the overhead of data transmission.

[0040] In a possible implementation method, a transmission overhead corresponding to the codebook indicated by the first information is less than a transmission overhead threshold.

[0041] In a possible implementation method, the transmission overhead includes codebook indication overhead and / or data transmission overhead after the first data is quantized.

[0042] In a possible implementation method, a quantization error of a codebook in the first codebook set is less than or equal to a quantization error threshold.

[0043] In the above solution, the quantization error of the codebooks in the first codebook set is less than or equal to the quantization error threshold. Therefore, the first codebook in the determined first codebook set is also less than or equal to the quantization error threshold, which helps to determine a suitable codebook and reduce the overhead of data transmission.

[0044] In a possible implementation method, the quantization error threshold is preconfigured or predefined.

[0045] In a possible implementation method, the quantization error threshold is related to the importance of the first data.

[0046] The above solution determines the quantization error threshold according to the importance of the first data, which helps to determine a suitable codebook for the first data.

[0047] In a possible implementation method, the method further includes: sending third information, where the third information is used to indicate the first codebook.

[0048] In the above solution, the second device determines the first codebook corresponding to the first data, which can ensure that the determined first codebook can meet the quantization performance and reduce the overhead of data transmission.

[0049] In a possible implementation method, the first codebook meets a preset modulation and coding scheme MCS requirement.

[0050] In a possible implementation method, the first data is associated with a first service, a first data radio bearer or a first logical channel.

[0051] In a possible implementation method, the method further includes: receiving second indication information, where the second indication information is used to indicate a second codebook, and the second codebook is used to quantize the second data; and determining the second codebook according to the second indication information.

[0052] In the above solution, the second data is quantized by using a codebook, which can reduce transmission resource overhead during data transmission.

[0053] In one possible implementation method, the first data and the second data correspond to the same artificial intelligence model.

[0054] In a third aspect, embodiments of the present application provide a communication device. The device has the function of implementing any of the implementation methods of aspects 1 to 2 above. The function can be implemented by hardware or by hardware executing corresponding software implementations. The hardware or software includes one or more modules corresponding to the above functions.

[0055] In a fourth aspect, an embodiment of the present application provides a communication device, comprising a unit or means for executing each step of any implementation method in the above-mentioned first to second aspects.

[0056] In a fifth aspect, an embodiment of the present application provides a communication device comprising at least one processor, wherein the processor is configured to execute any implementation method in the above-mentioned first to second aspects by at least one of the following: running computer instructions or programs, or logic circuits.

[0057] In one possible implementation, the processor is coupled to a memory that stores the aforementioned computer instructions or programs; alternatively, the memory stores a configuration file for a logic circuit. Alternatively, the memory may be located within the device, i.e., the device includes the memory; alternatively, the processor and memory are integrated together; alternatively, the memory may be located external to the device.

[0058] In a possible implementation, the communication device further includes: an interface circuit, the interface circuit being used to input and / or output signals. The processor is used to communicate with other devices through the interface circuit.

[0059] In a possible implementation, the communication device is a chip.

[0060] In a sixth aspect, an embodiment of the present application further provides a computer program product, which includes a computer program or instructions. When the computer program or instructions are run by a communication device, any implementation method in the above-mentioned first to second aspects is executed.

[0061] In the seventh aspect, an embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions, which, when run on a communication device, enables any implementation method in the above-mentioned first to second aspects to be executed.

[0062] In an eighth aspect, an embodiment of the present application further provides a chip system, comprising: a processor for executing any implementation method of the above-mentioned first to second aspects.

[0063] In a ninth aspect, an embodiment of the present application further provides a communication system, comprising a first device for executing any implementation method of the first aspect, and a second device for executing any implementation method of the second aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] FIG1 is a schematic diagram of the architecture of a communication system used in an embodiment of the present application;

[0065] FIG2 is a flow chart of a communication method provided in an embodiment of the present application;

[0066] FIG3 is an example diagram of a scalar quantization method;

[0067] FIG4 is an example diagram of two-dimensional vector quantization;

[0068] FIG5 is an example diagram of a grouping method;

[0069] FIG6 is an example diagram of another grouping method;

[0070] Figure 7 is a schematic diagram of the CDF distribution slope;

[0071] Figure 8 is a schematic diagram of the slope of the PDF distribution;

[0072] FIG9 is a flow chart of a communication method provided in an embodiment of the present application;

[0073] FIG10 is a flow chart of a communication method provided in an embodiment of the present application;

[0074] FIG11 is a schematic structural diagram of a communication device provided in an embodiment of the present application;

[0075] FIG12 is a schematic diagram of the structure of a communication device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0076] Figure 1 is a schematic diagram of the architecture of a communication system used in an embodiment of the present application. The communication system includes a first device and a second device.

[0077] In one implementation method, the first device is a network device or a module for a network device, and the second device is a terminal device or a module for a terminal device. The first device and the second device communicate with each other via an air interface.

[0078] In another implementation method, the first device is a terminal device or a module for a terminal device, and the second device is a network device or a module for a network device. The first device and the second device communicate with each other via an air interface.

[0079] In another implementation method, the first device is a network device or a module for a network device, and the second device is a network device or a module for a network device. The first device and the second device communicate with each other via an air interface or a wired manner.

[0080] In another implementation method, the first device is a terminal device or a module for a terminal device, and the second device is a terminal device or a module for a terminal device. The first device and the second device communicate with each other via an air interface.

[0081] Of course, the first device and the second device in the embodiment of the present application can also be other types of devices. For example, the first device can also be a cloud device or a cloud server and the second device can be a cloud device or a cloud server. This application does not limit this.

[0082] In the implementation of this application, a terminal device is a device with wireless transceiver capabilities, and may specifically refer to user equipment (UE), access terminal, subscriber unit, user station, mobile station, remote station, remote terminal, mobile device, user terminal, wireless communication device, user agent, or user device. The terminal device can be deployed on land, including indoors or outdoors, handheld or vehicle-mounted; can also be deployed on water (such as ships); and can also be deployed in the air (such as aircraft, balloons, and satellites). The terminal device can be a cellular phone, a mobile phone, a tablet computer (pad), a wireless data card, a wireless modem, a satellite terminal, a vehicle (e.g., a car, a bicycle, an electric car, an airplane, a ship, a train, a high-speed rail, etc.) onboard equipment, a robotic arm, a workshop equipment, a wearable device (e.g., a smart watch, a smart bracelet, a pedometer, etc.), a drone, a robot, a smart point of sale (POS) machine, a customer-premises equipment (CPE), a computer with a wireless transceiver function, a virtual reality (VR) terminal device, an augmented reality (AR) terminal device, a terminal device in industrial control, a terminal device in self-driving, a terminal device in remote medical care, a terminal device in a smart grid, a terminal in transportation safety, a terminal device in a smart city, a terminal in a smart home (e.g., a refrigerator, a television, an air conditioner, an electric meter, and other smart home devices). The terminal device can also be other devices with terminal functions. The embodiments of this application do not limit the device form factor of the terminal. The device used to implement the functions of the terminal device can be the terminal device; it can also be a device that supports the terminal device to implement the functions, such as a chip system. The device can be installed in the terminal device or used in conjunction with the terminal device. In the embodiments of this application, the chip system can be composed of chips or include chips and other discrete devices.

