WO2023179800A1 - 通信接收方法及其装置 - Google Patents

通信接收方法及其装置 Download PDF

Info

Publication number
WO2023179800A1
WO2023179800A1 PCT/CN2023/088851 CN2023088851W WO2023179800A1 WO 2023179800 A1 WO2023179800 A1 WO 2023179800A1 CN 2023088851 W CN2023088851 W CN 2023088851W WO 2023179800 A1 WO2023179800 A1 WO 2023179800A1
Authority
WO
WIPO (PCT)
Prior art keywords
semantic
decoder
information
semantic decoder
parameters
Prior art date
Application number
PCT/CN2023/088851
Other languages
English (en)
French (fr)
Inventor
董辰
陈梦颖
许晓东
韩书君
王碧舳
Original Assignee
北京邮电大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 北京邮电大学 filed Critical 北京邮电大学
Publication of WO2023179800A1 publication Critical patent/WO2023179800A1/zh

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/30—Semantic analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0464—Convolutional networks [CNN, ConvNet]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0475—Generative networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/048—Activation functions
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
    • G10L19/04—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using predictive techniques
    • G10L19/16—Vocoder architecture
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L1/00—Arrangements for detecting or preventing errors in the information received
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L1/00—Arrangements for detecting or preventing errors in the information received
    • H04L1/0001—Systems modifying transmission characteristics according to link quality, e.g. power backoff
    • H04L1/0006—Systems modifying transmission characteristics according to link quality, e.g. power backoff by adapting the transmission format
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L69/00—Network arrangements, protocols or services independent of the application payload and not provided for in the other groups of this subclass
    • H04L69/22—Parsing or analysis of headers
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W28/00—Network traffic management; Network resource management
    • H04W28/02—Traffic management, e.g. flow control or congestion control
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W28/00—Network traffic management; Network resource management
    • H04W28/02—Traffic management, e.g. flow control or congestion control
    • H04W28/06—Optimizing the usage of the radio link, e.g. header compression, information sizing, discarding information

