Method and device for correcting semantic error by using semantic channel equalizer

Abstract

A method according to an embodiment of the present specification comprises the steps of: receiving a first message including information related to a first task; transmitting a second message related to a semantic representation; and receiving a response related to the second message. On the basis of whether a semantic error related to the semantic representation has occurred or has been corrected, the response indicates an ACK or NACK. The second message includes semantic language information generated on the basis of the semantic representation. A soft value is determined for each performed task on the basis of the semantic language information, and whether the semantic error has occurred or whether the semantic error has been corrected is determined on the basis of the soft value.

Description

Method and apparatus for semantic error correction using a semantic channel equalizer

[0001] This specification relates to a method and apparatus for semantic error correction using a semantic channel equalizer.

[0002] Mobile communication systems were developed to provide voice services while ensuring user mobility. However, mobile communication systems have expanded their scope to include data services as well as voice. Currently, due to the explosive increase in traffic leading to resource shortages and users demanding higher-speed services, more advanced mobile communication systems are required.

[0003] The requirements for next-generation mobile communication systems largely include the ability to accommodate explosive data traffic, a dramatic increase in transmission rates per user, a significantly increased number of connected devices, very low end-to-end latency, and high energy efficiency. To achieve this, various technologies are being researched, such as dual connectivity, massive multiple input multiple output (MMIMO), in-band full duplex, non-orthogonal multiple access (NOMA), super wideband support, and device networking.

[0004] Meanwhile, the following technical considerations are taken into account in semantic communication systems. Semantic error correction techniques are essential to ensure the reliability of communication between a Source and a Destination. Existing technologies for detecting semantic errors require the simultaneous transmission of redundancy—which allows for additional verification of information regarding semantic errors—when sending semantic information.

[0005] A semantic channel equalizer can be utilized for semantic communication based on multiple tasks. Received semantic language information can be transformed, based on the semantic channel equalizer, into semantic language information for a different task (e.g., task i) than the task (e.g., task k) of the semantic language information. According to existing methods, it is impossible to determine whether a semantic error exists in the transformed semantic representation, and it is also impossible to correct semantic errors.

[0006] The purpose of this specification is to propose a method for detecting semantic errors in semantic communication using a semantic channel equalizer. Additionally, another purpose of this specification is to propose a method for correcting said semantic errors.

[0007] The technical problems to be solved in this specification are not limited to those mentioned above, and other technical problems not mentioned will be clearly understood by those skilled in the art to which this specification belongs from the description below.

[0008] A method according to one embodiment of the present specification includes the steps of receiving a first message containing information related to a first task from a second wireless device, transmitting a second message related to a semantic representation to the second wireless device, and receiving a response related to the second message from the second wireless device.

[0009] Based on whether a semantic error related to the above semantic expression has occurred or been corrected, the response indicates ACK or NACK.

[0010] The second message above includes semantic language information generated based on the semantic representation.

[0011] A soft value is determined for each of the tasks performed based on the above semantic language information, and the occurrence or correction of the above semantic error is determined based on the soft value.

[0012] The above semantic language information can be transformed into semantic language information for each of the above tasks based on a semantic channel equalizer.

[0013] Based on the converted semantic language information above, the results of each of the above tasks can be obtained.

[0014] The above soft value may be based on the largest value among the probability values ​​for the above results.

[0015] i) Among the above results, a higher result related to the soft value and ii) a first value can be determined based on the soft value.

[0016] It can be determined that the semantic error has occurred based on the fact that the first value is smaller than the threshold value.

[0017] The above tasks may include i) the first task and ii) the second task. The above result and the soft value may be based on the second task.

[0018] A second value may be determined based on the above upper result. A first value may be determined based on the product of the above soft value and the above second value.

[0019] The above tasks may include i) the first task and ii) the second tasks. Based on the occurrence of the semantic error, at least one of the second tasks may be selected for the correction of the semantic error.

[0020] A third value may be determined for each of the above second tasks. Among the above second tasks, a second task in which the third value is greater than a threshold value may be selected for the correction.

[0021] The third value for a specific second task among the above second tasks can be determined based on the remaining second tasks among the above second tasks, excluding the specific second task.

[0022] The third value for the specific second task above may be determined based on i) the soft value of each of the remaining second tasks and ii) the top result associated with the soft value.

[0023] A fourth value can be determined based on the upper result of each of the remaining second tasks mentioned above.

[0024] The above third value can be determined based on the product of the above soft value and the above fourth value.

[0025] The above fourth value may be determined based on whether the above upper result is included in one or more results determined based on the mapping relationship.

[0026] The above mapping relationship can be defined between i) the results of each of the remaining second tasks and ii) the results of the specific second task.

[0027] A mapping relationship can be defined between the results of the first task and the results of each of the second tasks.

[0028] Based on the top result of each of the at least one second task mentioned above, one or more results related to the first task can be determined from the mapping relationship.

[0029] A ratio may be determined based on the probability that each of the above one or more results is a correct result. A loss value may be determined based on the above ratio.

[0030] The above semantic representation can be updated based on the partial derivative of the above loss value.

[0031] The above partial derivative value may be related to gradient descent for the above update.

[0032] The above NACK may be a first NACK or a second NACK.

[0033] Based on the fact that a task related to the correction of the semantic error cannot be determined from the above semantic expression, the response may represent the first NACK.

[0034] Based on the response indicating the first NACK, the second message associated with the semantic expression may be retransmitted.

[0035] A task related to the correction of the semantic error is determined from the above semantic expression, and based on the failure of the correction through the determined task: the response may indicate the second NACK.

[0036] Based on the response indicating the second NACK, information regarding i) a third message and ii) a task related to the generation of the third message related to the semantic representation may be transmitted.

[0037] The task related to the generation of the third message may be different from the task for the generation of the second message.

[0038] A first wireless device according to another embodiment of the present specification includes one or more transceivers, one or more processors, and one or more memories connected to the one or more processors and storing instructions.

[0039] The above instructions are characterized by setting the one or more processors to perform all steps of any one of the above methods based on execution by the one or more processors.

[0040] An apparatus according to another embodiment of the present specification comprises one or more memories and one or more processors functionally connected to the one or more memories, wherein the one or more memories store instructions that set the one or more processors to perform all steps of any one of the methods based on execution by the one or more processors.

[0041] One or more non-transitory computer-readable storage media according to another embodiment of the present specification store instructions.

[0042] The above instructions, executable by one or more processors, are characterized by setting one or more processors to perform all steps of any one of the above methods.

[0043] A method according to another embodiment of the present specification includes the steps of transmitting a first message containing information related to a first task to a first wireless device, receiving a second message related to a semantic representation from the first wireless device, and transmitting a response related to the second message to the first wireless device.

[0044] Based on whether a semantic error related to the above semantic expression has occurred or been corrected, the response indicates ACK or NACK.

[0045] The second message above includes semantic language information generated based on the semantic representation.

[0046] A soft value is determined for each of the tasks performed based on the above semantic language information, and the occurrence or correction of the above semantic error is determined based on the soft value.

[0047] A second wireless device according to another embodiment of the present specification includes one or more transceivers, one or more processors, and one or more memories connected to the one or more processors and storing instructions.

[0048] The above instructions are characterized by setting the one or more processors to perform all steps of the method based on execution by the one or more processors.

[0049] In conventional methods utilizing a semantic channel equalizer for semantic communication based on multiple tasks, the detection and correction of semantic errors are impossible. According to the embodiments of this specification, the occurrence of a semantic error is determined based on a soft-value for each task performed based on semantic language information. Therefore, semantic error detection is possible without signaling separate information for semantic error detection. The reliability of semantic communication can be improved compared to conventional methods. Furthermore, the signaling overhead required to support robust semantic communication can be reduced compared to existing methods.

[0050] According to the embodiments of this specification, semantic error detection and correction are performed based on soft values ​​and results associated with said soft values. Therefore, semantic error detection / semantic error correction performance can be improved compared to cases where hard values ​​are utilized.

[0051] The effects obtainable in this specification are not limited to those mentioned above, and other unmentioned effects will be clearly understood by those skilled in the art to which this specification belongs from the description below.

[0052] The drawings attached below are intended to aid in understanding the present specification and may provide embodiments of the present specification along with detailed descriptions. However, the technical features of the present specification are not limited to specific drawings, and features disclosed in each drawing may be combined with one another to form new embodiments. Reference numerals in each drawing may denote structural elements.

[0053] FIG. 1 is a drawing showing an example of a communication system applicable to the present specification.

[0054] FIG. 2 is a drawing showing an example of a wireless device applicable to the present specification.

[0055] FIG. 3 is a diagram illustrating a method for processing a transmission signal applicable to the present specification.

[0056] FIG. 4 is a drawing showing another example of a wireless device applicable to the present specification.

[0057] FIG. 5 is a drawing showing an example of a portable device applicable to the present specification.

[0058] FIG. 6 is a diagram showing physical channels applicable to the present specification and a signal transmission method using them.

[0059] Figure 7 is a figure showing an example of a perceptron structure.

[0060] Figure 8 shows an example of a multilayer perceptron structure.

[0061] Figure 9 is a figure showing an example of a deep neural network.

[0062] Figure 10 is a figure showing an example of a convolutional neural network.

[0063] Figure 11 is a figure showing an example of a filter operation in a convolutional neural network.

[0064] Figure 12 shows an example of a neural network structure in which a circular loop exists.

[0065] Figure 13 shows an example of the operational structure of a recurrent neural network.

[0066] FIG. 14 is a diagram illustrating a level-by-level communication model to which the embodiments proposed in this specification can be applied.

[0067] Figure 15 is a diagram illustrating a semantic error.

[0068] FIG. 16 illustrates a semantic channel equalizer according to an embodiment of the present specification.

[0069] FIG. 17 shows an example of mapping between tasks according to an embodiment of the present specification.

[0070] FIG. 18 shows another example of mapping between tasks according to an embodiment of the present specification.

[0071] FIG. 19 illustrates semantic communication based on multi-tasks according to an embodiment of the present specification.

[0072] FIG. 20 illustrates a method for detecting semantic errors in semantic communication based on multi-tasks according to an embodiment of the present specification.

[0073] FIG. 21 is a diagram illustrating the process of obtaining the result of a task that can be performed according to an embodiment of the present specification.

[0074] FIG. 22 illustrates a process for verifying the reliability of the results of a task that can be performed according to an embodiment of the present specification.

[0075] FIG. 23 illustrates a correction rate for the label of a target task using the results of a reliable task according to an embodiment of the present specification.

[0076] FIG. 24 illustrates a semantic error correction procedure according to an embodiment of the present specification.

[0077] FIG. 25 illustrates a signaling procedure according to an embodiment of the present specification.

[0078] FIG. 26 is a flowchart illustrating a method according to one embodiment of the present specification.

[0079] FIG. 27 is a flowchart illustrating a method according to another embodiment of the present specification.

[0080] The following embodiments are combinations of the components and features of this specification in a predetermined form. Each component or feature may be considered optional unless otherwise explicitly stated. Each component or feature may be implemented in a form not combined with other components or features. Additionally, some components and / or features may be combined to constitute the embodiments of this specification. The order of operations described in the embodiments of this specification may be changed. Some components or features of any embodiment may be included in another embodiment, or may be replaced with corresponding components or features of another embodiment.

[0081] In the description of the drawings, procedures or steps that could obscure the gist of the specification have not been described, nor have procedures or steps that are understandable to those skilled in the art been described.