[0083] In the implementation of this application, the network device is a device with wireless transceiver functions, which is used to communicate with the terminal device or other network devices; it can also be a device that can access the terminal device to the wireless network, such as a radio access network (RAN) device or node. The network devices in the embodiments of the present application may include various forms of base stations, such as: base stations, evolved NodeBs (eNodeBs), next generation NodeBs (gNBs), macro base stations, micro base stations (also known as small stations), relay stations, access points, devices that implement base station functions in communication systems evolved after the fifth generation (5G) technology, access points (APs) in wireless local area networks (WLAN) systems, integrated access and backhaul (IAB) nodes, transmission points (TRPs), transmitting points (TPs), mobile switching centers, and devices that perform base station functions in device-to-device (D2D), vehicle-to-everything (V2X), and machine-to-machine (M2M) communications, etc., and may also include network devices in non-terrestrial network (NTN) communication systems, that is, they can be deployed on high-altitude platforms or satellites. In some possible scenarios, different network devices implement part of the functions of the base station respectively. For example, the network device can be a centralized unit (CU), a distributed unit (DU), a CU-control plane (CP), a CU-user plane (UP), or a radio unit (RU). The CU and DU can be set separately, or they can be included in the same network element, such as a baseband unit (BBU). The RU can be included in a radio frequency device or radio frequency unit, such as a remote radio unit (RRU), an active antenna unit (AAU), or a remote radio head (RRH).It is understood that the network device may be a CU node, a DU node, or a device including a CU node and a DU node. In addition, the CU may be classified as a network device in the access network RAN, or the CU may be classified as a network device in the core network (CN), without limitation herein.

[0084] In different systems, CU (or CU-CP and CU-UP), DU or RU may also have different names, but those skilled in the art can understand their meanings. For example, in an open RAN (open RAN, ORAN) system, CU may also be called O-CU (open CU), DU may also be called O-DU, CU-CP may also be called O-CU-CP, CU-UP may also be called O-CU-UP, and RU may also be called O-RU. For the convenience of description, this application uses CU, CU-CP, CU-UP, DU and RU as examples for description. Any unit of CU (or CU-CP, CU-UP), DU and RU in this application can be implemented by a software module, a hardware module, or a combination of a software module and a hardware module.

[0085] Quantization is the process of converting raw data into quantized data. Quantization includes scalar quantization and vector quantization. Scalar quantization can also be considered a one-dimensional vector quantization, meaning each raw data point is independently quantized. Vector quantization groups the raw data according to the dimension d (d greater than 1) of the quantized data point. Each group contains d raw data points, and each group is then quantized individually.

[0086] FIG2 is a flow chart of a communication method provided in an embodiment of the present application. The method includes the following steps:

[0087] In step 201 , a first device determines a first codebook, where the first codebook is used to quantize first data.

[0088] The first codebook includes the number of codewords (represented by k), the dimension of the quantized data (represented by d), and the mapping relationship between the codewords and the quantized data. The number of codewords is also referred to as the codebook size, and the dimension of the quantized data is also referred to as the codebook dimension. The quantized data mentioned in this application refers to the quantized data.

[0089] In one implementation method, the first codebook corresponding to the first data determined by the first device meets the preset modulation and coding scheme (MCS) requirement. The MCS requirement here can be determined by the MCS and the transmission resources corresponding to the first data. Specifically, the code rate can be determined by the MCS, and then the maximum transmission overhead corresponding to the first data can be determined based on the code rate and the transmission resources corresponding to the first data, such as the maximum number of bits that can be transmitted on the transmission resources corresponding to the first data. Or the MCS requirement directly corresponds to the maximum transmission overhead corresponding to the first data. The first codebook corresponding to the first data meets the preset MCS requirement, which means that the transmission resources corresponding to the first data can carry the first codebook and / or use the first codebook to quantize the first data to obtain quantized data.

[0090] Figure 3 shows an example of a scalar quantization scheme. In this example, raw data in the range {-1, 1} is mapped to 0, corresponding to the codeword 00; raw data in the range {1, 3} is mapped to 2, corresponding to the codeword 01; raw data in the range {3, 5} is mapped to 4, corresponding to the codeword 10; and raw data in the range {5, 7} is mapped to 6, corresponding to the codeword 11. Table 1 shows an example of a codebook.

[0091] Table 1

[0092] Assuming that the first data includes 0.1, 1.5, 2.8, 5.8, 6.9, and 4.8, if the first device determines that the first codebook corresponding to the first data is as shown in Table 1, then after mapping the first data to quantized data, the codebook shown in Table 1 is used to determine the codeword corresponding to each quantized data. Ultimately, the first device sends the codeword to the second device instead of the first data. Specifically, according to the example of Figure 3 and Table 1, 0.1 is mapped to 0, and the corresponding codeword is 00; 1.5 is mapped to 2, and the corresponding codeword is 01; 2.8 is mapped to 2, and the corresponding codeword is 01; 5.8 is mapped to 6, and the corresponding codeword is 11; 6.9 is mapped to 6, and the corresponding codeword is 11; 4.8 is mapped to 4, and the corresponding codeword is 10.

[0093] If the first device sends the first data to the second device, to ensure the accuracy of data expression, each data generally occupies 32 bits, so transmitting 0.1, 1.5, 2.8, 5.8, 6.9, 4.8 requires a total of 32*6=192 bits.

[0094] If the above quantization method is used, the codewords sent by the first device to the second device include 00, 01, 01, 11, 11, 10, which occupy a total of 2*6=12 bits. Therefore, compared with directly sending the first data, the method of sending the codeword corresponding to the quantized data can reduce transmission overhead.

[0095] After receiving the codeword sent by the first device, the second device can map the received codeword into quantized data according to the first codebook. That is, the second device determines that the quantized data corresponding to the first data includes 0, 2, 2, 6, 6, and 4. This quantization method loses some data accuracy but can reduce data transmission overhead.

[0096] In vector quantization, assume that the number of codewords included in the codebook is k, and the dimension of each codeword is d. Each group of data is mapped to one of the k codewords. For example, groups 1 to 10 are mapped to codeword 1, groups 11 to 20 are mapped to codeword 2, and so on. The number of bits required for k codewords is log2(k), so log2(k) bits can be used to represent a group of data. Figure 4 shows an example of two-dimensional vector quantization. The original data is grouped according to dimension d = 2, resulting in multiple two-dimensional data. Each two-dimensional data set includes two original data. Each black dot in Figure 4 represents a two-dimensional data set, and each star in Figure 4 represents a codeword. There are a total of k codewords (i.e., k star points). According to the quantization algorithm, the region consisting of multiple two-dimensional data (i.e., all black dots in Figure 4) can be divided into multiple subregions. The two-dimensional data in each subregion is mapped to the same codeword. Referring to Figure 4, all black dots in each subregion are mapped to the star points in that subregion. Table 2 is another example of a codebook. In this example, k=256, each codeword occupies 8 bits, and d=16, that is, every 16 original data are divided into one group.