Definitions

  • the present disclosure relates to the field of communication technology, and in particular, to a communication receiving method and a device thereof.
  • network nodes tend to be intelligent.
  • the intelligence of network nodes has led to the rapid expansion of information space and even dimensionality disaster. It has made it difficult to match traditional network service capabilities with high-dimensional information space, and the amount of data transmitted is excessive. It is too large and cannot continue to meet people's complex, diverse and intelligent information transmission needs.
  • Using artificial intelligence models to encode, disseminate and decode information can significantly reduce the amount of data transmission in communication services and greatly improve the efficiency of information transmission. These models are relatively stable, reusable and disseminating. The dissemination and reuse of models will help enhance network intelligence, while reducing overhead and resource waste, forming an intelligent network with extremely intelligent nodes and extremely simple networks.
  • the receiving end is an indispensable part of the communication system. After receiving the signal from the channel, after signal demodulation, channel decoding, source decoding and other steps are required to restore the signal transmitted by the channel to Information sent by the source. During channel coding and channel decoding, redundancy is added in order to reduce the bit error rate during transmission. Source encoding and source decoding reduce redundancy in order to reduce the amount of data transmission.
  • the structure of the traditional communication receiving end is redundant. Secondly, depending on the received signals, the receiving end must adapt to different powers, different modulation methods, different carrier signals, etc.
  • the receiving end of the communication system not only needs to receive physical bit information, but also needs to be able to receive models and model parameters. Therefore, it is necessary to The receiving end of the unified communication system is improved so that it can not only receive physical bits, but also need to receive models, model parameters, and model-encoded information.
  • the present disclosure provides a communication receiving method and a device thereof.
  • a communication receiving method including:
  • the semantic information is input into the semantic decoder for data restoration to obtain source information, where the data type of the source information corresponds to the semantic decoder.
  • a communication receiving device including:
  • the receiving module receives the bit stream transmitted by the channel
  • a parsing module that parses the bit stream to obtain semantic information and semantic decoder parameters
  • a restoration module that inputs the semantic information into the semantic decoder for data restoration to obtain source information, where the data type of the source information corresponds to the semantic decoder.
  • an electronic device including:
  • a memory communicatively connected to at least one processor; wherein,
  • the memory stores instructions that can be executed by at least one processor, and the instructions are executed by at least one processor, so that at least one processor can perform the above method.
  • a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause the computer to perform the above method.
  • the present invention realizes a method that can receive both physical bits from the sending end and models, model parameters and model-encoded information in an intelligent simplified network. methods and devices. By encoding and decoding information based on artificial intelligence models, the data transmission volume at the communication sender can be significantly reduced and communication efficiency improved.
  • Figure 1 is a schematic flowchart of a communication receiving method provided according to Embodiment 1;
  • Figure 2 is a data flow diagram at the receiving end provided according to Embodiment 1;
  • Figure 3 is a data flow diagram of the semantic encoder provided according to Embodiment 1;
  • Figure 4 is a second data flow diagram at the receiving end provided according to Embodiment 1;
  • Figure 5 is a schematic structural diagram of a communication receiving device provided according to Embodiment 2.
  • FIG. 6 is a block diagram of the electronic device of this embodiment.
  • the sending end device uses a preconfigured first model to extract the first service information and obtain the second service information to be transmitted; the sending end device transmits the second service information to the receiving end device.
  • the receiving end device receives the second service information, separates the parameters related to the second model from the second service information, and then uses the improved second model to restore the second service information to obtain the third service information;
  • the third business information restored by the second model will have some quality differences compared with the original first business information, but the two are consistent in content, and the user experience is almost the same.
  • the method further includes: an update module determines whether the receiving end device needs to update the second model, and when it determines that an update is needed, the update module
  • the receiving end device transmits a preconfigured third model, and the receiving end device uses the third model to update the second model.
  • Model slices can be distributed and stored on multiple network nodes. When a network node finds that it is missing or needs to update a certain model or a certain model slice, it can request it from surrounding nodes that may have the slice.
  • the transmission of the business information and the transmission of the model all occur in the communication network, and communication and transmission are performed based on network protocols.
  • the network nodes passing along the path of transmitting the business information and transmitting the model include intelligent simplified routers.
  • the functions of the intelligent router include but are not limited to business information transmission, model transmission, absorption model self-update, security protection and other functions.
  • the transmission function of the intelligent router involves transmitting business information or models from the source node to the sink node. There are multiple paths between the source node and the sink node.
  • the model transmission function of the Intelligent Router can transmit model slices. By rationally arranging the model slices to take multiple paths, the model slices can be multi-channel transmitted to improve the model transmission rate.
  • Figure 1 shows a communication receiving method provided by this embodiment. As shown in Figure 1, it includes:
  • Step S101 receive the bit stream transmitted by the channel
  • Step S102 parse the bit stream to obtain semantic information and semantic decoder parameters
  • Step S103 Call the semantic decoder of the receiving end according to the semantic decoder parameters
  • Step S104 Input the semantic information into the semantic decoder for data restoration to obtain source information.
  • the data type of the source information corresponds to the semantic decoder.
  • the semantic information and semantic decoder parameters are obtained, which not only protects the transmitted data; it also facilitates the receiving end to correctly determine the data type corresponding to the bit stream based on the semantic decoder parameters, thereby improving the correct semantic decoder.
  • a semantic decoder is used to decode the semantic information and restore the semantic information of the bit stream, thereby converting the physical bit encoding into human understandable information, where the physical bit encoding is a bit stream and the human understandable information is semantic information.
  • the steps of the corresponding method are not necessarily performed in the order shown and described in this specification.
  • the method may include more or fewer steps than described in this embodiment.
  • a single step described in this embodiment may be broken down into multiple steps for description in other embodiments; and the multiple steps described in this embodiment may also be combined into a single step in other embodiments. Describe the steps.
  • This embodiment provides a possible implementation, in which the data types of the source information include: natural language, voice, image, and video.
  • the receiving end has multiple types of first semantic decoders such as a first natural language semantic decoder, a first image semantic decoder, a first video semantic decoder, etc.; the source end has a second natural language semantic decoder.
  • first semantic decoders such as a first natural language semantic decoder, a first image semantic decoder, a first video semantic decoder, etc.
  • second semantic encoders such as language semantic encoder, second image semantic encoder, second video semantic encoder, etc.
  • the source end selects the first semantic encoder of the corresponding type to encode it.
  • the type of semantic information encoded by the source end is also different, such as natural language semantics, image semantics, and video semantics.
  • the source end also has multiple types of second semantic decoders such as a second natural language semantic decoder, a second image semantic decoder, and a second video semantic decoder; the above-mentioned second semantic decoder and the above-mentioned second semantic encoder Generative adversarial network training is performed together to obtain the trained second semantic decoder and the second semantic encoder.
  • the parameters of the second semantic decoder above are used as semantic decoding
  • the transmitter parameters are also transmitted to the receiving end.
