Device and method for performing semantic communication in wireless communication system

Abstract

The objective of the present invention is to perform semantic communication in a wireless communication system. A method performed by a terminal may comprise the steps of: receiving first control information from a base station; receiving a first data signal from the base station on the basis of the first control information; transmitting, to the base station, feedback information for the first data signal; receiving second control information from the base station; and receiving a second data signal from the base station on the basis of the second control information.

Description

Device and method for performing semantic communication in a wireless communication system

[0001] The following description relates to a wireless communication system, and to a device and method for performing semantic communication in a wireless communication system.

[0002] Wireless access systems are widely deployed to provide various types of communication services, such as voice and data. Typically, wireless access systems are multiple access systems that support communications with multiple users by sharing available system resources (e.g., bandwidth, transmission power). Examples of multiple access systems include code division multiple access (CDMA), frequency division multiple access (FDMA), time division multiple access (TDMA), orthogonal frequency division multiple access (OFDMA), and single-carrier frequency division multiple access (SC-FDMA).

[0003] In particular, as numerous communication devices demand greater communication capacity, enhanced mobile broadband (eMBB) communication technologies are being proposed, improving upon existing radio access technology (RAT). Furthermore, massive machine type communications (mMTC), which connects multiple devices and objects to provide diverse services anytime and anywhere, as well as communication systems that consider reliability and latency-sensitive services / user equipment (UE), are being proposed. Various technological configurations are being proposed for these solutions.

[0004] The present disclosure can provide a device and method for effectively transmitting and receiving semantic information of a message in a wireless communication system.

[0005] The present disclosure may provide a device and method for synchronizing background knowledge for semantic communication in a wireless communication system.

[0006] The present disclosure may provide a device and method for dividing background knowledge into a plurality of partial background knowledge in a wireless communication system.

[0007] The present disclosure may provide a device and method for changing at least one partial background knowledge based on feedback information in a wireless communication system.

[0008] The present disclosure may provide a device and method for reducing at least one partial background knowledge based on feedback information in a wireless communication system.

[0009] The present disclosure can provide a device and method for transmitting semantic information based on a semantic diversity technique in a wireless communication system.

[0010] The present disclosure may provide a device and method for generating a plurality of semantic features for source data based on background knowledge synchronized with background knowledge of a destination in a wireless communication system.

[0011] The present disclosure may provide a device and method for generating a synthetic semantic feature by combining a plurality of semantic features based on feedback information in a wireless communication system.

[0012] The present disclosure may provide a device and method for determining a coupling ratio for a plurality of semantic features in a wireless communication system.

[0013] The present disclosure may provide a device and method for determining a combination ratio for a plurality of semantic features based on the size of the intersection of each of partial background knowledge of a source and background knowledge of a destination in a wireless communication system.

[0014] The technical objectives to be achieved in the present disclosure are not limited to those mentioned above, and other technical tasks not mentioned can be considered by a person having ordinary skill in the technical field to which the technical configuration of the present disclosure is applied from the embodiments of the present disclosure described below.

[0015] As an example of the present disclosure, a method performed by a terminal in a wireless communication system includes the steps of receiving first control information from a base station, receiving a first data signal from the base station based on the first control information, transmitting feedback information for the first data signal to the base station, receiving second control information from the base station, and receiving a second data signal from the base station based on the second control information, wherein the feedback information includes information used to determine similarities between each of the partial background knowledges derived from background knowledge information of the base station and background knowledge possessed by the terminal based on a plurality of semantic features generated based on the partial background knowledges, and the second data signal may include a synthesized semantic feature that combines the plurality of semantic features based on the feedback information.

[0016] As an example of the present disclosure, a method performed by a base station in a wireless communication system includes the steps of transmitting first control information to a terminal, transmitting a first data signal to the terminal based on the first control information, receiving feedback information for the first data signal from the terminal, transmitting second control information to the terminal, and transmitting a second data signal from the terminal based on the second control information, wherein the feedback information includes information used to determine similarities between each of the partial background knowledges derived from background knowledge information of the base station and background knowledge possessed by the terminal based on a plurality of semantic features generated based on the partial background knowledges, and the second data signal may include a synthesized semantic feature that combines a plurality of semantic features based on the feedback information.

[0017] As an example of the present disclosure, in a wireless communication system, a terminal includes a transceiver and a processor connected to the transceiver, wherein the processor controls to receive first control information from a base station, receive a first data signal from the base station based on the first control information, transmit feedback information for the first data signal to the base station, receive second control information from the base station, and receive a second data signal from the base station based on the second control information, wherein the feedback information includes information used to determine similarities between each of the partial background knowledges derived from background knowledge information of the base station and background knowledge possessed by the terminal based on a plurality of semantic features generated based on the partial background knowledges, and the second data signal may include a synthesized semantic feature that combines a plurality of semantic features based on the feedback information.

[0018] As an example of the present disclosure, in a wireless communication system, a base station includes a transceiver and a processor connected to the transceiver, wherein the processor transmits first control information to a terminal, transmits a first data signal to the terminal based on the first control information, receives feedback information on the first data signal from the terminal, transmits second control information to the terminal, and controls the terminal to transmit a second data signal based on the second control information, wherein the feedback information includes information used to determine similarities between each of the partial background knowledges derived from background knowledge information of the base station and background knowledge possessed by the terminal based on a plurality of semantic features generated based on the partial background knowledges, and the second data signal may include a synthesized semantic feature that combines a plurality of semantic features based on the feedback information.

[0019] As an example of the present disclosure, a communication device includes at least one processor, and at least one computer memory connected to the at least one processor and storing instructions that, when executed by the at least one processor, direct operations, the operations include: receiving first control information from a base station; receiving a first data signal from the base station based on the first control information; transmitting feedback information for the first data signal to the base station; receiving second control information from the base station; and receiving a second data signal from the base station based on the second control information, wherein the feedback information includes information used to determine similarities between each of the partial background knowledges derived from background knowledge information of the base station and background knowledge possessed by the communication device based on a plurality of semantic features generated based on the partial background knowledges, and wherein the second data signal may include a synthetic semantic feature that combines the plurality of semantic features based on the feedback information.

[0020] As an example of the present disclosure, a non-transitory computer-readable medium storing at least one instruction includes at least one instruction executable by a processor, the at least one instruction controlling a device to receive a first data signal from the base station based on first control information, transmit feedback information for the first data signal to the base station, receive second control information from the base station, and receive a second data signal from the base station based on the second control information, wherein the feedback information includes information used to determine similarities between each of the partial background knowledges derived from background knowledge information of the base station and background knowledge possessed by a terminal based on a plurality of semantic features generated based on the partial background knowledges, and the second data signal may include a synthetic semantic feature that combines the plurality of semantic features based on the feedback information.

[0021] The above-described aspects of the present disclosure are only some of the preferred embodiments of the present disclosure, and various embodiments reflecting the technical features of the present disclosure can be derived and understood by a person having ordinary skill in the art based on the detailed description of the present disclosure to be described below.

[0022] The following effects may be achieved by embodiments based on the present disclosure.

[0023] According to the present disclosure, the performance of semantic communication can be improved.

[0024] The effects that can be obtained from the embodiments of the present disclosure are not limited to the effects mentioned above, and other effects not mentioned can be clearly derived and understood by those skilled in the art to which the technical configuration of the present disclosure is applied, from the description of the embodiments of the present disclosure below. In other words, unintended effects resulting from implementing the configuration described in the present disclosure can also be derived from the embodiments of the present disclosure by those skilled in the art.

[0025] The accompanying drawings are intended to aid understanding of the present disclosure and, together with detailed descriptions, may provide embodiments of the present disclosure. However, the technical features of the present disclosure are not limited to specific drawings, and the features disclosed in each drawing may be combined with each other to form new embodiments. Reference numerals in each drawing may indicate structural elements.

[0026] Figure 1 illustrates an example of a communication system applicable to the present disclosure.

[0027] FIG. 2 illustrates an example of a wireless device applicable to the present disclosure.

[0028] FIG. 3 illustrates a method for processing a transmission signal applicable to the present disclosure.

[0029] Figure 4 illustrates a communication procedure between a terminal and a base station applicable to the present disclosure.

[0030] FIG. 5 illustrates an example of a communication structure that can be provided in a 6G (6th generation) system applicable to the present disclosure.

[0031] Figure 6 illustrates an electromagnetic spectrum applicable to the present disclosure.

[0032] FIG. 7 illustrates a THz wireless communication transceiver applicable to the present disclosure.

[0033] Figure 8 illustrates a THz signal generation method applicable to the present disclosure.

[0034] FIG. 9 illustrates a wireless communication transceiver applicable to the present disclosure.

[0035] Figure 10 illustrates a transmitter structure applicable to the present disclosure.

[0036] Figure 11 illustrates a modulator structure applicable to the present disclosure.

[0037] Figure 12 illustrates the structure of a perceptron included in an artificial neural network applicable to the present disclosure.

[0038] Figure 13 illustrates an artificial neural network structure applicable to the present disclosure.

[0039] Figure 14 illustrates an example of a functional framework for application of artificial intelligence technology applicable to the present disclosure.

[0040] Figure 15 illustrates an example of a procedure for utilizing an artificial intelligence model applicable to the present disclosure.

[0041] Figure 16 illustrates another example of a procedure for utilizing an artificial intelligence model applicable to the present disclosure.

[0042] Figure 17 illustrates another example of a procedure for utilizing an artificial intelligence model applicable to the present disclosure.

[0043] Figure 18 illustrates an AI technology-based communication procedure applicable to the present disclosure.

[0044] Figure 19 illustrates a communication model applicable to the present disclosure.

[0045] Figure 20 illustrates an example of a semantic communication framework applicable to the present disclosure.

[0046] FIG. 21 illustrates an example of a process of background knowledge synchronization for semantic communication according to one embodiment of the present disclosure.

[0047] FIG. 22 illustrates an example of a semantic diversity technique based on multi-feature transmission according to one embodiment of the present disclosure.

[0048] FIG. 23 illustrates an example of a partial background knowledge-based semantic diversity technique according to one embodiment of the present disclosure.

[0049] FIG. 24 illustrates an example of a contextualizing encoder structure according to one embodiment of the present disclosure.

[0050] FIG. 25 illustrates an example of determining whether there is an intersection of background knowledge according to one embodiment of the present disclosure.

[0051] FIG. 26 illustrates an example of a feedback injection encoder structure according to one embodiment of the present disclosure.

[0052] FIG. 27 illustrates an example of a semantic diversity technique based on a feedback injection encoder according to one embodiment of the present disclosure.

[0053] FIG. 28 illustrates an example of a combination ratio control considering a downstream task according to one embodiment of the present disclosure.

[0054] FIG. 29 illustrates an example of a procedure for receiving a data signal for semantic communication according to one embodiment of the present disclosure.

[0055] FIG. 30 illustrates an example of a procedure for transmitting a data signal for semantic communication according to one embodiment of the present disclosure.

[0056] FIG. 31 illustrates an example of a synchronization and coupling ratio control procedure for background knowledge according to one embodiment of the present disclosure.

[0057] FIG. 32 illustrates an example of a synchronization and coupling ratio control procedure for background knowledge according to one embodiment of the present disclosure.

[0058] FIG. 33 illustrates an example of a synchronization and coupling ratio control procedure for background knowledge according to one embodiment of the present disclosure.

[0059] FIG. 34 illustrates an example of a synchronization and coupling ratio control procedure for background knowledge according to one embodiment of the present disclosure.

[0060] FIG. 35 illustrates an example of a background knowledge synchronization and semantic communication procedure based on synchronized background knowledge according to one embodiment of the present disclosure.

[0061] The following embodiments combine the components and features of the present disclosure in a predetermined form. Each component or feature may be considered optional unless explicitly stated otherwise. Each component or feature may be implemented without being combined with other components or features. Furthermore, some components and / or features may be combined to form embodiments of the present disclosure. The order of operations described in the embodiments of the present disclosure 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.

[0062] In the description of the drawings, procedures or steps that may obscure the gist of the present disclosure are not described, and procedures or steps that can be understood by a person skilled in the art are also not described.

