HARQ Communication Method, Apparatus, Electronic Device, and Storage Medium for Semantic Awareness

Abstract

The present application provides a semantic awareness HARQ communication method, apparatus, electronic device, and storage medium. By the receiving end performing retransmission detection on multiple received transmission semantic features, when it is determined that the retransmission detection result fails the detection, the distortion degree evaluation data generated by the semantic feature distortion evaluation network model is fed back to the sending end for the sending end to retransmit the identified distorted transmission semantic features, so that the multiple transmission semantic features received at the receiving end can meet a certain accuracy. Furthermore, by using the prediction algorithms of at least one preset type of semantic task to perform prediction processing on the decoding features corresponding to the multiple transmission semantic features, relatively accurate at least one prediction processing result can be obtained, thus ensuring the accuracy of the executed semantic tasks.

Description

Technical Field

[0001] The present application relates to the field of communication technologies, and in particular, to a semantic awareness HARQ communication method, apparatus, electronic device, and storage medium. Background Art

[0002] Semantic communication is a way of conveying meaning different from syntactic communication, aiming to transmit the meaning of information and improve communication efficiency, and can meet the needs of intelligent communication. However, affected by the noisy transmission channel, the accuracy of the transmitted information received at the receiving end will be reduced, thus affecting the accuracy of the semantic tasks executed in the corresponding application scenarios. Summary of the Invention

[0003] In view of this, the purpose of the present application is to propose a semantic awareness HARQ communication method, apparatus, electronic device, and storage medium to solve the above technical problems.

[0004] Based on the above purpose, the first aspect of the present application provides a semantic awareness HARQ communication method, which is applied to a semantic awareness HARQ communication system. The system includes a sending end and a receiving end. The method includes:

[0005] The sending end obtains a transmission information source data set, generates a plurality of transmission semantic features based on the transmission information source data set, and sends the plurality of transmission semantic features to the receiving end;

[0006] The receiving end performs retransmission detection based on the received plurality of transmission semantic features to obtain a retransmission detection result;

[0007] In response to the receiving end determining that the retransmission detection result fails the detection, the receiving end inputs the received plurality of transmission semantic features into a pre-trained semantic feature distortion evaluation network model to obtain distortion degree evaluation data, and converts the distortion degree evaluation data into a binary vector and transmits it to the sending end, so that the sending end determines the transmission semantic features to be retransmitted based on the binary vector and resends them to the receiving end. At the same time, when it is determined that the transmission channel meets the preset transmission conditions, a new plurality of transmission semantic features are incrementally transmitted to the receiving end, and the receiving end repeats the retransmission detection process based on the received plurality of transmission semantic features until it is determined that the retransmission detection result passes the detection, or the preset maximum retransmission times are reached; or,

[0008] In response to the receiving end determining that the retransmission detection result passes the detection, the receiving end determines decoding features corresponding to at least one preset type of semantic task based on the received plurality of transmission semantic features, and performs prediction processing on the decoding features using the prediction algorithms of at least one preset type of semantic task to obtain at least one prediction processing result.

[0009] Optionally, sending the multiple transmission semantic features to the receiving end includes:

[0010] The sending end obtains the importance degree values corresponding to the multiple transmission semantic features;

[0011] The sending end determines the multiple transmission semantic features to be transmitted in each transmission process according to the importance degree values of the features, and sends the determined multiple transmission semantic features to be transmitted to the receiving end in each transmission process.

[0012] Optionally, the receiving end performs retransmission detection based on the received multiple transmission semantic features to obtain a retransmission detection result, including:

[0013] The receiving end determines the received signal-to-noise ratio of the received multiple transmission semantic features and the cascaded semantic features corresponding to the received multiple transmission semantic features;

[0014] The receiving end determines the value of the uncertainty corresponding to the received multiple transmission semantic features based on the cascaded semantic features through an uncertainty algorithm;

[0015] The receiving end converts the value of the uncertainty into a comparison value corresponding to the received signal-to-noise ratio by using a preset conversion coefficient;

[0016] The receiving end compares the received signal-to-noise ratio and the comparison value to obtain a comparison result, and determines the retransmission detection result according to the comparison result.

[0017] Optionally, determining the retransmission detection result according to the comparison result includes:

[0018] In response to the receiving end determining that the comparison result is that the received signal-to-noise ratio is less than or equal to the comparison value, it is determined that the retransmission detection result fails the detection; or,

[0019] In response to the receiving end determining that the comparison result is that the received signal-to-noise ratio is greater than the comparison value, it is determined that the retransmission detection result passes the detection.

[0020] Optionally, the training process of the semantic feature distortion evaluation network model includes:

[0021] The receiving end obtains the original semantic features for training and the received semantic features for training corresponding to the original semantic features for training, and constructs an initial model, where the initial model includes an initial distortion measurement module and a gradient estimator module;

[0022] The receiving end inputs the original semantic features for training and the received semantic features for training into the initial distortion metric module, and the initial distortion metric module outputs a first contribution degree parameter and a second contribution degree parameter;

[0023] The receiving end constructs a training objective function based on the first contribution degree parameter and the second contribution degree parameter, and trains and adjusts the initial distortion metric module with the maximization result obtained during the maximization process of the training objective function to obtain a trained initial distortion metric module;

[0024] The receiving end obtains the received semantic features for training, and performs backpropagation training on the trained initial distortion metric module based on the received semantic features for training through the gradient estimator module to obtain the semantic feature distortion evaluation network model.

[0025] Optionally, the sending end determines the transmission semantic features to be retransmitted based on the binary vector, including:

[0026] The sending end searches for the transmission semantic features with the binary vector being a first preset value from each transmission semantic feature, and takes the transmission semantic features with the binary vector being the first preset value as the transmission semantic features to be retransmitted.

[0027] Optionally, after the receiving end determines that the retransmission detection result fails the detection, the method further includes:

[0028] The receiving end inputs the received multiple transmission semantic features into a pre-trained semantic feature distortion evaluation network model to obtain distortion degree evaluation data, and converts the distortion degree evaluation data into a binary vector and transmits it to the sending end, so that the sending end searches for the transmission semantic features with the binary vector being a second preset value from each transmission semantic feature, and takes the transmission semantic features with the binary vector being the second preset value as the transmission semantic features not to be retransmitted.

[0029] Based on the same inventive concept, a second aspect of the present application provides a semantic awareness HARQ communication device, the device is arranged in a semantic awareness HARQ communication system, the system includes a sending end and a receiving end, and the device includes:

[0030] The sending end is configured to obtain a transmission source data set, generate multiple transmission semantic features based on the transmission source data set, and send the multiple transmission semantic features to the receiving end;

[0031] The receiving end is configured to perform retransmission detection based on multiple received transmission semantic features to obtain a retransmission detection result; in response to determining that the retransmission detection result fails the detection, the receiving end inputs the multiple received transmission semantic features into a pre-trained semantic feature distortion evaluation network model to obtain distortion degree evaluation data, and converts the distortion degree evaluation data into a binary vector and transmits it to the sending end, so that the sending end determines the transmission semantic features to be retransmitted based on the binary vector and retransmits them to the receiving end. At the same time, when it is determined that the transmission channel meets the preset transmission conditions, a new set of multiple transmission semantic features is incrementally transmitted to the receiving end, and the receiving end repeats the retransmission detection process based on the received multiple transmission semantic features until it is determined that the retransmission detection result passes the detection or the preset maximum retransmission times are reached; or, in response to determining that the retransmission detection result passes the detection, the receiving end determines decoding features corresponding to at least one preset type of semantic task based on the received multiple transmission semantic features, and uses the prediction algorithms of at least one preset type of semantic task to perform prediction processing on the decoding features to obtain at least one prediction processing result.