  • the receiving end has multiple types of first semantic decoders such as a first natural language semantic decoder, a first image semantic decoder, and a first video semantic decoder;
  • the above-mentioned first semantic decoder may be preset at the receiving end and has the same structure as the above-mentioned second semantic decoder. Therefore, as long as the parameters of the trained second semantic decoder are passed to the first semantic decoder, the receiving end can quickly have the same semantic decoding capability.
  • the first semantic decoder of the receiving end can also be transmitted to the receiving end through model transmission, for example, from a third-party node, or from the above-mentioned source end.
  • the first semantic decoder at the receiving end may be deployed in advance.
  • the receiving end has a channel decoder that decodes the bit stream to obtain a channel decoding result.
  • the channel decoding results include semantic information and semantic decoder parameters.
  • Semantic information can be of different types, such as natural language semantics, image semantics, and video semantics.
  • the first semantic decoder at the source end and the second semantic decoder at the receiving end may be decoders that decode different types of semantic information. Therefore, the first semantic decoder and the second semantic decoder also correspond to different types and can process different types of data. Therefore, the parameters of the first semantic decoder and the second semantic decoder are also related to the type of data.
  • the parameters of the second semantic decoder may directly include information indicating the data type.
  • the parameters may include type information indicating the natural language type, image type or video type; the parameters of the second semantic decoder may also be derived from Extract type information.
  • the semantic decoder parameters can be used to determine which type of first semantic decoder the semantic decoder parameters are to be written to.
  • the channel decoding result decoded from the bit stream includes semantic information and the semantic decoder parameters. Assume that the semantic decoder parameters contain information indicating "image type", so it can be determined that the above semantic information is semantic information of "image type"; and the semantic decoder parameters need to be written into the first image semantic decoder.
  • the first image semantic decoder has the ability to parse image semantic information after writing the semantic decoder parameters. At this time, the above semantic information (actually image semantic information) is input into the first image semantic decoder for data restoration to obtain restored image source information.
  • the above semantic information (actually image semantic information) is input into the first image semantic decoder for data restoration to obtain restored image source information.
  • the source end will train the semantic encoder in advance.
  • the loss function of the semantic encoder and the semantic decoder corresponding to the semantic decoder can include mean square error (MSE) and semantic error.
  • MSE mean square error
  • semantic error semantic error
  • ⁇ is used to represent the weighting factor
  • L MSE is used to represent the mean square error, which can be further expressed as Usually, a large proportion of L MSE helps the semantic decoder to converge quickly;
  • L SE is used to represent the semantic error, which can be further expressed as
  • S is used to represent the original source information input to the transmitter at the source end, Source information used for receiver restoration;
  • using the mean square error can make the original source information and the restored source information as similar as possible during the training of the semantic decoder;
  • the semantic error is the semantic error between the original source information and the restored source information
  • the original source information may be an original image
  • the restored source information may be a restored image
  • step S102 includes the following steps:
  • Step S1021 Analyze the bit stream based on a channel decoder to obtain a channel decoding result
  • Step S1022 Analyze the channel decoding result according to predefined bit stream decoding rules to obtain semantic information and semantic decoder parameters.
  • the semantic information and semantic decoder parameters can be parsed out of the bit stream, so there are corresponding processing steps in the transmitter at the source end to encode the semantic information and semantic decoder parameters into a bit stream; the bit stream decoding algorithm corresponding to the analysis , and the transmitter at the corresponding source end has a corresponding bit stream coding algorithm.
  • bit stream decoding algorithm is obtained based on the bit stream encoding algorithm, and the bit stream decoding algorithm is used to analyze the bit stream, thereby achieving the purpose of protecting the transmitted data.
  • Generative adversarial networks include a generator and a classifier, where the generator tries to produce source information that is closer to the original, and accordingly, the classifier tries to more perfectly distinguish between the original source information and the restored source information.
  • GAN Generative adversarial networks
  • the generator and the classifier progress in the confrontation, and continue to compete after progress.
  • the data obtained by the generator will become more and more perfect, approaching the original source information, thus realizing the use of both the first model and the second model.
  • a fully trained generator in a generative adversarial network can improve the accuracy of the restored source information.
  • step S1022 the reverse rule can be used to obtain the bit stream decoding algorithm corresponding to the bit stream encoding algorithm.
  • bit stream decoding algorithm corresponding to the bit stream encoding algorithm is obtained through the reverse rule, and then the bit stream is parsed according to the bit stream decoding algorithm to obtain semantic information and semantic decoder parameters; thereby making the semantic information closer
  • semantic information on the source side and the semantic decoder parameters are closer to the semantic encoder parameters on the source side to improve the accuracy of data restoration.
  • x is used to represent the bit stream transmitted by the source
  • l is used to represent semantic information
  • ⁇ is used to represent the semantic decoder parameters
  • SR is the bit stream coding rule, specifically SR( ⁇ );
  • bit stream decoding algorithm of the receiving end corresponding to the bit stream encoding algorithm can be obtained, as shown in the following formula:
  • y is used to represent the bit stream received by the receiving end
  • SR -1 is the bit stream decoding rule.
  • the bit stream x transmitted by the source is input into the channel, so that the channel output bit stream y enters the receiving end, and then undergoes demodulation and channel decoder to obtain semantic information and semantic decoder parameters.
  • a semantic decoder is obtained based on the semantic decoder parameters, which in this embodiment is a semantic decoder. Input semantic information into the semantic decoder to obtain restored source information
  • bit stream can be input to the channel decoder in step S1023, so that the channel decoder parses the bit stream according to the bit stream decoding algorithm to obtain semantic information and semantic decoder parameter.
  • the semantic encoder includes:
  • the semantic decoder consists of:
  • the encoding fully connected layer corresponds to the decoding fully connected layer
  • the first encoding convolutional layer corresponds to the third decoding deconvolutional layer
  • the second encoding convolutional layer corresponds to the second decoding deconvolutional layer
  • the third encoding convolutional layer corresponds to the first decoding deconvolutional layer.
  • the number of layers of the first encoding convolutional layer, the second encoding convolutional layer and the third encoding convolutional layer depends on the data size.
  • This embodiment provides a possible implementation, in which the activation functions of the first coding convolution layer, the second coding convolution layer and the third coding convolution layer are all ReLU functions, and the convolution step sizes are all 2 and The number of output channels is 16.
  • This embodiment provides a possible implementation, in which the activation functions of the first decoding deconvolution layer and the second decoding deconvolution layer are both ReLU functions, the convolution step size is both 2, and the number of output channels is both 16;
  • the activation function of the third decoding deconvolution layer is the Sigmoid function, the convolution step size is 2 and the number of output channels is 1.
  • the images in the MNIST data set are first semantically extracted through three convolutional layers, which are the first encoding convolutional layer, the second encoding convolutional layer and the third encoding convolutional layer;
  • downsampling is performed by setting the convolution step size to 2, and the feature map with a channel number of 16 and a length and width value of 2 is output.
  • the feature map output by the convolutional layer is further compressed through the fully connected layer.
  • the decoder is symmetrical to the encoder.
  • the semantic information is decompressed by decoding the fully connected layer, and the feature map of [16, 2, 2] size is restored through the shaping operation.
  • the feature map is processed through three deconvolution layers to restore the size of the source image, where the three deconvolution layers are the first decoding deconvolution layer, the second decoding deconvolution layer and the third deconvolution layer. Decode the deconvolution layer.