[0063] Throughout the specification, when a part is said to "comprising" or "including" a component, this does not mean that other components are excluded, but rather that other components can be included, unless specifically stated otherwise. In addition, terms such as "part," "unit," and "module" described in the specification mean a unit that processes at least one function or operation, which may be implemented by hardware, software, or a combination of hardware and software. In addition, the words "a" or "an," "one," "the," and similar related words may be used in the singular and plural sense in the context of describing the present disclosure (especially in the context of the claims below) unless otherwise indicated herein or clearly contradicted by context.

[0064] Embodiments of the present disclosure described herein focus on the data transmission and reception relationship between a base station and a mobile station. Here, the base station is understood as a terminal node of a network that directly communicates with the mobile station. Certain operations described herein as being performed by the base station may, in some cases, be performed by an upper node of the base station.

[0065] 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, the term '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.

[0066] Additionally, in the embodiments of the present disclosure, 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).

[0067] Additionally, a transmitter refers to a fixed and / or mobile node that provides data or voice services, and a receiver refers to a fixed and / or mobile node that receives data or voice services. Therefore, for uplink, a mobile station can be the transmitter, and a base station can be the receiver. Similarly, for downlink, a mobile station can be the receiver, and a base station can be the transmitter.

[0068] Embodiments of the present disclosure may be supported by standard documents disclosed in at least one of wireless access systems, such as IEEE 802.xx system, 3rd Generation Partnership Project (3GPP) system, 3GPP Long Term Evolution (LTE) system, 3GPP 5th generation (5G) NR (New Radio) system and 3GPP2 system, and in particular, embodiments of the present disclosure may be supported by 3GPP TS (technical specification) 38.211, 3GPP TS 38.212, 3GPP TS 38.213, 3GPP TS 38.321 and 3GPP TS 38.331 documents.

[0069] Furthermore, the embodiments of the present disclosure can be applied to other wireless access systems and are not limited to the systems described above. For example, they can be applied to systems implemented after the 3GPP 5G NR system and are not limited to a specific system.

[0070] That is, obvious steps or parts not described in the embodiments of the present disclosure can be explained by referring to the above documents. In addition, all terms disclosed in this document can be explained by the above standard documents.

[0071] Hereinafter, preferred embodiments according to the present disclosure will be described in detail with reference to the accompanying drawings. The detailed description set forth below, together with the accompanying drawings, is intended to illustrate exemplary embodiments of the present disclosure and is not intended to represent the only embodiments in which the technical configurations of the present disclosure may be implemented.

[0072] Additionally, specific terms used in the embodiments of the present disclosure are provided to aid in understanding of the present disclosure, and the use of such specific terms may be changed to other forms without departing from the technical spirit of the present disclosure.

[0073] 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).

[0074] For clarity, the following description is based on 3GPP communication systems (e.g., LTE, NR, etc.), but the technical spirit of the present disclosure is not limited thereto. LTE may refer to technology after 3GPP TS 36.xxx Release 8. Specifically, LTE technology after 3GPP TS 36.xxx Release 10 may be referred to as LTE-A, and LTE technology after 3GPP TS 36.xxx Release 13 may be referred to as LTE-A pro. 3GPP NR may refer to technology after TS 38.xxx Release 15. 3GPP 6G may refer to technology after TS Release 17 and / or Release 18. "xxx" refers to a standard document detail number. LTE / NR / 6G may be collectively referred to as a 3GPP system.

[0075] For background information, terms, abbreviations, etc. used in this disclosure, reference may be made to standard documents published prior to this disclosure. For example, reference may be made to standard documents 36.xxx and 38.xxx.

[0076] Communication system applicable to the present disclosure

[0077] Although not limited thereto, the various descriptions, functions, procedures, proposals, methods and / or operational flowcharts of the present disclosure disclosed in this document may be applied to various fields requiring wireless communication / connectivity (e.g., 5G) between devices.

[0078] Hereinafter, more specific examples will be provided with reference to the drawings. In the drawings / descriptions below, the same drawing reference numerals may represent identical or corresponding hardware blocks, software blocks, or functional blocks, unless otherwise described.

[0079] Figure 1 illustrates an example of a communication system applied to the present disclosure.

[0080] Referring to FIG. 1, a communication system (100) applied to the present disclosure includes a wireless device, a base station, and a network. Here, the wireless device refers to a device that performs communication using a wireless access technology (e.g., LTE, LTE-A, LTE-A pro, NR, 5G, 5G-A, 6G) and may be referred to as a communication / wireless / 5G device. Although not limited thereto, the 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 Things) device (100f), and an AI (artificial intelligence) device / server (100g). For example, the vehicle may include a vehicle equipped with a wireless communication function, an autonomous vehicle, a vehicle capable of performing vehicle-to-vehicle communication, etc. Here, the vehicles (100b-1, 100b-2) may include unmanned aerial vehicles (UAVs) (e.g., drones). The XR devices (100c) include augmented reality (AR) / virtual reality (VR) / mixed reality (MR) devices, and may be implemented in the form of head-mounted devices (HMDs), head-up displays (HUDs) installed in vehicles, televisions, smartphones, computers, wearable devices, home appliances, digital signage, vehicles, robots, etc. The portable devices (100d) may include smartphones, smart pads, wearable devices (e.g., smartwatches, smart glasses), computers (e.g., laptops, etc.), etc. The home appliances (100e) may include TVs, refrigerators, washing machines, etc. The IoT devices (100f) may include sensors, smart meters, etc.For example, the base station (120) and the network (130) may also be implemented as wireless devices, and a specific wireless device (120a) may act as a base station / network node to other wireless devices.

[0081] Wireless devices (100a to 100f) can be connected to a network (130) via a base station (120). AI technology can be applied to the wireless devices (100a to 100f), and the wireless devices (100a to 100f) can be connected to an AI server (100g) via a network (130). The network (130) can be configured using a 3G network, a 4G (e.g., LTE) network, a 5G (e.g., NR), or a 6G network. The wireless devices (100a to 100f) can communicate with each other via the base station (120) / network (130), but can 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). Additionally, an IoT device (100f) (e.g., a sensor) can communicate directly with another IoT device (e.g., a sensor) or another wireless device (100a to 100f).

[0082] Wireless communication / connection (150a, 150b, 150c) can be established between wireless devices (100a to 100f) / base stations (120), and base stations (120) / base stations (120). Here, the wireless communication / connection can be established through various wireless access technologies such as uplink / downlink communication (150a), sidelink communication (150b) (or D2D communication), and base station-to-base station communication (150c) (e.g., relay, IAB (integrated access backhaul)). Through the wireless communication / connection (150a, 150b, 150c), the wireless device and base station / wireless device, and base stations and base stations can transmit / receive wireless signals to / from each other. For example, the wireless communication / connection (150a, 150b, 150c) can transmit / receive signals through various physical channels. To this end, based on various proposals of the present disclosure, at least some of various configuration information setting processes for transmitting / receiving wireless signals, various signal processing processes (e.g., channel encoding / decoding, modulation / demodulation, resource mapping / demapping, etc.), resource allocation processes, etc. may be performed.

[0083] Devices applicable to the present disclosure

[0084] FIG. 2 illustrates an example of a wireless device applicable to the present disclosure.

[0085] Referring to FIG. 2, the wireless device (200) can transmit and receive wireless signals via various wireless access technologies (e.g., LTE, LTE-A, LTE-A pro, NR, 5G, 5G-A, 6G). The wireless device (200) includes at least one processor (202) and at least one memory (204), and may additionally include at least one transceiver (206) and / or at least one antenna (208).

[0086] The processor (202) controls the memory (204) and / or the transceiver (206), and may be configured to implement the descriptions, functions, procedures, proposals, methods, and / or operational flowcharts disclosed in this document. For example, the processor (202) may process information in the memory (204) to generate first information / signal, and then transmit a wireless signal including the first information / signal via the transceiver (206). In addition, the processor (202) may receive a wireless signal including second information / signal via the transceiver (206), and then store information obtained from signal processing of the second information / signal in the memory (204). The memory (204) may be connected to the processor (202) and may store various information related to the operation of the processor (202). For example, the memory (204) may store software code including instructions for performing some or all of the processes controlled by the processor (202), or for performing the descriptions, functions, procedures, proposals, methods, and / or operational flowcharts disclosed herein. Here, the processor (202) and the memory (204) may be part of a communication modem / circuit / chip designed to implement wireless communication technology. The transceiver (206) may be connected to the processor (202) and may transmit and / or receive wireless signals via at least one antenna (208). The transceiver (206) may include a transmitter and / or a receiver. The transceiver (206) may be used interchangeably with an RF (radio frequency) unit. In the present disclosure, a wireless device may also mean a communication modem / circuit / chip.

[0087] Hereinafter, the hardware elements of the wireless device (200) will be described in more detail. Although not limited thereto, at least one protocol layer may be implemented by at least one processor (202). For example, at least one processor (202) may implement at least one layer (e.g., a functional layer such as physical (PHY), media access control (MAC), radio link control (RLC), packet data convergence protocol (PDCP), radio resource control (RRC), and service data adaptation protocol (SDAP)). At least one processor (202) may generate at least one Protocol Data Unit (PDU) and / or at least one Service Data Unit (SDU) according to the descriptions, functions, procedures, proposals, methods, and / or operation flowcharts disclosed in this document. At least one processor (202) may generate a message, control information, data, or information according to the descriptions, functions, procedures, proposals, methods, and / or operation flowcharts disclosed in this document. At least one processor (202) can generate a signal (e.g., a baseband signal) including a PDU, an SDU, a message, control information, data or information according to the functions, procedures, proposals and / or methods disclosed in this document, and provide the signal to at least one transceiver (206). At least one processor (202) can receive a signal (e.g., a baseband signal) from at least one transceiver (206) and obtain the PDU, SDU, message, control information, data or information according to the descriptions, functions, procedures, proposals, methods and / or operational flowcharts disclosed in this document.

[0088] At least one processor (202) may be referred to as a controller, a microcontroller, a microprocessor, or a microcomputer. The at least one processor (202) may be implemented by hardware, firmware, software, or a combination thereof. For example, at least one application specific integrated circuit (ASIC), at least one digital signal processor (DSP), at least one digital signal processing device (DSPD), at least one programmable logic device (PLD), or at least one field programmable gate array (FPGA) may be included in the at least one processor (202). The descriptions, functions, procedures, proposals, methods, and / or operation flowcharts disclosed in this document may be implemented using firmware or software, and the firmware or software may be implemented to include modules, procedures, functions, etc. The descriptions, functions, procedures, proposals, methods, and / or operation flowcharts disclosed in this document may be included in the at least one processor (202), or may be stored in at least one memory (204) and driven by the at least one processor (202). The descriptions, functions, procedures, suggestions, methods and / or flowcharts disclosed in this document may be implemented using firmware or software in the form of code, instructions and / or sets of instructions.

[0089] At least one memory (204) can be connected to at least one processor (202) and can store various forms of data, signals, messages, information, programs, codes, instructions and / or commands. The at least one memory (204) can be configured as a read only memory (ROM), a random access memory (RAM), an erasable programmable read only memory (EPROM), a flash memory, a hard drive, a register, a cache memory, a computer readable storage medium and / or a combination thereof. The at least one memory (204) can be located internally and / or externally to the at least one processor (202). In addition, the at least one memory (204) can be connected to the at least one processor (202) via various technologies such as a wired or wireless connection.