[0097] Table 2

[0098] The first data is grouped into groups of 16 data to obtain multiple groups of data. If the first device determines that the first codebook corresponding to the first data is as shown in Table 2, then after mapping the first data to quantized data, each quantized data is a group of data containing 16 original data. Then, the codebook shown in Table 2 is used to determine the codeword corresponding to each quantized data. Ultimately, multiple codewords are sent to the second device, each codeword occupying 8 bits.

[0099] Taking the first data including 1600 original data as an example, if the first device sends the first data to the second device, in order to ensure the accuracy of data expression, each data generally occupies 32 bits, then a total of 32*1600=51200 bits are required to transmit the 1600 original data.

[0100] If the above quantization method is used, the 1600 original data are divided into 100 groups, each group containing 16 original data. The quantized data corresponding to each group is then determined, resulting in 100 quantized data. A codeword corresponding to each quantized data is then determined. The codeword sent by the first device to the second device occupies a total of 8 * 100 = 800 bits. Therefore, compared to directly sending the first data, this method of sending the codeword corresponding to the quantized data can reduce transmission overhead.

[0101] After receiving the codeword sent by the first device, the second device can map the received codeword into quantized data according to the first codebook. This quantization method loses a certain amount of data accuracy but can reduce data transmission overhead.

[0102] The following describes two different grouping methods for vector quantization, assuming that the number of original data in the first data is N and the dimension of the quantized data is d.

[0103] Method 1: According to the dimension d of the quantized data, N original data in the first data are sequentially divided into m groups, where m=N / d, and each group includes d original data.

[0104] Figure 5 is an example diagram of a grouping method. It can be seen that the first group of data includes the 1st to dth original data in the first data, the second group of data includes the d+1th to 2dth original data in the first data, and so on.

[0105] Method 2: According to the dimension d of the quantized data, the N original data in the first data are divided into m groups by interleaving, each group includes d original data, and the interleaving depth is m=N / d.

[0106] Figure 6 is an example diagram of another grouping method. Data is read in by column and read out by row. By interleaving, the distribution of the original data can be transformed, and the distribution characteristics required by vector quantization can be obtained. Specifically, the source of gain of vector quantization compared to scalar quantization mainly depends on the distribution characteristics of the data. When the distribution of data is more concentrated, vector quantization can bring smaller transmission overhead. It can be seen that the first group of data includes the first original data in the first data, the m+1th original data, the 2m+1th original data, ..., the (d-1)m+1th original data, the second group of data includes the second original data in the first data, the m+2th original data, the 2m+2th original data, ..., the (d-1)m+2th original data, and so on.

[0107] For the above method 1 and method 2, when grouping the original data in the first data, if N is not an integer multiple of d, any of the following processing methods can be used:

[0108] Processing method 1: The first device discards the last group of data, and the number of original data in this group of data is less than d. The second device automatically fills the corresponding number of preset values, such as zero, based on the number N of original data in the first data and the dimension d of the quantized data.

[0109] Processing method 2: The first device supplements the last group of data with a corresponding number of preset values, such as zeros, so that the number of original data in the group of data is equal to d. The second device deletes the redundant data in the last group of data based on the number N of original data in the first data and the dimension d of the quantized data.

[0110] In step 202, the first device sends first indication information to the second device. Correspondingly, the second device receives the first indication information.

[0111] The first indication information is used to indicate a first codebook.

[0112] In one implementation method, if the first device has multiple codebooks and an index for each codebook, or if the first device and the second device have multiple codebooks and an index for each codebook, and the first codebook is one of the multiple codebooks, then the first indication information may be the index of the first codebook. Exemplarily, the multiple codebooks and the index for each codebook on the first device and the second device may be predefined by a protocol or preconfigured, and this application does not limit this.

[0113] In another implementation method, the first indication information includes a first codebook, that is, the first indication information includes the number of codewords, the dimension of the quantized data, and a mapping relationship, where the mapping relationship is a mapping relationship between the codewords and the quantized data. Exemplarily, the first codebook can be as shown in Table 1 or Table 2.

[0114] Step 203: The second device determines a first codebook according to the first indication information.

[0115] In one implementation method, the first indication information is an index of a first codebook, and the second device obtains the first codebook from a plurality of codebooks predefined or preconfigured on the second device according to the index.

[0116] In another implementation method, the first indication information includes a first codebook, and the second device obtains the first codebook from the first indication information.

[0117] In the above solution, the first data is quantized by using a codebook, which can reduce transmission resource overhead during data transmission.

[0118] In one implementation method, the first device further quantizes the first data to obtain quantized data, then determines a codeword corresponding to the quantized data based on a first codebook, and sends the codeword corresponding to the quantized data to the second device. The codeword corresponding to the quantized data can be sent simultaneously with the aforementioned first indication information, or separately. After receiving the codeword corresponding to the quantized data, the second device determines the quantized data based on the first codebook and the codeword corresponding to the quantized data. That is, using the first codebook, the received codeword is mapped to quantized data. The quantized data is the quantized data corresponding to the first data.

[0119] In one implementation method, the first data is associated with the first service, that is, the first data is data of the first service.

[0120] In one implementation method, the first data is associated with a first data radio bearer, that is, the codeword corresponding to the quantized data of the first data is sent to the second device through the first data radio bearer.

[0121] In one implementation method, the first data is associated with the first logical channel, that is, the codeword corresponding to the quantized data of the first data is sent to the second device through the first logical channel.

[0122] In one implementation method, the first device determines the first codebook corresponding to the first data in step 201. Specifically, the first device may determine a first codebook set based on the distribution characteristics of the first data, where the first codebook set includes at least one codebook. The first device then determines a first codebook in the first codebook set, i.e., selects the first codebook from the first codebook set. This method first determines a codebook set and then selects the first codebook corresponding to the first data from the codebook set, which facilitates accurate determination of the first codebook corresponding to the first data.

[0123] In one implementation method, the quantization error of a codebook in the first codebook set is less than or equal to a quantization error threshold, and the quantization error threshold is preconfigured or predefined. Exemplarily, the quantization error threshold is related to the importance of the first data (such as packet loss rate, latency). For example, when the packet loss rate corresponding to the first data is lower than the first threshold, the quantization error threshold is lower than the second threshold; when the packet loss rate corresponding to the first data is lower than the third threshold, the quantization error threshold is lower than the fourth threshold, where the third threshold is lower than the first threshold, and the fourth threshold is lower than the second threshold.

[0124] In one implementation method, a first device determines a first codebook set based on distribution characteristics of first data. Specifically, the method includes: the first device determines, based on the distribution characteristics of the first data and a mapping relationship between the distribution characteristics and the codebooks, a first codebook set corresponding to the distribution characteristics of the first data in a mapping relationship. The mapping relationship is preconfigured or predefined. By determining the first codebook corresponding to the first data from the preconfigured or predefined mapping relationship, the method can quickly determine the first codebook.