[0082] Throughout the specification, when a part is described as "comprising" or "including" a component, this means that, unless specifically stated otherwise, it does not exclude other components but may include additional components. Furthermore, terms such as "...part," "...unit," and "module" as used in the specification refer to a unit that performs at least one function or operation, and this may be implemented in hardware, software, or a combination of hardware and software. Additionally, "one (a or an)," "one," "the," and similar related terms may be used in the context describing this specification (particularly in the context of the following claims) to include both singular and plural forms, unless otherwise indicated in this specification or clearly contradicted by the context.

[0083] The embodiments of this specification have been described with a focus on the data transmission and reception relationship between a base station and a mobile station. Here, the base station refers to a terminal node of a network that communicates directly with a mobile station. Specific operations described herein as being performed by a base station may, in some cases, be performed by an upper node of the base station.

[0084] That is, in a network consisting of multiple network nodes including a base station, various operations performed for communication with a mobile station may be performed by the base station or other network nodes other than the base station. In this case, 'base station' may be replaced by terms such as fixed station, Node B, eNB (eNode B), gNB (gNode B), ng-eNB, advanced base station (ABS), or access point.

[0085] Additionally, in the embodiments of this specification, the term terminal may be replaced with terms such as user equipment (UE), mobile station (MS), subscriber station (SS), mobile subscriber station (MSS), mobile terminal, or advanced mobile station (AMS).

[0086] Furthermore, the transmitting end refers to a fixed and / or mobile node that provides data or voice services, and the receiving end refers to a fixed and / or mobile node that receives data or voice services. Therefore, in the case of the uplink, a mobile station can be the transmitting end and a base station can be the receiving end. Similarly, in the case of the downlink, a mobile station can be the receiving end and a base station can be the transmitting end.

[0087] The embodiments of this specification may be supported by standard documents disclosed in at least one of the wireless access systems, such as IEEE 802.xx systems, 3GPP (3rd Generation Partnership Project) systems, 3GPP LTE (Long Term Evolution) systems, 3GPP 5G (5th generation) NR (New Radio) systems and 3GPP2 systems, and in particular, the embodiments of this specification may be supported by the documents 3GPP TS (technical specification) 38.211, 3GPP TS 38.212, 3GPP TS 38.213, 3GPP TS 38.321 and 3GPP TS 38.331.

[0088] In addition, the embodiments of this specification may be applied to other wireless access systems and are not limited to the systems described above. For example, they may be applicable to systems applied after the 3GPP 5G NR system and are not limited to specific systems.

[0089] That is, obvious steps or parts not described in the embodiments of this specification may be described by referring to the aforementioned documents. Additionally, all terms disclosed in this specification may be explained by the aforementioned standard documents.

[0090] Hereinafter, preferred embodiments according to the present specification will be described in detail with reference to the accompanying drawings. The detailed description disclosed below, together with the accompanying drawings, is intended to describe exemplary embodiments of the present specification and is not intended to represent the only embodiment in which the technical configuration of the present specification can be implemented.

[0091] Additionally, specific terms used in the embodiments of this specification are provided to aid in understanding this specification, and the use of such specific terms may be modified in other forms without departing from the technical spirit of this specification.

[0092] The following technology can be applied to various wireless access systems such as CDMA (code division multiple access), FDMA (frequency division multiple access), TDMA (time division multiple access), OFDMA (orthogonal frequency division multiple access), and SC-FDMA (single carrier frequency division multiple access).

[0093] For the sake of clarity in the following description, the explanation is based on 3GPP communication systems (e.g., LTE, NR, etc.), but the technical concept of the present invention is not limited thereto. LTE may refer to technology from 3GPP TS 36.xxx Release 8 onwards. Specifically, LTE technology from 3GPP TS 36.xxx Release 10 onwards is referred to as LTE-A, and LTE technology from 3GPP TS 36.xxx Release 13 onwards may be referred to as LTE-A pro. 3GPP NR may refer to technology from TS 38.xxx Release 15 onwards. 3GPP 6G may refer to technology from TS Release 17 and / or Release 18 onwards. "xxx" indicates a standard document detail number. LTE / NR / 6G may be collectively referred to as 3GPP systems.

[0094] Regarding the background technology, terms, abbreviations, etc. used in this specification, reference may be made to matters described in standard documents published prior to the present invention. For example, reference may be made to standard documents 36.xxx and 38.xxx.

[0095] Communication systems applicable to the present specification

[0096] Although not limited thereto, the various descriptions, functions, procedures, proposals, methods, and / or flowcharts of operation disclosed in this specification may be applied to various fields requiring wireless communication / connection (e.g., 5G) between devices.

[0097] Examples are provided in more detail below with reference to the drawings. In the following drawings and descriptions, the same reference numerals may represent the same or corresponding hardware blocks, software blocks, or function blocks unless otherwise described.

[0098] FIG. 1 is a drawing illustrating an example of a communication system to which the present specification applies. Referring to FIG. 1, the communication system (100) to which the present specification applies includes a wireless device, a base station, and a network. Here, a wireless device refers to a device that performs communication using wireless access technology (e.g., 5G NR, LTE) and may be referred to as a communication / wireless / 5G device. Although not limited thereto, a wireless device may include a robot (100a), a vehicle (100b-1, 100b-2), an XR (extended reality) device (100c), a hand-held device (100d), a home appliance (100e), an IoT (Internet of Thing) device (100f), and an AI (artificial intelligence) device / server (100g). For example, a vehicle may include a vehicle equipped with wireless communication capabilities, an autonomous vehicle, a vehicle capable of performing inter-vehicle communication, etc. Here, the vehicle (100b-1, 100b-2) may include an unmanned aerial vehicle (UAV) (e.g., a drone). The XR device (100c) includes an augmented reality (AR) / virtual reality (VR) / mixed reality (MR) device and may be implemented in the form of a head-mounted device (HMD), a head-up display (HUD) equipped in a vehicle, a television, a smartphone, a computer, a wearable device, a home appliance, digital signage, a vehicle, a robot, etc. The portable device (100d) may include a smartphone, a smart pad, a wearable device (e.g., a smartwatch, smart glasses), a computer (e.g., a laptop, etc.). The home appliance (100e) may include a TV, a refrigerator, a washing machine, etc. The IoT device (100f) may include a sensor, a smart meter, etc.For example, the base station (120) and network (130) may also be implemented as wireless devices, and a specific wireless device (120a) may act as a base station / network node for other wireless devices.

[0099] Wireless devices (100a to 100f) can be connected to a network (130) through a base station (120). AI technology may be applied to the wireless devices (100a to 100f), and the wireless devices (100a to 100f) can be connected to an AI server (100g) through the network (130). The network (130) can be configured using a 3G network, a 4G (e.g., LTE) network, or a 5G (e.g., NR) network. The wireless devices (100a to 100f) may communicate with each other through the base station (120) / network (130), but they may also communicate directly (e.g., sidelink communication) without going through the base station (120) / network (130). For example, vehicles (100b-1, 100b-2) can communicate directly (e.g., V2V (vehicle to vehicle) / V2X (vehicle to everything) communication). Also, IoT devices (100f) (e.g., sensors) can communicate directly with other IoT devices (e.g., sensors) or other wireless devices (100a to 100f).

[0100] Wireless communication / connection (150a, 150b, 150c) can be established between wireless devices (100a~100f) / base station (120) and between base station (120) / base station (120). Here, wireless communication / connection can be established through various wireless access technologies (e.g., 5G NR), such as uplink / downlink communication (150a), sidelink communication (150b) (or D2D communication), and communication between base stations (150c) (e.g., relay, IAB (integrated access backhaul)). Through wireless communication / connection (150a, 150b, 150c), wireless devices and base stations / wireless devices, and base stations and base stations can transmit / receive wireless signals to / from each other. For example, wireless communication / connection (150a, 150b, 150c) can transmit / receive signals through various physical channels. To this end, based on the various proposals of this specification, at least some of the following may be performed: a process for setting various configuration information for transmitting / receiving wireless signals, a process for various signal processing (e.g., channel encoding / decoding, modulation / demodulation, resource mapping / demapping, etc.), a resource allocation process, etc.

[0101] Communication systems applicable to the present specification

[0102] FIG. 2 is a drawing illustrating an example of a wireless device that can be applied to the present specification.

[0103] Referring to FIG. 2, the first wireless device (200a) and the second wireless device (200b) can transmit and receive wireless signals through various wireless access technologies (e.g., LTE, NR). Here, {the first wireless device (200a), the second wireless device (200b)} may correspond to {the wireless device (100x), the base station (120)} and / or {the wireless device (100x), the wireless device (100x)} of FIG. 1.

[0104] The first wireless device (200a) includes one or more processors (202a) and one or more memories (204a), and may additionally include one or more transceivers (206a) and / or one or more antennas (208a). The processor (202a) controls the memory (204a) and / or transceivers (206a) and may be configured to implement the descriptions, functions, procedures, proposals, methods and / or sequences of operation disclosed herein. For example, the processor (202a) may process information within the memory (204a) to generate a first information / signal and then transmit a wireless signal containing the first information / signal through the transceiver (206a). Additionally, the processor (202a) may receive a wireless signal containing a second information / signal through the transceiver (206a) and then store information obtained from the signal processing of the second information / signal in the memory (204a). Memory (204a) may be connected to the processor (202a) and may store various information related to the operation of the processor (202a). For example, memory (204a) may store software code including instructions for performing some or all of the processes controlled by the processor (202a) or for performing the descriptions, functions, procedures, proposals, methods, and / or sequences of operation disclosed in this specification. Here, the processor (202a) and memory (204a) may be part of a communication modem / circuit / chip designed to implement wireless communication technology (e.g., LTE, NR). A transceiver (206a) may be connected to the processor (202a) and may transmit and / or receive wireless signals through one or more antennas (208a). The transceiver (206a) may include a transmitter and / or receiver. The transceiver (206a) may be combined with an RF (radio frequency) unit. In this specification, a wireless device may refer to a communication modem / circuit / chip.

[0105] The second wireless device (200b) includes one or more processors (202b) and one or more memories (204b), and may additionally include one or more transceivers (206b) and / or one or more antennas (208b). The processor (202b) controls the memory (204b) and / or transceivers (206b) and may be configured to implement the descriptions, functions, procedures, proposals, methods and / or sequences of operation disclosed herein. For example, the processor (202b) may process information within the memory (204b) to generate a third information / signal and then transmit a wireless signal containing the third information / signal through the transceiver (206b). Additionally, the processor (202b) may receive a wireless signal containing a fourth information / signal through the transceiver (206b) and then store information obtained from the signal processing of the fourth information / signal in the memory (204b). The memory (204b) may be connected to the processor (202b) and may store various information related to the operation of the processor (202b). For example, the memory (204b) may store software code including instructions for performing some or all of the processes controlled by the processor (202b) or for performing the descriptions, functions, procedures, proposals, methods, and / or sequence diagrams of operation disclosed in this specification. Here, the processor (202b) and the memory (204b) may be part of a communication modem / circuit / chip designed to implement wireless communication technology (e.g., LTE, NR). The transceiver (206b) may be connected to the processor (202b) and may transmit and / or receive wireless signals through one or more antennas (208b). The transceiver (206b) may include a transmitter and / or receiver. The transceiver (206b) may be used in combination with an RF unit. In this specification, a wireless device may refer to a communication modem / circuit / chip.