[0090] At least one transceiver (206) can transmit user data, control information, wireless signals / channels, etc., mentioned in the methods and / or flowcharts of this document to at least one other device. At least one transceiver (206) can receive user data, control information, wireless signals / channels, etc. mentioned in the descriptions, functions, procedures, proposals, methods and / or flowcharts disclosed in this document from at least one other device. For example, at least one transceiver (206) can be connected to at least one processor (202) and can transmit and receive wireless signals. For example, at least one processor (202) can control at least one transceiver (206) to transmit user data, control information, or wireless signals to at least one other device. Furthermore, at least one processor (202) can control at least one transceiver (206) to receive user data, control information, or wireless signals from at least one other device. In addition, at least one transceiver (206) may be connected to at least one antenna (208), and at least one transceiver (206) may be configured to transmit and receive user data, control information, wireless signals / channels, etc. mentioned in the descriptions, functions, procedures, proposals, methods and / or operation flowcharts disclosed in this document through at least one antenna (208). In this document, at least one antenna may be a plurality of physical antennas or a plurality of logical antennas (e.g., antenna ports). At least one transceiver (206) may convert the received wireless signals / channels, etc. from RF band signals to baseband signals in order to process the received user data, control information, wireless signals / channels, etc. using at least one processor (202). At least one transceiver (206) may convert the processed user data, control information, wireless signals / channels, etc. from baseband signals to RF band signals using at least one processor (202).For this purpose, at least one transceiver (206) may include an (analog) oscillator and / or filter.

[0091] The components of the wireless device described with reference to FIG. 2 may be referred to by different terms in terms of functionality. For example, the processor (202) may be referred to as a control unit, the transceiver (206) as a communication unit, and the memory (204) as a storage unit. In some cases, the communication unit may be used to mean at least a portion of the processor (202) and the transceiver (206).

[0092] The structure of the wireless device described with reference to FIG. 2 can be understood as the structure of at least a portion of various devices. For example, the structure of the wireless device illustrated in FIG. 2 can be at least a portion of various devices described with reference to FIG. 1 (e.g., a robot (100a), a vehicle (100b-1, 100b-2), an XR device (100c), a portable device (100d), a home appliance (100e), an IoT device (100f), an AI device / server (100g)). Furthermore, according to various embodiments, in addition to the components illustrated in FIG. 2, the device may further include other components.

[0093] For example, the device may be a portable device such as a smartphone, a smart pad, a wearable device (e.g., a smart watch, smart glasses), or a portable computer (e.g., a laptop, etc.). In this case, the device may further include at least one of a power supply unit that supplies power and includes a wired / wireless charging circuit, a battery, etc., an interface unit that includes at least one port for connection with another device (e.g., an audio input / output port, a video input / output port), and an input / output unit for inputting and outputting image information / signals, audio information / signals, data, and / or information input from a user.

[0094] For example, the device may be a mobile device such as a mobile robot, a vehicle, a train, an aerial vehicle (AV), a ship, etc. In this case, the device may further include at least one of a driving unit including at least one of an engine, a motor, a power train, wheels, brakes, and a steering unit of the device, a power supply unit including a wired / wireless charging circuit, a battery, etc. that supplies power, a sensor unit that senses status information, environmental information, and user information of the device or its surroundings, an autonomous driving unit that performs functions such as path maintenance, speed control, and destination setting, and a position measurement unit that obtains location information of the mobile device through a global positioning system (GPS) and various sensors.

[0095] For example, the device may be an XR device such as an HMD, a head-up display (HUD) installed in a vehicle, a television, a smartphone, a computer, a wearable device, a home appliance, a digital signage, a vehicle, a robot, etc. In this case, the device may further include at least one of a power supply unit that supplies power and includes a wired / wireless charging circuit, a battery, etc., an input / output unit that obtains control information, data, etc. from the outside and outputs the generated XR object, and a sensor unit that senses status information, environmental information, and user information of the device or the surroundings of the device.

[0096] For example, the device may be a robot that can be classified into industrial, medical, household, military, etc. types depending on the purpose or field of use. In this case, the device may further include at least one of a sensor unit that senses status information, environmental information, and user information of the device or its surroundings, and a driving unit that performs various physical actions, such as moving the robot joints.

[0097] For example, the device may be an AI device such as a TV, a projector, a smartphone, a PC, a laptop, a digital broadcasting terminal, a tablet PC, a wearable device, a set-top box (STB), a radio, a washing machine, a refrigerator, digital signage, a robot, a vehicle, etc. In this case, the device may further include at least one of an input unit that acquires various types of data from the outside, an output unit that generates output related to sight, hearing, or touch, a sensor unit that senses status information, environmental information, and user information of the device or its surroundings, and a training unit that trains a model composed of an artificial neural network using learning data.

[0098] The structure of the wireless device illustrated in FIG. 2 may be understood as a part of a RAN node (e.g., a base station, DU, RU, RRH, etc.). That is, the device illustrated in FIG. 2 may be a RAN node. In this case, the device may further include a wired transceiver for front haul and / or back haul communications. However, if the front haul and / or back haul communications are based on wireless communications, at least one transceiver (206) illustrated in FIG. 2 may be used for front haul and / or back haul communications, and a wired transceiver may not be included.

[0099] FIG. 3 illustrates a method for processing a transmission signal applicable to the present disclosure. For example, the transmission signal may be processed by a signal processing circuit. At this time, 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). At this time, as an example, the operations / functions of FIG. 3 may be performed in the processor (202) and / or the transceiver (206) of FIG. 2. Furthermore, as an example, the hardware elements of FIG. 3 may be implemented in the processor (202) and / or the transceiver (206) of FIG. 2. As an example, blocks 310 to 360 may be implemented in the processor (202) of FIG. 2. Additionally, blocks 310 to 350 may be implemented in the processor (202) of FIG. 2, and block 360 may be implemented in the transceiver (206) of FIG. 2, and are not limited to the above-described embodiment.

[0100] The 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 transport block (e.g., a UL-SCH transport block, a DL-SCH transport block). The wireless signal may be transmitted through various physical channels (e.g., a PUSCH, a PDSCH). Specifically, the codeword can be converted into a bit sequence scrambled by a scrambler (310). The scramble sequence used for scrambling is generated based on an initialization value, and the initialization value may include ID information of the wireless device, etc. The scrambled bit sequence can be modulated into a modulation symbol sequence by a modulator (320). The modulation scheme may include pi / 2-binary phase shift keying (pi / 2-BPSK), m-phase shift keying (m-PSK), m-quadrature amplitude modulation (m-QAM), etc.

[0101] A complex modulation symbol sequence can be mapped to at least one transmission layer by a layer mapper (330). The modulation symbols of each transmission layer can be mapped to corresponding antenna port(s) by a precoder (340). The output z of the precoder (340) can be obtained by multiplying the output y of the layer mapper (330) by a precoding matrix W of NХM, 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., discrete Fourier transform (DFT) transform) on the complex modulation symbols. Additionally, the precoder (340) can perform precoding without performing transform precoding.

[0102] The resource mapper (350) can map modulation symbols of each antenna port to time-frequency resources. The time-frequency resources can include multiple symbols (e.g., CP-OFDMA symbols, DFT-s-OFDMA symbols) in the time domain and multiple subcarriers in the frequency domain. The signal generator (360) generates a wireless signal from the mapped modulation symbols, and the generated wireless signal can be transmitted to another device through each antenna. To this end, the signal generator (360) can include an inverse fast Fourier transform (IFFT) module, a cyclic prefix (CP) inserter, a digital-to-analog converter (DAC), a frequency uplink converter, and the like.

[0103] The signal processing process for a received signal in a wireless device may be configured in reverse order of the signal processing process (310 to 360) of FIG. 3. For example, a wireless device (e.g., 200 of FIG. 2) may receive a wireless signal from the outside through an antenna port / transceiver. The received wireless signal may be converted into a baseband signal through a signal restorer. For this purpose, 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. Thereafter, the baseband signal may be restored to a codeword through a resource demapper process, a postcoding process, a demodulation process, and a descrambling process. The codeword may be restored to the original information block through decoding. Therefore, a signal processing circuit (not shown) for a received signal may include a signal restorer, a resource demapper, a postcoder, a demodulator, a descrambler, and a decoder.

[0104] Figure 4 illustrates a communication procedure between a terminal and a base station applicable to the present disclosure. Figure 4 illustrates operations of a terminal (410) and a base station (420) transmitting and / or receiving data and operations performed prior thereto.

[0105] Referring to FIG. 4, in step 401, the terminal (410) and the base station (420) perform synchronization. For example, the terminal (410) performs an initial cell search operation. Specifically, the terminal (410) can detect at least one synchronization signal transmitted from the base station (420) according to a predefined rule. Here, the synchronization signal can include multiple synchronization signals classified according to structure or purpose (e.g., primary synchronization signal, secondary synchronization signal). Through this, the terminal (410) can check the boundary of the frame, subframe, slot, and / or symbol of the base station (420) and obtain information about the base station (420) (e.g., cell identifier).

[0106] In step 403, the terminal (410) obtains system information transmitted from the base station (420). The system information is information related to the properties, characteristics, and / or capabilities of the base station (420) required to access the base station (420) and use the service, and may be classified according to the content (e.g., whether it is essential for access), transmission structure (e.g., channel used, whether provided on-demand), etc., and may be classified into, for example, a master information block (MIB) and a system information block (SIB). If necessary, the terminal (410) may transmit a signal requesting system information before receiving the system information. However, the request and provision of the system information may be performed after the random access procedure described below.

[0107] In step 405, the terminal (410) and the base station (420) perform a random access procedure. The terminal (410) may transmit and / or receive at least one message (e.g., a random access preamble, a random access response (RAR) message, etc.) for the random access procedure based on information related to the random access channel of the base station (420) obtained through system information (e.g., channel position, channel structure, supported preamble structure, etc.). For example, the terminal (410) may transmit a preamble (e.g., MSG1) through the random access channel, receive an RAR message (e.g., MSG2), transmit a message (e.g., MSG3) including information related to the terminal (410) (e.g., identification information) to the base station (420) using scheduling information included in the RAR message, and receive a message (e.g., MSG4) for contention resolution and / or connection establishment. As another example, MSG1 and MSG3 may be sent and received as one message, or MSG2 and MSG4 may be sent and received as one message.

[0108] In step 407, the terminal (410) and the base station (420) perform signaling of control information. Here, the control information may be defined in various layers, such as a layer that controls a connection (e.g., a radio resource control (RRC) layer), a layer that handles mapping between logical channels and transport channels (e.g., a media access control (MAC) layer), and a layer that handles physical channels (e.g., a physical (PHY) layer). For example, the terminal (410) and the base station (420) may perform at least one of signaling for establishing a connection, signaling for determining settings related to communication, and signaling for indicating allocated resources.

[0109] In step 409, the terminal (410) and the base station (420) transmit and / or receive data. In other words, the terminal (410) and the base station (420) can process, transmit, and / or receive data based on the signaling of the control information. For example, when transmitting data, the terminal (410) or the base station (420) can perform at least one of channel encoding, rate matching, scrambling, constellation mapping, layer mapping, waveform modulation, antenna mapping, and resource mapping on the information bits. Conversely, when receiving data, the terminal (410) or the base station (420) can perform at least one of signal extraction from resources, waveform demodulation for each antenna, signal arrangement considering layer mapping, constellation demapping, descrambling, and channel decoding.

[0110] 6G communication systems and core implementation technologies of 6G systems

[0111] The 5G system defines various operating bands within FR1 (frequency range 1), which covers 410 MHz to 7125 MHz, and FR2 (frequency range 2), which covers 24,250 MHz to 71,000 MHz. Various frequencies are being discussed as operating bands for the subsequent 6G system, and the use of higher frequencies than 5G systems is also being considered for wider bandwidth and higher transmission speeds. One such band is the THz (terahertz) frequency band, which covers approximately 100 GHz to 10 THz. The THz frequency band is a band that has both the transparency of radio waves and the straightness of light waves, and communications using the THz frequency band are expected to play a transitional role from existing radio-centered communications to lightwave-based communications.

[0112] 6G systems utilizing the THz frequency band have the following goals: 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 divided into four aspects: “intelligent connectivity,” “deep connectivity,” “holographic connectivity,” and “ubiquitous connectivity,” and the 6G system can be designed to satisfy the requirements as shown in [Table 1] below.