[0125] The distribution characteristics of the first data include the data distribution slope of the first data (expressed by ρ) and / or the data expectation of the first data (expressed by ε). The data distribution slope can be the distribution slope of the cumulative distribution function (CDF) (expressed by ρ). CDF denoted by ρ) or the probability density function (PDF) distribution slope (denoted by ρ PDF express).

[0126] Figure 7 is a schematic diagram of the CDF distribution slope. Point A in the figure is the highest reference probability C h The intersection point with the CDF curve of the first data, point B is the lowest reference probability C l The intersection of the CDF curve of the first data and point A is connected with point B to obtain the CDF distribution slope of the first data, that is, ρ CDF Among them, C h is a number close to 1, C l is a number close to 0.

[0127] Figure 8 is a schematic diagram of the PDF distribution slope. Point A in the figure is the highest absolute probability P h is the peak of the PDF curve of the first data, and point B is the lowest reference probability P l The first intersection point with the PDF curve of the first data, point C, is the lowest reference probability P l The last intersection point with the PDF curve of the first data. Connect point A and point B to get the first CDF distribution slope of the first data, that is, ρ PDF_l Connecting point A and point C, we can get the second CDF distribution slope of the first data, which is ρ PDF_r Among them, P l is a number close to 0. The PDF distribution slope of the first data is ρ PDF =min(|ρ PDF_l |,|ρ PDF_r |), i.e. ρ PDF_r The absolute value of ρ PDF_l The smaller of the absolute values ​​of is used as the CDF distribution slope of the first data.

[0128] The distribution characteristics of the first data will affect the quantization performance, which can be specifically measured by quantization error, data loss rate, and data accuracy. Taking quantization error as an example, the specific impact of the distribution characteristics of the first data on the quantization performance is given below. Under a given distribution slope ρ, the higher the dimension d of the codeword, the greater the slope of the quantization error curve. Under the same codeword dimension d, the larger the codebook, the smaller the quantization error. Table 3 shows the corresponding relationship between the distribution slope ρ, the dimension d of the codeword, and the codebook size k when the quantization error of the codebook is lower than the quantization error threshold of 0.02.

[0129] Table 3

[0130] Table 3 shows that, considering quantization performance, when the distribution slope is low, smaller codeword dimensions will first fall below a certain quantization error threshold. As the distribution slope increases, larger codeword dimensions will also fall below the certain quantization error threshold. When the distribution slope is high, medium-sized codeword dimensions will finally fall below the certain quantization error threshold. In other words, the codeword dimensions that meet a certain quantization performance trend from the two extremes to the middle as the distribution slope increases.

[0131] For example, Table 4 shows the dimension d of the codeword corresponding to different value ranges of the distribution slope ρ.

[0132] Table 4

[0133] As can be seen from Table 4, the larger the distribution slope of the first data, that is, the data distribution is relatively concentrated, and when a certain quantization performance is achieved (that is, below a certain quantization error), more codeword dimensions can be used. The smaller the distribution slope of the first data, that is, the relatively dispersed data distribution, and when a certain quantization performance is achieved (that is, below a certain quantization error), fewer codeword dimensions can be used.

[0134] For example, Table 5 shows the distribution slope ρ, the codeword dimension d, the expected data ε, the codebook size k (i.e., the number of codewords in the codebook), and the corresponding relationship between codebooks. Furthermore, the codebooks in Table 5 all meet certain quantization error requirements, for example, the quantization error is less than or equal to the quantization error threshold of 0.02. Table 5 uses ρ ≤ 0.1 as an example; similar tables exist for other values ​​of ρ.

[0135] Table 5

[0136] Referring to Table 5, when the distribution slope ρ of the first data is ≤0.1, the dimension d of the codeword corresponding to the first data is 2, and when the data expectations of the first data are ε1 and ε2 respectively, a corresponding codebook can be selected.

[0137] For example, assuming that the distribution slope of the first data is ρ≤0.1 and the data expectation is ε1, the first codebook set corresponding to the first data includes codebook 1, codebook 2, codebook 3, and codebook 4. Optionally, the quantization error of the codebooks in the first codebook set is less than or equal to the quantization error threshold.

[0138] Three different implementation methods for the first apparatus to determine the first codebook in the first codebook set are described below.

[0139] In a first approach, a first device determines a first codebook from a first codebook set. Specifically, the first device determines the first codebook based on transmission overheads corresponding to the codebooks in the first codebook set. For example, the codebook with the lowest transmission overhead is selected from the first codebook set as the first codebook corresponding to the first data. The transmission overhead corresponding to the codebook includes a codebook indication overhead and / or a data transmission overhead after quantization of the first data.

[0140] The codebook indication overhead refers to the indication overhead of the first codebook selected by the first device to the second device. If the first indication information sent by the first device in the above step 202 is the index of the first codebook, the codebook indication overhead is the number of bits occupied by the index of the first codebook. If the first indication information sent by the first device in the above step 202 includes the first codebook, the codebook indication overhead is equal to k*d*32, where k represents the number of codewords in the first codebook, d represents the codeword dimension of each codeword (that is, the dimension of the quantized data), and 32 represents the number of bits occupied by each data in the quantized data. Taking the above Table 1 as an example, k=4, d=1, then the codebook indication overhead is equal to 128 bits. Taking the above Table 2 as an example, k=128, d=16, then the codebook indication overhead is equal to 65536 bits.

[0141] The data transmission overhead after the first data is quantized refers to the transmission overhead of the codeword corresponding to the quantized data after the first data is quantized. The codeword transmission overhead is equal to (N / d)*log2(k), where N is the number of original data in the first data, d is the dimension of the quantized data (i.e., the dimension of the codeword), so N / d represents the number of quantized data, that is, the number of codewords required to be used, k is the number of codewords in the first codebook, and log2(k) represents the number of bits occupied by each codeword. Taking Table 1 above as an example, k=4, d=1, assuming N=1600, the data transmission overhead after the first data is quantized is equal to 3200 bits. Taking Table 2 above as an example, k=256, d=16, assuming N=1600, the data transmission overhead after the first data is quantized is equal to 800 bits.

[0142] In a second manner, a first device determines a first codebook in a first codebook set. Specifically, the first device sends first information to a second device. The first information is used to indicate some or all codebooks in the first codebook set. For example, the first information includes transmission overheads corresponding to some or all codebooks in the first codebook set. The transmission overhead includes codebook indication overhead and / or data transmission overhead after quantization of the first data. For another example, the first information includes indexes of some or all codebooks in the first codebook set. That is, the first information actually includes the transmission overheads or codebook indexes corresponding to some or all codebooks in the first codebook set. The transmission overheads or codebook indexes corresponding to some or all codebooks in the first codebook set are used to indicate some or all codebooks in the first codebook set. Optionally, the transmission overhead corresponding to the codebook indicated by the first information is less than a transmission overhead threshold. That is, the first device selects a codebook from the first codebook set whose transmission overhead is less than the transmission overhead threshold and indicates the selected codebook through the first information. After receiving the first information, the second device determines whether to adopt (or use or adopt) the codebook indicated by the first information. The second device then sends second information to the first device, where the second information indicates whether to adopt (or use or adopt) some or all of the codebooks indicated by the first information. The first device determines the first codebook based on the second information. For example, it may select the codebook indicated by the second information that has the lowest transmission overhead, or the codebook indicated by the second information that has the lowest quantization error.