[0106] Hereinafter, hardware elements of the wireless device (200a, 200b) will be described in more detail. Although not limited thereto, one or more protocol layers may be implemented by one or more processors (202a, 202b). For example, one or more processors (202a, 202b) may implement one or more layers (e.g., functional layers such as PHY (physical), MAC (media access control), RLC (radio link control), PDCP (packet data convergence protocol), RRC (radio resource control), and SDAP (service data adaptation protocol). One or more processors (202a, 202b) may generate one or more PDUs (Protocol Data Units) and / or one or more SDUs (service data units) according to the descriptions, functions, procedures, proposals, methods, and / or flowcharts of operation disclosed herein. One or more processors (202a, 202b) may generate messages, control information, data, or information according to the descriptions, functions, procedures, proposals, methods, and / or flowcharts of operation disclosed in this specification. One or more processors (202a, 202b) may generate a signal (e.g., baseband signal) including a PDU, SDU, message, control information, data, or information according to the functions, procedures, proposals, and / or methods disclosed in this specification and provide it to one or more transceivers (206a, 206b). One or more processors (202a, 202b) may receive a signal (e.g., baseband signal) from one or more transceivers (206a, 206b) and may obtain a PDU, SDU, message, control information, data, or information according to the descriptions, functions, procedures, proposals, methods, and / or flowcharts of operation disclosed in this specification.

[0107] One or more processors (202a, 202b) may be referred to as a controller, microcontroller, microprocessor, or microcomputer. One or more processors (202a, 202b) may be implemented by hardware, firmware, software, or a combination thereof. For example, one or more application-specific integrated circuits (ASICs), one or more digital signal processors (DSPs), one or more digital signal processing devices (DSPDs), one or more programmable logic devices (PLDs), or one or more field programmable gate arrays (FPGAs) may be included in one or more processors (202a, 202b). Descriptions, functions, procedures, proposals, methods, and / or flowcharts disclosed herein may be implemented using firmware or software, and the firmware or software may be implemented to include modules, procedures, functions, etc. Firmware or software configured to perform the descriptions, functions, procedures, proposals, methods, and / or operation sequences disclosed in this specification may be included in one or more processors (202a, 202b) or stored in one or more memories (204a, 204b) and driven by one or more processors (202a, 202b). The descriptions, functions, procedures, proposals, methods, and / or operation sequences disclosed in this specification may be implemented using firmware or software in the form of code, instructions, and / or sets of instructions.

[0108] One or more memories (204a, 204b) may be connected to one or more processors (202a, 202b) and may store various forms of data, signals, messages, information, programs, codes, instructions, and / or commands. One or more memories (204a, 204b) may be composed of ROM (read-only memory), RAM (random access memory), EPROM (erasable programmable read-only memory), flash memory, hard drives, registers, cache memory, computer-readable storage media, and / or combinations thereof. One or more memories (204a, 204b) may be located inside and / or outside of one or more processors (202a, 202b). Additionally, one or more memories (204a, 204b) may be connected to one or more processors (202a, 202b) through various technologies such as wired or wireless connections.

[0109] One or more transceivers (206a, 206b) may transmit user data, control information, wireless signals / channels, etc., as mentioned in the methods and / or operation flowcharts, etc., of this specification to one or more other devices. One or more transceivers (206a, 206b) may receive user data, control information, wireless signals / channels, etc., as mentioned in the descriptions, functions, procedures, proposals, methods and / or operation flowcharts, etc., disclosed in this specification from one or more other devices. For example, one or more transceivers (206a, 206b) may be connected to one or more processors (202a, 202b) and may transmit and receive wireless signals. For example, one or more processors (202a, 202b) may control one or more transceivers (206a, 206b) to transmit user data, control information, or wireless signals to one or more other devices. Additionally, one or more processors (202a, 202b) may control one or more transceivers (206a, 206b) to receive user data, control information, or wireless signals from one or more other devices. Additionally, one or more transceivers (206a, 206b) may be connected to one or more antennas (208a, 208b), and one or more transceivers (206a, 206b) may be configured to transmit and receive user data, control information, wireless signals / channels, etc., as described in the descriptions, functions, procedures, proposals, methods, and / or flowcharts of operation disclosed herein through one or more antennas (208a, 208b). In this specification, one or more antennas may be a plurality of physical antennas or a plurality of logical antennas (e.g., antenna ports). One or more transceivers (206a, 206b) can convert the received wireless signal / channel, etc. from an RF band signal to a baseband signal in order to process the received user data, control information, wireless signal / channel, etc. using one or more processors (202a, 202b).One or more transceivers (206a, 206b) can convert user data, control information, wireless signals / channels, etc. processed using one or more processors (202a, 202b) from baseband signals to RF band signals. To this end, one or more transceivers (206a, 206b) may include (analog) oscillators and / or filters.

[0110] FIG. 3 is a diagram illustrating a method for processing a transmission signal applicable to the present specification. For example, the transmission signal may be processed by a signal processing circuit. In this case, the signal processing circuit (300) may include a scrambler (310), a modulator (320), a layer mapper (330), a precoder (340), a resource mapper (350), and a signal generator (360). In this case, for example, the operation / function of FIG. 3 may be performed in the processor (202a, 202b) and / or transceiver (206a, 206b) of FIG. 2. Also, for example, the hardware element of FIG. 3 may be implemented in the processor (202a, 202b) and / or transceiver (206a, 206b) of FIG. 2. For example, blocks 310 to 350 may be implemented in the processor (202a, 202b) of FIG. 2, and block 360 may be implemented in the transceiver (206a, 206b) of FIG. 2, but are not limited to the above-described embodiment.

[0111] A codeword can be converted into a wireless signal through the signal processing circuit (300) of FIG. 3. Here, the codeword is an encoded bit sequence of an information block. The information block may include a transmission block (e.g., UL-SCH transmission block, DL-SCH transmission block). The wireless signal may be transmitted through various physical channels (e.g., PUSCH, PDSCH) of FIG. 6. Specifically, the codeword can be converted into a scrambled bit sequence by a scrambler (310). The scrambled sequence used for scrambling is generated based on an initialization value, which may include ID information of a wireless device, etc. The scrambled bit sequence may be modulated into a modulation symbol sequence by a modulator (320). The modulation method may include pi / 2-BPSK (pi / 2-binary phase shift keying), m-PSK (m-phase shift keying), m-QAM (m-quadrature amplitude modulation), etc.

[0112] A complex modulation symbol sequence can be mapped to one or more transmission layers by a layer mapper (330). The modulation symbols of each transmission layer can be mapped to the corresponding antenna port(s) by a precoder (340) (precoding). The output z of the precoder (340) can be obtained by multiplying the output y of the layer mapper (330) by an N*M precoding matrix W, where N is the number of antenna ports and M is the number of transmission layers. Here, the precoder (340) can perform precoding after performing transform precoding (e.g., a discrete Fourier transform (DFT)) on the complex modulation symbols. Alternatively, the precoder (340) can perform precoding without performing transform precoding.

[0113] A resource mapper (350) can map the modulation symbols of each antenna port to a time-frequency resource. The time-frequency resource may include multiple symbols (e.g., CP-OFDMA symbols, DFT-s-OFDMA symbols) in the time domain and multiple subcarriers in the frequency domain. A signal generator (360) generates a radio signal from the mapped modulation symbols, and the generated radio signal can be transmitted to another device through each antenna. To this end, the signal generator (360) may include an inverse fast Fourier transform (IFFT) module, a cyclic prefix (CP) inserter, a digital-to-analog converter (DAC), a frequency uplink converter, etc.

[0114] The signal processing process for a received signal in a wireless device can be configured as the inverse of the signal processing process (310–360) of FIG. 3. For example, a wireless device (e.g., 200a, 200b of FIG. 2) can receive a wireless signal from the outside through an antenna port / transceiver. The received wireless signal can be converted into a baseband signal through a signal restorer. To this end, the signal restorer may include a frequency downlink converter, an analog-to-digital converter (ADC), a CP remover, and a fast Fourier transform (FFT) module. Subsequently, the baseband signal can be restored into a codeword through a resource de-mapper process, a postcoding process, a demodulation process, and a de-scrambling process. The codeword can be restored into the original information block through decoding. Accordingly, a signal processing circuit (not shown) for a received signal may include a signal restorer, a resource de-mapper, a postcoder, a demodulator, a de-scrambler, and a decoder.

[0115] Wireless device structure applicable to the present specification

[0116] FIG. 4 is a drawing illustrating another example of a wireless device to which the present specification applies.

[0117] Referring to FIG. 4, the wireless device (400) corresponds to the wireless device (200a, 200b) of FIG. 2 and may be composed of various elements, components, units / parts, and / or modules. For example, the wireless device (400) may include a communication unit (410), a control unit (420), a memory unit (430), and additional elements (440). The communication unit may include a communication circuit (412) and transceiver(s) (414). For example, the communication circuit (412) may include one or more processors (202a, 202b) and / or one or more memories (204a, 204b) of FIG. 2. For example, the transceiver(s) (414) may include one or more transceivers (206a, 206b) and / or one or more antennas (208a, 208b) of FIG. 2. The control unit (420) is electrically connected to the communication unit (410), the memory unit (430), and additional elements (440) and controls the general operation of the wireless device. For example, the control unit (420) may control the electrical / mechanical operation of the wireless device based on a program / code / command / information stored in the memory unit (430). Additionally, the control unit (420) may transmit information stored in the memory unit (430) to an external (e.g., another communication device) via a wireless / wired interface through the communication unit (410), or store information received from an external (e.g., another communication device) via a wireless / wired interface through the communication unit (410) in the memory unit (430).

[0118] The additional element (440) can be configured in various ways depending on the type of wireless device. For example, the additional element (440) may include at least one of a power unit / battery, an input / output unit, a driving unit, and a computing unit. Although not limited thereto, the wireless device (400) may be implemented in the form of a robot (Fig. 1, 100a), a vehicle (Fig. 1, 100b-1, 100b-2), an XR device (Fig. 1, 100c), a portable device (Fig. 1, 100d), a home appliance (Fig. 1, 100e), an IoT device (Fig. 1, 100f), a digital broadcasting terminal, a hologram device, a public safety device, an MTC device, a medical device, a fintech device (or financial device), a security device, a climate / environment device, an AI server / device (Fig. 1, 140), a base station (Fig. 1, 120), a network node, etc. Wireless devices can be used in a movable or fixed location depending on the use—e.g., service.

[0119] In FIG. 4, various elements, components, units / parts, and / or modules within the wireless device (400) may be entirely interconnected via a wired interface, or at least partially connected via a communication unit (410). For example, within the wireless device (400), the control unit (420) and the communication unit (410) may be connected via a wire, and the control unit (420) and the first unit (e.g., 430, 440) may be connected wirelessly via the communication unit (410). Additionally, each element, component, unit / part, and / or module within the wireless device (400) may include one or more additional elements. For example, the control unit (420) may be composed of one or more sets of processors. For example, the control unit (420) may be composed of a set of a communication control processor, an application processor, an electronic control unit (ECU), a graphics processing processor, a memory control processor, etc. As another example, the memory unit (430) may be composed of RAM, DRAM (dynamic RAM), ROM, flash memory, volatile memory, non-volatile memory and / or a combination thereof.

[0120] Mobile devices to which this specification applies

[0121] FIG. 5 is a drawing illustrating an example of a portable device to which the present specification applies.

[0122] FIG. 5 illustrates a portable device to which the present specification applies. The portable device may include a smartphone, a smartpad, a wearable device (e.g., a smart watch, smart glasses), a portable computer (e.g., a laptop, etc.). The portable device may be referred to as an MS (mobile station), UT (user terminal), MSS (mobile subscriber station), SS (subscriber station), AMS (advanced mobile station), or WT (wireless terminal).

[0123] Referring to FIG. 5, the portable device (500) may include an antenna unit (508), a communication unit (510), a control unit (520), a memory unit (530), a power supply unit (540a), an interface unit (540b), and an input / output unit (540c). The antenna unit (508) may be configured as part of the communication unit (510). Blocks 510 to 530 / 540a to 540c correspond to blocks 410 to 430 / 440 of FIG. 4, respectively.