[0113] Per device peak data rate1 TbpsE2E latency1 msMaximum spectral efficiency100 bps / HzMobility supportup to 1000 km / hrSatellite integrationFullyAIFullyAutonomous vehicleFullyXRFullyHaptic CommunicationFully

[0114] At this time, the 6G system may have key factors such as enhanced mobile broadband (eMBB), ultra-reliable low latency communications (URLLC), massive machine type communications (mMTC), AI integrated communication, tactile internet, high throughput, high network capacity, high energy efficiency, low backhaul and access network congestion, and enhanced data security. FIG. 5 illustrates an example of a communication structure that can be provided in a 6G system applicable to the present disclosure. Referring to FIG. 5, the 6G system is expected to have simultaneous wireless communication connectivity that is 50 times higher than that of a 5G wireless communication system. URLLC, a key feature of 5G, is expected to become even more crucial in 6G communications by providing end-to-end latency of less than 1 ms. Furthermore, 6G systems will boast significantly higher volumetric spectral efficiency than the commonly used area spectral efficiency. 6G systems can offer extremely long battery life and advanced battery technologies for energy harvesting, potentially eliminating the need for separate charging for mobile devices in 6G systems.

[0115] As core implementation technologies of the 6G system, technologies such as artificial intelligence (AI), THz (terahertz) communication, optical wireless technology, FSO backhaul network, massive MIMO technology, blockchain, 3D networking, quantum communication, unmanned aerial vehicles, cell-free communication, wireless information and energy transfer (WIET), integration of sensing and communication, integration of access backhaul networks, holographic beamforming, big data analysis, and large intelligent surface (LIS) can be adopted.

[0116] For example, THz communication is a communication that utilizes a spectrum in a frequency band between 0.1 THz and 10 THz with a corresponding wavelength in the range of 0.03 mm to 3 mm as shown in Fig. 6, and can be implemented using circuit elements having a structure as shown in Fig. 7. In addition, optical wireless technology is a technology that generates and modulates THz signals using optical elements, and can be implemented based on devices having structures as shown in Figs. 8, 9, 10, and 11.

[0117] In addition, AI can be implemented based on various models such as neural networks and machine learning (machine models). For example, an AI model of a neural network structure can be based on the structure of a perceptron as shown in Fig. 12. Referring to Fig. 12, an artificial neural network can be composed of multiple perceptrons. According to the structure of the perceptron, when an input vector x={x1, x2, …, xd} is input, each component is multiplied by a weight {W1, W2, …, Wd}, and all the results are added, and then an activation function σ(·) is applied. A large artificial neural network structure can be formed by extending the simplified perceptron structure illustrated in Fig. 12, and the input vector can be applied to perceptrons of different dimensions. When perceptrons are stacked, a neural network having an input layer, a hidden layer, and an output layer as shown in Fig. 13 can be configured.

[0118] AI technology utilizing the structure of the neural network as described above can be operated based on a functional framework as shown in FIG. 14 below. FIG. 14 illustrates an example of a functional framework for application of AI technology applicable to the present disclosure. First, the data collection block (1410) performs data preparation on input data collected from objects (e.g., UE, RAN node, network node, etc.) to generate training data (1411) and / or inference data (1411) including processed input data. The model training block (1420) performs training on an AI model using the training data (1411) and provides information on the trained model to the model inference block (1430). The model inference block (1430) generates an output (1416) by performing inference and / or prediction using the inference data (1411). Additionally, the model inference block (1430) can provide model performance feedback (1414) to the model training block (1420). Here, the output (1416) refers to the inference output of the AI ​​model generated by the model inference block (1430), and the details of the inference output may vary depending on the use case. The actor block (1440) triggers or performs a specified task / action based on the output (1416). The actor block (1440) can trigger a task / action for another object (e.g., at least one UE, at least one RAN node, at least one network node, etc.) or for itself. Any one of the functions illustrated in FIG. 14 described above may be performed by two or more entities including the RAN, the network node, the network operator's OAM, or the UE in collaboration. This may be referred to as a split AI operation.

[0119] FIG. 15 illustrates an example of a procedure for utilizing an AI model applicable to the present disclosure. FIG. 15 illustrates a case where a model training function (e.g., a function of a model training block (1420)) is included in a network node, and a model inference function (e.g., a function of a model inference block (1430)) is included in a RAN node. Referring to FIG. 15 , in step 1, RAN node 1 and RAN node 2 transmit input data (e.g., training data) for training an AI model to the network node. Here, RAN node 1 and RAN node 2 may also transmit data collected from the UE (e.g., measurements of the UE related to RSRP, RSRQ, SINR of the serving cell and neighboring cells, the UE's position, velocity, etc.) to the network node. In step 2, the network node trains the AI ​​model using the received training data. In step 3, the network node distributes / updates the AI ​​model to RAN node 1 and / or RAN node 2. RAN node 1 and / or RAN node 2 may continue model training based on the received AI model. In this procedure, it is assumed that the AI ​​model is deployed / updated only to RAN node 1. In step 4, RAN node 1 receives input data (e.g., inference data) for AI model inference from the UE and RAN node 2. In step 5, RAN node 1 performs inference based on the AI ​​model using the received inference data to generate output data (e.g., a prediction or a decision). In step 6, if applicable, RAN node 1 may send model performance feedback to network nodes. In step 7, RAN node 1, RAN node 2, and UE (or 'RAN node 1 and UE', or 'RAN node 1 and RAN node 2') perform actions based on the output data. For example, in case of load balancing operation, the UE may move from RAN node 1 to RAN node 2.In step 8, RAN node 1 and RAN node 2 transmit feedback information to the network nodes.

[0120] FIG. 16 illustrates another example of a procedure for utilizing an AI model applicable to the present disclosure. FIG. 16 illustrates a case where a model training function (e.g., a function of a model training block (1420)) and a model inference function (e.g., a function of a model inference block (1430)) are included in a RAN node. Referring to FIG. 16, in step 1, a UE and a RAN node 2 transmit input data (e.g., training data) for training an AI model to a RAN node 1. In step 2, RAN node 1 trains an AI model using the received training data. In step 3, RAN node 1 receives input data (e.g., inference data) for AI model-based inference from the UE and RAN node 2. In step 4, RAN node 1 performs AI model-based inference using the received inference data to generate output data (e.g., a prediction or decision). In step 5, RAN node 1, RAN node 2, and UE (or 'RAN node 1 and UE', or 'RAN node 1 and RAN node 2') perform actions based on the output data. For example, in case of load balancing operation, the UE may move from RAN node 1 to RAN node 2. In step 6, RAN node 2 transmits feedback information to RAN node 1.

[0121] FIG. 17 illustrates another example of a procedure for utilizing an AI model applicable to the present disclosure. FIG. 17 illustrates a case where a model training function (e.g., a function of a model training block (1420)) is included in a RAN node, and a model inference function (e.g., a function of a model inference block (1430)) is included in a UE. Referring to FIG. 17, in step 1, a UE transmits input data (e.g., training data) for training an AI model to a RAN node. Here, the RAN node may collect data from various UEs and / or other RAN nodes. In step 2, the RAN node trains an AI model using the received training data. In step 3, the RAN node distributes / updates the AI ​​model to the UE. The UE may continue model training based on the received AI model. In step 4, input data (e.g., inference data) for inference based on the AI ​​model is received from the UE, the RAN node, and / or other UEs. In step 5, the UE generates output data (e.g., predictions or decisions) by performing AI model-based inference using the received inference data. In step 6, if applicable, the UE may transmit model performance feedback to the RAN node. In step 7, the UE and the RAN node perform actions based on the output data. In step 8, the UE transmits feedback information to the RAN node.

[0122] According to the aforementioned framework and procedures, an AI model can be trained and utilized in a wireless communication system. In the aforementioned framework and procedures, various data, such as input data, training data, and inference data, are introduced. The specific content of the aforementioned data may vary depending on the task for which the AI ​​model is utilized. For example, information used in various embodiments of the present disclosure described below may be included in the aforementioned data.

[0123] Figure 18 illustrates an AI technology-based communication procedure applicable to the present disclosure. The detailed procedures illustrated in Figure 18 can be combined with various embodiments of the present disclosure described below. For example, data generated according to various embodiments of the present disclosure can be used for operations (e.g., configuration, training, inference, and / or data transmission / reception) in at least one of the detailed procedures illustrated in Figure 18. As another example, the results of the inference illustrated in Figure 18 can be used to transmit and / or receive data according to various embodiments of the present disclosure.

[0124] Referring to FIG. 18, in step 1801, at least one of a UE (1810), a RAN node (1820), and a network node (1830) performs an initial access procedure. For example, in this step, at least one of an initial cell search operation, a system information acquisition operation, a random access operation, and a registration operation may be performed. In step 1803, at least one of the UE (1810), the RAN node (1820), and the network node (1830) performs a configuration procedure. Through the configuration procedure, parameters, resources, connections, and / or entities necessary for performing subsequent procedures in layers between the UE (1810) and the RAN node (1820) and / or in at least one layer between the UE (1810) and the network node (1830) may be determined and / or created. In this case, the configuration procedure may be performed based on information, status, and / or characteristics of an AI model used for subsequent training and inference.

[0125] At step 1805, at least one of the UE (1810), the RAN node (1820), and the network node (1830) performs a model training procedure. At least one of the UE (1810), the RAN node (1820), and the network node (1830) may collect training data and perform learning using the training data. For example, the model training procedure may be performed as described with reference to FIG. 15, FIG. 16, or FIG. 17. If an offline-trained model is used, this step may be omitted.

[0126] At step 1807, at least one of the UE (1810), the RAN node (1820), and the network node (1830) performs a task using the trained model. That is, the task may be performed based on the results of inference and / or prediction using the trained model. For example, the task may be a procedure belonging to a communication protocol, and may be a preparatory operation for subsequent data transmission and / or reception, or may be related to data transmission and / or reception, or may be related to data processing (e.g., encoding, decoding, etc.).

[0127] In step 1809, at least one of the UE (1810), the RAN node (1820), and the network node (1830) transmits and / or receives data. At this time, the result of the task performed in step 1807 may be used. In some cases, the task performed in step 1807 may include transmitting and / or receiving data, in which case this step may be omitted as it is part of step 1807.

[0128] Specific embodiments of the present invention

[0129] The present disclosure relates to a technology for transmitting and receiving semantic features in a wireless communication system that supports semantic communication. Specifically, the present disclosure relates to a technology for synchronizing background knowledge of a source and a destination in a wireless communication system, and performing semantic communication based on the synchronized background knowledge. Here, the semantic feature may be referred to as a semantic representation. Furthermore, the background knowledge refers to data sets required to generate and interpret semantic representations for semantic communication. The background knowledge applicable to the present disclosure may be defined in various forms, and for example, may be defined as at least one data set in the form of a graph representing relationships between nodes.

[0130] Problems related to communication can be divided into three levels, as illustrated in Figure 19, according to the philosophy of Shannon and Weaver. Figure 19 illustrates a communication model applicable to the present disclosure. The problem at Level A (1910) is a technical problem, which concerns how accurately symbols in communication can be conveyed. The problem at Level B (1920) is a semantic problem, which concerns how accurately the conveyed symbols convey the desired meaning. The problem at Level C (1930) is an effectiveness problem, which concerns how effectively the received meaning influences the operation in the desired manner.

[0131] Shannon's information theory focuses only on technical problems at Level A (1910) and does not consider communication from a semantic perspective. However, Weaver argues that Shannon's information theory is sufficiently general to consider problems at Level B (1920) and Level C (1930) by adding semantic transmitters, semantic receivers, and semantic noise to Shannon's communication model.

[0132] Meanwhile, one of the various goals of 6G communications is to enable a variety of new services that interconnect people and machines. Therefore, there is a need to provide a semantic communication method that goes beyond simply considering the technical issues of Level A (1910) in communication systems and considers the semantic issues of Level B (1920).

[0133] Typically, when communicating information between people, each word is associated with its corresponding "meaning." Considering this in relation to the communication model in Figure 19, if the concept associated with the message sent from the source is accurately interpreted by the destination, then proper semantic communication can be considered to have occurred.

[0134] For semantic communication, a source can generate semantic features based on given or collected raw data and transmit the generated semantic features to a destination. The destination interprets and infers the received semantic features according to the source's intent. At this time, semantic communication is not primarily about reducing reconstruction errors that occur during the process of restoring received semantic representations to their original raw data. Rather, it requires an approach that assesses whether downstream tasks performed by the destination utilize the received semantic features to operate according to the source's intent, i.e., whether the interpretation and / or inference of semantic features is successful. Therefore, the destination utilizes its own background knowledge when performing inference operations, and it is desirable for the background knowledge contained in the data transmitted from the source to be reflected in the destination's background knowledge to ensure correct interpretation results.