[0032] Based on the same inventive concept, a third aspect of the present application provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable by the processor. When the processor executes the computer program, the method described in the first aspect above is implemented.

[0033] Based on the same inventive concept, a fourth aspect of the present application provides a non-transitory computer-readable storage medium storing computer instructions for causing a computer to execute the method described in the first aspect above.

[0034] As can be seen from the above, for the semantic awareness HARQ communication method, device, electronic device, and storage medium provided by the present application, the receiving end performs retransmission detection on multiple received transmission semantic features. When it is determined that the retransmission detection result fails the detection, on this basis, the distortion degree evaluation data generated by the semantic feature distortion evaluation network model is fed back to the sending end for the sending end to retransmit the identified distorted transmission semantic features to be retransmitted, so that the multiple transmission semantic features received at the receiving end can meet a certain accuracy. Furthermore, the prediction algorithms of at least one preset type of semantic task are used to perform prediction processing on the decoding features corresponding to the multiple transmission semantic features, and relatively accurate at least one prediction processing result can be obtained, thus ensuring the accuracy of the executed semantic task. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] To more clearly illustrate the technical solutions in the present application or related technologies, the following will briefly introduce the accompanying drawings required for use in the description of the embodiments or related technologies. Obviously, the accompanying drawings in the following description are only the embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can also be obtained based on these drawings.

[0036] Figure 1A Schematic diagram of the framework of the semantic awareness HARQ communication system according to an embodiment of the present application;

[0037] Figure 1B Flowchart of the semantic awareness HARQ communication method according to an embodiment of the present application;

[0038] Figure 1C Flowchart of the semantic awareness HARQ mechanism according to an embodiment of the present application;

[0039] Figure 1D Schematic diagram of the training process of the semantic feature distortion evaluation network model according to an embodiment of the present application;

[0040] Figure 1E Schematic diagram of the comparison of vehicle re-identification simulation performance according to an embodiment of the present application;

[0041] Figure 1F Schematic diagram of the comparison of vehicle color classification simulation performance according to an embodiment of the present application;

[0042] Figure 1G Schematic diagram of the comparison of vehicle type classification simulation performance according to an embodiment of the present application;

[0043] Figure 2 Structural block diagram of the semantic awareness HARQ communication device according to an embodiment of the present application;

[0044] Figure 3 Schematic diagram of the electronic device according to an embodiment of the present application. Detailed implementation manners

[0045] To make the objectives, technical solutions, and advantages of the present application more clear and understandable, the following further elaborates on the present application in detail with reference to specific embodiments and the accompanying drawings.

[0046] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the embodiments of this application should have the ordinary meanings understood by those of ordinary skill in the field to which this application belongs. The "first", "second" and similar terms used in the embodiments of this application do not denote any order, quantity or importance, but are only used to distinguish different components. Words such as "including" or "comprising" mean that the elements or objects appearing before this word cover the elements or objects listed after this word and their equivalents, without excluding other elements or objects. Words such as "connected" or "coupled" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Upper", "lower", "left", "right", etc. are only used to represent relative position relationships, and when the absolute position of the object being described changes, the relative position relationship may also change accordingly.

[0047] It can be understood that before using the technical solutions of the various embodiments of this application, the types, usage scopes, usage scenarios, etc. of the personal information involved will be informed to the user in an appropriate manner, and the user's authorization will be obtained.

[0048] For example, when responding to receiving an active request from the user, a prompt message is sent to the user to clearly prompt the user that the operation requested by the user will require obtaining and using the user's personal information. Thus, the user can autonomously choose whether to provide personal information to software or hardware such as an electronic device, an application program, a server, or a storage medium that performs the operations of the technical solutions of this application according to the prompt message.

[0049] As an optional but non-limiting implementation manner, the way of sending a prompt message to the user in response to receiving an active request from the user can be, for example, in the form of a pop-up window, and the prompt message can be presented in text in the pop-up window. In addition, the pop-up window can also carry a selection control for the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0050] It can be understood that the above process of notifying and obtaining the user's authorization is only illustrative and does not limit the implementation manner of this application. Other ways that meet relevant laws and regulations can also be applied to the implementation manner of this application.

[0051] In the related art, a semantic communication system based on incremental transmission, a retransmission system aware of data importance, a wireless communication transmission (Hybrid Automatic Repeat Request, HARQ) semantic communication framework considering incremental transmission, and a general implementation method of a semantic communication system are proposed.

[0052] Among them, for the semantic communication system based on incremental transmission: The system transmits the encoded semantic features at a preset code rate. When the upper limit of the preset transmission times is not reached or the semantic task is not completed, the receiving end feeds back a negative acknowledgement (NACK) to the sending end. At this time, the sending end does not consider the specific distortion situation of the semantic features, but only sends new incremental semantic features at a certain code rate until the upper limit of the transmission times is reached. Finally, the semantic features obtained from multiple transmissions are spliced and then input into the decoder to execute the intelligent semantic task. However, each time the system transmits, it only incrementally transmits new semantic features, without considering the situation that the received semantic features may not be successfully decoded and need to be retransmitted, resulting in low utilization and flexibility of communication resources. In addition, each transmission of the system corresponds to different codecs, resulting in a complex system structure and the need for retraining each time, which increases the training complexity.

[0053] In addition, for the retransmission system with data importance awareness: After the information is transmitted to the receiving end, it is judged whether retransmission is needed according to the uncertainty of the received information. If the amount of received information does not reach the threshold, the sending end is notified to retransmit the data, and each transmission is a repeated transmission. The system repeatedly transmits the received information and cannot identify the information that has been successfully decoded, which may cause waste of communication resources. At the same time, it does not consider transmitting incremental information, resulting in limited semantic information received and low efficiency in implementing semantic tasks.

[0054] Regarding the wireless communication transmission (Hybrid Automatic Repeat Request, HARQ) semantic communication framework that considers incremental transmission: This framework designs multiple pairs of codecs for multiple transmissions, and each transmission requires re-training. At the same time, Comparative Document 3 refers to the traditional Cyclic Redundancy Check (CRC) algorithm and proposes the Sim32 retransmission detection algorithm. Specifically, the sender adds 32-bit redundant information when encoding semantic features, and at the receiver, an additional Sim32 decoder is used to jointly decode the semantic features and the redundant information, and outputs the value (0 or 1) of the positive feedback (Acknowledgement, ACK), which is used as the basis for whether to re-transmit the incremental information. However, each transmission in this framework corresponds to a different codec, resulting in a complex system structure, and each transmission requires re-training, increasing the training complexity. And each time only the new semantic features are incrementally transmitted, without considering the situation that the received semantic features may not be successfully decoded and need to be re-transmitted, resulting in low communication resource utilization and flexibility. Moreover, the Sim32 retransmission detection algorithm in it needs to add redundant information after the semantic features, increasing the transmission burden, and designing an additional network is required to perform retransmission detection using the redundant information, increasing the system complexity. In addition, this detection algorithm can only output an instruction on whether retransmission is needed and cannot detect the specific distortion situation of the semantic features, with weak error detection ability.