[0124] The communication unit (510) can transmit and receive signals (e.g., data, control signals, etc.) with other wireless devices and base stations. The control unit (520) can control the components of the portable device (500) to perform various operations. The control unit (520) may include an application processor (AP). The memory unit (530) can store data / parameters / programs / code / commands required for the operation of the portable device (500). Additionally, the memory unit (530) can store input / output data / information, etc. The power supply unit (540a) supplies power to the portable device (500) and may include wired / wireless charging circuits, batteries, etc. The interface unit (540b) can support the connection between the portable device (500) and other external devices. The interface unit (540b) may include various ports (e.g., audio input / output ports, video input / output ports) for connection with external devices. The input / output unit (540c) can receive or output video information / signals, audio information / signals, data, and / or information input from a user. The input / output unit (540c) may include a camera, a microphone, a user input unit, a display unit (540d), a speaker and / or a haptic module, etc.

[0125] For example, in the case of data communication, the input / output unit (540c) acquires information / signals (e.g., touch, text, voice, image, video) input by the user, and the acquired information / signals can be stored in the memory unit (530). The communication unit (510) converts the information / signals stored in the memory into wireless signals and can directly transmit the converted wireless signals to another wireless device or to a base station. Additionally, the communication unit (510) can receive wireless signals from another wireless device or base station and then restore the received wireless signals to their original information / signals. The restored information / signals are stored in the memory unit (530) and then can be output in various forms (e.g., text, voice, image, video, haptic) through the input / output unit (540c).

[0126] Physical channels and general signal transmission

[0127] FIG. 6 is a diagram illustrating physical channels applicable to the present specification and a signal transmission method using them.

[0128] When a terminal is turned on again after being turned off, or when it newly enters a cell, it performs initial cell search operations, such as synchronizing with the base station, in step S611. To do this, the terminal receives the primary synchronization channel (P-SCH) and secondary synchronization channel (S-SCH) from the base station to synchronize with the base station and obtain information such as the cell ID.

[0129] Subsequently, the terminal can obtain in-cell broadcast information by receiving a physical broadcast channel (PBCH) signal from the base station. Meanwhile, during the initial cell search phase, the terminal can check the downlink channel status by receiving a Downlink Reference Signal (DL RS). After completing the initial cell search, the terminal can obtain more specific system information by receiving the physical downlink control channel (PDCCH) and the physical downlink shared channel (PDSCH) based on the physical downlink control channel information in step S612.

[0130] Subsequently, the terminal may perform a random access procedure, such as steps S613 through S616, to complete the connection to the base station. To this end, the terminal transmits a preamble through a physical random access channel (PRACH) (S613) and receives a random access response (RAR) for the preamble through a physical downlink control channel and a corresponding physical downlink shared channel (S614). The terminal transmits a physical uplink shared channel (PUSCH) using scheduling information within the RAR (S615) and performs a contention resolution procedure, such as receiving a physical downlink control channel signal and a corresponding physical downlink shared channel signal (S616).

[0131] A terminal that has performed the procedure described above may subsequently perform the reception of a physical downlink control channel signal and / or a physical downlink shared channel signal (S617) and the transmission of a physical uplink shared channel (PUSCH) signal and / or a physical uplink control channel (PUCCH) signal (S618) as a general uplink / downlink signal transmission procedure.

[0132] Control information transmitted by a terminal to a base station is collectively referred to as uplink control information (UCI). UCI includes HARQ-ACK / NACK (hybrid automatic repeat and request acknowledgment / negative-ACK), SR (scheduling request), CQI (channel quality indication), PMI (precoding matrix indication), RI (rank indication), BI (beam indication) information, etc. In this case, UCI is generally transmitted periodically via PUCCH, but depending on the embodiment (e.g., when control information and traffic data need to be transmitted simultaneously), it may be transmitted via PUSCH. Additionally, the terminal may transmit UCI non-periodically via PUSCH in response to a request or instruction from the network.

[0133] 6G communication system

[0134] The 6G (wireless communication) system aims for (i) very high data rates per device, (ii) a very large number of connected devices, (iii) global connectivity, (iv) very low latency, (v) reduced energy consumption of battery-free IoT devices, (vi) ultra-reliable connectivity, and (vii) connected intelligence with machine learning capabilities. The vision of the 6G system can be seen in four aspects, such as "intelligent connectivity," "deep connectivity," "holographic connectivity," and "ubiquitous connectivity," and the 6G system can satisfy the requirements shown in Table 1 below. In other words, Table 1 represents the requirements of the 6G system.

[0135]

[0136] At this time, 6G systems can have key factors such as enhanced mobile broadband (eMBB), ultra-reliable low latency communications (URLC), mMTC (massive machine type communications), AI integrated communication, tactile internet, high throughput, high network capacity, high energy efficiency, low backhaul and access network congestion, and enhanced data security.

[0137] Core implementation technology of 6G systems

[0138] Artificial Intelligence

[0139] The most critical and newly introduced technology for 6G systems is AI. AI was not involved in 4G systems. 5G systems will support AI partially or to a very limited extent. However, 6G systems will be supported by AI for complete automation. Advancements in machine learning will create more intelligent networks for real-time communication in 6G. Introducing AI into communications can streamline and enhance real-time data transmission. AI can determine how complex target tasks are performed using numerous analyses. In other words, AI can increase efficiency and reduce processing latency.

[0140] Time-consuming tasks such as handover, network selection, and resource scheduling can be performed instantly by using AI. AI can also play a significant role in M2M, machine-to-human, and human-to-machine communication. Furthermore, AI can enable rapid communication in Brain-Computer Interfaces (BCI). AI-based communication systems can be supported by metamaterials, intelligent structures, intelligent networks, intelligent devices, intelligent cognitive radios, self-sustaining wireless networks, and machine learning.

[0141] Recently, attempts to integrate AI with wireless communication systems have emerged, but these have primarily focused on the application layer and network layer, particularly deep learning in the field of wireless resource management and allocation. However, such research is increasingly advancing toward the MAC layer and physical layer, with attempts to combine deep learning with wireless transmission, particularly at the physical layer. AI-based physical layer transmission refers to the application of signal processing and communication mechanisms based on AI drivers rather than traditional communication frameworks in terms of fundamental signal processing and communication mechanisms. Examples include deep learning-based channel coding and decoding, deep learning-based signal estimation and detection, deep learning-based MIMO mechanisms, and AI-based resource scheduling and allocation.

[0142] Machine learning can be used for channel estimation and channel tracking, and for power allocation and interference cancellation in the physical layer of the downlink (DL). In addition, machine learning can be used for antenna selection, power control, and symbol detection in MIMO systems.

[0143] Below, we will take a closer look at machine learning.

[0144] Machine learning refers to a series of operations for training machines to create machines capable of performing tasks that humans can or find difficult to do. Machine learning requires data and learning models. Data learning methods in machine learning can be broadly classified into three types: supervised learning, unsupervised learning, and reinforcement learning.

[0145] The purpose of neural network training is to minimize output errors. It is a process that repeatedly inputs training data into a neural network, calculates the error between the network's output and the target for the training data, and updates the weights of each node by backpropagating the error from the output layer to the input layer in a direction that reduces the error.

[0146] Supervised learning uses training data with correct answers labeled, whereas unsupervised learning may not have correct answers labeled. That is, for example, in the case of supervised learning regarding data classification, the training data may consist of data where each training data point is labeled with a category. Labeled training data is input into a neural network, and an error can be calculated by comparing the network's output (category) with the labels of the training data. The calculated error is backpropagated within the neural network (i.e., from the output layer to the input layer), and the connection weights of each node in each layer of the neural network can be updated according to this backpropagation. The amount of change in the connection weights of each node being updated can be determined by the learning rate. The neural network's calculations on the input data and the backpropagation of the error can constitute a learning cycle (epoch). The learning rate can be applied differently depending on the number of iterations of the neural network's learning cycle. For example, efficiency can be increased by using a high learning rate in the early stages of neural network training to enable the network to quickly achieve a certain level of performance, and accuracy can be increased by using a low learning rate in the later stages of training.

[0147] The learning method may vary depending on the characteristics of the data. For example, if the goal is to accurately predict data transmitted from the transmitting end at the receiving end in a communication system, it is preferable to perform learning using supervised learning rather than unsupervised learning or reinforcement learning.

[0148] Learning models correspond to the human brain, and while the most basic linear models can be considered, a machine learning paradigm that uses highly complex neural network structures, such as artificial neural networks, as learning models is called deep learning.

[0149] The neural network cores used for learning methods are broadly classified into deep neural networks (DNN), convolutional deep neural networks (CNN), and recurrent Boltzmann machines (RNN).

[0150] An artificial neural network is an example of connecting multiple perceptrons.

[0151] Referring to Fig. 7, the entire process of inputting an input vector x=(x1,x2,...,xd), multiplying each component by a weight (W1,W2,...,Wd), summing all the results, and then applying an activation function σ() is called a perceptron. A large artificial neural network structure can extend the simplified perceptron structure illustrated in Fig. 7 to apply input vectors to different multi-dimensional perceptrons. For convenience of explanation, input or output values ​​are referred to as nodes.

[0152] Meanwhile, the perceptron structure illustrated in Fig. 7 can be described as consisting of a total of three layers based on input and output values. An artificial neural network can be represented as shown in Fig. 8, in which there are H (d+1) dimensional perceptrons between the 1st layer and the 2nd layer, and K (H+1) dimensional perceptrons between the 2nd layer and the 3rd layer.

[0153] The layer where the input vector is located is called the input layer, the layer where the final output value is located is called the output layer, and all layers located between the input and output layers are called hidden layers. Although the example in Fig. 8 shows three layers, the input layer is excluded when counting the actual number of layers in an artificial neural network, so it can be viewed as having a total of two layers. An artificial neural network is constructed by connecting perceptrons of basic blocks in a two-dimensional manner.

[0154] The aforementioned input layer, hidden layer, and output layer can be applied not only to multilayer perceptrons but also to various artificial neural network structures such as CNNs and RNNs, which will be described later. As the number of hidden layers increases, the artificial neural network becomes deeper, and the machine learning paradigm that uses a sufficiently deep artificial neural network as a learning model is called Deep Learning. In addition, the artificial neural network used for Deep Learning is called a Deep Neural Network (DNN).

[0155] The deep neural network illustrated in Fig. 9 is a multilayer perceptron composed of eight hidden and output layers. The structure of the multilayer perceptron is referred to as a fully-connected neural network. In a fully-connected neural network, there are no connections between nodes located in the same layer, and connections exist only between nodes located in adjacent layers. A DNN possesses a fully-connected neural network structure and is composed of a combination of multiple hidden layers and activation functions; it can be effectively applied to identify correlation characteristics between inputs and outputs. Here, correlation characteristics may refer to the joint probability of the input and output. Fig. 9 is a diagram illustrating an example of a deep neural network.

[0156] Meanwhile, depending on how multiple perceptrons are connected to each other, various artificial neural network structures different from the aforementioned DNN can be formed.

[0157] In a DNN, nodes located within a single layer are arranged in a one-dimensional vertical direction. However, Figure 10 assumes a case where nodes are arranged two-dimensionally, with w nodes horizontally and h nodes vertically (the convolutional neural network structure of Figure 10). In this case, since a weight is applied for each connection during the connection process from a single input node to a hidden layer, a total of h × w weights must be considered. Since there are h × w nodes in the input layer, a total of h²w² weights are required between two adjacent layers.

[0158] Figure 10 is a figure showing an example of a convolutional neural network.

[0159] The convolutional neural network of Fig. 10 has a problem in which the number of weights increases exponentially depending on the number of connections. Therefore, instead of considering all mode connections between adjacent layers, it is assumed that there are small filters, and weighted sum and activation function operations are performed on the parts where filters overlap as in Fig. 10.