[0135] As mentioned above, semantic features generated at the source and delivered to the destination must be created with consideration for the downstream tasks that will be performed at the destination. Therefore, a task-oriented semantic communication system is needed to facilitate semantic communication. This system preserves task-relevant information while introducing useful invariances for downstream tasks.

[0136] The semantic communication characteristics of Level B of FIG. 19 can be expressed as in FIG. 20. FIG. 20 illustrates an example of a semantic communication framework applicable to the present disclosure. Referring to FIG. 20, the following definition can be applied to a message x transmitted from a source (2020) to a destination (2010).

[0137] world model The Shannon entropy H(W) is expressed as [Mathematical Formula 1] below.

[0138]

[0139] In [Equation 1], is a world model As Shannon entropy for , it can be referred to as model entropy of semantic source. Also, refers to the probability distribution for the model.

[0140] World Model The probability distribution is It is called a set of interpretations, is the probability distribution for the model, The corresponding model for which x is "true" When a set of its models is called a message, logical probability is expressed as in mathematical formula 2.

[0141]

[0142] In [Equation 2], is a message is the logical probability, is a world model, Is The model that is "true" is a set for, refers to the probability distribution for the model. Here, stands for a general propositional satisfaction relation. is 'entailed'. It is also called 'model', which semantically means 'entails the following result' or 'is a stronger condition'. That is, reveals relevance from a semantic point of view.

[0143] message Semantic entropy of is defined as in [Mathematical Formula 3].

[0144]

[0145] In [Equation 3], is a message is the semantic entropy of, is a message is the logical probability.

[0146] At this time, when considering the background knowledge K, the set of possible worlds of [Equation 2] and [Equation 3] is It is limited to a set compatible with . Therefore, the set of possible worlds in [Equation 2] and [Equation 3] is expressed as a conditional logical probability as in [Equation 4] and [Equation 5].

[0147]

[0148] In [Equation 4], is background knowledge Message in case of means the conditional logical probability of, is a world model, Is The model that is "true" is a set for, refers to the probability distribution for the model. Here, refers to a general propositional satisfaction relation.

[0149]

[0150] In [Equation 5], is a message when the background knowledge is K is the semantic entropy of, is background knowledge Message in case of It means the conditional logical probability of .

[0151] For example, assume that the statistical probabilities are p and the background knowledge is K, and the truth table is given as in [Table 2] below.

[0152] [Table 2] is the truth table for p(A)=p(B)=0.5 and K={A→B}.

[0153] #ABA→Bprobability10010.2520110.2531000.2541110.25

[0154] At this point, the possible worlds are reduced to a series of truth assignments where A → B is true. That is, the possible worlds can be reduced to Case 1, Case 2, and Case 4 where A → B is true.

[0155] Therefore, the conditional logical probability can be expressed as in [Equation 6], [Equation 7], and [Equation 8].

[0156]

[0157]

[0158]

[0159] Here, background knowledge Since there exists a conditional logical probability, it is different from the a priori statistical probability, and in the new distribution, A and B are no longer logically independent. That is, am.

[0160] When background knowledge K exists, the new distribution for the model set can be expressed as in [Equation 9] and [Equation 10].

[0161]

[0162] In [Equation 9], is the new distribution of the model set, is the probability distribution for the model. Also, refers to a general propositional satisfaction relation.

[0163]

[0164] In [Equation 10], is background knowledge In this case, the world model is the model entropy, is a new distribution for that model.

[0165] When background knowledge is not considered, the model entropy of the source is expressed as in [Mathematical Formula 11], and when background knowledge is considered, the model entropy of the source is expressed as in [Mathematical Formula 12].

[0166]

[0167] In [Equation 11], World model without considering background knowledge is the model entropy for .

[0168]

[0169] In [Equation 12], H(W|K) is the model entropy for the world model W when the background knowledge is K.

[0170] As shown in [Equations 11] and

[12] , shared background knowledge helps prevent information loss during the compression of messages intended to be conveyed from the source, and allows the destination to obtain maximum information from the source by transmitting and receiving only short messages. Thus, communication at the semantic level can provide performance improvements compared to existing technical levels because it takes background knowledge into account. In other words, as described above, when generating and conveying semantic features from the source to the destination while considering the downstream tasks of the destination, utilizing background knowledge can be seen as consistent with the purpose of performing semantic communication.

[0171] To implement semantic communication encompassing all of the aforementioned components, a new layer called the semantic layer can be added, which governs the overall operation of semantic data and messages. When building a task-oriented semantic communication system, the semantic layer can be included in both the source and destination. For communication between the semantic layers of the source and destination, a protocol, which is a convention between the layers, and a definition of a series of operations are required.

[0172] That is, to perform accurate representation and reasoning in semantic communication within the newly defined semantic layer, a process of consensus on background knowledge information between the source and destination is required. Specifically, in situations where the source possesses information related to the destination's background knowledge, a process of fine-tuning this background knowledge is necessary to perform semantic communication. Therefore, this disclosure proposes a method for synchronizing background knowledge between the source and destination.

[0173] Figure 21 illustrates an example of a synchronization process for background knowledge for semantic communication according to one embodiment of the present disclosure. Referring to Figure 21, a source (2120) and a destination (2110) each possess background knowledge. The source (2120) may possess a large amount of background knowledge, while the destination (2110) may possess a small amount of background knowledge. In particular, the background knowledge of the destination (2110) may correspond to a portion of the background knowledge of the source (2120).

[0174] The source (2120) and the destination (2110) can synchronize the background knowledge to be used for semantic communication. Specifically, as in step 1 (2101), the source (2120) can estimate information about the background knowledge of the destination (2110) based on information obtained from the destination (2110). That is, the source (2120) can determine which part of the background knowledge possessed by the source (2120) is most similar to the background knowledge of the destination (2110). However, the part of the background knowledge of the source (2120) determined to be most similar to the background knowledge of the destination (2110) may not be identical to the actual background knowledge of the destination (2110). Therefore, as in step 2, the source (2120) can perform fine tuning of the background knowledge of the destination (2110) using additional information obtained from the destination (2110). That is, the source (2120) can perform a procedure to more accurately identify information about the background knowledge of the destination (2110) in order to improve the performance of semantic communication.

[0175] As described above, the present disclosure proposes a scheme for synchronizing background knowledge between a source and a destination, and more specifically, a scheme for fine-tuning information about the background knowledge of the destination at the source. To this end, the source according to an embodiment of the present disclosure utilizes a semantic diversity scheme based on multiple feature transmission. The semantic diversity scheme applies the channel diversity scheme used in existing wireless communication systems to semantic communication. Existing channel diversity techniques are techniques for improving the reliability of received signals by transmitting and receiving the same signal using different wireless channels. Existing channel diversity techniques can obtain diversity gains by combining corresponding signals based on channel information at the transmitter or receiver. On the other hand, the diversity technique in semantic communication differs from the existing channel diversity technique in that it treats background knowledge (2222a, 2222b, 2222c) as channels between a source (2220) and a destination (2210), as illustrated in FIG. 22. FIG. 22 illustrates an example of a semantic diversity technique based on multi-feature transmission according to an embodiment of the present disclosure. Referring to FIG. 22, a source (2220) can generate multiple semantic features for source data using multiple background knowledge (2222a, 2222b, 2222c) according to the semantic diversity technique, and transmit the generated multiple semantic features to the destination (2210). The multiple semantic features can be referred to as multiple semantic features.

[0176] The behavior of the source and destination according to semantic diversity techniques may vary depending on whether the source and destination share information about background knowledge. For example, if the source and destination share information about background knowledge, a closed-loop semantic diversity technique can be used to set a combination ratio for multiple semantic features and obtain semantic diversity gains for downstream tasks. Conversely, if the source and destination do not share information about background knowledge, an open-loop semantic diversity technique can be used to select a single feature from multiple semantic features and perform partial operations on the downstream task using the single feature.

[0177] The present disclosure assumes a situation where a source and destination share information about background knowledge. Therefore, the present disclosure proposes a semantic diversity technique that includes a method for fine-tuning the shared background knowledge information during initial connection for semantic communication, and combining the fine-tuned background knowledge information to improve the performance of semantic communication.

[0178] Specifically, the synchronization process for background knowledge according to various embodiments of the present disclosure includes a feedback process for determining whether there is an intersection between partial background knowledge used to generate multiple semantic features in a semantic encoder of a source and background knowledge of a destination in order to fine-tune information about shared background knowledge, a process for reducing the partial background knowledge of the source to be used to generate multiple semantic features based on feedback information from the destination to the intersection portion with the background knowledge of the destination, and a process for improving the performance of a downstream task through combining ratio control for the multiple semantic features generated based on the reduced partial background knowledge. Here, a feedback injection encoder that uses feedback information from the destination can be used for reducing the partial background knowledge of the source. Hereinafter, the present disclosure proposes a semantic layer protocol and procedure required for the synchronization process for background knowledge of semantic communication as described above.

[0179] The background knowledge of the source and destination may have a knowledge graph structure. The background knowledge of the source may be divided into N knowledge partitions based on graph clustering. The source may perform semantic communication using a semantic diversity technique based on the knowledge partitions. Specifically, the source generates multiple semantic features for a single source data using multiple contextual encoders that each utilize each knowledge partition as attention, and transmits the generated multiple semantic features to the destination. In this disclosure, the knowledge partition may be referred to as partial knowledge, partial background knowledge, or other terms having equivalent technical meaning.

[0180] FIG. 23 illustrates an example of a partial background knowledge-based semantic diversity technique according to one embodiment of the present disclosure.

[0181] Referring to FIG. 23, a source (2320) generates multiple semantic features for source data (2302) using a plurality of contextualizing encoders (2322-1, 2322-2, …, 2322-N) and transmits the generated multiple semantic features to a destination (2310).

[0182] Each of the plurality of contextual encoders (2322-1, 2322-2, …, 2322-N) generates a semantic feature corresponding to the input source data (2302) by using partial background knowledge as attention. The partial background knowledge can be obtained by dividing the background knowledge of the source into multiple parts as described above. That is, the n-th contextual encoder generates an embedding (x) corresponding to the i-th graph component for the graph representation of the source data (2302). i ) using the nth partial background knowledge to contextualize the feature. Encodes as. Contextualized features may be referred to as semantic features or contextualized semantic features. For example, contextualized encoder #1 (2322-1) uses partial background knowledge #1 as attention to generate semantic features corresponding to source data (2302). (2304-1) generates, and contextual encoder #2 (2322-2) uses partial background knowledge #2 as attention to generate semantic features corresponding to the source data (2302). (2304-2) generates, and contextual encoder #N (2322-N) uses partial background knowledge #N as attention to generate semantic features corresponding to the source data (2302). Create (2304-N).

[0183] Multiple semantic features generated using multiple contextual encoders (2322-1, 2322-2, …, 2322-N) may be transmitted to the destination (2310) at the same time or at different times. For example, each of the multiple semantic features may be transmitted to the destination (2310) sequentially.

[0184] FIG. 24 illustrates an example of a semantic encoder structure according to one embodiment of the present disclosure.

[0185] Referring to FIG. 24, a semantic encoder (2400) includes a graph representation block (2410) and at least one contextualizing encoder (2420).

[0186] The graph representation block (2410) converts source data into input data having a graph structure and provides the converted input data to the contextualization encoder (2420).

[0187] The contextualizing encoder (2420) is a graph transformer-based encoder that utilizes background knowledge of the graph structure as attention. Specifically, it generates contextualized features for input data by utilizing partial background knowledge as attention. To this end, the contextualizing encoder (2420) may be configured to include L self-attention encoders (2422) and an additional attention encoder (2424).