[0055] In addition, for the implementation method of a general semantic communication system: After obtaining the semantic task requirements, constant bitrate coding transmission is adopted to transmit all semantic features without discrimination, and then directly decoded at the receiver, without further considering the retransmission gain. And this method cannot control the transmission bitrate according to the communication conditions, with low communication efficiency, and cannot select semantic features highly relevant to the semantic task for transmission at the sender, and cannot achieve elastic coding highly coupled with the semantic task. The receiver only sets a decoder to perform error correction and detection on the received semantic features, without considering the role of HARQ transmission in improving the system's error correction and detection ability, and without utilizing the performance gain brought by retransmission.

[0056] Compared with the semantic communication system based on incremental transmission and the wireless communication transmission (Hybrid Automatic Repeat Request, HARQ) semantic communication framework that considers incremental transmission in the related art, the semantic-aware HARQ communication system built in this application combines feature retransmission and incremental transmission in the order of semantic importance, re-transmits the distorted features while supplemented by incremental transmission, rather than simply continuously transmitting incremental information, improving the reception accuracy of important features, and thus improving the multi-task performance.

[0057] Compared with the retransmission system with data importance awareness in the related art, in this application, a semantic awareness HARQ mechanism is designed. Different from the retransmission system with data importance awareness in the related art that retransmits all features, in this application, instead of directly retransmitting all features, features with severe distortion that truly need to be retransmitted are selected according to the distortion conditions of different features, greatly improving the retransmission efficiency.

[0058] In addition, compared with the implementation method of the general semantic communication system in the related art, in this application, elastic retransmission is performed on the distorted features. Different from the implementation method of the general semantic communication system in the related art that does not consider the gain of retransmission, in this application, the error correction and detection capabilities and transmission efficiency of the system are greatly improved, thus achieving better task performance.

[0059] An embodiment of this application provides a semantic awareness HARQ communication method. The receiving end performs retransmission detection on multiple received transmission semantic features. When it is determined that the retransmission detection result fails the detection, on this basis, the distortion degree evaluation data generated by the semantic feature distortion evaluation network model is fed back to the sending end for the sending end to retransmit the identified distorted transmission semantic features to be retransmitted, so that the multiple received transmission semantic features that pass the detection at the receiving end can meet a certain accuracy. Furthermore, by using the prediction algorithm of at least one preset type of semantic task to perform prediction processing on the decoded features corresponding to the multiple transmission semantic features, relatively accurate at least one prediction processing result can be obtained, thus ensuring the accuracy of the executed semantic task.

[0060] The method of this embodiment is applied to a semantic awareness HARQ communication system, such as Figure 1A shows a schematic framework diagram of the semantic awareness HARQ communication system provided by the embodiment of this application. The system includes a sending end and a receiving end. The sending end is composed of a multi-task joint semantic encoder, a source-channel joint encoder, a HARQ system elastic feature selector, and a normalization layer. The receiving end is composed of a retransmission detection module, a source-source joint decoder, semantic decoders corresponding to multiple tasks, and a task executor. The method of this application can be applied to multiple semantic communication scenarios with limited channel resources, such as intelligent transportation scenarios, smart home scenarios, vehicle-to-everything application scenarios, etc., and can support the transmission of various modal source data such as images, texts, audio and video, and solve various semantic tasks such as classification, recognition, and prediction.

[0061] As Figure 1B shown, the method of this embodiment includes:

[0062] Step 101, the sending end obtains a transmission source data set, generates multiple transmission semantic features based on the transmission source data set, and sends the multiple transmission semantic features to the receiving end.

[0063] In this step, as Figure 1A shown, taking the vehicle pictures in the intelligent transportation scenario as an example, the sender obtains the input transmission source data set. At this time, the transmission source data set is the vehicle picture set x. After the multi-task joint semantic encoder extracts the source semantic features to obtain the joint transmission semantic features s, and then through the source-channel joint encoder for channel coding to obtain multiple transmission semantic features f. Then, through the HARQ system elastic feature selector, combined with the feature importance value, elastic feature transmission under the guidance of channel capacity is performed. When performing retransmission and incremental transmission, important features are selected for priority transmission to obtain multiple selected transmission semantic features After passing through the normalization layer for power control, multiple transmission semantic features z that need to be transmitted finally are obtained j , and they are sent to the receiver through the channel, where j is the number of retransmission times.

[0064] Among them, the source semantic feature extraction can be regarded as a source transformation, that is, transforming the transmission source data set from the original attribute space to the feature attribute space. The source semantic feature extraction is the basis of semantic communication. It uses a deep neural network to combine low-level features to form more abstract high-level semantic features, improving the source compression efficiency while retaining the source semantic meaning.

[0065] Step 102, the receiver performs retransmission detection based on the received multiple transmission semantic features to obtain a retransmission detection result.

[0066] In this step, as Figure 1A shown, the receiver uses the retransmission detection module to perform retransmission detection on the received multiple transmission semantic features to obtain a retransmission detection result to determine whether to trigger the next retransmission.

[0067] Step 103, the receiver determines whether the retransmission detection result is a pass.

[0068] In this step, as Figure 1A shown, the receiver quickly determines whether the multiple transmission semantic features received by the receiver pass the detection based on the retransmission detection result obtained by the retransmission detection module.

[0069] Step 104, in response to the receiver determining that the retransmission detection result fails the detection, the receiver inputs the received multiple transmission semantic features into a pre-trained semantic feature distortion evaluation network model to obtain distortion evaluation data, and converts the distortion evaluation data into a binary vector and transmits it to the sender.

[0070] In this step, as Figure 1A shown, when the receiver determines that the retransmission detection result fails the detection, the received multiple transmission semantic features Input semantic feature distortion evaluation network model (FDE network), and use the semantic feature distortion evaluation network model to evaluate multiple received transmission semantic features for distortion measurement to obtain distortion degree evaluation data, and on this basis, implement HARQ for semantic distortion awareness.

[0071] Step 105, the sending end determines the transmission semantic features to be retransmitted based on the binary vector, and at the same time, when it is determined that the transmission channel meets the preset transmission conditions, incrementally transmit multiple new transmission semantic features to the receiving end.

[0072] In this step, the sending end retransmits the identified distorted features (i.e., the transmission semantic features to be retransmitted) according to the distortion degree evaluation data p generated by the semantic feature distortion evaluation network model, and at the same time, when the transmission channel meets the preset transmission conditions, incrementally transmit multiple new transmission semantic features to the receiving end, so that the multiple received transmission semantic features detected at the receiving end can meet a certain accuracy. j Among them, the preset transmission conditions indicate the transmission performance of the transmission channel, and the preset transmission conditions can be that there is surplus in the channel conditions (such as bandwidth) of the transmission channel.