[0160] A single filter has weights corresponding to its size, and the weights can be trained to extract and output specific features on an image as factors. In Figure 10, a 3×3 filter is applied to the top-left 3×3 area of ​​the input layer, and the output value resulting from the weighted sum and activation function operation for the corresponding node is stored in z22.

[0161] The above filter performs weighted sum and activation function operations while scanning the input layer and moving by a fixed interval horizontally and vertically, and places the output value at the current filter position. This method of operation is similar to the convolution operation on images in the field of computer vision, so a deep neural network with this structure is called a convolutional neural network (CNN), and the hidden layer generated as a result of the convolution operation is called a convolutional layer. In addition, a neural network having multiple convolutional layers is called a deep convolutional neural network (DCNN).

[0162] Figure 11 is a figure showing an example of a filter operation in a convolutional neural network.

[0163] In the convolution layer, the number of weights can be reduced by calculating a weighted sum that includes only the nodes located within the area covered by the filter, starting from the node where the current filter is located. As a result, a single filter can be utilized to focus on features of a local area. Accordingly, CNNs can be effectively applied to image data processing where physical distance in a 2D area serves as an important judgment criterion. Meanwhile, multiple filters can be applied immediately before the convolution layer in a CNN, and multiple output results can be generated through the convolution operation of each filter.

[0164] Meanwhile, depending on the data attributes, there may be data where sequence characteristics are important. A structure that applies a method to an artificial neural network in which elements of the data sequence are input one by one at each timestep, taking into account the length variability and sequence relationships of such sequence data, and the output vector (hidden vector) of the hidden layer output at a specific timestep is input along with the next element in the sequence is called a recurrent neural network structure.

[0165] Referring to Fig. 12, the recurrent neural network (RNN) is structured such that, in the process of inputting elements (x1(t), x2(t), ..., xd(t)) of a time point t in a data sequence into a fully connected neural network, the previous time point t-1 is input along with the hidden vector (z1(t-1), z2(t-1), ..., zH(t-1)), and a weighted sum and activation function are applied. The reason for passing the hidden vector to the next time point in this manner is that the information in the input vectors from previous time points is considered to be accumulated in the hidden vector of the current time point.

[0166] Figure 12 shows an example of a neural network structure in which a circular loop exists.

[0167] Referring to Fig. 12, the recurrent neural network operates on the input data sequence in a predetermined time sequence.

[0168] When the input vector (x1(t), x2(t), ..., xd(t)) at time point 1 is input into the recurrent neural network, the hidden vector (z1(1), z2(1), ..., zH(1)) is input together with the input vector (x1(2), x2(2), ..., xd(2)) at time point 2, and the vector (z1(2), z2(2), ..., zH(2)) of the hidden layer is determined through a weighted sum and activation function. This process is performed repeatedly up to time point 2, time point 3, ..., time point T.

[0169] Figure 13 shows an example of the operational structure of a recurrent neural network.

[0170] Meanwhile, when multiple hidden layers are placed within a recurrent neural network, it is called a deep recurrent neural network (DRNN). Recurrent neural networks are designed to be usefully applied to sequence data (e.g., natural language processing).

[0171] In addition to DNN, CNN, and RNN, it includes various deep learning techniques such as Restricted Boltzmann Machine (RBM), Deep Belief Networks (DBN), and Deep Q-Network as neural network cores used for learning, and can be applied in fields such as computer vision, speech recognition, natural language processing, and speech / signal processing.

[0172] The symbols / abbreviations / terms used in this specification are as follows.

[0173] - CE: Checking Entropy

[0174] - OT: Other Tasks

[0175] - SCE: Semantic Channel Equalizer

[0176] - SR: Semantic Representation

[0177] Below, we will examine the technical problems related to the embodiments proposed in this specification with reference to FIG. 14.

[0178] FIG. 14 is a diagram illustrating a level-by-level communication model to which the embodiments proposed in this specification can be applied.

[0179] Referring to Fig. 14, the communication model can be defined at three levels (A to C).

[0180] Level A relates to how accurately symbols (technical messages) can be transmitted between a transmitter and a receiver. This can be considered when the communication model is understood from a technical perspective.

[0181] Level B relates to how accurately symbols transmitted between a transmitter and a receiver convey meaning. This can be considered when the communication model is understood in terms of semantics.

[0182] Level C relates to how effectively the meaning received at the destination contributes to subsequent actions. This can be considered when the communication model is assessed in terms of effectiveness.

[0183] In the design of communication models, not all Level A to Level C perspectives are considered, and this may vary depending on the implementation method.

[0184] For example, a communication model implemented with a focus on Level A, similar to a communication model based on conventional technology, may be considered. As another example, a communication model that considers not only Level A but also Level B (and Level C) to support semantic communication may be considered. In such a communication model, the transmitter and receiver may be referred to as a semantic transmitter and a semantic receiver, and semantic noise may be additionally considered.

[0185] The following describes in detail the considerations regarding graph neural networks.

[0186] One of the various goals of 6G communication is to enable diverse new services that interconnect humans and machines possessing various levels of intelligence. It is necessary to consider not only existing technical problems (e.g., Fig. 14A) but also semantic problems (e.g., Fig. 14B). Semantic communication will be explained in detail below, using communication between people as an example.

[0187] Words used to exchange information (word information) are related to "meaning." Upon hearing the speaker's words, the listener can interpret the meaning or concept represented by the speaker's words. When this is linked to the communication model in Fig. 14, for semantic communication to be supported, the concept related to the message sent from the Source needs to be correctly interpreted at the Destination.

[0188] Using data given to the Source or collected raw data, the Source can generate a semantic representation (SR). The Source can transmit the SR to the Destination. In this case, during the process of recovering the SR received by the Destination into raw data, an approach is required that focuses on whether the interpretation and reasoning were done well, rather than an approach focused on the conventional goal of reducing reconstruction error.

[0189] Specifically, in the process of restoring the SR received by the Destination back to the original raw data, access is required to determine whether the downstream task performed at the Destination was executed in accordance with the intent conveyed by the Source using the received SR.

[0190] When performing inference operations, the Destination operates by utilizing the background knowledge it possesses. To this end, the background knowledge contained in the data transmitted from the Source must be able to be reflected in the Destination's background knowledge.

[0191] To perform semantic communication, it is important for the Source to properly generate an SR containing semantic information about the data to be transmitted and send it to the Destination, and for the Destination to accurately interpret the semantic information contained in the received SR.

[0192] When conducting semantic communication through a communication channel containing noise, the message received at the destination may contain errors due to channel noise or a mismatch in the receiver's background knowledge. Such errors may occur at the technical level (level A in FIG. 14) or the semantic level (level B in FIG. 14). In this case, since the ultimate goal of the technical level is to preserve the syntactic relationship between the transmitted message and the received message, a technical error can be defined as a syntactic difference between the transmitted message and the received message.

[0193] In contrast, the semantic level does not aim for the preservation of the syntactic relationship between the transmitted and received messages. The ultimate goal of the semantic level is to reason similarly between the semantics contained in the transmitted message and the semantics contained in the received message. Therefore, a semantic error can be defined as a semantic difference between the transmitted message and the received message. For example, in the Source, "Monday tuesday wednesday thursday The sentence "Friday" was sent, and in Destination " Saturday It can be assumed that the sentence "Sunday" is received. Since the same semantic can be obtained based on background knowledge regarding the day of the week, it can be considered that semantic similarity has been maintained; however, because a syntactic difference has occurred in terms of syntactic preservation, it can be interpreted as a technical error having occurred. Another example of a semantic error will be explained in detail below with reference to Fig. 15.

[0194] Figure 15 is a diagram illustrating a semantic error.

[0195] Referring to Fig. 15, it can be assumed that a sound (e.g., voice) is used to convey the word "copy machine" from the Source to the Destination. Since the "p" sound is very similar to the "ff" sound, the following semantic error may occur. Regarding the Source's pronunciation of the word "copy machine," the Destination may misinterpret it as the word "coffee machine." In such cases, it can be considered that a semantic error has occurred regarding the word intended to be sent.

[0196] To prevent such semantic errors, the Source can send the word "Xerox" instead of "copy machine" as a message. This prevents the "p" sound from being misinterpreted as "ff," thereby blocking semantic errors.

[0197] When performing semantic communication in this manner, a semantic mismatch between the Source and Destination occurs due to semantic errors. To ensure reliable semantic communication, a semantic error check to detect the aforementioned semantic mismatch can be performed by the Destination. If a semantic error is detected, the Destination can obtain the accurate semantic information intended by the Source by performing semantic error correction and requesting retransmission from the Source.

[0198] Semantic error detection technology based on a semantic channel equalizer (SCE) can be used as a method to detect such semantic errors. The SCE enables the interpretation of SRs generated based on each of the different semantic languages. Specifically, the semantic channel equalizer plays the role of transforming the SR into each semantic language.

[0199] FIG. 16 illustrates a semantic channel equalizer according to an embodiment of the present specification.

[0200] Referring to Fig. 16, the following notation is used for mathematical explanation.

[0201] Language generator

[0202] Language interpreter

[0203] The language generator is an SR for message m that the transmitting end intends to send. Creates. The created SR from using the language interpreter result ...can be obtained. In this case, considering two tasks (task k, task l), if there is a language generator and a language interpreter for each, the following situation can be considered.

[0204] Here Is The th transmitter (or It refers to a language generator for the (th) task, Is The th transmitter (or It refers to the language interpreter for the (th) task.

[0205] Fig. 16 , is the input message image This is a language generator and interpreter configured to select which number from 0 to 9 the number written in is. An image with the number 3 written in it. language generator Using as input, SR This can be obtained. The corresponding SR This language interpreter When it is entered as input, You can obtain the result with a high probability.

[0206] Also, of Fig. 16 , is a language generator and interpreter configured to select whether the number written in the input message image m is odd or even. An image with the number 8 written on it. language generator SR as input This can be obtained. The corresponding SR This language interpreter When it is entered as input, You can obtain the result with a high probability.

[0207] However, when the tasks of the language generator and the language interpreter are different, it is difficult to predict the outcome of an SR-based language interpreter. Specifically, the language generator SR generated using language interpreter The result when entered into It can be expressed as. In this case, It is impossible to predict whether it will be odd or even. This is because the two pairs of language generators and interpreters are configured independently, so the SR mapping and interpreter result mapping methods for the same input value differ. In such cases, the SR generated based on task k is language interpreter Additional processing is required to obtain accurate results through this.

[0208] In the above situation, the Semantic channel equalizer is cast It transforms into. Specifically, the language interpreter through equalizing by a semantic channel equalizer from You can obtain. In other words, the Semantic channel equalizer is It enables normal operation between transmitters and receivers (or tasks) with different semantic languages ​​through this. The transformation mapping of the semantic channel equalizer It is written as .

[0209] In order to mathematically express the correlation between results obtained using a language interpreter Uses. Referring to FIGS. 17 and 18 below, a mapping showing the correlation between the results of the tasks ( Explain ) in detail.

[0210] FIG. 17 illustrates an example of mapping between tasks according to an embodiment of the present specification. Specifically, FIG. 17 illustrates a mapping between a task determining digits (0 to 9 digits) and a task determining even / odd (even / odd). It represents ).

[0211] task is a task that distinguishes digits from 0 to 9, and the task It is a task that distinguishes between even and odd. The results (0~9) of the task The results (even, odd) can be mapped as follows.

[0212] task {0, 2, 4, 6, 8} is task It maps to the even of the task {1, 3, 5, 7, 9} is task It maps to the odds of. Fig. 17 is an example of a many-to-one mapping. Mapping between the results of tasks ( The mapping can be performed as a one-to-many mapping. This will be explained below with reference to Fig. 18.