  • the activation function of the third decoding deconvolution layer is the Sigmoid function.
  • Sigmoid function helps map the image to the [0,1] interval to improve the accuracy of data restoration.
  • Figure 5 shows the communication receiving device 200 provided in this embodiment. As shown in Figure 5, it includes:
  • the receiving module 201 receives the bit stream transmitted by the channel
  • Parsing module 202 parses the bit stream to obtain semantic information and semantic decoder parameters
  • Call module 203 to call the semantic decoder of the device according to the semantic decoder parameters
  • the restoration module 204 inputs the semantic information into the semantic decoder for data restoration to obtain source information.
  • the data type of the source information corresponds to the semantic decoder.
  • This disclosure can reduce the amount of information transmission by acquiring the bit stream; by parsing the bit stream to obtain the semantic decoder parameters and semantic information, obtain the semantic decoder according to the semantic decoder parameters, and obtain the semantic decoder by using Decode the semantic information and restore the semantic information of the semantic information, thereby converting the physical bit encoding into human understandable information, and can significantly increase the communication rate of the receiving end.
  • the present disclosure also provides an electronic device and a readable storage medium.
  • the electronic equipment includes:
  • a memory communicatively connected to at least one processor; wherein,
  • the memory stores instructions executable by at least one processor, and the instructions are executed by at least one processor Execution to enable at least one processor to execute the above method.
  • the electronic device can obtain the bit stream to reduce the amount of information transmission; parse the bit stream to obtain semantic decoder parameters and semantic information, obtain the semantic decoder according to the semantic decoder parameters, and obtain the semantic decoding
  • the device decodes the semantic information and restores the semantic information of the semantic information, thereby converting the physical bit encoding into human understandable information, and can significantly increase the communication rate of the receiving end.
  • This non-transient computer-readable storage medium stores computer instructions, and the computer instructions are used to cause the computer to execute the method provided in this embodiment.
  • the readable storage medium can obtain a bit stream to reduce the amount of information transmitted; parse the bit stream to obtain semantic decoder parameters and semantic information, call the semantic decoder according to the semantic decoder parameters, and adopt a perfect
  • the subsequent semantic decoder decodes the semantic information and restores the semantic information of the semantic information, thereby converting the physical bit encoding into human understandable information, and can significantly increase the communication rate of the receiving end.
  • FIG. 6 shows a schematic block diagram of an example electronic device 300 that may be used to implement embodiments of the present disclosure.
  • Electronic devices are intended to mean various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers.
  • Electronic devices may also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, wearable devices, and other similar computing devices.
  • the components shown herein, their connections and relationships, and their functions are examples only and are not intended to limit implementations of the disclosure described and/or claimed herein.
  • the device 300 includes a computing unit 301 that can execute according to a computer program stored in a read-only memory (ROM) 302 or loaded from a storage unit 308 into a random access memory (RAM) 303 Various appropriate actions and treatments.
  • ROM read-only memory
  • RAM random access memory
  • various programs and data required for the operation of the device 300 can also be stored.
  • Computing unit 301, ROM 302 and RAM 303 are connected to each other via bus 304.
  • Input/output (I/O) interface 307 is also connected to Bus 304.
  • I/O interface 306 Multiple components in the device 300 are connected to the I/O interface 306, including: input unit 306, such as a keyboard, mouse, etc.; output unit 307, such as various types of displays, speakers, etc.; storage unit 308, such as a magnetic disk, optical disk, etc. ; and communication unit 309, such as a network card, modem, wireless communication transceiver, etc.
  • the communication unit 309 allows the device 300 to exchange information/data with other devices through computer networks such as the Internet and/or various telecommunications networks.
  • Computing unit 301 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of the computing unit 301 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units that run machine learning model algorithms, digital signal processing processor (DSP), and any appropriate processor, controller, microcontroller, etc.
  • the computing unit 301 performs various methods and processes described above, such as the method communication receiving method.
  • the method communication receiving method may be implemented as a computer software program that is tangibly embodied in a machine-readable medium, such as the storage unit 307.
  • part or all of the computer program may be loaded and/or installed onto device 300 via ROM 302 and/or communication unit 309.
  • the computer program When the computer program is loaded into the RAM 303 and executed by the computing unit 301, one or more steps of the above-described method communication receiving method may be performed.
  • the computing unit 301 may be configured to perform the method communication receiving method in any other suitable manner (eg, by means of firmware).
  • Various implementations of the systems and techniques described above may be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on a chip implemented in a system (SOC), load programmable logic device (CPLD), computer hardware, firmware, software, and/or a combination thereof.
  • FPGAs field programmable gate arrays
  • ASICs application specific integrated circuits
  • ASSPs application specific standard products
  • SOC system
  • CPLD load programmable logic device
  • computer hardware firmware, software, and/or a combination thereof.
  • Various embodiments may include implementation in one or more computer programs executable and/or interpreted on a programmable system including at least one programmable processor, the programmable processor It may be a special purpose or general purpose programmable processor, which may be configured from a memory system, at least one input device, and at least one output device receive data and instructions, and transmit the data and instructions to the storage system, the at least one input device, and the at least one output device.
  • Program code for implementing the methods of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that the program codes, when executed by the processor or controller, cause the functions specified in the flowcharts and/or block diagrams/ The operation is implemented.
  • the program code may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.
  • a machine-readable medium may be a tangible medium that may contain or store a program for use by or in connection with an instruction execution system, apparatus, or device.
  • the machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium.
  • Machine-readable media may include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or devices, or any suitable combination of the foregoing.
  • machine-readable storage media would include one or more wire-based electrical connections, laptop disks, hard drives, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination of the above.
  • RAM random access memory
  • ROM read only memory
  • EPROM or flash memory erasable programmable read only memory
  • CD-ROM portable compact disk read-only memory
  • magnetic storage device or any suitable combination of the above.
  • the systems and techniques described herein may be implemented on a computer having a display device (eg, a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user ); and a keyboard and pointing device (eg, a mouse or a trackball) through which a user can provide input to the computer.
  • a display device eg, a CRT (cathode ray tube) or LCD (liquid crystal display) monitor
  • a keyboard and pointing device eg, a mouse or a trackball
  • Other kinds of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and may be provided in any form, including Acoustic input, voice input or tactile input) to receive input from the user.
  • the systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., A user's computer having a graphical user interface or web browser through which the user can interact with implementations of the systems and technologies described herein), or including such backend components, middleware components, or any combination of front-end components in a computing system.
  • the components of the system may be interconnected by any form or medium of digital data communication (eg, a communications network). Examples of communication networks include: local area network (LAN), wide area network (WAN), and the Internet.
  • Computer systems may include clients and servers.
  • Clients and servers are generally remote from each other and typically interact over a communications network.
  • the relationship of client and server is created by computer programs running on corresponding computers and having a client-server relationship with each other.
  • the server can be a cloud server, a distributed system server, or a server combined with a blockchain.