[0073] Step 106, in response to the receiving end determining that the retransmission detection result is passed, the receiving end determines decoding features corresponding to at least one preset type of semantic task based on the multiple received transmission semantic features, and uses the prediction algorithms of at least one preset type of semantic task to perform prediction processing on the decoding features to obtain at least one prediction processing result.

[0074] In this step, as

[0075] shown, when the receiving end determines that the retransmission detection result is passed, then based on the multiple retransmitted transmission semantic features received Figure 1A cascade all the feature vectors of multiple transmissions cached at the receiving end as decoded by the source-channel joint decoder as restored and distributed by multiple semantic decoders corresponding to multiple tasks as where K is the total number of multiple tasks. Finally, after passing through K task executors, each task executor corresponds to the prediction algorithm of the preset type of semantic task, and generates K probability prediction vectors c1,..., c (i.e., the prediction processing results) based on the prediction algorithms of the preset type of semantic tasks corresponding to the K task executors, which are used to execute different intelligent multi-semantic tasks, such as vehicle re-identification, vehicle color classification, etc., to ensure the accuracy of the executed semantic tasks. K (i.e., the prediction processing results), which are used to execute different intelligent multi-semantic tasks, such as vehicle re-identification, vehicle color classification, etc., to ensure the accuracy of the executed semantic tasks.

[0076] Through the above solution, the receiving end performs retransmission detection on multiple received transmission semantic features. When it is determined that the retransmission detection result fails the detection, on this basis, the distortion degree evaluation data generated by the semantic feature distortion evaluation network model is fed back to the sending end, so that the sending end can retransmit the identified distorted transmission semantic features to be retransmitted, enabling the multiple received transmission semantic features that pass the detection at the receiving end to meet a certain accuracy. Furthermore, by using the prediction algorithms of at least one preset type of semantic task to perform prediction processing on the decoding features corresponding to the multiple transmission semantic features, relatively accurate at least one prediction processing result can be obtained, thus ensuring the accuracy of the executed semantic task.

[0077] In some embodiments, in step 101, sending the multiple transmission semantic features to the receiving end includes:

[0078] Step A1, the sending end obtains the feature importance degree values corresponding to the multiple transmission semantic features.

[0079] Step A2, the sending end determines the multiple transmission semantic features to be transmitted in each transmission process according to the feature importance degree values, and sends the determined multiple transmission semantic features to be transmitted to the receiving end in each transmission process.

[0080] In the above solution, as Figure 1A shown, the HARQ system elastic feature selector in the sending end of the present application aims to improve the reliability and efficiency of data transmission through the transmission technology (Hybrid Automatic Repeat Request, HARQ) of the wireless communication system. By combining Automatic Repeat Request (ARQ) and incremental transmission, HARQ performs error detection and retransmission during data transmission to reduce the error rate. The present application adopts a method that combines information retransmission and incremental transmission, increasing the transmission efficiency while improving the system robustness.

[0081] The HARQ system elastic feature selector performs elastic feature transmission under the guidance of channel capacity in combination with the feature importance degree values. When performing retransmission and incremental transmission, it preferentially transmits important features, and when selecting features for transmission, it selects the most important multiple features according to the channel conditions and sends them to the receiving end.

[0082] In some embodiments, step 102 includes:

[0083] Step B1, the receiving end determines the received signal-to-noise ratio of the multiple received transmission semantic features, and the cascaded semantic features corresponding to the multiple received transmission semantic features.

[0084] Step B2, the receiving end determines the uncertainty value corresponding to the received multiple transmission semantic features based on the cascaded semantic features through an uncertainty algorithm.

[0085] Step B3, the receiving end converts the uncertainty value into a comparison value corresponding to the received signal-to-noise ratio by using a preset conversion coefficient.

[0086] Step B4, the receiving end compares the received signal-to-noise ratio and the comparison value to obtain a comparison result, and determines the retransmission detection result according to the comparison result.

[0087] In the above solution, as shown in Figure 1A and Figure 1C , first, after multiple transmission semantic features z j are transmitted through the channel to the receiving end, a retransmission detection module is used to perform retransmission detection on the received multiple transmission semantic features to decide whether to trigger the next retransmission.

[0088] This application makes a retransmission decision using the uncertainty corresponding to the received multiple transmission semantic features . Since the received signal-to-noise ratio increases with the increase of the number of transmissions, when the received signal-to-noise ratio is small, the system needs to trigger retransmission until it can compensate for the uncertainty of the multiple transmission semantic features received by the receiving end. Specifically, for the concatenated feature vector after the jth transmission (i.e., the cascaded semantic feature), the receiving end continuously triggers retransmission until as follows:

[0089]

[0090] where is the received signal-to-noise ratio, represents the comparison value, is the uncertainty value, L(·) is a monotonically increasing function, this application uses L(x) = 1 + x, θ0 is the conversion coefficient between the uncertainty value and the received SNR (i.e., the preset conversion coefficient), and θ SNR is the preset maximum value of the received signal-to-noise ratio SNR. The uncertainty value is measured by the average entropy of the posterior probability distribution corresponding to different tasks. The uncertainty corresponding to the kth task is given by the following formula:

[0091]

[0092] Among them, c represents the class label of the classification task, and γ are all the parameters involved in the decoder and the task executor. After the retransmission detection is completed, if the detection fails, it is input into the FDE network (i.e., the semantic feature distortion evaluation network model) to trigger retransmission. The sender retransmits the distorted features and, according to the channel state, uses the remaining channel resources to incrementally transmit new features in order of importance; if the detection passes, all the features are concatenated and then decoded and the task is executed. Thus, the HARQ process of semantic awareness ends.

[0093] In some embodiments, in step B4, the determining the retransmission detection result according to the comparison result includes:

[0094] Step C1, in response to the receiving end determining that the comparison result is that the received signal-to-noise ratio is less than or equal to the comparison value, determining that the retransmission detection result fails the detection. Or,

[0095] Step C2, in response to the receiving end determining that the comparison result is that the received signal-to-noise ratio is greater than the comparison value, determining that the retransmission detection result passes the detection.

[0096] In the above solution, as shown in Figure 1A and Figure 1C , first, after multiple transmitted semantic features z j are transmitted through the channel to the receiving end, a retransmission detection module is used to perform retransmission detection on the received multiple transmitted semantic features to decide whether to trigger the next retransmission.

[0097] This application uses the uncertainty corresponding to the received multiple transmitted semantic features to make a retransmission decision. Since the received signal-to-noise ratio increases with the increase in the number of transmissions, when the received signal-to-noise ratio is small, the system needs to trigger retransmission until it can compensate for the uncertainty of the multiple transmitted semantic features received at the receiving end. Specifically, for the concatenated feature vector after the j-th transmission (i.e., the concatenated semantic feature), the receiving end continuously triggers retransmission until as follows:

[0098]

[0099] Among them, is the received signal-to-noise ratio, represents the comparison value, is the value of the uncertainty, L(·) is a monotonically increasing function, this application uses L(x) = 1 + x, θ0 is the conversion coefficient between the value of the uncertainty and the received SNR (i.e., the preset conversion coefficient), and θ SNR is the preset maximum value of the received signal-to-noise ratio SNR.