[0213] FIG. 18 shows another example of mapping between tasks according to an embodiment of the present specification.

[0214] Specifically, FIG. 18 shows a mapping between a task determining a Modulus 3 value and a task determining digits (0 to 9 digits). It represents ).

[0215] In FIG. 24, task k is a task that determines the value of Modulus 3, and task l is a task that determines digits 0 through 9. The results of task k (0, 1, 2) can be mapped to the results of task l (0 through 9) as follows.

[0216] 0 of task k can be mapped to {0, 3, 6, 9} of task l. 1 of task k can be mapped to {1, 4, 7} of task l. 2 of task k can be mapped to {2, 5, 8} of task l.

[0217] To explain the method for detecting semantic errors, the following notation is defined.

[0218] : Task semantic language generator for ( )

[0219] : Task semantic language interpreter for ( )

[0220] : Transformation from task to task (Operation by Semantic channel equalizer)

[0221] : Task and task Correlation map for the labels of

[0222] SCE-based semantic error detection technology , , , and ( ) considers the given situation. In a multi-task semantic communication situation, a situation like that shown in Fig. 19 can be considered.

[0223] FIG. 19 illustrates semantic communication based on multi-tasks according to an embodiment of the present specification.

[0224] Referring to Fig. 19, when semantic communication begins, the destination has information about the task it intends to perform. Sends to source.

[0225] Source is based on the received task information Select . Source is based on the selected task The message to be sent using SR regarding Create . Then, the source becomes the destination and transmits. At this time, It is assumed that it is information that can be transmitted without errors. is a channel error This added state It is received as the destination.

[0226] Destination is the received SR Regarding semantic channel equalizer (transformation language interpreter using ) SR that can be interpreted as Obtains. Finally, Destination is task As a result of You can obtain.

[0227] If semantic communication as described above is performed, the final task obtained Results regarding It is not possible to verify whether it is an accurate result. In other words, the received SR It is not possible to determine whether a semantic error exists.

[0228] Operations for determining a semantic error that occurred in the received SR are explained with reference to FIG. 20.

[0229] FIG. 20 illustrates a method for detecting semantic errors in semantic communication based on multi-tasks according to an embodiment of the present specification.

[0230] Referring to FIG. 20, the destination is a task from the received SR ( task results for ) ...can be obtained. Using this result, a metric for determining semantic error, such as the following mathematical formula 1, can be defined.

[0231]

[0232] is as an indicator function It has a value of 1 if true and 0 otherwise. Specifically, task ( The result for ) is a value determined based on the mapping relationship of the received SR. In cases belonging to, becomes 1. Otherwise, It becomes 0.

[0233] At this time, a specific threshold value If given, If this condition is satisfied, it can be determined that a semantic error occurred in the received SR. task The result of and other tasks Through the correlation between the results satisfying Because the number is too small, the final task obtained Since the results regarding it cannot be trusted, it can be determined that a semantic error has occurred.

[0234] For example, consider a case where three tasks are given. Assume that for a number image, Task 1 distinguishes digits from 0 to 9, Task 2 distinguishes odd / even, and Task 3 calculates the value of the modulus 3. Also, Source is an image with 3 drawn on it. Regarding SR as the language generator for task 2 It was generated and sent to the destination, and the destination is SR It is assumed that the case where is received is.

[0235] Received SR Regarding the result for task 1 In the case obtained, the following actions may be performed to check whether a semantic error occurred for the result '4' for task 1.

[0236] Destination is the received SR Based on this, obtain the results for task 2 and task 3. At this time, the result for task 2 is and the result for task 3 is Assuming the case, as follows: This can be decided.

[0237] Regarding task 2 since, =0. For task 3 since, =0. Therefore, It is determined to be 0. If we say, Therefore, it can be determined that a semantic error occurred in the received SR.

[0238] Above Appropriate threshold value for the value To obtain, regarding the following two hypotheses The probability distribution of must be modeled.

[0239] The first hypothesis is "task obtained from Destination" The result for is the original message task for It can be set to "identical to the result of", and if expressed as a formula, it is " It can be written as ". Here The message task for It means the correct result of.

[0240] The second hypothesis is "task obtained from Destination" The result for is the original message task for It can be set to "not identical to the result of", and expressed as a formula, " It can be written as "

[0241] Regarding the two hypotheses When following each hypothesis, the distribution can be set as follows.

[0242] i) Case following the first hypothesis:

[0243]

[0244] ii) Case following the second hypothesis:

[0245]

[0246] In the above equation, ~ means that it follows the probability distribution formed by the corresponding equation, and Is All that satisfy It refers to the convolution operation on the distribution. is probability It refers to the Bernoulli distribution with as a parameter. Also is, is a value that can be evaluated when a language generator, language interpreter, and transformation (channel equalizer) are given and multiple data are given. is, Is It can be calculated as. and If defined as such, using the two distributions, the given Value and appropriate threshold value You can perform a hypothesis test as follows using .

[0247]

[0248] At this time, the threshold value The false alarm probability can be determined by [this]. The false alarm probability is equal to the sum of the type 1 error probability and the type 2 error probability. The false alarm probability can be defined as follows.

[0249]

[0250] represents the probability of a false alarm.

[0251] is the type 2 error probability value. silver Actually A value sampled by, but It means the probability of judging that it follows. is the type 1 error probability value. Is Actually A value sampled by, but It means the probability of judging that it follows.

[0252] Therefore, to minimize the probability of false alarms You can select and use it. That is, Like It can be calculated. However, depending on the system design / implementation method, satisfying the above equation Larger or smaller than You can choose.

[0253] When it is determined that a semantic error has occurred using the SCE-based semantic error detection technology described above, correction of the semantic error is not supported according to existing methods. Therefore, this specification proposes a technology capable of correcting semantic errors when they are determined to have occurred based on SCE. A method for performing semantic error correction based on SCE will be described in detail below with reference to FIGS. 21 to 24.

[0254] To perform highly reliable semantic communication, it is essential to identify semantic errors that may occur due to the channel and to correct them if they do occur. The following describes a technique that detects semantic errors based on SCE and performs SCE-based semantic error correction when an error is identified.

[0255] 1) Soft-value of Task Result

[0256] In semantic communication scenarios, when performing tasks using the destination interpreter, probability values ​​for the results of each task can be calculated using soft-max values. Soft-values, such as the aforementioned probability values, are used to perform SCE-based semantic error correction.

[0257] for example, task for nth result label If we define it as the soft-max value for The largest of them label with value task It can be obtained as a result of. At this time task Named as the top-1 result for It can be written as. is task It refers to a set containing the result labels, am. It satisfies the following mathematical equation 2.

[0258]

[0259] The SCE-based semantic error correction proposed in this specification can be performed based on the top-1 results of tasks and the probabilities thereof.

[0260] 2) SCE-based Semantic Error Detection by using Soft-value

[0261] According to the aforementioned SCE-based semantic error detection technique, the indicator function (i.e., of Equation 1) Semantic error is detected based solely on the inclusion relationships of the results of each task using ). Below, a technique for detecting semantic error based on soft values ​​is described.

[0262] A metric that can verify whether a semantic error has occurred can be defined using soft-values ​​as shown in the following mathematical formula 3.

[0263]

[0264] The above metric for semantic error detection is based on the indicator function in existing metrics It takes the form of a weighted summation where the values ​​are multiplied. When the metric is transformed in this way, the threshold used to compare whether a semantic error has occurred. The derivation method for this can be changed. To this end, the null hypothesis and alternative hypothesis can be re-established as follows.

[0265] i) Case following the first hypothesis:

[0266]

[0267] ii) Case following the second hypothesis:

[0268]

[0269] At this time, It can be defined as follows.

[0270]

[0271] Using this, the probability of false alarms is similar to what was explained earlier. The value that can be minimized You can find a value and use it as a threshold.

[0272] The aforementioned semantic error detection method can conduct tests by pre-configuring a look-up table for threshold values. However, when performing semantic error detection based on soft values, a difference arises in that the threshold value must be calculated every time the semantic error detection technique is executed. By using soft values ​​in this way, more accurate semantic error detection performance can be obtained. The destination is based on Equation 1 according to the situation and Performing semantic error detection using or modeled based on soft values and Semantic error detection can be performed using .

[0273] 3) Identification of Reliable Task

[0274] When SCE-based semantic error detection is performed at the Destination, if it is determined that a semantic error has occurred, the Destination can achieve reliable communication by performing correction for the said semantic error. This will be explained below with reference to Fig. 21.

[0275] FIG. 21 is a diagram illustrating the process of obtaining the result of a task that can be performed according to an embodiment of the present specification.

[0276] To perform SCE-based semantic error correction at the Destination, a task with a reliable result is selected from among the tasks other than the Target task (Other Tasks; OT), and semantic error correction is performed based on it. For example, OT may refer to M-1 tasks excluding the Target task i out of a total of M tasks.

[0277] Referring to 21A in Fig. 21, soft reliability can be obtained using the softmax values ​​of the results. The soft reliability value of v-task's result (Top-1 result) of the result of the v-th task is It can be expressed as.

[0278] Referring to 21B in Fig. 21, )Is It can be obtained for.

[0279] Determining the reliability of the OT results can be performed in the same manner as performing semantic error detection on the OT results. The process of checking the reliability of the OT results is explained with reference to Fig. 22.

[0280] FIG. 22 illustrates a process for verifying the reliability of the results of a task that can be performed according to an embodiment of the present specification. Specifically, FIG. 22 illustrates a process for verifying the reliability of the results of OT 1 (other task 1) among OT (other tasks).

[0281] Referring to Fig. 22, the result of OT 1 In order to verify the reliability of Maps are utilized. For example, ) is the result of task j( It can mean the result(s) of task v mapped to ).

[0282] Similar to the semantic error detection process value threshold If it has a larger value compared to the value, it can be determined as a reliable result.

[0283] As a specific example, =2 for It can be decided. represents the soft value of task 2. Result of OT 1 The result of task 2 mapped to ( Based on whether it belongs to ). ga is determined to be 1 or 0. Likewise, the remainder For (= 3, .. M) is determined.

[0284] The sum of the values ​​determined for ( This threshold If it is greater than or equal to the value, it is the result of task 1 It can be determined as a reliable result.

[0285] The sum of the values ​​determined for ( This threshold If it is smaller than the value, it is the result of task 1 It can be determined as an incorrect result.

[0286] For example, threshold It can be obtained similarly to the process of calculating a threshold in semantic error detection. For example, the threshold It can vary depending on the task.

[0287] If no tasks are determined to have a reliable result through the reliable test, the destination can no longer perform semantic error correction. Therefore, the destination terminates the semantic error correction procedure and sends a NACK1 signal to the source to indicate that no reliable result exists. The source's response to the NACK1 signal is explained in detail in 5) Signaling. If one or more reliable results are determined in the reliable test, the results of those tasks are used to perform the next semantic error correction procedure.

[0288] 4) Update Semantic Representation

[0289] Semantic representation updates can be performed using the top-1 result of tasks identified as reliable results and their corresponding soft-values.

[0290] FIG. 23 illustrates the correction ratio for the labels of a target task using the results of a reliable task according to an embodiment of the present specification. Specifically, FIG. 23 shows the correction ratio for the labels of a target task using the top-1 result of a reliable task.

[0291] Referring to Fig. 23, it is assumed that the results of task 1 and task 3 are determined to be reliable results. In this case, the result of task 1 and task 3 result of map(= You can obtain the target task label using map. for The result for the map is am. for The result for the map is is. To put it differently, Determined based on the map target task mapped to The result of And, Based on the map target task mapped to The result of is. At this time, is target task It means a set containing all the labels of.