Abstract

一种通信接收方法及其装置（200），涉及通信技术领域。具体实现方案为：接收由信道传输的比特流(S101)；对比特流进行解析，以得到语义信息和语义解码器参数(S102)；根据语义解码器参数调用接收端的语义解码器(S103)；将语义信息输入到语义解码器中进行数据还原，以得到源信息(S104)。

Description

通信接收方法及其装置
本公开要求于2022年03月21日提交中国专利局、申请号为2022102791280、发明名称为"通信接收方法及其装置"的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本公开涉及通信技术领域，尤其涉及一种通信接收方法及其装置。
背景技术
在未来的万物智联网络中，网络节点趋向于智能化，网络节点智能化导致了信息空间快速扩张、甚至维度灾难，导致传统的网络服务能力与高维信息空间难以匹配，信息传输的数据量过大，无法持续满足人们复杂、多样和智能化信息传输的需求。而通过人工智能模型来编码、传播、解码信息，可显著降低通信业务中的数据传输量，极大地提升信息传输效率。这些模型相对稳定，并具有复用性、传播性。模型的传播和复用将有助于增强网络智能，同时降低开销和资源浪费，形成节点极智、网络极简的智简网络。
在数字通信系统中，接收端是通信系统中不可或缺的一部分，从信道中接收到信号之后，经过信号解调制，要通过信道解码、信源解码等步骤，才能将信道传输的信号恢复为信源发送的信息。其中信道编码和信道解码时，为了降低传输过程中的误码率增添了冗余。而信源编码和信源解码，则为了减少数据传输量，而减少冗余。传统的通信接收端的结构冗余，其次根据接收的信号不同，接收端要适应不同的功率、不同的调制方法、不同的载波信号等。
这些和未来高速率、高准确度并且支持人工智能模型编码、传输、解码的网络架构并不匹配。随着智简网络的到来，通信系统接收端不仅需要接收物理比特信息，还需要能够接收模型以及模型参数等，因此需要对传 统通信系统的接收端做出改进，使其不仅能够接收物理比特，在此基础上还需要接收模型、模型参数、以及经模型编码的信息等。
发明内容
本公开提供了一种通信接收方法及其装置。
根据本公开的第一方面，提供了一种通信接收方法，包括：
接收由信道传输的比特流；
对所述比特流进行解析，以得到语义信息和语义解码器参数；
根据所述语义解码器参数调用接收端的语义解码器；
将所述语义信息输入到所述语义解码器中进行数据还原，以得到源信息，其中，所述源信息的数据类型与所述语义解码器相对应。
根据本公开的第二方面，提供了一种通信接收装置，包括：
接收模块，接收由信道传输的比特流；
解析模块，对所述比特流进行解析，以得到语义信息和语义解码器参数；
调用模块，根据所述语义解码器参数调用所述装置的语义解码器；
还原模块，将所述语义信息输入到所述语义解码器中进行数据还原，以得到源信息，其中，所述源信息的数据类型与所述语义解码器相对应。
根据本公开的第三方面，提供了一种电子设备，包括：
至少一个处理器；以及
与至少一个处理器通信连接的存储器；其中，
存储器存储有可被至少一个处理器执行的指令，指令被至少一个处理器执行，以使至少一个处理器能够执行上述方法。
根据本公开的第四方面，提供了一种存储有计算机指令的非瞬时计算机可读存储介质，其中，计算机指令用于使计算机执行上述方法。
本发明通过上述技术方案中的通信接收方法、装置、电子设备以及存储介质，实现了一种在智简网络中即可以从发送端接收物理比特又可以接收模型、模型参数以及经模型编码的信息的方法和装置。通过基于人工智能模型对信息进行编码和解码，可以显著地降低通信发送端的数据传输量，提升通信效率。
应当理解，本部分所描述的内容并非旨在标识本公开的实施例的关键或 重要特征，也不用于限制本公开的范围。本公开的其他特征将通过以下的说明书而变得容易理解。
附图说明
附图用于更好地理解本方案，不构成对本公开的限定。其中：
图1是根据本实施例一提供的通信接收方法的流程示意图；
图2是根据本实施例一提供的接收端的数据流通图一；
图3是根据本实施例一提供的语义编码器的数据流通图；
图4是根据本实施例一提供的接收端的数据流通图二；
图5是根据本实施例二提供的通信接收装置的结构示意图；
图6是本实施例的电子设备的框图。
具体实施方式
以下结合附图对本公开的示范性实施例做出说明，其中包括本实施例的各种细节以助于理解，应当将它们认为仅仅是示范性的。因此，本领域普通技术人员应当认识到，可以对这里描述的实施例做出各种改变和修改，而不会背离本公开的范围和精神。同样，为了清楚和简明，以下的描述中省略了对公知功能和结构的描述。
智简网络中主要通过人工智能模型传播业务信息，通过使用人工智能模型将待传播的第一业务信息压缩为与所述人工智能模型相关的第二业务信息，极大地降低了网络中的数据通信量，压缩效率远超传统的压缩算法。其中，发送端设备利用预先配置的第一模型对所述第一业务信息进行提取并得到待传输的第二业务信息；所述发送端设备向接收端设备传输所述第二业务信息。接收端设备接收所述第二业务信息，从第二业务信息中分离出有关第二模型的参数，而后利用完善后的第二模型对所述第二业务信息进行恢复处理得到第三业务信息；经第二模型恢复的第三业务信息比起原先的第一业务信息会有些许质量上的差异，但两者在内容上是一致的，给用户的体验几乎是无差 异的。在所述发送端设备向接收端设备传输所述第二业务信息之前，还包括：更新模块判断所述接收端设备是否需要对所述第二模型进行更新，并在判断需要更新时向所述接收端设备传输预先配置的第三模型，所述接收端设备利用所述第三模型对所述第二模型进行更新。通过预先训练好的人工智能模型对业务信息进行处理，可显著降低通信业务中的数据传输量，极大地提升了信息传输效率。这些模型相对稳定，并具有复用性、传播性。模型的传播和复用将有助于增强网络智能，同时降低开销和资源浪费。所述模型能够根据不同切分规则切分为若干个模型切片，上述模型切片也可以在不同的网络节点之间传输，模型切片可以组装成模型。模型切片可以分散存储在多个网络节点上。当网络节点请发现自己缺少或需要更新某模型或某模型切片时，可以通过请求的方式，向周围可能具有该切片的节点请求。
传输所述业务信息、传输所述模型均发生在通信网络中，基于网络协议进行通信传输。传输所述业务信息、传输所述模型的路径上经过的网络节点包括智简路由器。智简路由器的功能包括但不限于业务信息传输、模型传输，吸收模型自我更新，安全保护等功能。智简路由器的传输功能，涉及将业务信息或模型从信源节点传输到信宿节点，信源节点和信宿节点之间存在多个路径。智简路由器的模型传输功能，可以对模型切片进行传输，通过合理安排模型切片走多个路径，多路传输模型切片，提高模型传输速率。
实施例一
图1示出了本实施例提供的一种通信接收方法，如图1所示，包括：
步骤S101，接收由信道传输的比特流；
步骤S102，对所述比特流进行解析，以得到语义信息和语义解码器参数；
步骤S103，根据所述语义解码器参数调用接收端的语义解码器；
步骤S104，将所述语义信息输入到所述语义解码器中进行数据还原，以得到源信息。
其中，所述源信息的数据类型与所述语义解码器相对应。
比特流解析之后得到语义信息和语义解码器参数，不仅能起到保护传输数据的目的；而且便于接收端根据语义解码器参数，正确判断比特流对应的数据类型，从而完善正确的语义解码器。采用语义解码器对语义信息进行译码，恢复比特流的语义信息，从而实现将物理比特编码转换为人类可理解信息，其中该物理比特编码为比特流，该人类可理解信息为语义信息。
需要说明的是：在其他实施例中并不一定按照本说明书示出和描述的顺序来执行相应方法的步骤。在一些其他实施例中，其方法所包括的步骤可以比本实施例所描述的更多或更少。此外，本实施例中所描述的单个步骤，在其他实施例中可能被分解为多个步骤进行描述；而本实施例中所描述的多个步骤，在其他实施例中也可能被合并为单个步骤进行描述。
本实施例提供了一种可能的实现方式，其中，源信息的数据类型包括：自然语言、语音、图像、视频。