[0188] Each of the L self-attention encoders (2422) includes a multi-head self-attention block (2422a), an addition and normalization block (2422b), a multi-layer perceptron (MLP) block (2422c), and an addition and normalization block (2422d). The multi-head self-attention block (2422a) includes K attention heads, and uses multiple query vectors, key vectors, and value vectors to represent each input vector in a different representation space for each purpose, and obtains a self-attention value for each input vector. The addition and normalization block (2422b) is a hierarchical normalization block that adds and normalizes the input and output of the multi-head self-attention block (2422a). The MLP block (2422c) is a fully connected layer that performs detailed learning on each input value. The summation and normalization block (2422d) is a hierarchical normalization block that adds and normalizes the input and output of the MLP block (2422c) and outputs the result.

[0189] The additional attention encoder (2424) includes a multi-head attention block (2424a), an addition and normalization block (2424b), a multi-layer perceptron (MLP) block (2424c), and an addition and normalization block (2422d). The multi-head attention block (2424a) receives the output of the self-attention encoder (2422) and the graph representation (2404) of partial background knowledge as input, performs multi-head attention, and outputs the result. The addition and normalization block (2424b) is a layer normalization block that adds and normalizes the output of the graph representation block (2410) and the output of the multi-head attention block (2424a) and outputs the result. The MLP block (2424c) is a fully connected layer that performs detailed learning for each input value. The summation and normalization block (2424d) is a hierarchical normalization block that adds and normalizes the input and output of the MLP block (2424c) and outputs them.

[0190] As described above, the contextual encoder (2420) utilizes a graph transformer structure that uses L self-attention encoders (2422) with K heads to generate self-attention-based semantic features. Generates. Self-attention-based semantic features can be determined through operations such as determining multi-head attention-based features based on the similarity to past self-attention-based features, normalizing the sum of multi-head attention-based features and past self-attention-based features, and aggregating multi-head attention. For example, self-attention-based semantic features can be generated as in [Mathematical Equation 13] and [Mathematical Equation 14].

[0191]

[0192] In [Equation 13], is the output feature of the multi-head attention block in the lth self-attention encoder in the tth assimilation process for the ith graph component, and K represents the head of the self-attention encoder (2622). It is a function that normalizes all input values ​​to values ​​between 0 and 1 and outputs them, and has the characteristic that the sum of the output values ​​is always 1. is the attention function used in the kth attention head, which calculates the similarity between input factors. is the final output feature of the lth self-attention encoder in the tth assimilation process for the ith graph element, is the final output feature of the lth self-attention encoder in the tth assimilation process for the jth graph component, is the weight matrix used in the kth attention head.

[0193]

[0194] In [Equation 14], is the final output feature of the (l+1)th self-attention encoder in the t-th assimilation process for the j-th graph component, is the layer normalization function, is the final output feature of the multi-head attention block in the l+1-th self-attention encoder in the t-th assimilation process for the j-th graph component, is the final output feature of the lth self-attention encoder in the tth assimilation process for the jth graph component, is a function that aggregates the results of multi-head attention. can be implemented as a multi-layered feed forward neural network.

[0195] Contextual encoder (2420) is a self-attention based semantic feature and attention values ​​between the embedding vectors corresponding to partial background knowledge. Determine the attention value can be determined based on the similarity between semantic features and embedding vectors based on self-attention. For example, the attention value can be calculated as shown in [Mathematical Formula 15] below. The contextual encoder (2420) uses an additional attention encoder (2424) to represent the graph of the source data. For , an assimilation process including attention values ​​is performed, and the result is used as input to a self-attention encoder (2422). The contextualization encoder (2420) repeats the assimilation process using an additional attention encoder (2424) T times, and the contextualized feature Creates.

[0196]

[0197] In [Equation 15], is the i-th attention value, is a function that normalizes all input values ​​to values ​​between 0 and 1 and outputs them. is an inner product-based attention function that can measure the similarity between input factors, is the self-attention feature for the i-th graph component that has passed through a total of L self-attention encoder layers in the t-th assimilation process, refers to an embedding vector whose index is c among the graph components of partial background knowledge of the source.

[0198]

[0199] In [Equation 16], is a semantic feature that goes back to the initial input of the self-attention encoder after going through the tth assimilation, is the layer normalization function, is the i-th attention value, is the ith graph representation of the source data, is a function that aggregates the results of multi-head attention. can be implemented as a multi-layered feed forward neural network.

[0200]

[0201] In [Equation 17], is a contextualized feature for the i-th source data, is a semantic feature generated through L self-attention encoding and T assimilation processes for the i-th graph component, and is the feature finally output from the contextual encoder.

[0202] Through the aforementioned process, the source generates multiple semantic features using multiple partial background knowledge as attention for a single source data and transmits the generated multiple semantic features to the destination. The multiple semantic features may be transmitted simultaneously or sequentially.

[0203] The destination receives multiple semantic features from the source and performs reasoning on the multiple semantic features based on its background knowledge. Attention values ​​between the embedding vectors of the background knowledge and the multiple semantic features are determined based on the similarity between the multiple semantic features and the embedding vectors. For example, attention values is calculated as shown in [Mathematical Formula 18] below.

[0204]

[0205] In [Equation 18], denotes the attention value for the nth multi-semantic feature of the i-th index, is an inner product-based attention function that can measure the similarity between input factors, means the nth multi-semantic feature of the i-th index, refers to a component whose index is c among the node embeddings of background knowledge held by the destination.

[0206] As in [Equation 18], the attention value In the process of calculating can be referred to as the attention coefficient. The attention coefficient is a value used for the classification of the ith index. If a component with an index c among the components of the background knowledge held by the destination overlaps with the nth partial background knowledge of the source, the attention coefficient has a relatively large value. However, if a component with an index c among the components of the background knowledge held by the destination does not overlap with the nth partial background knowledge of the source, the attention coefficient has a relatively small value. In other words, the attention coefficient can be treated as information indicating the degree of agreement between the corresponding partial background knowledge of the source and the background knowledge held by the destination.

[0207] If there is an intersection between the background knowledge of the destination and the nth partial background knowledge of the source, the calculated attention value Since is expressed as a weighted sum of the node embeddings of the corresponding intersection, it contains information about the background knowledge actually possessed by the destination in the nth partial background knowledge. On the other hand, if there is no intersection between the background knowledge possessed by the destination and the nth partial background knowledge of the source, the calculated attention value does not have any meaning for intersection. Therefore, the destination can calculate the attention coefficient as shown in Fig. 25, and determine whether there is an intersection between the background knowledge possessed by the destination and the nth background knowledge of the source based on the calculated attention coefficient.

[0208] FIG. 25 illustrates an example of determining whether there is an intersection of background knowledge according to one embodiment of the present disclosure. Referring to FIG. 25, a destination (2510) calculates an attention coefficient between multiple semantic features generated based on the first partial background knowledge (2522) and the second partial background knowledge (2524) of a source and an embedding vector of the background knowledge (2512) of the destination (2510). At this time, if the variance of the attention coefficient calculated based on the component index of the first semantic feature among the multiple semantic features and the background knowledge of the destination is higher than a threshold, the destination (2510) determines that there is an intersection (2532) between the background knowledge (2512) of the destination and the first partial background knowledge (2522) of the source used to generate the corresponding semantic feature. On the other hand, if the variance of the attention coefficient calculated based on the component index of the second semantic feature among the multiple semantic features and the background knowledge of the destination is lower than the threshold, the destination (2510) determines that there is no intersection (2534) between the background knowledge (2512) of the destination and the second partial background knowledge (2524) of the source used to generate the corresponding semantic feature, and reports to the source that there is no intersection with the second partial background knowledge (2524). In this case, the source can adjust the contextual encoder that uses the second partial background knowledge (2524) as attention to use other partial background knowledge as attention, and transmit the multiple semantic features based on the other partial background knowledge. That is, the source can control the contextual encoder to use the second partial background knowledge and other partial background knowledge as attention if it has previously used the second partial background knowledge as attention.

[0209] By performing the process described above on the destination and the source, the partial background knowledge utilized by each of the contextual encoders of the source all has an intersection with the background knowledge held by the destination. The source can generate multiple semantic features for the source data using the contextual encoders that use the partial background knowledge that has an intersection with the background knowledge held by the destination as attention, and transmit the generated multiple semantic features to the destination. The destination calculates attention values ​​between the node embeddings of the received multiple semantic features and the background knowledge held by the destination, and feeds back the calculated attention values ​​to the source. The source performs an assimilation process on the multiple semantic features based on the attention values ​​fed back from the destination using a feedback injection encoder as illustrated in FIG. 26.

[0210] FIG. 26 illustrates an example of a feedback injection encoder structure according to one embodiment of the present disclosure.

[0211] Referring to FIG. 26, the source (2620) may additionally include a feedback injection encoder (2622).

[0212] The feedback injection encoder (2622) generates and outputs transmission features (2623) by encoding contextualized features (2621) based on attention values ​​(2612) fed back from the destination (2610). To this end, the feedback injection encoder (2622) is configured to include L self-attention encoders (2622a), summation and normalization (2622b), MLPs (2622c), and summation and normalization (2622c).

[0213] The L self-attention encoders (2622a) can be configured identically to the self-attention encoders (2622) of Fig. 24. The summation and normalization (2622b) is a hierarchical normalization block, which receives the attention values ​​(2612) from the destination (2610). ) and contextualized features (2621)( ) and normalizes it to output. The MLP (2622c) is a fully connected layer that performs detailed learning on the input values ​​from the summation and normalization (2622b). The summation and normalization (2622c) is a layer normalization block that adds and normalizes the input and output of the MLP block (2622c), and then provides the resulting values ​​to L self-attention encoders (2622a).

[0214] The source (2622) can reduce the partial background knowledge used as attention when generating multiple semantic features to a portion that overlaps with the background knowledge possessed by the destination through the feedback injection encoder (2622) configured as described above. Furthermore, the source (2620) and the destination (261) can set a feedback cycle for the attention value. The feedback cycle for the attention value can be determined based on performance metrics resulting from the performance of downstream tasks at the destination (2620).

[0215] FIG. 27 illustrates an example of a semantic diversity technique based on a feedback injection encoder according to one embodiment of the present disclosure.

[0216] Referring to Fig. 27, the source (2720) encodes the source data (2302) through contextualized encoders (2722-1, 2722-2, …, 2722-N) that utilize different partial background knowledge as attention, thereby generating contextualized features. , , … , . Afterwards, the source (2720) receives the attention values ​​fed back from the destination (2710) through the feedback injection encoders (2724-1, 2724-2, …, 2724-N). , , … , Contextualized features based on , , … , By performing encoding for , transmission features that have as attention the part of the source's partial background knowledge that overlaps with the destination's background knowledge , , … , generates. The source (2720) is the generated transmission features , , … , is transmitted to the destination (2710) as a multi-semantic feature.

[0217] According to an embodiment of the present disclosure, a source has a transmission structure including a feedback injection encoder as described above, thereby performing semantic communication based on multi-feature transmission using N-partially divided partial background knowledge as attention. Multi-feature transmission is a closed-loop semantic diversity technique in which the source and destination retain information about each other's background knowledge, and by setting a combining ratio for each of the multiple semantic features, the performance of downstream tasks can be improved.

[0218] According to one embodiment, a combining ratio control operation for multiple semantic features can be performed at the source. That is, the source controls the combining ratio based on the size of the partial background knowledge used as attention to generate multiple semantic features. The size of the partial background knowledge can be determined based on the difference in the attention values ​​fed back from the destination. That is, if the difference in the attention values ​​fed back from the destination at each feedback period for a specific semantic feature among the multiple semantic features is small, the source can determine that the size of the overlap, i.e., the size of the intersection, between the partial background knowledge corresponding to the specific semantic feature and the background knowledge held by the destination is large. Conversely, if the difference in the attention values ​​fed back from the destination at each feedback period for a specific semantic feature among the multiple semantic features is large, the source can determine that the intersection between the partial background knowledge corresponding to the corresponding attention value and the background knowledge held by the destination is small.

[0219] The combination ratio for multiple semantic features is set as shown in [Mathematical Formula 19] below.

[0220]

[0221] In [Equation 19], is the combined ratio for the semantic features generated based on the n-th contextual encoder, is the attention value fed back at time t+1 for the semantic feature generated by the n-th contextual encoder, is the attention value fed back at time t for the semantic feature generated by the n-th contextual encoder.