[0143] The following is an example for explanation. For example, the first codebook set includes codebook 1, codebook 2, codebook 3, and codebook 4, wherein the transmission overhead corresponding to codebook 1, codebook 2, and codebook 3 are all less than the transmission overhead threshold, and the transmission overhead corresponding to codebook 4 is greater than the transmission overhead threshold.

[0144] In one example, a first device sends first information to a second device that includes transmission overhead 1 corresponding to codebook 1, transmission overhead 2 corresponding to codebook 2, and transmission overhead 3 corresponding to codebook 3. The second device determines that transmission overhead 1 and transmission overhead 2 are permitted, but transmission overhead 3 is not permitted, and then sends second information to the first device. The second information includes transmission overhead 1 corresponding to codebook 1 and transmission overhead 2 corresponding to codebook 2, but does not include transmission overhead 3 corresponding to codebook 3. After receiving the second information, the first device determines that the codebooks indicated by the second information as permitted for use by the second device include codebook 1 and codebook 2. The first device then selects a codebook from codebook 1 and codebook 2, namely, the first codebook.

[0145] In another example, the first information sent by the first device to the second device includes transmission overhead 1 corresponding to codebook 1, transmission overhead 2 corresponding to codebook 2, and transmission overhead 3 corresponding to codebook 3. The second device determines that transmission overhead 1, transmission overhead 2, and transmission overhead 3 are all not allowed, and then sends indication information to the first device, indicating that the transmission overhead corresponding to the codebook indicated in the first information is not allowed. Upon receiving the indication information, the first device determines that it cannot select a codebook, that is, it cannot select the first codebook corresponding to the first data. After this, the first device can adjust the MCS requirement and / or quantization error threshold, re-determine the first codebook set, and then re-determine the first codebook according to the above method.

[0146] In another example, the first information sent by the first device to the second device includes index 1 of codebook 1, index 2 of codebook 2, and index 3 of codebook 3. The second device determines that codebook 1 and codebook 2 are permitted to be used, but codebook 3 is not permitted to be used. The second device then sends second information to the first device, where the second information includes index 1 of codebook 1 and index 2 of codebook 2, but does not include index 3 of codebook 3. After receiving the second information, the first device determines that the codebooks permitted to be used by the second device, as indicated by the second information, include codebook 1 and codebook 2. The first device then selects a codebook from codebook 1 and codebook 2, i.e., a first codebook.

[0147] In another example, the first information sent by the first device to the second device includes index 1 of codebook 1, index 2 of codebook 2, and index 3 of codebook 3. The second device determines that codebook 1, codebook 2, and codebook 3 are all not allowed, and then sends indication information to the first device, indicating that the codebooks indicated in the first information are not allowed. Upon receiving the indication information, the first device determines that it cannot select a codebook, that is, it cannot select the first codebook corresponding to the first data. After this, the first device can adjust the MCS requirement and / or quantization error threshold, re-determine the first codebook set, and then re-determine the first codebook according to the above method.

[0148] In the third approach, the first device determines the first codebook in the first codebook set. Specifically, the first device receives third information from the second device, where the third information indicates the first codebook in the first codebook set. In other words, the second device selects the first codebook from the first codebook set and indicates the first codebook to the first device. Of course, the second device may also independently determine the first codebook corresponding to the first data and indicate the first codebook to the first device, rather than selecting the first codebook from the first codebook set.

[0149] In one implementation method, the following steps 204 to 206 may also be executed, and the execution order of steps 204 to 206 and the aforementioned steps 201 to 203 is not limited.

[0150] In step 204 , the first device determines a second codebook, where the second codebook is used to quantize the second data.

[0151] The second codebook includes the number of codewords, the dimension of the quantized data, and the mapping relationship between the codewords and the quantized data. The number of codewords is also called the codebook size.

[0152] The manner of determining the second codebook is similar to the manner of determining the first codebook, and reference may be made to the foregoing description.

[0153] In one implementation method, the second codebook is the same as the first codebook. That is, the number of codewords in the second codebook is the same as the number of codewords in the first codebook, the dimension of the quantized data in the second codebook is the same as the dimension of the quantized data in the first codebook, and the mapping relationship between the codewords in the second codebook and the quantized data is the same as the mapping relationship between the codewords in the first codebook and the quantized data.

[0154] In another implementation method, the second codebook is different from the first codebook, that is, it satisfies at least one of the following conditions:

[0155] 1) The number of codewords in the second codebook is different from the number of codewords in the first codebook;

[0156] 2) The dimension of the quantized data in the second codebook is different from the dimension of the quantized data in the first codebook;

[0157] 3) The mapping relationship between codewords and quantized data in the second codebook is different from the mapping relationship between codewords and quantized data in the first codebook.

[0158] The second data is different from the first data.

[0159] In one implementation method, when an embodiment of the present application is applied to a scenario where data of an artificial intelligence (AI) model is transmitted, the data of the artificial intelligence model can be divided into multiple data, each data corresponding to a codebook. The multiple data include the above-mentioned first data and second data, so the above-mentioned first data and second data correspond to the same artificial intelligence model. The present application does not limit the specific type of AI model. Exemplarily, the AI ​​model includes a logistic regression model, a decision tree model, a support vector machine (SVM) model, a random forest model, a neural network model, a convolutional neural network (CNN) model, a recurrent neural network (RNN) model, a regression neutral network (RNN) model, a deep neural network (DNN) model or one or more of a generative adversarial network (GAN) model.

[0160] The following introduces different implementation methods for dividing the data of artificial intelligence models into multiple data.

[0161] Implementation method 1: divide one layer of the artificial intelligence model into one data.

[0162] For example, the first data or the second data mentioned above is the data of a layer in an artificial intelligence model.

[0163] This application does not limit the specific types of layers of the AI ​​model. For example, the layers of the AI ​​model include but are not limited to: one or more of: convolutional layers, pooling layers, fully connected (FC) layers, impulse function layers, or batch normalization (BN) layers.

[0164] Implementation method 2: divide multiple layers of the artificial intelligence model into one data.

[0165] For example, the first data or the second data mentioned above is data of multiple layers in an artificial intelligence model.

[0166] For example, multiple layers with the same structure and appearing continuously in an artificial intelligence model can be divided into a layer group, and the data of the layer group can be divided into one data.

[0167] Implementation method 3: divide the data of a layer of the artificial intelligence model into multiple data groups, and divide the data of each data group into one data.