在一种具体实施方式中，接收端具有第一自然语言语义解码器、第一图像语义解码器、第一视频语义解码器等多种类型的第一语义解码器；信源端具有第二自然语言语义编码器、第二图像语义编码器、第二视频语义编码器等多种类型的第二语义编码器；
信源端根据源信息的数据类型的不同，选择对应类型的第一语义编码器对其编码。
根据源信息的数据类型的不同，信源端对其编码得到的语义信息的类型也是不同的，例如自然语言语义、图像语义、视频语义。
信源端还具有第二自然语言语义解码器、第二图像语义解码器、第二视频语义解码器等多种类型的第二语义解码器；上述第二语义解码器与上述第二语义编码器一起进行生成式对抗网络训练，从而获得训练好的上述第二语义解码器和上述第二语义编码器。上述第二语义解码器的参数作为语义解码 器参数也传输给接收端。
所述接收端具有第一自然语言语义解码器、第一图像语义解码器、第一视频语义解码器等多种类型的第一语义解码器；
上述第一语义解码器可以是预置在接收端的，具有和上述第二语义解码器相同的结构。因此只要将训练好的第二语义解码器的参数传递给第一语义解码器，就可以在接收端迅速地具备相同的语义解码能力。
上述接收端的第一语义解码器也可以是通过模型传输的方式传递给接收端的，例如从第三方节点传输过来，也可以从上述信源端传输过来。接收端的第一语义解码器可以是提前进行部署的。
接收端具有信道解码器，对所述比特流进行解码，以得到信道解码结果。所述信道解码结果包括语义信息和语义解码器参数。
语义信息可以是不同的类型，例如自然语言语义、图像语义、视频语义。
由于信源端的第一语义解码器和接收端的第二语义解码器可以是对不同类型的语义信息进行解码的解码器。因此第一语义解码器、第二语义解码器也对应不同的类型，可以处理不同类型的数据。因此，第一语义解码器、第二语义解码器的参数也与数据的类型相关。
所述第二语义解码器的参数，可以直接包含指示数据类型的信息，例如参数中具有指示自然语言类型、图像类型或视频类型的类型信息；所述第二语义解码器的参数，也可以从中提取出类型信息。
由于接收端具有多种类型的第一语义解码器，因此可以通过所述语义解码器参数，来判断要将所述语义解码器参数写入哪个类型的第一语义解码器。
从比特流中解码出来的信道解码结果，包含语义信息和所述语义解码器参数。假设所述语义解码器参数中包含指示“图像类型”的信息，因此可以判断上述语义信息是“图像类型”的语义信息；而所述语义解码器参数需要写入第一图像语义解码器。
所述第一图像语义解码器，在写入所述语义解码器参数后，具备了解析图像语义信息的能力。此时将上述语义信息(实际为图像语义信息)输入到所述第一图像语义解码器中进行数据还原，以得到还原的图像源信息。
本实施例中的信源端会提前训练好语义编码器，其中，语义编码器和语义解码器对应的语义解码器的损失函数可以包括均方误差(MSE)和语义误差，损失函数如下述公式所示：
其中，用于表示损失函数；
γ用于表示加权因子；
LMSE用于表示均方误差，可以进一步可以表示为通常LMSE占比大有助于语义解码器快速收敛；
LSE用于表示语义误差，可以进一步可以表示为
S用于表示输入到信源端的发射机的原始的源信息，用于接收端还原的源信息；
其中，采用均方误差可以使得在语义解码器的训练中使原始的源信息和还原的源信息尽可能相似；
其中，语义误差为原始的源信息与还原的源信息之间的在语义上的误差；
需要说明的是，原始的源信息可以为原始图像，还原的源信息可以为还原图像。
本实施例提供了一种可能的实现方式，其中，步骤S102包括以下步骤：
步骤S1021，基于信道解码器对所述比特流进行解析，以得到信道解码结果；
步骤S1022，根据预定义比特流解码规则对所述信道解码结果进行解析，以得到语义信息和语义解码器参数。
比特流中可以解析出语义信息和语义解码器参数，因此在信源端的发射机中具有对应的处理步骤，将语义信息和语义解码器参数编码为比特流；所述解析对应的比特流解码算法，而在对应的信源端的发射机中，具有对应的比特流编码算法。
本公开中根据比特流编码算法获取得到比特流解码算法，并采用比特流解码算法对比特流进行解析，从而起到保护传输数据的目的。
生成式对抗网络(GAN)包括生成器和分类器，其中生成器试图产生更接近原始的源信息，相应地，分类器试图更完美地分辨原始的源信息与还原的源信息。由此，生成器和分类器在对抗中进步，在进步后继续对抗，由生成器得的数据也就越来越完美，逼近原始的源信息，从而实现将第一模型和第二模型均采用完成训练的生成式对抗网络中的生成器可以提高还原的源信息的精确度。
本实施例提供了一种可能的实现方式，其中，步骤S1022中可以采用反向规则获取得到与比特流编码算法相对应的比特流解码算法。
具体地，通过反向规则获取得到与比特流编码算法相对应的比特流解码算法，随后根据比特流解码算法对比特流进行解析，以得语义信息和语义解码器参数；从而使得语义信息更接近信源端的语义信息，语义解码器参数更接近信源端的语义编码器参数，以提高数据还原的精确度。
示例地，信源端的比特流编码算法如下述公式所示：
x＝SR(l，β)；
其中，x用于表示信源端发射的比特流；
l用于表示语义信息；
β用于表示语义解码器参数；
SR为比特流编码规则，具体为SR(·)；
因此，根据反向规则可以获取得到与比特流编码算法相对应的接收端的比特流解码算法，具体如下述公式所示：
其中，用于表示解析获得的语义解码器参数；
用于表示解析获得的语义信息；
y用于表示接收端接收到的比特流；
SR-1为比特流解码规则。
如图2所示，信源端发射的比特流x输入到信道中，使得信道输出比特流y进入接收端，随后依次经过解调和信道解码器得到语义信息和语义解码器参数。根据语义解码器参数获得语义解码器，在本实施例中语义解码器是语义解码器。将语义信息输入到语义解码器中，以得到还原的源信息
本实施例提供了一种可能的实现方式，其中，步骤S1023中可以将比特流输入到信道解码器，使得信道解码器根据比特流解码算法对比特流进行解析，以得语义信息和语义解码器参数。
本实施例提供了一种可能的实现方式，其中，如图3所示，语义编码器依次包括：
第一编码卷积层、第二编码卷积层、第三编码卷积层和编码全连接层；其中，将原始数据输入到语义编码器中，编码全连接层能够输出原始语义信息；
如图4所示，语义解码器依次包括：
解码全连接层、第一解码反卷积层、第二解码反卷积层和第三解码反卷积层；其中，将语义信息输入到语义解码器中，第三解码反卷积层能够输出源信息；
其中，编码全连接层与解码全连接层相对应；
第一编码卷积层与第三解码反卷积层相对应；
第二编码卷积层与第二解码反卷积层相对应；
第三编码卷积层与第一解码反卷积层相对应。
具体地，第一编码卷积层、第二编码卷积层和第三编码卷积层的层数视数据大小而定。
本实施例提供了一种可能的实现方式，其中，第一编码卷积层、第二编码卷积层和第三编码卷积层的激活函数均为ReLU函数、卷积步长均为2和输出通道数均为16。
本实施例提供了一种可能的实现方式，其中，第一解码反卷积层和第二解码反卷积层的激活函数均为ReLU函数、卷积步长均为2和输出通道数均为16；第三解码反卷积层的激活函数为Sigmoid函数、卷积步长为2和输出通道数为1。
示例性地，在信源端：
MNIST数据集的图像首先通过三个卷积层进行语义提取，该三个卷积层依次为第一编码卷积层、第二编码卷积层和第三编码卷积层；
然后通过设置卷积步长为2进行下采样，输出通道数为16、长度和宽度值为2的特征图。
在整形操作后，将卷积层输出的特征图进一步通过全连接层压缩。
在接收端，解码器与编码器对称。
首先，通过解码全连接层对接语义信息进行解压缩，通过整形操作恢复[16,2,2]大小的特征映射。
随后，通过三层反卷积层对特征图进行处理，以恢复源图像的大小，其中，三层反卷积层依次为第一解码反卷积层、第二解码反卷积层和第三解码反卷积层。
其中，第三解码反卷积层的激活函数是Sigmoid函数，采用Sigmoid函数有助于将图像映射到[0,1]区间，以提高数据还原的精确度。
实施例二
以上是对通信接收方法进行的描述，下面将对通信接收装置进行描述。
图5示出了本实施例提供的通信接收装置200，如图5所示，包括：