[0100] That is, at this time, it means that the receiving end determines that the comparison result is that the receiving signal-to-noise ratio is greater than the comparison value, and then determines that the retransmission detection result is passed.

[0101] The value of uncertainty is measured by the average entropy of the posterior probability distribution corresponding to different tasks, where the uncertainty corresponding to the kth task is given by the following formula:

[0102]

[0103] Among them, C represents the category label of the classification task, and γ is all the parameters involved in the decoder and task executor. After the retransmission detection is completed, if it fails to pass the detection, it will be input into the FDE network (i.e., the semantic feature distortion evaluation network model) to trigger retransmission. The transmitter retransmits the distorted features and uses the remaining channel resources to incrementally transmit new features in order of importance according to the channel state; if it passes the detection, all features are cascaded for decoding and task execution, and the semantic-aware HARQ process ends.

[0104] When it is as follows:

[0105]

[0106] That is, at this time, it means that the receiving end determines that the comparison result is that the receiving signal-to-noise ratio is greater than the comparison value, and then determines that the retransmission detection result is passed.

[0107] The retransmission detection result can be quickly determined by receiving the signal-to-noise ratio and the comparison value.

[0108] In some embodiments, in step 104, the training process of the semantic feature distortion assessment network model includes:

[0109] In step D1, the receiving end obtains original semantic features for training and received semantic features for training corresponding to the original semantic features for training, and constructs an initial model, wherein the initial model includes an initial distortion metric module and a gradient estimator module.

[0110] In step D2, the receiving end inputs the original semantic features for training and the received semantic features for training into the initial distortion measurement module, and the initial distortion measurement module outputs a first contribution degree parameter and a second contribution degree parameter.

[0111] Step D3, the receiving end constructs a training objective function based on the first contribution degree parameter and the second contribution degree parameter, and trains and adjusts the initial distortion measurement module according to the maximization processing result obtained in the process of maximizing the training objective function to obtain a trained initial distortion measurement module.

[0112] Step D4: The receiving end obtains the received semantic features for training, and based on the received semantic features for training, performs backpropagation training on the trained initial distortion metric module through the gradient estimator module to obtain the semantic feature distortion evaluation network model.

[0113] In the above solution, as Figure 1D shown, for the initial distortion metric module

[0114] To obtain the distortion degree of the received semantic features (i.e., the distortion evaluation data of multiple transmitted semantic features received), this application sets an initial distortion metric module consisting of four dense neural network layers, and evaluates the semantic feature distortion degree during the optimization of the dense layer parameters.

[0115] Specifically, in order to efficiently improve the distortion evaluation ability, before the end-to-end overall training, this application freezes the initial parameters of the remaining modules in the system and performs separate training on . Since mutual information can represent the dependence degree between random variables, therefore, in order to learn the difference between the semantic features at the sending and receiving ends and enable to learn the correlation between the received semantic features of all categories in the dataset and the original undistorted semantic features, this application uses the mutual information between s (i.e., the original semantic features for training) in Figure 1A and (i.e., the received semantic features for training) as the training objective of , and the specific optimization objective is as follows:

[0116]

[0117] Among them, is the trainable parameter of ; on this basis, the value of the output vector can represent the contribution degree of the corresponding input vector to the mutual information during the training process, thereby representing its correlation with s. If the input semantic features are severely distorted during transmission, they cannot play a role in the mutual information optimization process and have a low correlation with the undistorted vectors of the corresponding categories. Therefore, the corresponding output value is small. Based on the above analysis, when using mutual information as the training objective, the output of can indirectly reflect the distortion degree of the input features, that is, this output can form a corresponding distortion evaluation vector. After the independent training is completed, the system further performs overall end-to-end training. The receiving end inputs the feature vector received after the jth retransmission (i.e., the received semantic features for training) into The output of the dense layer 3 therein is combined with Vectors in the same dimension and convert them into a distortion metric vector d j , where (i.e., distortion evaluation data).

[0118] To estimate the mutual information of the semantic features at the transceiver as accurately as possible, this application adopts the mutual information neural estimation technology. Mutual information neural estimation: Mutual information is a measure of the degree of association between two random variables. Since mutual information can be regarded as the relative entropy (Kullback-Leibler, KL divergence) between the joint distribution and the independent distribution of two random variables, estimating mutual information only requires estimating the value of this KL divergence. The specific method is to give a lower bound of the KL divergence and use a neural network to continuously optimize this lower bound value to approximate the true value of the KL divergence, thereby estimating the value of mutual information. Specifically, first, the mutual information calculation formula between s and is as follows:

[0119]

[0120] where is the joint probability distribution of s and , and p(s) and are the marginal probability distributions of s and respectively. According to the derivation in the mutual information neural estimation technology, the lower bound of the mutual information is:

[0121]

[0122] where T can be any integrable function or approximated by a neural network, and the expectation in the formula can be calculated after sampling. At this time, maximizing the mutual information is to maximize the lower bound in, and this application adopts the multi-layer neural network in as T in

[0123]

[0124] In summary, when training the FDE network, taking as the training objective, optimizing the parameters of the dense layer in an unsupervised learning manner, where the input of is the sampling of the joint distribution and the marginal distributions s, respectively. After passing through 4 dense layers, the system substitutes the values output by the dense layer 4 into

[0125] where the values output by the dense layer 4 are and represents the first contribution degree parameter, representing the second contribution degree parameter.

[0126] For the gradient estimator (Gumbel-softmax) module:

[0127] To utilize the distortion evaluation vector d j (i.e., the distortion degree evaluation data) to guide the next retransmission, this application sets a retransmission threshold t, setting the values in d j greater than the threshold to 1 and vice versa to 0, thereby quantizing d j into a binary vector p j+1 , which is used to guide the (j + 1)-th feature retransmission, where 1 indicates that the semantic feature at this position needs to be retransmitted, and vice versa.