[0292] thus obtained The specific gravity for the map can be calculated as follows. An event that satisfies , An event that satisfies When it was said, The weight of the map is , , and It can be determined based on . At this time, Since this case is an event dealing with a situation where all reliable results are incorrect, is excluded.

[0293] In the case of It becomes, In the case of It becomes, In the case of This becomes the case. Using this, the ratio can be determined as the probability that each label of the target task is the correct label. That is, the ratio of each target task label can be set as follows.

[0294]

[0295]

[0296]

[0297] To make the total ratio equal to 1, each of the above ratios is a proportionality constant Divide by . At this time, is the reliable result among the labels of the target task. It refers to the set containing all labels existing in the map (in this example, (means).

[0298] As shown above, the semantic representation can be updated based on the ratio of target task labels. First, the Target task The loss function used when performing training on using the semantic error correction loss as shown in the following mathematical equation 4 can be calculated.

[0299]

[0300] At this time, represents the SR received from the destination and is a true label and the output top-1 result is Task for the case It refers to the loss value. In other words, the above semantic error correction loss equation represents the loss for the correct label candidates of the target task obtained as reliable results. It becomes the loss value summed according to the weighting of . Using this, SR Partial differentiation is possible, and this can be expressed as the following mathematical equation 5.

[0301]

[0302] If the SR is updated using the partial differential above, it can be expressed as Equation 6.

[0303]

[0304] At this time, is semantic error correction iteration It means SR when and represents the step-size. When the SR is updated as shown above, the updated SR increases the selection probability for correct label candidates of the target task. In other words, through SCE-based semantic error correction, the SR [corrects] the results of candidate labels that could be correct labels to the task The probability of obtaining the top-1 result can increase.

[0305] 5) Signaling

[0306] FIG. 24 illustrates a semantic error correction procedure according to an embodiment of the present specification. Specifically, FIG. 24 illustrates an SCE-based semantic error correction procedure.

[0307] Referring to Fig. 24, the destination receives an SR and first performs semantic error detection. If it is determined that there are no semantic errors in the received SR, the destination can perform the target task using the received SR. If it is determined that there are semantic errors in the received SR, the destination [can] the current semantic error correction iteration is the maximum iteration If you haven't passed it ( ), can perform semantic error correction procedures.

[0308] The first step of SCE-based semantic error correction is to identify reliable results. Specifically, reliable result(s) are identified among the results of OT (Other Tasks) other than the target task. If it is determined through this process that there are no reliable results, the destination can feed back a NACK1 signal to the source. If it is determined that there is one or more reliable results, the process can proceed to the next step.

[0309] The second procedure is to select the correct label candidates. Specifically, reliable results are used to calculate the candidates for the correct label of the target task and the proportion for each candidate label. The proportion for each correct label candidate is used in the following procedure.

[0310] The third step is to calculate the gradient descent for updating the semantic representation. Specifically, the semantic error correction loss value (Equation 4) is obtained using the proportion of correct label candidates and the loss function of the target task. Based on the semantic error correction loss value, the gradient for updating the semantic representation can be calculated.

[0311] The final step is the semantic representation update step. Specifically, the SR update can be performed using a gradient based on the semantic error correction loss (Equation 5) (Equation 6). The updated SR can then be checked for semantic errors through semantic error detection, at which point the number of semantic error correction iterations increases by 1.

[0312] FIG. 25 illustrates a signaling procedure according to an embodiment of the present specification.

[0313] Specifically, FIG. 25 illustrates a signaling procedure for SCE-based semantic error correction.

[0314] In S25010, initialization for semantic communication is performed.

[0315] In S25020, the destination transmits the information of target task i to the source.

[0316] In S25030, the source selects task k. The source is The largest value Select .

[0317] In S25040, source performs message encoding. Specifically, from message m, a language generator SR (i.e., message x) is generated by.

[0318] In S25050, the source is SR (message x) and Sends information about to the destination.

[0319] In S25060, the results of tasks that can be performed based on the received SR at the destination and the soft-value regarding this You can obtain.

[0320] In S25070, the destination performs a semantic error check using soft-values.

[0321] If it is determined that there is no semantic error, the destination sends an ACK 0 signal to the source and performs the target task based on the corresponding SR (S25080).

[0322] If it is determined that a semantic error has occurred as a result of the semantic error check, the destination performs a semantic error correction procedure for the corresponding SR as shown in FIG. 24 (S25090).

[0323] In S25100, if a semantic error is corrected during the semantic error correction procedure, the destination sends ACK 0 to the source. If a semantic error is not corrected during the semantic error correction procedure, the destination sends NACK 1 to the source or NACK 2 It is sent to the source as follows. At this time, is a set of indices of reliable tasks obtained from the first step of semantic error correction. It means a subset of

[0324] A NACK 1 signal indicates that there are no reliable results, and thus a NACK 1 signal may signify a case where a strong semantic error has occurred. In this case, the source that receives NACK 1 can retransmit the previous SR to the destination.

[0325] In the case of a NACK 2 signal, it may mean that while a reliable result exists, it contains insufficient information for semantic error correction. A source receiving NACK 2 is can maximize Select You can create a new SR and send it to the destination.

[0326] Additionally, if the following conditions are satisfied while the destination is performing semantic error correction, the semantic error correction procedure may be stopped early and a NACK 2 signal may be transmitted to the source.

[0327] - Checking entropy test fail

[0328] Checking entropy (CE) It can be defined as in the following mathematical formula 7 and used as a measure to check whether semantic error correction is proceeding well.

[0329]

[0330] Here It refers to a set of reliable tasks. is task It refers to the binary entropy of the result. A specific iteration ( From ) above If the phenomenon of increasing occurs, the destination can terminate semantic error correction and transmit a NACK 2 signal to the source. That is, If this condition is satisfied, it can be determined that the uncertainty regarding reliable results has increased; therefore, the destination can terminate semantic error correction and transmit a NACK 2 signal to the source. Here is iteration It means checking entropy in.

[0331] - Gradient vanishing

[0332] To find for updating SR If the value is too small Even if an update is performed through this, the change in value does not appear significantly. In this case, the destination may determine that it is difficult to proceed with semantic error correction further, terminate semantic error correction, and send a NACK 2 signal to the source.

[0333] According to the embodiments of the present specification described above, the following effects are derived.

[0334] To perform highly reliable semantic communication, it is necessary to check whether a semantic error has occurred in the signal received at the destination. If a semantic error has occurred, semantic error correction must be performed and information regarding this transmitted to the source, and if necessary, additional procedures to overcome the semantic error must be carried out. Through the technology proposed in this specification, detection and correction of semantic errors that may occur in semantic communication situations can be performed, thereby enabling highly reliable semantic communication.

[0335] In particular, the proposed technology can detect and correct semantic errors without transmitting additional redundancy from the source to the destination by considering semantic communication situations capable of performing multiple tasks. To this end, this technology utilizes a transformation corresponding to a semantic channel equalizer and a map capable of verifying the correlation between the results of each task. It utilized [this]. Through this, it is possible to design a semantic communication system capable of overcoming semantic errors without consuming additional resources for redundancy.

[0336] In terms of implementation, operations according to the embodiments described above (e.g., operations related to the detection and correction of semantic errors) can be processed by the device of FIGS. 1 to 5 described above (e.g., the processor (202a, 202b) of FIG. 2).

[0337] In addition, operations according to the above-described embodiments (e.g., operations related to the detection and correction of semantic errors) may be stored in memory (e.g., 204a, 204b of FIG. 2) in the form of instructions / programs (e.g., instructions, executable code) for driving at least one processor (e.g., processor (202a, 202b) of FIG. 2).

[0338] The embodiments described above will be explained in detail below with reference to FIGS. 26 and FIGS. 27 in terms of the operation of a wireless device (e.g., the first wireless device (200a) and the second wireless device (200b) of FIG. 2). The methods described below are distinguished only for convenience of explanation, and it is understood that a part of one method may be substituted with a part of another method or combined with one another and applied.

[0339] FIG. 26 is a flowchart illustrating a method according to one embodiment of the present specification.

[0340] Referring to FIG. 26, a method according to one embodiment of the present specification includes a first message receiving step (S2610), a second message transmitting step (S2620), and a response receiving step (S2630) related to the second message.

[0341] In the following, the first wireless device may mean a Source in semantic communication, and the second wireless device may mean a Destination in semantic communication.

[0342] In S2610, the first wireless device receives a first message from the second wireless device containing information related to the first task (e.g., task i). As an example, S2610 may be based on S25020 of FIG. 25.

[0343] In S2620, the first wireless device transmits a second message related to a semantic representation to the second wireless device. As an example, S2620 may be based on S25050 of FIG. 25.

[0344] For example, the second message may include semantic language information. The semantic language information may be generated based on a language generator associated with a specific task (e.g., task k). For example, the second message may further include information (e.g., k) about a task associated with the generation of the second message.

[0345] In S2630, the first wireless device receives a response related to the second message from the second wireless device. Based on whether a semantic error related to the semantic representation occurred or was corrected, the response indicates an ACK or a NACK. For example, S2630 may be based on S25080 or S25100 of FIG. 25.

[0346] According to one embodiment, a soft-value may be determined for each of the tasks performed based on the semantic language information. Whether or not the semantic error occurs or is corrected may be determined based on the soft-value. This embodiment may be based on embodiments (1) to 5) described above) related to SCE-based semantic error detection using soft-values.

[0347] Examples related to semantic error detection are described in detail below.

[0348] According to one embodiment, the semantic language information can be transformed into semantic language information for each of the tasks based on a semantic channel equalizer. Based on the transformed semantic language information, the results of each of the tasks can be obtained.

[0349] According to one embodiment, the soft value (e.g., ) are the above results (e.g., , The largest value among the probability values ​​for ) (e.g., the largest It can be based on the value.

[0350] According to one embodiment, i) among the results, the upper result related to the soft value (e.g., ) and ii) the above soft value (e.g., A first value can be determined based on ). The first value is of Equation 3 It can be based on. The above first value is a threshold value (e.g., It can be determined that the above semantic error occurred based on being smaller than ).

[0351] According to one embodiment, the tasks may include i) the first task (e.g., task i) and ii) second tasks (e.g., task j≠i). The upper result and the soft value may be based on the second tasks.

[0352] According to one embodiment, a second value may be determined based on the upper result. For example, referring to Equation 3, the second value It may mean... Based on the product of the above soft value and the above second value, the above first value (e.g., of Equation 3) ) can be determined.

[0353] Examples related to semantic error correction are described in detail below.

[0354] According to one embodiment, the tasks may include i) the first task (e.g., task i) and ii) second tasks (e.g., task j (≠i)). Based on the occurrence of the semantic error, at least one of the second tasks may be selected for the correction of the semantic error.

[0355] According to one embodiment, a third value may be determined for each of the second tasks. Among the second tasks, a second task in which the third value is greater than a threshold value may be selected for the correction. For example, the third value is as described in FIG. 22. It can be based on. The above threshold is It can be based on.

[0356] According to one embodiment, the third value for a specific second task (e.g., task 1 of FIG. 22) among the second tasks may be determined based on the remaining second tasks (e.g., task v(≠1) of FIG. 22) excluding the specific second task among the second tasks. The third value for the specific second task may be determined based on i) the soft value of each of the remaining second tasks and ii) a top result associated with the soft value.

[0357] According to one embodiment, a fourth value may be determined based on the upper result of each of the remaining second tasks. The third value may be determined based on the product of the soft value and the fourth value. For example, the soft value Based on, and the above upper result It can be based on.

[0358] According to one embodiment, the fourth value may be determined based on whether the upper result is included in one or more results determined based on a mapping relationship. The mapping relationship (e.g., ) can be defined between i) the results of each of the remaining second tasks and ii) the results of the specific second task. For example, the fourth value It can be based on. One or more results determined based on the above mapping relationship are It can be based on. For example, the above third value for task 1 is It can be expressed as.