接收模块201，接收由信道传输的比特流；
解析模块202，对所述比特流进行解析，以得到语义信息和语义解码器参数；
调用模块203，根据所述语义解码器参数调用所述装置的语义解码器；
还原模块204，将所述语义信息输入到所述语义解码器中进行数据还原，以得到源信息。
其中，所述源信息的数据类型与所述语义解码器相对应。
本公开能通过获取比特流，以减少信息量的传输；通过对比特流进行解析，以得到语义解码器参数和语义信息，根据语义解码器参数获取得到语义解码器，并采用获取得到语义解码器对语义信息进行译码，恢复语义信息的语义信息，从而实现将物理比特编码转换为人类可理解信息，并且可以显著地提升接收端的通信速率。
对于本实施例，其实现的有益效果同上述通信接收方法实施例，此处不再赘述。
根据本公开的实施例，本公开还提供了一种电子设备和一种可读存储介质。
该电子设备，包括：
至少一个处理器；以及
与至少一个处理器通信连接的存储器；其中，
存储器存储有可被至少一个处理器执行的指令，指令被至少一个处理器 执行，以使至少一个处理器能够执行上述方法。
该电子设备能通过获取比特流，以减少信息量的传输；通过对比特流进行解析，以得到语义解码器参数和语义信息，根据语义解码器参数获取得到语义解码器，并采用获取得到语义解码器对语义信息进行译码，恢复语义信息的语义信息，从而实现将物理比特编码转换为人类可理解信息，并且可以显著地提升接收端的通信速率。
该存储有计算机指令的非瞬时计算机可读存储介质，计算机指令用于使计算机执行本实施例提供的方法。
该可读存储介质能通过获取比特流，以减少信息量的传输；通过对比特流进行解析，以得到语义解码器参数和语义信息，根据语义解码器参数调用所述语义解码器，并采用完善后的语义解码器对语义信息进行译码，恢复语义信息的语义信息，从而实现将物理比特编码转换为人类可理解信息，并且可以显著地提升接收端的通信速率。
图6示出了可以用来实施本公开的实施例的示例电子设备300的示意性框图。电子设备旨在表示各种形式的数字计算机，诸如，膝上型计算机、台式计算机、工作台、个人数字助理、服务器、刀片式服务器、大型计算机、和其他适合的计算机。电子设备还可以表示各种形式的移动装置，诸如，个人数字处理、蜂窝电话、智能电话、可穿戴设备和其他类似的计算装置。本文所示的部件、它们的连接和关系、以及它们的功能仅仅作为示例，并且不意在限制本文中描述的和/或者要求的本公开的实现。
如图6所示，设备300包括计算单元301，其可以根据存储在只读存储器(ROM)302中的计算机程序或者从存储单元308加载到随机访问存储器(RAM)303中的计算机程序，来执行各种适当的动作和处理。在RAM 303中，还可存储设备300操作所需的各种程序和数据。计算单元301、ROM 302以及RAM 303通过总线304彼此相连。输入/输出(I/O)接口307也连接至 总线304。
设备300中的多个部件连接至I/O接口306，包括：输入单元306，例如键盘、鼠标等；输出单元307，例如各种类型的显示器、扬声器等；存储单元308，例如磁盘、光盘等；以及通信单元309，例如网卡、调制解调器、无线通信收发机等。通信单元309允许设备300通过诸如因特网的计算机网络和/或各种电信网络与其他设备交换信息/数据。
计算单元301可以是各种具有处理和计算能力的通用和/或专用处理组件。计算单元301的一些示例包括但不限于中央处理单元(CPU)、图形处理单元(GPU)、各种专用的人工智能(AI)计算芯片、各种运行机器学习模型算法的计算单元、数字信号处理器(DSP)、以及任何适当的处理器、控制器、微控制器等。计算单元301执行上文所描述的各个方法和处理，例如方法通信接收方法。例如，在一些实施例中，方法通信接收方法可被实现为计算机软件程序，其被有形地包含于机器可读介质，例如存储单元307。在一些实施例中，计算机程序的部分或者全部可以经由ROM 302和/或通信单元309而被载入和/或安装到设备300上。当计算机程序加载到RAM 303并由计算单元301执行时，可以执行上文描述的方法通信接收方法的一个或多个步骤。备选地，在其他实施例中，计算单元301可以通过其他任何适当的方式(例如，借助于固件)而被配置为执行方法通信接收方法。
本文中以上描述的系统和技术的各种实施方式可以在数字电子电路系统、集成电路系统、场可编程门阵列(FPGA)、专用集成电路(ASIC)、专用标准产品(ASSP)、芯片上系统的系统(SOC)、负载可编程逻辑设备(CPLD)、计算机硬件、固件、软件、和/或它们的组合中实现。这些各种实施方式可以包括：实施在一个或者多个计算机程序中，该一个或者多个计算机程序可在包括至少一个可编程处理器的可编程系统上执行和/或解释，该可编程处理器可以是专用或者通用可编程处理器，可以从存储系统、至少一个输入装置、 和至少一个输出装置接收数据和指令，并且将数据和指令传输至该存储系统、该至少一个输入装置、和该至少一个输出装置。
用于实施本公开的方法的程序代码可以采用一个或多个编程语言的任何组合来编写。这些程序代码可以提供给通用计算机、专用计算机或其他可编程数据处理装置的处理器或控制器，使得程序代码当由处理器或控制器执行时使流程图和/或框图中所规定的功能/操作被实施。程序代码可以完全在机器上执行、部分地在机器上执行，作为独立软件包部分地在机器上执行且部分地在远程机器上执行或完全在远程机器或服务器上执行。
在本公开的上下文中，机器可读介质可以是有形的介质，其可以包含或存储以供指令执行系统、装置或设备使用或与指令执行系统、装置或设备结合地使用的程序。机器可读介质可以是机器可读信号介质或机器可读储存介质。机器可读介质可以包括但不限于电子的、磁性的、光学的、电磁的、红外的、或半导体系统、装置或设备，或者上述内容的任何合适组合。机器可读存储介质的更具体示例会包括基于一个或多个线的电气连接、便携式计算机盘、硬盘、随机存取存储器(RAM)、只读存储器(ROM)、可擦除可编程只读存储器(EPROM或快闪存储器)、光纤、便捷式紧凑盘只读存储器(CD-ROM)、光学储存设备、磁储存设备、或上述内容的任何合适组合。
为了提供与用户的交互，可以在计算机上实施此处描述的系统和技术，该计算机具有：用于向用户显示信息的显示装置(例如，CRT(阴极射线管)或者LCD(液晶显示器)监视器)；以及键盘和指向装置(例如，鼠标或者轨迹球)，用户可以通过该键盘和该指向装置来将输入提供给计算机。其它种类的装置还可以用于提供与用户的交互；例如，提供给用户的反馈可以是任何形式的传感反馈(例如，视觉反馈、听觉反馈、或者触觉反馈)；并且可以用任何形式(包括声输入、语音输入或者、触觉输入)来接收来自用户的输入。
可以将此处描述的系统和技术实施在包括后台部件的计算系统(例如，作为数据服务器)、或者包括中间件部件的计算系统(例如，应用服务器)、或者包括前端部件的计算系统(例如，具有图形用户界面或者网络浏览器的用户计算机，用户可以通过该图形用户界面或者该网络浏览器来与此处描述的系统和技术的实施方式交互)、或者包括这种后台部件、中间件部件、或者前端部件的任何组合的计算系统中。可以通过任何形式或者介质的数字数据通信(例如，通信网络)来将系统的部件相互连接。通信网络的示例包括：局域网(LAN)、广域网(WAN)和互联网。
计算机系统可以包括客户端和服务器。客户端和服务器一般远离彼此并且通常通过通信网络进行交互。通过在相应的计算机上运行并且彼此具有客户端-服务器关系的计算机程序来产生客户端和服务器的关系。服务器可以是云服务器，也可以为分布式系统的服务器，或者是结合了区块链的服务器。
应该理解，可以使用上面所示的各种形式的流程，重新排序、增加或删除步骤。例如，本发公开中记载的各步骤可以并行地执行也可以顺序地执行也可以不同的次序执行，只要能够实现本公开公开的技术方案所期望的结果，本文在此不进行限制。
上述具体实施方式，并不构成对本公开保护范围的限制。本领域技术人员应该明白的是，根据设计要求和其他因素，可以进行各种修改、组合、子组合和替代。任何在本公开的精神和原则之内所作的修改、等同替换和改进等，均应包含在本公开保护范围之内。