[0128] Since the quantization process is discrete and non-differentiable, this makes the network non-differentiable here and difficult to optimize using backpropagation. The currently commonly used straight-through estimator method treats the binarization function as an identity mapping for backpropagation, which may cause differences between forward propagation and backpropagation, resulting in high variance in gradient estimation. Therefore, this application considers directly discretizing d j in the forward propagation, and adopting a gradient estimator in the backpropagation, using a differentiable sampling process to replace the non-differentiable discretization process, which can effectively improve the efficiency and stability of gradient estimation. Specifically, this application proposes a gradient estimator based on the multi-dimensional Gumbel-softmax distribution, the Gumbel-softmax technique: Gumbel-softmax is a technique for approximating discrete sampling, commonly used in probability models and generative models in deep learning. To solve the problem that traditional discrete sampling methods are usually non-differentiable and cannot perform backpropagation, it estimates the gradient of the system backpropagation during discrete sampling by introducing a combination of the Gumbel distribution and the Softmax function, achieving an approximation of continuousization of the discrete distribution. Gumbel-softmax is trained using gradient optimization methods in discrete problems, enabling operations on discrete variables to be carried out in a continuous manner, which makes Gumbel-softmax widely applicable in training deep generative models, reinforcement learning, and other tasks that require discrete variables. Each sample is a differentiable proxy for the corresponding discrete sample. First, assuming there are R semantic features to be retransmitted, this application derives R probability vectors based on d j and forms R Gumbel-softmax distributions based on these probability vectors, where the r-th probability vector can be expressed as:

[0129]

[0130] Among them, is obtained by setting the largest r - 1 values in d j to 0. It represents the distortion situation of the remaining features when not considering the r - 1 features with the largest distortion degree. On this basis, from after being converted into a probability vector through the softmax function Then the value with the largest probability in it corresponds to the feature with the r - th largest distortion degree. After generating R probability vectors according to the above method, the corresponding R Gumbel - softmax sampling vectors can be obtained. Each random variable in this vector follows the Gumbel - softmax distribution. The r - th sampling vector can be expressed as:

[0131]

[0132] Among them, g r is a sampling vector of the independently distributed Gumbel distribution. The softmax function is a continuously differentiable approximation of the argmax function, and τ is the sampling temperature. The system sets τ to gradually decrease during the training process. At this time, all R Gumbel - softmax distributions will tend to be discretized. At the same time, the sampling vector in it will also tend to be a one - hot vector (that is, only one value is 1 and the rest are 0). At this time, it is equivalent to approximately sampling out the value with the largest probability in it and marking 1 at the corresponding position in the sampling vector. Therefore, by combining the R sampling vectors, a continuous approximation vector of p j+1 can be obtained for the backpropagation of the neural network, that is:

[0133]

[0134] In some embodiments, in step 105, the sender determines the transmission semantic features to be retransmitted based on the binarized vector, including:

[0135] The sender searches for the transmission semantic features with the binarized vector being the first preset value from each transmission semantic feature, and takes the transmission semantic features with the binarized vector being the first preset value as the transmission semantic features to be retransmitted.

[0136] In the above solution, in order to use the distortion evaluation vector d j (that is, the distortion degree evaluation data) to guide the next retransmission, this application sets a retransmission threshold t, sets the values in d j greater than the threshold to 1 (that is, the first preset value), and vice versa to 0 (that is, the second preset value), so as to quantize d j into a binarized vector pj+1 , which is used to guide the (j + 1)-th feature retransmission.

[0137] Among them, 1 indicates that the transmission semantic feature at this position needs to be retransmitted, otherwise it does not need to be retransmitted.

[0138] The binary vector quantized by the distortion evaluation data and the retransmission threshold can quickly determine the transmission semantic features that need to be retransmitted.

[0139] In some embodiments, in step 104, after the receiving end determines that the retransmission detection result fails the detection, the method further includes:

[0140] The receiving end inputs the received multiple transmission semantic features into a pre-trained semantic feature distortion evaluation network model to obtain distortion evaluation data, and converts the distortion evaluation data into a binary vector and transmits it to the sending end, so that the sending end can find the transmission semantic features whose binary vector is the second preset value from each transmission semantic feature, and use the transmission semantic features whose binary vector is the second preset value as the transmission semantic features that do not need to be retransmitted.

[0141] In the above solution, in order to use the distortion evaluation vector d j (i.e., the distortion evaluation data) to guide the next retransmission, this application sets a retransmission threshold t, and sets the value in d j greater than the threshold to 1 (i.e., the first preset value), otherwise to 0 (i.e., the second preset value), so as to quantize d j into a binary vector p j+1 , which is used to guide the (j + 1)-th feature retransmission.

[0142] Among them, 1 indicates that the transmission semantic feature at this position needs to be retransmitted, otherwise it does not need to be retransmitted.

[0143] The binary vector quantized by the distortion evaluation data and the retransmission threshold can quickly determine the transmission semantic features that do not need to be retransmitted, avoiding the problem of reducing the channel resource utilization rate caused by incorrect retransmission.

[0144] In some embodiments, in combination with Figure 1A , Figure 1C and Figure 1DAs shown, in the present application, under limited bandwidth and latency, a semantic-aware multi-task HARQ semantic-aware HARQ communication system is realized, enabling the receiving end to have the ability to perform distortion evaluation in the semantic feature dimension, improving the robustness of the system. Further, by combining distortion feature retransmission and feature incremental transmission, the transmission efficiency and task implementation performance of the system are improved. The present invention is simulated and compared with three cases: a semantic-aware HARQ system that only considers semantic incremental transmission (IK-S-HARQ), a semantic-aware HARQ system that only considers full feature retransmission (RT-S-HARQ), and a traditional type II HARQ system (Traditional II-HARQ). Taking the three tasks of vehicle re-identification, vehicle color identification, and vehicle type identification in an intelligent transportation system as examples, the performance simulation comparison results are as Figure 1E 、 Figure 1F and Figure 1G shown.

[0145] It should be noted that the method of the embodiment of the present application can be executed by a single device, such as a computer or a server. The method of this embodiment can also be applied to a distributed scenario and completed by multiple devices cooperating with each other. In this case of a distributed scenario, one of the multiple devices can only execute one or more steps of the method of the embodiment of the present application, and these multiple devices will interact with each other to complete the described method.

[0146] It should be noted that some embodiments of the present application are described above. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be executed in a different order than in the above embodiments and still achieve the desired result. Additionally, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired result. In certain embodiments, multi-semantic task processing and parallel processing are also possible or may be advantageous.

[0147] Based on the same inventive concept, corresponding to the method of any of the above embodiments, the present application also provides a semantic-aware HARQ communication device.

[0148] Referring to Figure 2 , the semantic-aware HARQ communication device is applied to a semantic-aware HARQ communication system, which includes a sending end and a receiving end. The device includes:

[0149] The sending end 201 is configured to obtain a transmission source data set, generate multiple transmission semantic features based on the transmission source data set, and send the multiple transmission semantic features to the receiving end;

[0150] The receiving end 202 is configured to perform retransmission detection based on multiple received transmission semantic features to obtain a retransmission detection result; in response to determining that the retransmission detection result fails the detection, the receiving end inputs the multiple received transmission semantic features into a pre-trained semantic feature distortion evaluation network model to obtain distortion evaluation data, and converts the distortion evaluation data into a binary vector for transmission to the sending end, so that the sending end determines the transmission semantic features to be retransmitted based on the binary vector and retransmits them to the receiving end. At the same time, when it is determined that the transmission channel meets the preset transmission conditions, a new set of multiple transmission semantic features is incrementally transmitted to the receiving end, and the receiving end repeats the retransmission detection process based on the received multiple transmission semantic features until it is determined that the retransmission detection result passes the detection or the preset maximum retransmission times are reached; or, in response to determining that the retransmission detection result passes the detection, the receiving end determines decoding features corresponding to at least one preset type of semantic task based on the received multiple transmission semantic features, and uses the prediction algorithms of at least one preset type of semantic task to perform prediction processing on the decoding features to obtain at least one prediction processing result.