[0168] For example, the data of a certain layer of the artificial intelligence model is divided into three data groups, and the data of the three data groups are divided into three data, each of which corresponds to a codebook. For example, the first data is the first data of the three data, and the second data is the second data of the three data.

[0169] In step 205, the first device sends second indication information to the second device, and the second device receives the second indication information accordingly.

[0170] The second indication information is used to indicate a second codebook.

[0171] In one implementation method, if multiple codebooks and an index of each codebook are predefined or preconfigured on the first device, or multiple codebooks and an index of each codebook are predefined or preconfigured on the first device and the second device, when the second codebook is one of the multiple codebooks, then the second indication information may be an index of the second codebook.

[0172] In another implementation method, the second indication information includes a second codebook, that is, the second indication information includes the number of codewords, the dimension of the quantized data, and the mapping relationship between the codewords and the quantized data. Exemplarily, the second codebook can be as shown in Table 1 or Table 2.

[0173] Step 206: The second device determines a second codebook according to the second indication information.

[0174] In one implementation method, the second indication information is an index of the second codebook, and the second device obtains the second codebook from multiple codebooks predefined or preconfigured on the second device according to the index.

[0175] In another implementation method, the second indication information includes a second codebook, and the second device obtains the second codebook from the second indication information.

[0176] In the above solution, the second data is quantized by using a codebook, which can reduce transmission resource overhead during data transmission.

[0177] Figure 9 is a flow chart of another communication method provided in an embodiment of the present application. This method is a specific implementation of determining the first codebook corresponding to the first data in the embodiment of Figure 2. In this method, the first device independently determines the first codebook corresponding to the first data. Exemplarily, this method is applicable to scenarios where a terminal device independently determines transmission resources in a sidelink, or to scenarios where downlink transmission occurs.

[0178] The method comprises the following steps:

[0179] Step 901: The first device determines a first codebook set according to a distribution characteristic of first data and a quantization error threshold.

[0180] The codebooks in the first codebook set are all smaller than the quantization error threshold.

[0181] For a specific implementation of determining the first codebook set according to the distribution characteristics of the first data and the quantization error threshold, reference may be made to the description of the embodiment in FIG. 2 .

[0182] Step 902: The first device determines a codebook c with the smallest transmission overhead in the first codebook set.

[0183] The method for determining the transmission overhead of the codebooks in the first codebook set may refer to the description of the embodiment in FIG. 2 .

[0184] Step 903: The first device determines whether the transmission overhead corresponding to the codebook c meets the MCS requirement.

[0185] If the MCS requirement is met, go to step 904 ; if the MCS requirement is not met, go to step 905 .

[0186] In step 904 , the first apparatus determines that the codebook c is the first codebook corresponding to the first data.

[0187] That is, the codebook c is subsequently used to transmit the quantized data of the first data.

[0188] Step 905: The first device increases the quantization error threshold by x.

[0189] Since the transmission overhead corresponding to codebook c does not meet the MCS requirement, the first device increases the quantization error threshold by x, which helps the first device determine a codebook that meets the MCS requirement in the next loop process.

[0190] In step 906 , the first device determines whether the quantization error threshold is greater than the quantization error upper limit.

[0191] If the adjusted quantization error threshold is greater than the quantization error upper limit, go to step 907 ; if the adjusted quantization error threshold is less than or equal to the quantization error upper limit, go to step 901 .

[0192] Step 907 : The first device re-determines the MCS requirement, the quantization error upper limit, and the quantization error threshold.

[0193] Since the quantization error threshold has exceeded the quantization error upper limit, the first device needs to reset the MCS requirement, the quantization error upper limit, and the quantization error threshold to re-perform the process of determining the codebook corresponding to the first data.

[0194] FIG10 is a flow chart of another communication method provided in an embodiment of the present application. This method is a specific implementation of determining the first codebook corresponding to the first data in the embodiment of FIG2 . In this method, the first device and the second device jointly determine the first codebook corresponding to the first data. Exemplarily, this method is applicable to uplink transmission scenarios. The method includes the following steps:

[0195] Step 1001: A first device determines a first codebook set according to a distribution characteristic of first data and a quantization error threshold.

[0196] The codebooks in the first codebook set are all smaller than the quantization error threshold.

[0197] For a specific implementation of determining the first codebook set according to the distribution characteristics of the first data and the quantization error threshold, reference may be made to the description of the embodiment in FIG. 2 .

[0198] Step 1002: The first device sends first information to the second device.

[0199] The first information is used to indicate part or all of the codebooks in the first codebook set. For example, the first information indicates a codebook in the first codebook set whose transmission overhead is less than a transmission overhead threshold.

[0200] The first information includes transmission overheads corresponding to part or all of the codebooks in the first codebook set, where the transmission overhead includes a codebook indication overhead and / or a data transmission overhead after the first data is quantized.

[0201] Alternatively, the first information includes indexes of some or all codebooks in the first codebook set.

[0202] For the specific implementation method of the first information, please refer to the description of the embodiment in FIG2 .

[0203] Step 1003: The first device receives second information from the second device.

[0204] The second information is used to indicate that part or all of the codebooks indicated by the first information are adopted (or used or adopted).

[0205] For the specific implementation method of the second information, please refer to the description of the embodiment in FIG2 .

[0206] Step 1004: The first device determines a first codebook according to the second information.

[0207] For example, the first device selects a codebook with the smallest corresponding transmission overhead among the codebooks indicated by the second information, or selects a codebook with the smallest corresponding quantization error among the codebooks indicated by the second information.

[0208] It is understandable that in order to implement the functions in the above embodiments, the first device or the second device includes hardware structures and / or software modules corresponding to the execution of each function. It should be readily apparent to those skilled in the art that, in combination with the units and method steps of each example described in the embodiments disclosed in this application, the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is executed in the form of hardware or computer software driving hardware depends on the specific application scenario and design constraints of the technical solution.

[0209] Figures 11 and 12 are schematic diagrams of the structures of possible communication devices provided in embodiments of the present application. These communication devices can be used to implement the functions of the first device or the second device in the above-mentioned method embodiments, thereby also achieving the beneficial effects of the above-mentioned method embodiments. In the embodiments of the present application, the communication device can be the first device or the second device shown in Figure 1.

[0210] The communication device 1100 shown in Figure 11 includes a processing unit 1110 and a transceiver unit 1120. The communication device 1100 is used to implement the functions of the first device or the second device in the above method embodiment.

[0211] When the communication device 1100 is used to implement the function of the first device in the above method embodiment, the processing unit 1110 is used to determine a first codebook, where the first codebook is used to quantize the first data; and the transceiver unit 1120 is used to send first indication information, where the first indication information is used to indicate the first codebook.

[0212] In a possible implementation method, the processing unit 1110 is configured to determine a first codebook, specifically including: determining a first codebook set according to a distribution characteristic of the first data, the first codebook set including at least one codebook; and determining the first codebook in the first codebook set.

[0213] In a possible implementation method, the processing unit 1110 is configured to determine the first codebook in the first codebook set, specifically including: determining the first codebook according to transmission overhead corresponding to the codebook in the first codebook set.