[0222] As described above, for a semantic diversity technique based on multi-feature transmission, a source can set a combining ratio for each of a plurality of semantic features, and transmit a synthetic semantic feature that combines the plurality of semantic features based on the set combining ratio to a destination.

[0223] In one embodiment, the combining ratio control operation for multiple semantic features can be performed at the destination. The destination generates an embedding for performing a downstream task by utilizing its background knowledge of multiple semantic features. The destination calculates the similarity between the task-specific embedding and the weighted sum of the multiple semantic features, and measures the importance of each semantic feature based on the calculated similarity. The destination can calculate a combining weight for each of the multiple semantic features by normalizing the importance of each semantic feature. In this case, the combining weight is calculated as shown in [Mathematical Equation 20] below.

[0224]

[0225] In [Equation 20], is the joint weight for the semantic feature generated by the n-th contextual encoder, is the result of adding the weights for each of the multiple semantic features, is a function that generates an embedding for a task operation.

[0226] That is, the destination feeds back the attention values ​​for the diversity technique based on multi-feature transmission, and then first combines the weights of each feature for the downstream task operation. , and the final feature, i.e., the synthetic semantic feature, is generated by randomly setting the combined weights. At this time, the embedding-generating MLP and the coupling ratio can be updated through learning based on the performance of the downstream task based on the synthetic semantic features. The destination can utilize the MLP and coupling ratio learned through this process for the downstream task.

[0227] FIG. 28 illustrates an example of a combination ratio control considering a downstream task according to one embodiment of the present disclosure.

[0228] Referring to FIG. 28, the destination (2810) may include a plurality of MLP blocks (2812-1 to 2812-N), a plurality of joint weight calculation blocks (2814-1 to 2814-N), a plurality of multipliers (2816-1 to 2816-N), and an adder (2818). The destination (2810) performs detailed learning on a plurality of semantic features received from a source using the plurality of ML blocks (2812-1 to 2812-N), and calculates a joint weight for each semantic feature using the plurality of joint weight calculation blocks (2814-1 to 2814-N). At this time, the joint weight may be calculated as in [Mathematical Formula 20]. The destination (2810) can perform a downstream task based on the final value obtained by multiplying the combined weights for each semantic feature calculated by the combined weight calculation blocks (2814-1 to 2814-N) by the corresponding semantic features using the multipliers (2816-1 to 2816-N), and adding the results from the multipliers (2816-1 to 2816-N) using the adder (2818). At this time, the initial value of the weight during the calculation of the combined ratio of the destination can be set by receiving the weight calculated from the source or can be set to an arbitrary value. In addition, the embedding generation MLP and the combined ratio can be updated through learning according to the results of the downstream task performed by the destination.

[0229] According to one embodiment, a source receives a combination ratio for a plurality of semantic features from a destination, and adjusts at least one partial background knowledge of the source based on the fed-back combination ratio. The combination ratio fed back from the destination may include a combination weight calculated at the destination. Specifically, the source calculates a difference between the combination ratio calculated at the source and the fed-back combination ratio, and if the calculated difference is greater than a threshold, performs re-partitioning of the background knowledge or changes to at least one partial background knowledge. By performing a combination ratio control operation based on the re-partitioned partial background knowledge or the changed partial background knowledge, the source can set a combination ratio suitable for a downstream task of the destination.

[0230] FIG. 29 illustrates an example of a procedure for receiving a data signal for semantic communication according to one embodiment of the present disclosure. FIG. 29 illustrates a method performed by a terminal operating as a destination.

[0231] Referring to FIG. 29, in step S2901, the terminal receives configuration information from the base station. The configuration information may include at least one of control information related to establishing a connection between the terminal and the base station, control information for determining settings related to communication, and information for indicating resources for communication. For example, the terminal may receive control information for establishing a semantic communication connection, control information for determining settings related to semantic communication, or configuration information for indicating resources for semantic communication.

[0232] In step S2903, the terminal receives a first data signal from the base station. At this time, the first data signal includes a plurality of semantic features generated based on partial background knowledge derived from background knowledge information of the base station. Each of the plurality of semantic features may be generated based on a plurality of contextual encoders that utilize at least a portion of different partial background knowledge for source data as attention. At this time, although not illustrated in FIG. 29, the terminal may receive first control information from the base station and receive the first data signal based on the first control information. The first control information may include at least one of information related to resources allocated for the first data signal and information necessary for interpreting the first data signal.

[0233] In step S2905, the terminal transmits feedback information to the base station. That is, the terminal generates feedback information for the first data signal and transmits the generated feedback information to the base station. The feedback information includes information used to determine similarities between each of the partial background knowledge and the background knowledge held by the terminal, based on a plurality of semantic features generated based on partial background knowledge derived from the background knowledge information of the base station. For example, the feedback information may include at least one of attention values ​​between the embedding vectors of the plurality of semantic features received from the base station and the background knowledge held by the terminal, attention coefficients calculated in the process of calculating the attention values, or background knowledge change request information generated based on the attention coefficients. The attention values ​​may be used to reduce the partial background knowledge used in the generation of the plurality of semantic features by the base station to an intersection portion with the background knowledge held by the terminal and to determine the size of the reduced intersection portion. Attention coefficients can be used to determine whether there is an intersection between the partial background knowledge used by the base station to generate multiple semantic features and the background knowledge of the terminal, or whether the partial background knowledge used to generate the semantic features needs to be changed. The change in the partial background knowledge can include at least one of re-segmentation of the background knowledge of the base station, or a change in at least one partial background knowledge used to generate the semantic features.

[0234] In step S2907, the terminal receives a second data signal. The second data signal includes a synthesized semantic feature that combines a plurality of semantic features based on feedback information. That is, the base station may calculate a combining ratio of each of the plurality of semantic features based on the attention values ​​fed back by the terminal to the base station, and transmit a synthesized semantic feature that combines the plurality of semantic features to the terminal according to the calculated combining ratio. At this time, although not illustrated in FIG. 29, the terminal may receive second control information from the base station and receive the second data signal based on the second control information. The second control information may include at least one of information related to resources allocated for the second data signal and information necessary for interpreting the second data signal.

[0235] FIG. 30 illustrates an example of a procedure for transmitting a data signal for semantic communication according to one embodiment of the present disclosure. FIG. 30 illustrates a method performed by a base station operating as a source.

[0236] Referring to FIG. 30, in step S3001, the base station transmits configuration information to the terminal. The configuration information may include at least one of control information related to establishing a connection between the terminal and the base station, control information for determining settings related to communication, and information for indicating resources for communication. For example, the base station may transmit control information for establishing a semantic communication connection, control information for determining settings related to semantic communication, or configuration information for indicating resources for semantic communication.

[0237] In step S3003, the base station transmits a first data signal to the terminal. The data signal includes a plurality of semantic features generated based on partial background knowledge derived from background knowledge information of the base station. Each of the plurality of semantic features may be generated based on a plurality of contextual encoders that utilize at least a portion of different partial background knowledge for source data as attention. At this time, although not illustrated in FIG. 30, the terminal may receive first control information from the base station and receive the first data signal based on the first control information. The first control information may include at least one of information related to resources allocated for the first data signal and information necessary for interpreting the first data signal.

[0238] In step S3005, the base station receives feedback information from the terminal. That is, the base station receives feedback information from the terminal in response to the data signal transmission. The feedback information includes information used to determine similarities between each of the partial background knowledge and the background knowledge held by the terminal, based on a plurality of semantic features generated based on partial background knowledge derived from the background knowledge information of the base station. For example, the feedback information may include at least one of attention values ​​between the embedding vectors of the plurality of semantic features received from the base station and the background knowledge held by the terminal, attention coefficients calculated in the process of calculating the attention values, or background knowledge change request information generated based on the attention coefficients. Based on the attention values, the base station may reduce the partial background knowledge used to generate the plurality of semantic features to an intersection portion with the background knowledge held by the terminal, and determine the size of the reduced intersection portion. The base station can determine, based on attention coefficients, whether there is an intersection between each of the partial background knowledge used to generate a plurality of semantic features and the background knowledge of the terminal, or whether the partial background knowledge to be used to generate the semantic features needs to be changed. The change in the partial background knowledge can include at least one of re-segmentation of the background knowledge of the base station, or a change in at least one partial background knowledge to be used to generate the semantic features.

[0239] In step S3007, the base station transmits a second data signal. The second data signal includes a synthesized semantic feature that combines a plurality of semantic features based on feedback information. That is, the base station may calculate a combining ratio of each of the plurality of semantic features based on attention values ​​fed back from the terminal, and transmit a synthesized semantic feature that combines the plurality of semantic features to the terminal according to the calculated combining ratio. At this time, although not illustrated in FIG. 30, the base station may transmit second control information to the terminal and transmit the second data signal based on the second control information. The second control information may include at least one of information related to resources allocated for the second data signal and information necessary for interpreting the second data signal.

[0240] FIG. 31 illustrates an example of a synchronization and binding ratio control procedure for background knowledge according to one embodiment of the present disclosure. FIG. 31 illustrates a method performed by a terminal operating as a destination. According to one embodiment, at least some of the operations of FIG. 31 may be understood as examples of steps S2903, S2905, and S2907 of FIG. 28.

[0241] Referring to FIG. 31, in step S3101, the terminal receives a plurality of semantic features. That is, the terminal receives a plurality of semantic features corresponding to specific source data from the base station. Each of the plurality of semantic features is generated using each of the contextual encoders that utilize the segmented partial background knowledge of the base station as attention. The partial background knowledge used for attention in each of the contextual encoders may differ from each other in at least part. For example, the first partial background knowledge corresponding to the first contextual encoder and the second partial background knowledge corresponding to the second contextual encoder may be configured to have at least one different graph node.

[0242] In step S3103, the terminal calculates attention coefficients. That is, the terminal calculates attention coefficients to determine whether there is an intersection between each of the partial background knowledge used to generate the multiple semantic features and the terminal's background knowledge. The attention coefficients can be obtained by calculating the attention value between the multiple received semantic features and the embedding vector of the background knowledge held by the terminal, as shown in Equation 18.

[0243] In step S3105, the terminal determines whether an intersection exists. That is, based on the calculated attention coefficients, the terminal determines whether an intersection exists between each of the partial background knowledge used to generate multiple semantic features and the terminal's background knowledge. For example, if the attention coefficient is greater than the reference coefficient, the terminal can determine that an intersection exists between the partial background knowledge and the terminal's background knowledge. On the other hand, if the attention coefficient is less than or equal to the reference coefficient, the terminal can determine that no intersection exists between the partial background knowledge and the terminal's background knowledge.

[0244] If no intersection exists, in step S3113, the terminal requests the base station to repartition the background knowledge. That is, the terminal reports to the base station that there is no intersection between each of the partial background knowledges used to generate multiple semantic features in the base station and the background knowledge of the terminal, thereby requesting a change to at least one partial background knowledge among the partial background knowledges of the base station that does not have an intersection with the background knowledge of the terminal.

[0245] If an intersection exists, the terminal provides attention values ​​in step S3107. The attention values ​​are calculated based on the multiple received semantic features and the embedding vector of the background knowledge held by the terminal, and can be calculated as shown in Equation 18.

[0246] In step S3109, the terminal receives a synthetic semantic feature from the base station. The synthetic semantic feature can be generated by combining multiple semantic features generated by the base station according to a combination ratio determined by the base station. The combination ratio can be determined based on the attention values ​​fed back by the terminal.

[0247] In step S3111, the terminal performs a downstream task. That is, the terminal performs a downstream task based on the synthesized semantic features received from the base station.

[0248] FIG. 32 illustrates an example of a synchronization and combination ratio control procedure for background knowledge according to one embodiment of the present disclosure. FIG. 32 illustrates a method performed by a base station operating as a source. According to one embodiment, at least some of the operations of FIG. 32 may be understood as examples of steps S3003, S3005, and S3007 of FIG. 30.

[0249] Referring to Figure 32, in step S3201, the base station divides background knowledge. The base station divides the entire background knowledge it possesses into multiple partial background knowledge pieces and sets all or some of the divided partial background knowledge pieces as candidates for background knowledge held by the terminal. At this time, the entire background knowledge held by the base station may be divided into multiple partial background knowledge pieces based on graph clustering.