[0151] In some embodiments, the sending end 201 is specifically configured to:

[0152] The sending end obtains the feature importance degree values corresponding to the multiple transmission semantic features;

[0153] The sending end determines the multiple transmission semantic features to be transmitted in each transmission process according to the feature importance degree values, and transmits the determined multiple transmission semantic features to the receiving end in each transmission process.

[0154] In some embodiments, the receiving end 202 includes:

[0155] A first determination unit configured to determine the received signal-to-noise ratio of the multiple received transmission semantic features and the cascaded semantic features corresponding to the multiple received transmission semantic features;

[0156] A second determination unit configured to determine the uncertainty value corresponding to the multiple received transmission semantic features based on the cascaded semantic features through an uncertainty algorithm;

[0157] A conversion unit configured to convert the uncertainty value into a comparison value corresponding to the received signal-to-noise ratio by using a preset conversion coefficient;

[0158] A comparison unit configured to compare the received signal-to-noise ratio with the comparison value to obtain a comparison result, and determine the retransmission detection result according to the comparison result.

[0159] In some embodiments, the comparison unit is specifically configured to:

[0160] In response to the receiving end determining that the comparison result is that the received signal-to-noise ratio is less than or equal to the comparison value, determine that the retransmission detection result fails the detection; or,

[0161] In response to the receiving end determining that the comparison result is that the received signal-to-noise ratio is greater than the comparison value, determine that the retransmission detection result passes the detection.

[0162] In some embodiments, the semantic awareness HARQ communication device further includes a training module, and the training module is specifically configured to:

[0163] The receiving end acquires original semantic features for training and received semantic features corresponding to the original semantic features for training, and constructs an initial model, where the initial model includes an initial distortion metric module and a gradient estimator module;

[0164] The receiving end inputs the original semantic features for training and the received semantic features for training into the initial distortion metric module, and the initial distortion metric module outputs a first contribution degree parameter and a second contribution degree parameter;

[0165] The receiving end constructs a training objective function based on the first contribution degree parameter and the second contribution degree parameter, and trains and adjusts the initial distortion metric module with the maximization result obtained during the maximization process of the training objective function to obtain a trained initial distortion metric module;

[0166] The receiving end acquires received semantic features for training, and performs backpropagation training on the trained initial distortion metric module through the gradient estimator module based on the received semantic features for training to obtain the semantic feature distortion evaluation network model.

[0167] In some embodiments, the sending end 201 is specifically configured to:

[0168] The sending end searches for the transmission semantic features whose binarized vectors are a first preset value from each transmission semantic feature, and uses the transmission semantic features whose binarized vectors are a first preset value as the transmission semantic features to be retransmitted.

[0169] In some embodiments, after the receiving end determines that the retransmission detection result fails the detection, the receiving end 202 is specifically configured to:

[0170] The receiving end inputs the received multiple transmission semantic features into a pre-trained semantic feature distortion evaluation network model to obtain distortion evaluation data, and converts the distortion evaluation data into a binary vector and transmits it to the sending end, so that the sending end can search for the transmission semantic feature whose binary vector is a second preset value from each transmission semantic feature, and use the transmission semantic feature whose binary vector is the second preset value as the transmission semantic feature that does not need to be retransmitted.

[0171] For the convenience of description, when describing the above device, it is divided into various modules according to functions and described separately. Of course, when implementing the present application, the functions of each module can be implemented in the same or multiple software and / or hardware.

[0172] The device in the above embodiment is used to implement the corresponding semantic awareness HARQ communication method in any of the foregoing embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be elaborated here.

[0173] Based on the same inventive concept, corresponding to the method in any of the above embodiments, the present application further provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the program, it implements the semantic awareness HARQ communication method described in any of the above embodiments.

[0174] Figure 3 FIG. shows a more specific schematic diagram of the hardware structure of the electronic device provided in this embodiment. The device may include: a processor 301, a memory 302, an input / output interface 303, a communication interface 304, and a bus 305. Among them, the processor 301, the memory 302, the input / output interface 303, and the communication interface 304 are communicatively connected to each other inside the device through the bus 305.

[0175] The processor 301 may be implemented in a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, etc., and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.

[0176] The memory 302 can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage devices, dynamic storage devices, etc. The memory 302 can store the operating system and other application programs. When implementing the technical solutions provided in the embodiments of this specification through software or firmware, the relevant program codes are stored in the memory 302 and are called and executed by the processor 301.

[0177] The input / output interface 303 is used to connect to the input / output module to achieve information input and output. The input / output module can be configured as a component in the device (not shown in the figure) or can be externally connected to the device to provide corresponding functions. Among them, the input device can include a keyboard, a mouse, a touch screen, a microphone, various sensors, etc., and the output device can include a display, a speaker, a vibrator, an indicator light, etc.

[0178] The communication interface 304 is used to connect to the communication module (not shown in the figure) to achieve communication interaction between this device and other devices. Among them, the communication module can achieve communication through wired means (such as USB, network cable, etc.) or can also achieve communication through wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0179] The bus 305 includes a path for transmitting information between various components of the device (such as the processor 301, the memory 302, the input / output interface 303, and the communication interface 304).

[0180] It should be noted that although the above device only shows the processor 301, the memory 302, the input / output interface 303, the communication interface 304, and the bus 305, in the specific implementation process, this device may also include other components necessary for normal operation. In addition, those skilled in the art can understand that the above device may also only include the components necessary to implement the solutions of the embodiments of this specification and do not necessarily include all the components shown in the figure.

[0181] The electronic device of the above embodiment is used to implement the corresponding semantic awareness HARQ communication method in any of the foregoing embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be elaborated here.

[0182] Based on the same inventive concept, corresponding to the method of any of the above embodiments, the present application also provides a non-transitory computer-readable storage medium, and the non-transitory computer-readable storage medium stores computer instructions, and the computer instructions are used to make the computer execute the semantic awareness HARQ communication method as described in any of the foregoing embodiments.

[0183] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette tapes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device.

[0184] The computer instructions stored in the storage medium of the above embodiment are used to cause the computer to execute the semantic awareness HARQ communication method described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be elaborated here.

[0185] Those of ordinary skill in the art should understand that the discussion of any of the above embodiments is only exemplary and is not intended to imply that the scope of the present application (including the claims) is limited to these examples; under the concept of the present application, the technical features in the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations in different aspects of the embodiments of the present application as described above, which are not provided in detail for the sake of brevity.

[0186] In addition, for simplicity of explanation and discussion, and in order not to make the embodiments of the present application difficult to understand, the well-known power / ground connections to integrated circuit (IC) chips and other components may or may not be shown in the provided drawings. In addition, the device may be shown in block diagram form to avoid making the embodiments of the present application difficult to understand, and this also takes into account the fact that the details of the implementation of these block diagram devices are highly dependent on the platform on which the embodiments of the present application are to be implemented (i.e., these details should be fully within the understanding of those skilled in the art). In the case where specific details (such as circuits) are set forth to describe the exemplary embodiments of the present application, it will be apparent to those skilled in the art that the embodiments of the present application can be implemented without these specific details or with variations of these specific details. Therefore, these descriptions should be considered illustrative rather than restrictive.