[0359] According to one embodiment, there is a mapping relationship between the results of the first task and the results of each of the second tasks (e.g., ) can be defined. Based on the top result of each of the at least one second task, one or more results related to the first task can be determined from the mapping relationship. For example, referring to FIG. 23, the top result of task 1 ( Based on ), mapping relationship( Results related to target task i from ) This is determined. The top result of task 3 ( Based on ), mapping relationship( Results related to target task i from ) This is determined. The above one or more results are It can mean.

[0360] According to one embodiment, the probability that each of the one or more results is a correct result (e.g., The ratio can be determined based on ). The loss value can be determined based on the above ratio. The loss value can be based on Equation 4.

[0361] According to one embodiment, the semantic expression is the partial derivative of the loss value (e.g., of Equation 5). It can be updated based on ). The above partial derivative value may be related to gradient descent for the update. For example, the updated semantic expression is of Equation 6 It can be based on.

[0362] The following describes embodiments related to NACK signaling based on whether a semantic error occurs and whether a semantic error is corrected.

[0363] According to one embodiment, the NACK may be a first NACK or a second NACK.

[0364] According to one embodiment, based on the fact that a task related to the correction of the semantic error cannot be determined from the semantic expression, the response may indicate the first NACK. Based on the response indicating the first NACK, the second message related to the semantic expression may be retransmitted.

[0365] According to one embodiment, a task related to the correction of the semantic error is determined from the semantic expression, and based on the failure of the correction through the determined task: the response may indicate the second NACK.

[0366] Based on the response indicating the second NACK, information regarding i) a third message associated with the semantic representation and ii) a task (e.g., k') associated with the generation of the third message may be transmitted. The task (e.g., k') associated with the generation of the third message may be different from the task (e.g., k) for the generation of the second message.

[0367] Operations based on the above-described S2610 to S2630 can be implemented by the device of FIG. 2. For example, the first wireless device (200a) can control one or more transceivers (206a) and / or one or more memories (204a) to perform operations based on S2610 to S2630.

[0368] The embodiments described above will be explained in detail below in terms of the operation of the second wireless device.

[0369] S2710 to S2730 described below correspond to S2610 to S2630 described in FIG. 26. Considering the above correspondence, redundant descriptions are omitted. That is, the specific description of the operation of the second wireless device described below may be replaced by the description / embodiment of FIG. 26 corresponding to the operation.

[0370] FIG. 27 is a flowchart illustrating a method according to another embodiment of the present specification.

[0371] Referring to FIG. 27, a method according to another embodiment of the present specification includes a first message transmission step (S2710), a second message reception step (S2720), and a response transmission step (S2730) related to the second message.

[0372] In S2710, the second wireless device transmits a first message to the first wireless device containing information related to the first task (e.g., task i). As an example, S2710 may be based on S25020 of FIG. 25.

[0373] In S2720, the second wireless device receives a second message related to a semantic representation from the first wireless device. As an example, S2720 may be based on S25050 of FIG. 25.

[0374] In S2730, the second wireless device transmits a response related to the second message to the first wireless device. Based on whether a semantic error related to the semantic representation occurred or was corrected, the response indicates an ACK or a NACK. For example, S2730 may be based on S25080 or S25100 of FIG. 25.

[0375] Operations based on S2710 to S2730 described above can be implemented by the device of FIG. 2. For example, the second wireless device (200b) can control one or more transceivers (206b) and / or one or more memories (204b) to perform operations based on S2710 to S2730.

[0376] Here, the wireless communication technology implemented in the wireless devices (200a, 200b) of this specification may include LTE, NR, and 6G, as well as Narrowband Internet of Things for low-power communication. For example, NB-IoT technology may be an example of LPWAN (Low Power Wide Area Network) technology and may be implemented according to standards such as LTE Cat NB1 and / or LTE Cat NB2, but is not limited to the names mentioned above. Additionally, or generally, the wireless communication technology implemented in the wireless devices (200a, 200b) of this specification may perform communication based on LTE-M technology. For example, LTE-M technology may be an example of LPWAN technology and may be referred to by various names such as eMTC (enhanced Machine Type Communication). For example, LTE-M technology may be implemented in at least one of various standards such as 1) LTE CAT 0, 2) LTE Cat M1, 3) LTE Cat M2, 4) LTE non-BL (non-Bandwidth Limited), 5) LTE-MTC, 6) LTE Machine Type Communication, and / or 7) LTE M, and is not limited to the names mentioned above. Additionally or generally, wireless communication technology implemented in the wireless devices (200a, 200b) of this specification may include at least one of ZigBee, Bluetooth, and Low Power Wide Area Network (LPWAN) for low-power communication, and is not limited to the names mentioned above. As an example, ZigBee technology can create personal area networks (PANs) related to small / low-power digital communication based on various standards such as IEEE 802.15.4, and may be referred to by various names.

[0377] The embodiments described above are combinations of the components and features of this specification in a specific form. Each component or feature should be considered optional unless otherwise explicitly stated. Each component or feature may be implemented in a form not combined with other components or features. Additionally, it is possible to construct embodiments of this specification by combining some components and / or features. The order of operations described in the embodiments of this specification may be changed. Some components or features of one embodiment may be included in another embodiment, or may be replaced with corresponding components or features of another embodiment. It is obvious that embodiments may be constructed by combining claims that do not have an explicit citation relationship in the claims, or that they may be included as new claims through amendments made after filing.

[0378] Embodiments according to the present specification may be implemented by various means, e.g., hardware, firmware, software, or a combination thereof. In the case of implementation by hardware, an embodiment of the present invention may be implemented by one or more ASICs (application specific integrated circuits), DSPs (digital signal processors), DSPDs (digital signal processing devices), PLDs (programmable logic devices), FPGAs (field programmable gate arrays), processors, controllers, microcontrollers, microprocessors, etc.

[0379] In the case of implementation by firmware or software, an embodiment of the present specification may be implemented in the form of a module, procedure, function, etc., that performs the functions or operations described above. The software code may be stored in memory and executed by a processor. The memory may be located inside or outside the processor and may exchange data with the processor by various known means.

[0380] It is obvious to those skilled in the art that this specification may be embodied in other specific forms without departing from the essential features of this specification. Accordingly, the detailed description set forth above should not be interpreted restrictively in all respects but should be considered illustrative. The scope of this specification shall be determined by a reasonable interpretation of the appended claims, and all modifications within the equivalent scope of this specification are included within the scope of this specification.

Claims

1. Regarding the method, A step of receiving a first message containing information related to a first task from a second wireless device; The step of transmitting a second message related to a semantic representation to the second wireless device; and The method includes the step of receiving a response related to the second message from the second wireless device; Based on whether a semantic error related to the above semantic expression has occurred or been corrected, the response indicates an ACK or a NACK, and The second message above includes semantic language information generated based on the semantic expression, and For each of the tasks performed based on the above semantic language information, a soft value is determined, and A method characterized in that the occurrence or correction of the above semantic error is determined based on the above soft value.

2. In Paragraph 1, The above semantic language information is transformed into semantic language information for each of the above tasks based on a semantic channel equalizer, and A method characterized by obtaining the results of each of the above tasks based on the above-described converted semantic language information.

3. In Paragraph 2, The above soft value is a method based on the largest value among the probability values ​​for the above results.

4. In Paragraph 3, i) a higher result among the above results related to the soft value and ii) a first value is determined based on the soft value, and A method characterized by determining that the semantic error has occurred based on the fact that the first value is smaller than a threshold value.

5. In Paragraph 4, The above tasks include i) the first task and ii) the second tasks, and A method characterized by the above upper result and above soft value being based on the above second tasks.

6. In Paragraph 5, Based on the above upper result, a second value is determined, and A method characterized by determining the first value based on the product of the soft value and the second value.

7. In Paragraph 1, The above tasks include i) the first task and ii) the second tasks, and A method characterized by selecting at least one of the second tasks for the correction of the semantic error based on the occurrence of the semantic error.

8. In Paragraph 7, For each of the above second tasks, a third value is determined, and A method characterized in that among the above second tasks, the second task in which the third value is greater than the threshold value is selected for the correction.

9. In Paragraph 8, The third value for a specific second task among the above second tasks is determined based on the remaining second tasks among the above second tasks, excluding the specific second task, and A method characterized in that the third value for the specific second task is determined based on i) the soft value of each of the remaining second tasks and ii) the top result associated with the soft value.

10. In Paragraph 9, A fourth value is determined based on the upper result of each of the remaining second tasks mentioned above, and A method characterized in that the above third value is determined based on the product of the above soft value and the above fourth value.

11. In Paragraph 10, The above fourth value is determined based on whether the above upper result is included in one or more results determined based on a mapping relationship, and A method characterized by the above mapping relationship being defined between i) the results of each of the remaining second tasks and ii) the results of the specific second task.

12. In Paragraph 7, A mapping relationship is defined between the results of the first task and the results of each of the second tasks, and A method characterized by determining one or more results related to the first task from the mapping relationship based on the top result of each of the at least one second task.

13. In Paragraph 12, The ratio is determined based on the probability that each of the above one or more results is a correct result, and A method characterized by determining a loss value based on the above ratio.

14. In Paragraph 13, A method characterized in that the above semantic expression is updated based on the partial derivative of the above loss value.

15. In Paragraph 14, A method characterized in that the above partial derivative value is related to gradient descent for the above update.

16. In Paragraph 1, A method characterized in that the above NACK is a first NACK or a second NACK.

17. In Paragraph 16, Based on the fact that a task related to the correction of the semantic error cannot be determined from the above semantic expression, the response represents the first NACK, and A method characterized by retransmitting the second message associated with the semantic expression based on the response indicating the first NACK.

18. In Paragraph 16, A task related to the correction of the semantic error is determined from the above semantic expression, and based on the failure of the correction through the determined task: the response indicates the second NACK, Based on the response indicating the second NACK, information regarding i) a third message related to the semantic representation and ii) a task related to the generation of the third message is transmitted, and A method characterized in that the task related to the generation of the third message is different from the task for the generation of the second message.

19. In the first wireless device, One or more transmitters / receivers; One or more processors; and It includes one or more memories connected to the above one or more processors and storing instructions, A first wireless device characterized by the above instructions being set so that the one or more processors perform all steps of the method according to any one of claims 1 to 18, based on execution by the one or more processors.

20. An apparatus comprising one or more memories and one or more processors functionally connected to the one or more memories, An apparatus characterized in that the above one or more memories store instructions that set the one or more processors to perform all steps of the method according to any one of claims 1 to 18, based on execution by the above one or more processors.

21. In one or more non-transitory computer-readable storage media storing instructions, One or more non-transitory computer-readable storage media characterized by instructions executable by one or more processors, wherein the instructions set the one or more processors to perform all steps of the method according to any one of claims 1 to 18.

22. Regarding the method, A step of transmitting a first message containing information related to a first task to a first wireless device; A step of receiving a second message related to a semantic representation from the first wireless device; and The method includes the step of transmitting a response related to the second message to the first wireless device; Based on whether a semantic error related to the above semantic expression has occurred or been corrected, the response indicates an ACK or a NACK, and The second message above includes semantic language information generated based on the semantic expression, and For each of the tasks performed based on the above semantic language information, a soft value is determined, and A method characterized in that the occurrence or correction of the above semantic error is determined based on the above soft value.

23. In the second wireless device, One or more transmitters / receivers; One or more processors; and It includes one or more memories connected to the above one or more processors and storing instructions, A second wireless device characterized by the above instructions being set so that the one or more processors perform all steps of the method according to claim 22, based on execution by the one or more processors.

Images