Claims (11)

  1. 一种通信接收方法，包括：
    接收由信道传输的比特流；
    对所述比特流进行解析，以得到语义信息和语义解码器参数；
    根据所述语义解码器参数调用接收端的语义解码器；
    将所述语义信息输入到所述语义解码器中进行数据还原，以得到源信息。
  2. 如权利要求1所述的通信接收方法，其特征在于，所述对所述比特流进行解析，以得到语义信息和语义解码器参数，包括：
    基于信道解码器对所述比特流进行解析，以得到信道解码结果；
    根据预定义解析规则对所述信道解码结果进行解析，以得到语义信息和语义解码器参数。
  3. 如权利要求1所述的通信接收方法，其特征在于，所述根据所述语义解码器参数完善语义解码器，包括：
    所述语义解码器在信源端具有对应的语义编码器；
    所述语义解码器参数是信源端的语义解码器在与信源端的语义编码器对抗训练后得到的参数。
  4. 如权利要求3所述的通信接收方法，其特征在于，所述语义编码器包括：
    任意卷积层和编码全连接层；其中，将源信息输入到所述语义编码器中，所述编码全连接层能够输出语义信息；
    所述语义解码器包括：
    解码全连接层、任意反卷积层；其中，将语义信息输入到所述语义解码器中，所述最终解码反卷积层能够输出源信息；
    其中，所述任意编码层与所述任意解码层相对应。
  5. 如权利要求1所述的通信接收方法，其特征在于，当所述源信息的数据类型包括自然语言或图像或视频。
  6. 如权利要求5所述的通信接收方法，其特征在于，当所述源信息的数据类型为自然语言时，所述语义解码器为处理自然语 言的模型；
    所述语义解码器包括残差网络或自然语言处理。
  7. 如权利要求5所述的通信接收方法，其特征在于，当所述源信息的数据类型为图像或视频时，所述语义解码器为处理图像或视频的模型；
    所述语义解码器包括以特征图为基的神经网络模型或自编码器。
  8. 如权利要求7所述的通信接收方法，其特征在于，所述根据所述语义解码器参数调用接收端的语义解码器,之后包括：
    将所述语义解码器参数写入所述语义解码器。
  9. 一种通信接收装置，包括：
    接收模块，接收由信道传输的比特流；
    解析模块，对所述比特流进行解析，以得到语义信息和语义解码器参数；
    调用模块，根据所述语义解码器参数调用所述装置的语义解码器；
    还原模块，将所述语义信息输入到所述语义解码器中进行数据还原，以得到源信息。
  10. 一种电子设备，包括：
    至少一个处理器；以及
    与所述至少一个处理器通信连接的存储器；其中，
    所述存储器存储有可被所述至少一个处理器执行的指令，所述指令被所述至少一个处理器执行，以使所述至少一个处理器能够执行权利要求1-8中任一项所述的方法。
  11. 一种存储有计算机指令的非瞬时计算机可读存储介质，其中，所述计算机指令用于使所述计算机执行根据权利要求1-8中任一项所述的方法。
PCT/CN2023/088851 2022-03-21 2023-04-18 通信接收方法及其装置 WO2023179800A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210279128.0A CN116863949A (zh) 2022-03-21 2022-03-21 通信接收方法及其装置
CN202210279128.0 2022-03-21

Publications (1)

Publication Number Publication Date
WO2023179800A1 true WO2023179800A1 (zh) 2023-09-28

Family

ID=88100033

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/088851 WO2023179800A1 (zh) 2022-03-21 2023-04-18 通信接收方法及其装置

Country Status (2)

Country Link
CN (1) CN116863949A (zh)
WO (1) WO2023179800A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117614584A (zh) * 2023-10-26 2024-02-27 北京邮电大学 信道可迁移的语义通信方法以及相关设备

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20070010984A (ko) * 2005-07-20 2007-01-24 주식회사 휴맥스 비트스트림 인코딩/디코딩 방법 및 장치
WO2020139770A1 (en) * 2018-12-27 2020-07-02 Telepathy Labs, Inc. System and method for communicating with a user with speech processing
CN112800247A (zh) * 2021-04-09 2021-05-14 华中科技大学 基于知识图谱共享的语义编/解码方法、设备和通信系统
CN113379040A (zh) * 2021-07-07 2021-09-10 东南大学 基于语义编码的混合重传方法
CN113705245A (zh) * 2021-09-01 2021-11-26 北京邮电大学 语义通信方法、装置、系统、计算机设备及存储介质
CN113839750A (zh) * 2021-11-25 2021-12-24 之江实验室 一种语义通信系统中的信息传输方法
CN114091673A (zh) * 2022-01-19 2022-02-25 华中科技大学 语义信息表示方法、语义编码器、解码器和语义通信框架

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20070010984A (ko) * 2005-07-20 2007-01-24 주식회사 휴맥스 비트스트림 인코딩/디코딩 방법 및 장치
WO2020139770A1 (en) * 2018-12-27 2020-07-02 Telepathy Labs, Inc. System and method for communicating with a user with speech processing
CN112800247A (zh) * 2021-04-09 2021-05-14 华中科技大学 基于知识图谱共享的语义编/解码方法、设备和通信系统
CN113379040A (zh) * 2021-07-07 2021-09-10 东南大学 基于语义编码的混合重传方法
CN113705245A (zh) * 2021-09-01 2021-11-26 北京邮电大学 语义通信方法、装置、系统、计算机设备及存储介质
CN113839750A (zh) * 2021-11-25 2021-12-24 之江实验室 一种语义通信系统中的信息传输方法
CN114091673A (zh) * 2022-01-19 2022-02-25 华中科技大学 语义信息表示方法、语义编码器、解码器和语义通信框架

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117614584A (zh) * 2023-10-26 2024-02-27 北京邮电大学 信道可迁移的语义通信方法以及相关设备

Also Published As

Publication number Publication date
CN116863949A (zh) 2023-10-10

Similar Documents

Publication Publication Date Title
US9787321B1 (en) Point cloud data compression using a space-filling curve
US10341048B2 (en) Channel encoding and decoding method and apparatus
Shi et al. A new communication paradigm: from bit accuracy to semantic fidelity
US9223765B1 (en) Encoding and decoding data using context model grouping
WO2023179800A1 (zh) 通信接收方法及其装置
US20220394791A1 (en) Method and apparatus for vehicle-machine interconnection
US11538197B2 (en) Channel-wise autoregressive entropy models for image compression
CN110518917A (zh) 基于Huffman编码的LZW数据压缩方法及系统
CN113889076B (zh) 语音识别及编解码方法、装置、电子设备及存储介质
CN114445831A (zh) 一种图文预训练方法、装置、设备以及存储介质
CN113963359B (zh) 文本识别模型训练方法、文本识别方法、装置及电子设备
US20140185950A1 (en) Progressive entropy encoding
CN112784897A (zh) 图像处理方法、装置、设备和存储介质
WO2023066183A1 (zh) 数据处理方法、装置、设备以及存储介质
CN106851733A (zh) 一种针对移动网络应用的自适应http消息压缩方法
US10248666B2 (en) Creation of hierarchical dictionary
CN113689868A (zh) 一种语音转换模型的训练方法、装置、电子设备及介质
CN114614829A (zh) 卫星数据帧的处理方法、装置、电子设备和可读存储介质
WO2014029081A1 (zh) 压缩方法及设备
CN115314121B (zh) 量子通信方法、装置及电子设备
CN110663062B (zh) 三角网格压缩中的引导遍历
WO2024045672A1 (zh) 编码方法、解码方法、编码装置、解码装置及电子设备
CN114461816A (zh) 基于知识图谱的信息补充语义通信系统的实现方法
WO2023138238A1 (zh) 基于智简网络的信息发送方法、装置、电子设备及介质
CN104113394B (zh) 通信调制信号的压缩及解压方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23774043

Country of ref document: EP

Kind code of ref document: A1