[0187] Although the present application has been described in connection with specific embodiments thereof, many alternatives, modifications, and variations of these embodiments will be apparent to those of ordinary skill in the art in light of the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may be used with the embodiments discussed.

[0188] Embodiments of the present application are intended to cover all such alternatives, modifications, and variations that fall within the broad scope of the appended claims. Accordingly, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the embodiments of the present application shall be included within the protection scope of the present application.

Claims

1. A semantically aware HARQ communication method, characterized in that: A semantically aware HARQ communication system, the system comprising a transmitting end and a receiving end, the method comprising: The sending end obtains a transmission source data set, generates a plurality of transmission semantic features based on the transmission source data set, and sends the plurality of transmission semantic features to the receiving end; The receiving end determines a received signal-to-noise ratio corresponding to the received multiple transmission semantic features, and a cascade semantic feature corresponding to the received multiple transmission semantic features; The receiving end determines, based on the cascade semantic feature, uncertainty values ​​corresponding to the received multiple transmission semantic features through an uncertainty algorithm; The receiving end converts the uncertainty value into a comparison value corresponding to the received signal-to-noise ratio using a preset conversion coefficient; The receiving end compares the received signal-to-noise ratio with the comparison value to obtain a comparison result, and determines the retransmission detection result according to the comparison result; In response to the receiving end determining that the retransmission detection result is failed, the receiving end inputs the received multiple transmission semantic features into a pre-trained semantic feature distortion assessment network model to obtain distortion assessment data, and converts the distortion assessment data into a binary vector and transmits it to the sending end, so that the sending end can determine the transmission semantic features to be retransmitted based on the binary vector and resend them to the receiving end. At the same time, when it is determined that the transmission channel meets the preset transmission conditions, multiple new transmission semantic features are incrementally transmitted to the receiving end. The receiving end repeatedly performs the retransmission detection process based on the received multiple transmission semantic features until it is determined that the retransmission detection result is passed, or the preset maximum number of retransmissions is reached; or, In response to the receiving end determining that the retransmission detection result is a passed detection, the receiving end determines a decoding feature corresponding to at least one preset type of semantic task based on the received multiple transmission semantic features, and uses a prediction algorithm of at least one preset type of semantic task to predict the decoding feature to obtain at least one prediction processing result.

2. The method according to claim 1, characterized in that Sending the plurality of transmission semantic features to the receiving end includes: The transmitting end obtains feature importance values ​​corresponding to the multiple transmission semantic features; The transmitting end determines a plurality of transmission semantic features to be transmitted in each transmission process according to the feature importance value, and sends the determined plurality of transmission semantic features to be transmitted to the receiving end in each transmission process.

3. The method according to claim 1, characterized in that The determining the retransmission detection result according to the comparison result includes: In response to the receiving end determining that the comparison result is that the received signal-to-noise ratio is less than or equal to the comparison value, determining that the retransmission detection result is a failure to pass the detection; or, In response to the receiving end determining that the comparison result is that the received signal-to-noise ratio is greater than the comparison value, the retransmission detection result is determined to be a pass detection.

4. The method according to claim 1, characterized in that The training process of the semantic feature distortion assessment network model includes: The receiving end obtains original semantic features for training and received semantic features for training corresponding to the original semantic features for training, and constructs an initial model, wherein the initial model includes an initial distortion metric module and a gradient estimator module; The receiving end inputs the original semantic features for training and the received semantic features for training into the initial distortion measurement module, and the initial distortion measurement module outputs a first contribution degree parameter and a second contribution degree parameter; The receiving end constructs a training objective function based on the first contribution degree parameter and the second contribution degree parameter, and performs training adjustment on the initial distortion metric module according to a maximization processing result obtained in the process of maximizing the training objective function, so as to obtain a trained initial distortion metric module; The receiving end obtains the received semantic features for training, and performs back-propagation training on the trained initial distortion metric module through the gradient estimator module based on the received semantic features for training, so as to obtain the semantic feature distortion assessment network model.

5. The method according to claim 1, characterized in that The transmitting end determines the transmission semantic feature to be retransmitted based on the binarized vector, including: The transmitting end searches for the transmission semantic feature whose binarization vector is a first preset value from various transmission semantic features, and uses the transmission semantic feature whose binarization vector is the first preset value as the transmission semantic feature to be retransmitted.

6. The method according to claim 1, characterized in that After the receiving end determines that the retransmission detection result is a failure to pass the detection, the method further includes: The receiving end inputs the received multiple transmission semantic features into a pre-trained semantic feature distortion assessment network model to obtain distortion assessment data, and converts the distortion assessment data into a binary vector and transmits it to the sending end, so that the sending end can search for the transmission semantic feature whose binarized vector is a second preset value from each transmission semantic feature, and use the transmission semantic feature whose binarized vector is the second preset value as the transmission semantic feature that is not retransmitted.

7. A semantically aware HARQ communication device, characterized in that: The device is arranged in a semantically aware HARQ communication system, the device comprises a transmitting end and a receiving end, and the device comprises: The transmitting end is configured to obtain a transmission source data set, generate a plurality of transmission semantic features based on the transmission source data set, and send the plurality of transmission semantic features to the receiving end; The receiving end is configured to perform retransmission detection based on the received multiple transmission semantic features to obtain a retransmission detection result; in response to determining that the retransmission detection result is not passed, the receiving end inputs the received multiple transmission semantic features into a pre-trained semantic feature distortion assessment network model to obtain distortion assessment data, and converts the distortion assessment data into a binary vector and transmits it to the sending end, so that the sending end determines the transmission semantic features to be retransmitted based on the binary vector and resends it to the receiving end, and when it is determined that the transmission channel meets the preset transmission conditions, incrementally transmits multiple new transmission semantic features to the receiving end, and the receiving end repeatedly performs the retransmission detection process based on the received multiple transmission semantic features until it is determined that the retransmission detection result is passed, or a preset maximum number of retransmissions is reached; or, in response to determining that the retransmission detection result is passed, the receiving end determines a decoding feature corresponding to at least one preset type of semantic task based on the received multiple transmission semantic features, and uses a prediction algorithm of at least one preset type of semantic task to predict the decoding feature to obtain at least one prediction processing result; The receiving end comprises: A first determining unit is configured to determine a received signal-to-noise ratio corresponding to the received multiple transmission semantic features and a cascade semantic feature corresponding to the received multiple transmission semantic features; A second determining unit is configured to determine uncertainty values ​​corresponding to the received multiple transmission semantic features through an uncertainty algorithm based on the cascade semantic feature; A conversion unit, configured to convert the value of the uncertainty into a comparison value corresponding to the received signal-to-noise ratio using a preset conversion coefficient; The comparison unit is configured to compare the received signal-to-noise ratio with the comparison value to obtain a comparison result, and determine the retransmission detection result according to the comparison result.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method according to any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium storing computer instructions, characterized in that: The computer instructions are used to enable a computer to execute the method according to any one of claims 1 to 6.

Images