[0214] In a possible implementation method, the transmission overhead includes codebook indication overhead and / or data transmission overhead after the first data is quantized.

[0215] In a possible implementation method, the processing unit 1110 is configured to determine the first codebook in the first codebook set, specifically including: sending first information through the transceiver unit 1120, where the first information is used to indicate some or all codebooks in the first codebook set; receiving second information through the transceiver unit 1120, where the second information is used to indicate some or all codebooks in the codebook indicated by the first information; and determining the first codebook based on the second information.

[0216] In a possible implementation method, a transmission overhead corresponding to the codebook indicated by the first information is less than a transmission overhead threshold.

[0217] In a possible implementation method, the processing unit 1110 is configured to determine the first codebook in the first codebook set, specifically including: receiving third information through the transceiver unit 1120, where the third information is used to indicate the first codebook.

[0218] In a possible implementation method, a quantization error of a codebook in the first codebook set is less than or equal to a quantization error threshold.

[0219] In a possible implementation method, the quantization error threshold is preconfigured or predefined.

[0220] In a possible implementation method, the quantization error threshold is related to the importance of the first data.

[0221] In one possible implementation method, the processing unit 1110 is configured to determine the first codebook set based on the distribution characteristics of the first data, specifically including: determining, based on the distribution characteristics of the first data and the mapping relationship between the distribution characteristics and the codebook, the first codebook set corresponding to the distribution characteristics of the first data in the mapping relationship.

[0222] In a possible implementation method, the distribution characteristics of the first data include data distribution slope and / or data expectation.

[0223] In a possible implementation method, the first codebook meets a preset modulation and coding scheme MCS requirement.

[0224] In a possible implementation method, the first data is associated with a first service, a first data radio bearer or a first logical channel.

[0225] In one possible implementation method, the processing unit 1110 is further used to quantize the first data to obtain quantized data; determine the codeword corresponding to the quantized data according to the first codebook; and the transceiver unit 1120 is further used to send the codeword corresponding to the quantized data.

[0226] In a possible implementation method, the processing unit 1110 is further configured to determine a second codebook, where the second codebook is used to quantize the second data; and the transceiver unit 1120 is further configured to send second indication information, where the second indication information is used to indicate the second codebook.

[0227] In one possible implementation method, the first data and the second data correspond to the same artificial intelligence model.

[0228] When the communication device 1100 is used to implement the function of the second device in the above method embodiment, the transceiver unit 1120 is used to receive first indication information, where the first indication information is used to indicate a first codebook, and the first codebook is used to quantize first data; and the processing unit 1110 is used to determine the first codebook based on the first indication information.

[0229] In a possible implementation method, the transceiver unit 1120 is further configured to receive a codeword corresponding to quantized data, where the quantized data is obtained by quantizing the first data.

[0230] In a possible implementation method, the processing unit 1110 is further configured to determine the quantized data according to the first codebook and a codeword corresponding to the quantized data.

[0231] In a possible implementation method, the transceiver unit 1120 is further configured to receive first information, where the first information is used to indicate some or all codebooks in a first codebook set; and send second information, where the second information is used to indicate some or all codebooks in the codebook indicated by the first information, where the codebook indicated by the second information includes the first codebook.

[0232] In a possible implementation method, a transmission overhead corresponding to the codebook indicated by the first information is less than a transmission overhead threshold.

[0233] In a possible implementation method, the transmission overhead includes codebook indication overhead and / or data transmission overhead after the first data is quantized.

[0234] In a possible implementation method, a quantization error of a codebook in the first codebook set is less than or equal to a quantization error threshold.

[0235] In a possible implementation method, the quantization error threshold is preconfigured or predefined.

[0236] In a possible implementation method, the quantization error threshold is related to the importance of the first data.

[0237] In a possible implementation method, the transceiver unit 1120 is further configured to send third information, where the third information is used to indicate the first codebook.

[0238] In a possible implementation method, the first codebook meets a preset modulation and coding scheme MCS requirement.

[0239] In a possible implementation method, the first data is associated with a first service, a first data radio bearer or a first logical channel.

[0240] In one possible implementation method, the transceiver unit 1120 is further configured to receive second indication information, where the second indication information is used to indicate a second codebook, and the second codebook is used to quantize the second data; and the processing unit 1110 is further configured to determine the second codebook based on the second indication information.

[0241] In one possible implementation method, the first data and the second data correspond to the same artificial intelligence model.

[0242] For a more detailed description of the processing unit 1110 and the transceiver unit 1120 , reference can be made to the relevant description in the above method embodiment, which will not be repeated here.

[0243] The communication device 1200 shown in Figure 12 includes a processor 1210 and an interface circuit 1220. The processor 1210 and the interface circuit 1220 are coupled to each other. It is understood that the interface circuit 1220 can be a transceiver or an input / output interface. Optionally, the communication device 1200 may further include a memory 1230 for storing instructions executed by the processor 1210, or storing input data required by the processor 1210 to execute instructions, or storing data generated after the processor 1210 executes instructions; or when the processor is a logic circuit (device), storing a configuration file of the logic circuit (device).

[0244] When the communication device 1200 is used to implement the above method embodiment, the processor 1210 is used to implement the functions of the above processing unit 1110 , and the interface circuit 1220 is used to implement the functions of the above transceiver unit 1120 .

[0245] It is understood that the processor in the embodiments of the present application may be a central processing unit (CPU), or may be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field programmable gate arrays (FPGA), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. The general-purpose processor may be a microprocessor or any conventional processor.

[0246] The method steps in the embodiments of the present application can be implemented by hardware or by a processor executing software instructions. The software instructions can be composed of corresponding software modules, and the software modules can be stored in a random access memory, a flash memory, a read-only memory, a programmable read-only memory, an erasable programmable read-only memory, an electrically erasable programmable read-only memory, a register, a hard disk, a mobile hard disk, a compact disc read-only memory (CD-ROM) or any other form of storage medium well known in the art. An exemplary storage medium is coupled to the processor so that the processor can read information from the storage medium and write information to the storage medium. Of course, the storage medium can also be an integral part of the processor. The processor and the storage medium can be located in an ASIC. In addition, the ASIC can be located in the first device or the second device. Of course, the processor and the storage medium can also be present in the first device or the second device as discrete components.

[0247] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer programs or instructions. A computer program refers to a set of instructions that instruct an electronic computer or other device with message processing capabilities to perform each step of the action, usually written in a certain programming language and running on a certain target architecture. When the computer program or instruction is loaded and executed on a computer, the process or function described in the embodiment of the present application is executed in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, a network device, a terminal device or other programmable device. The computer program or instruction can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer program or instruction can be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired or wireless means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium may be a magnetic medium, such as a floppy disk, a hard disk, or a magnetic tape; an optical medium, such as a digital video disk; or a semiconductor medium, such as a solid-state drive. The computer-readable storage medium may be a volatile or non-volatile storage medium, or may include both volatile and non-volatile types of storage media.

[0248] In the various embodiments of the present application, unless otherwise specified or there is a logical conflict, the terms and / or descriptions between different embodiments are consistent and can be referenced by each other. The technical features in different embodiments can be combined to form new embodiments according to their inherent logical relationships.

[0249] In this application, "at least one" means one or more, and "more" means two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone, where A and B can be singular or plural. In the text description of this application, the character " / " generally indicates that the previous and next related objects are in an "or" relationship; in the formulas of this application, the character " / " indicates that the previous and next related objects are in a "division" relationship.

[0250] It is understood that the various numbers used in the embodiments of this application are merely for ease of description and are not intended to limit the scope of the embodiments of this application. The order of the sequence numbers of the above-mentioned processes does not necessarily imply a specific order of execution; the order of execution of the processes should be determined by their functions and inherent logic.

Claims

1. A communication method, It is characterized in that include: Determine a first codebook, wherein the first codebook is used to quantize the first data; First indication information is sent, where the first indication information is used to indicate the first codebook.

2. The method according to claim 1, It is characterized in that The determining of the first codebook includes: Determine a first codebook set according to a distribution characteristic of the first data, where the first codebook set includes at least one codebook; The first codebook in the first codebook set is determined.

3. The method according to claim 2, It is characterized in that The determining the first codebook in the first codebook set includes: The first codebook is determined according to transmission overhead corresponding to a codebook in the first codebook set.

4. The method according to claim 3, It is characterized in that The transmission overhead includes a codebook indication overhead and / or a data transmission overhead after the first data is quantized.

5. The method according to claim 2, It is characterized in that The determining the first codebook in the first codebook set includes: Sending first information, where the first information is used to indicate part or all of the codebooks in the first codebook set; receiving second information, where the second information is used to indicate a part or all of the codebooks indicated by the first information; The first codebook is determined according to the second information.

6. The method according to claim 5, It is characterized in that The transmission overhead corresponding to the codebook indicated by the first information is less than a transmission overhead threshold.

7. The method according to claim 2, It is characterized in that The determining the first codebook in the first codebook set includes: Third information is received, where the third information is used to indicate the first codebook.

8. The method according to any one of claims 2 to 7, It is characterized in that A quantization error of a codebook in the first codebook set is less than or equal to a quantization error threshold.

9. The method according to claim 8, It is characterized in that The quantization error threshold is preconfigured or predefined.

10. The method according to claim 8 or 9, It is characterized in that The quantization error threshold is related to the importance of the first data.

11. The method according to any one of claims 2 to 10, It is characterized in that The determining a first codebook set according to the distribution characteristics of the first data includes: According to the distribution characteristics of the first data and the mapping relationship between the distribution characteristics and the codebook, the first codebook set corresponding to the distribution characteristics of the first data in the mapping relationship is determined.

12. The method according to any one of claims 2 to 11, It is characterized in that The distribution characteristics of the first data include data distribution slope and / or data expectation.

13. The method according to any one of claims 1 to 12, It is characterized in that The first codebook meets a preset modulation and coding scheme MCS requirement.

14. The method according to any one of claims 1 to 13, It is characterized in that The first data is associated with a first service, a first data radio bearer or a first logical channel.

15. The method according to any one of claims 1 to 14, It is characterized in that The method further comprises: quantizing the first data to obtain quantized data; Determining, according to the first codebook, a codeword corresponding to the quantized data; A codeword corresponding to the quantized data is sent.

16. The method according to any one of claims 1 to 15, It is characterized in that The method further comprises: determining a second codebook, wherein the second codebook is used to quantize the second data; Second indication information is sent, where the second indication information is used to indicate the second codebook.

17. The method of claim 16, It is characterized in that The first data and the second data correspond to the same artificial intelligence model.

18. A codebook determination method, It is characterized in that include: receiving first indication information, where the first indication information is used to indicate a first codebook, and the first codebook is used to quantize first data; The first codebook is determined according to the first indication information.

19. The method of claim 18, It is characterized in that The method further comprises: A code word corresponding to quantized data is received, where the quantized data is obtained by quantizing the first data.

20. The method of claim 19, It is characterized in that The method further comprises: The quantized data is determined according to the first codebook and a codeword corresponding to the quantized data.

21. The method according to any one of claims 18 to 20, It is characterized in that The method further comprises: receiving first information, where the first information is used to indicate part or all of the codebooks in the first codebook set; Second information is sent, where the second information is used to indicate that part or all of the codebooks indicated by the first information are used, and the codebooks indicated by the second information include the first codebook.

22. The method of claim 21, It is characterized in that The transmission overhead corresponding to the codebook indicated by the first information is less than a transmission overhead threshold.

23. The method of claim 22, It is characterized in that The transmission overhead includes a codebook indication overhead and / or a data transmission overhead after the first data is quantized.

24. The method according to any one of claims 21 to 23, It is characterized in that A quantization error of a codebook in the first codebook set is less than or equal to a quantization error threshold.

25. The method of claim 24, It is characterized in that The quantization error threshold is preconfigured or predefined.

26. The method of claim 24 or 25, It is characterized in that The quantization error threshold is related to the importance of the first data.

27. The method according to any one of claims 18 to 20, It is characterized in that The method further comprises: Send third information, where the third information is used to indicate the first codebook.

28. The method according to any one of claims 18 to 27, It is characterized in that The first codebook meets a preset modulation and coding scheme MCS requirement.

29. The method according to any one of claims 18 to 28, It is characterized in that The first data is associated with a first service, a first data radio bearer or a first logical channel.

30. The method according to any one of claims 18 to 29, It is characterized in that The method further comprises: receiving second indication information, where the second indication information is used to indicate a second codebook, and the second codebook is used to quantize second data; The second codebook is determined according to the second indication information.

31. The method of claim 30, It is characterized in that The first data and the second data correspond to the same artificial intelligence model.

32. A communication device, It is characterized in that include: At least one processor is used to execute at least one of the following: running computer instructions or programs, logic circuits, so that the communication device executes the data transmission method according to any one of claims 1 to 17, or executes the data transmission method according to any one of claims 18 to 31.

33. A communication system, It is characterized in that The method comprises a first device for executing the method according to any one of claims 1 to 17, and a second device for executing the method according to any one of claims 18 to 31.

34. A computer program product, It is characterized in that The computer program product comprises a computer program or instructions, and when the computer program or instructions are executed by a communication device, the method according to any one of claims 1 to 17 or any one of claims 18 to 31 is executed.

35. A computer readable storage medium, It is characterized in that The storage medium stores a computer program or an instruction. When the computer program or the instruction is executed by the communication device, the method as claimed in any one of claims 1 to 17 or the method as claimed in any one of claims 18 to 31 is implemented.

Citation Information

Patent Citations

  • Method and apparatus for table-based compression with embedded coding

    US20010017941A1

  • Systems and methods for distributed quantization of multimodal images

    US20180097557A1

  • System and method for image compression using systolic vector quantization

    WO2001059934A1

  • Method and system for multi-level feedback of channel information

    WO2011097881A1