[0250] In step S3203, the base station generates and transmits a plurality of semantic features. That is, the base station generates a plurality of semantic features by encoding the source data using each of a plurality of encoders that utilize each of a plurality of partial background knowledge sets as candidates as attention. The base station transmits the generated plurality of semantic features to the base station. At this time, the plurality of semantic features may be transmitted sequentially. According to one embodiment, the base station may transmit the first attention values ​​generated in the process of generating the semantic features to the terminal. The first attention values ​​are not necessarily transmitted and may be omitted depending on the embodiment.

[0251] In step S3205, the base station determines whether a request for background knowledge re-segmentation is received from the terminal. That is, the base station determines whether information indicating whether there is an intersection between each of the partial background knowledge used to generate the plurality of semantic features and the background knowledge of the terminal is received from the terminal. If information indicating a request for background knowledge re-segmentation is received from the terminal, or an attention coefficient smaller than or equal to a reference coefficient is received from the terminal, the base station may determine that a request for background knowledge re-segmentation has been received. On the other hand, if information indicating a request for background knowledge re-segmentation is not received from the terminal, or an attention coefficient larger than the reference coefficient is received from the terminal, the base station may determine that a request for background knowledge re-segmentation has not been received.

[0252] If a request for repartitioning background knowledge is received from the terminal, the base station repartitions the background knowledge in step S3213. That is, the base station changes at least one partial piece of background knowledge that does not intersect with the background knowledge held by the terminal. For example, the base station may change at least one partial piece of background knowledge that does not intersect with the background knowledge held by the terminal into another piece of partial background knowledge. As another example, the base station may repartition the entire background knowledge held by the base station into multiple pieces of partial background knowledge, as in step S3201. Of course, the method for repartitioning the background knowledge or the components used for repartitioning the background knowledge may differ from step S3201.

[0253] If a background knowledge re-segmentation request is not received from the terminal, in step S3207, the base station receives attention values ​​from the terminal. The attention values ​​received from the terminal include attention values ​​between a plurality of semantic features and an embedding vector of the terminal's background knowledge. The base station can receive the attention values ​​generated by the terminal as feedback information for the plurality of semantic features.

[0254] In step S3209, the base station sets a combining ratio. That is, the base station sets a combining ratio for each of a plurality of semantic features based on the attention values ​​received from the terminal for a semantic diversity technique based on multi-feature transmission. The combining ratio can be set according to the size of the intersection between the partial background knowledge used for transmitting the plurality of semantic features and the background knowledge possessed by the terminal. That is, the larger the intersection size, the larger the combining ratio for the corresponding semantic feature can be set, and the smaller the intersection size, the smaller the combining ratio for the corresponding semantic feature can be set.

[0255] In step S3211, the base station generates and transmits a synthetic semantic feature based on the attention values ​​and the combining ratio. That is, the base station can generate a plurality of semantic features using a feedback injection encoder that utilizes the attention values ​​fed back from the terminal, and can generate a synthetic semantic feature by combining the generated plurality of semantic features according to the combining ratio. At this time, since the feedback injection encoder utilizes the attention values ​​fed back from the terminal, the attention for the plurality of semantic features can be reduced to the intersection portion with the background knowledge possessed by the terminal among the corresponding partial background knowledge.

[0256] Figures 31 and 32 illustrate examples of cases where the combining ratio control operation for semantic diversity techniques is performed at the source. However, as previously described, the present disclosure is not limited thereto. That is, the combining ratio control operation for semantic diversity techniques may also be performed at the destination, as illustrated in Figures 33 and 34 below.

[0257] FIG. 33 illustrates an example of a synchronization and binding ratio control procedure for background knowledge according to one embodiment of the present disclosure. FIG. 33 illustrates a method performed by a terminal operating as a destination. According to one embodiment, at least some of the operations in FIG. 33 may be understood as examples of steps S2903, S2905, and S2907 of FIG. 28.

[0258] Referring to FIG. 33, in step S3301, the terminal receives a plurality of semantic features. That is, the terminal receives a plurality of semantic features corresponding to specific source data from the base station.

[0259] In step S3303, the terminal calculates attention coefficients. That is, the terminal calculates attention coefficients to determine whether there is an intersection between each of the partial background knowledge used to generate multiple semantic features and the terminal's background knowledge.

[0260] In step S3305, the terminal determines whether an intersection exists. That is, based on the calculated attention coefficients, the terminal determines whether there is an intersection between each of the partial background knowledge used to generate multiple semantic features and the terminal's background knowledge.

[0261] If no intersection exists, in step S3317, the terminal requests the base station to repartition the background knowledge. That is, the terminal reports to the base station that there is no intersection between each of the partial background knowledges used to generate multiple semantic features in the base station and the background knowledge of the terminal, thereby requesting a change to at least one partial background knowledge among the partial background knowledges of the base station that does not have an intersection with the background knowledge of the terminal.

[0262] If an intersection exists, the terminal provides attention values ​​in step S3307. The attention values ​​are calculated based on the multiple received semantic features and the embedded vector of background knowledge held by the terminal, and can be calculated as shown in Equation 18.

[0263] In step S3309, the terminal receives a plurality of semantic features from the base station. The plurality of semantic features can be generated using a contextual encoder that uses partial background knowledge from the base station as attention and a feedback injection encoder that uses attention values ​​fed back from the terminal. Here, by the feedback injection encoder using the attention values ​​fed back from the terminal, the attention for the plurality of semantic features can be reduced to the intersection of the partial background knowledge with the background knowledge possessed by the terminal.

[0264] In step S3311, the terminal calculates a combined weight for each of the multiple semantic features. That is, the terminal generates an embedding for performing a downstream task using the background knowledge of the destination for the multiple semantic features, and measures the importance of each of the multiple semantic features based on the similarity between the generated embedding and the weighted sum of the multiple semantic features. The terminal calculates a combined weight for each of the multiple semantic features by normalizing the importance of each of the multiple semantic features. For example, the terminal can calculate a combined weight for each of the multiple semantic features as shown in Equation 20.

[0265] In step S3313, the terminal generates a synthetic semantic feature based on the combined weights. That is, the terminal applies the combined weights for each of the multiple semantic features to each of the multiple semantic features, and then adds the resulting values ​​to which the combined weights have been applied, thereby generating a synthetic semantic feature.

[0266] In step S3315, the terminal performs a downstream task. That is, the terminal performs a downstream task based on the synthetic semantic features generated at the terminal.

[0267] FIG. 34 illustrates an example of a synchronization and combination ratio control procedure for background knowledge according to one embodiment of the present disclosure. FIG. 34 illustrates a method performed by a base station operating as a source. According to one embodiment, at least some of the operations of FIG. 34 may be understood as examples of steps S3003, S3005, and S3007 of FIG. 30.

[0268] Referring to Figure 34, in step S3401, the base station divides background knowledge. The base station divides the entire background knowledge it possesses into multiple partial background knowledge pieces, and sets all or some of the divided partial background knowledge pieces as candidates for background knowledge possessed by the terminal.

[0269] In step S3403, the base station generates and transmits multiple semantic features. That is, the base station generates multiple semantic features by performing encoding on the source data using each of multiple encoders that utilize each of the multiple partial background knowledge sets as candidates as attention.

[0270] In step S3405, the base station determines whether a background knowledge re-segmentation request is received from the terminal. That is, the base station determines whether information indicating whether an intersection exists between each of the partial background knowledge used to generate multiple semantic features and the terminal's background knowledge is received from the terminal.

[0271] When a request for background knowledge repartition is received from a terminal, the base station repartitions the background knowledge in step S3411. That is, the base station can change at least one partial background knowledge that does not have an intersection with the background knowledge held by the terminal to another partial background knowledge.

[0272] If a background knowledge re-segmentation request is not received from the terminal, in step S3407, the base station receives attention values ​​from the terminal. The attention values ​​received from the terminal include attention values ​​between multiple semantic features and the terminal's background knowledge embedding vector.

[0273] In step S3409, the base station regenerates and transmits a plurality of semantic features based on the attention values. That is, the base station can regenerate a plurality of semantic features using a feedback injection encoder that utilizes the attention values ​​fed back from the terminal, and transmit the generated plurality of semantic features to the terminal. At this time, since the feedback injection encoder utilizes the attention values ​​fed back from the terminal, the attention for the plurality of semantic features can be reduced to the intersection portion with the background knowledge possessed by the terminal among the corresponding partial background knowledge.

[0274] Figure 35 illustrates an example of a background knowledge synchronization and semantic communication procedure based on synchronized background knowledge according to one embodiment of the present disclosure. Figure 35 illustrates signaling between a base station and a terminal when the base station acts as a source and the terminal acts as a destination.

[0275] Referring to FIG. 35, in step S3501, the base station (3520) divides the background knowledge. That is, the base station (3520) divides the background knowledge into multiple partial background knowledge based on graph clustering.

[0276] In step S3503, the base station (3520) generates first semantic features using encoders corresponding to each of the partial background knowledges. That is, the base station (3520) performs encoding on the i-th source data using N encoders that utilize different partial background knowledges as attention, thereby generating first semantic features for the source data. Creates.

[0277] At step S3505, the base station (3520) provides the first semantic features is transmitted to the terminal (3510). At this time, the first semantic features can be sequentially transmitted to the terminal (3510).

[0278] In step S3507, the terminal (3510) calculates an attention coefficient. The attention coefficient indicates the similarity between the partial background knowledge of the base station used to generate the first semantic features and the background knowledge held by the terminal. The attention coefficient is obtained in the process of calculating the attention values ​​between the received first semantic features and the embedding vector of the background knowledge held by the terminal, as in mathematical expression 18. It could be worth it.

[0279] The terminal (3510) can perform an operation according to option A or an operation according to option B based on the calculated attention coefficient. Specifically, the terminal (3510) can determine whether there is an intersection between each of the partial background knowledge of the base station used to generate the first semantic features and the background knowledge of the terminal based on the calculated attention coefficient, and can perform an operation according to option A or an operation according to option B based on whether there is an intersection.

[0280] If there is no intersection between at least one partial background knowledge of the base station (3520) and the background knowledge of the terminal, the terminal (3510) performs an operation according to option A. That is, in step S3509, the terminal (3510) requests background knowledge repartitioning. That is, the terminal (3510) reports to the base station (3520) that at least one partial background knowledge among the partial background knowledge used by the base station (3520) for semantic communication with the terminal (3510) does not have an intersection with the background knowledge of the terminal (3510).

[0281] In step S3511, the base station (3520) redistributes the background knowledge. That is, in response to the terminal's request for background knowledge redistribution, the base station (3520) may replace at least one of the partial background knowledge pieces used for semantic communication with the terminal (3510) with other partial background knowledge pieces, or redistribute the entire background knowledge of the base station (3520) into multiple partial background knowledge pieces. Thereafter, the base station (3520) performs step S3503.

[0282] On the other hand, if there is an intersection between each of the plurality of partial background knowledges of the base station (3520) and the background knowledge of the terminal, the terminal (3510) performs an operation according to option B. That is, in step S3513, the terminal (3510) uses the attention values transmits to the base station. That is, the terminal (3510) receives the first semantic features from the base station (3520). Attention values ​​between the embedding vector of background knowledge held by the terminal (3510) and calculate the calculated attention values can be transmitted to the base station (3520).

[0283] In step S3515, the base station (3520) performs feedback injection encoding based on the attention values. That is, the base station (3520) additionally encodes the first semantic features through a feedback injection encoder that uses the attention values ​​fed back from the terminal (3510), thereby encoding the second semantic features. Generates second semantic features can be referred to as multiple transmission features. Second semantic features The terminal (3510) can only have attention to background knowledge that overlaps with the background knowledge it has.

[0284] At step S3517, the base station (3520) provides second semantic features transmits to the terminal (3510). According to one embodiment, the base station (3520) transmits the second semantic features The base station (3520) can transmit the combined synthetic semantic features to the terminal (3510). Specifically, the base station (3520) can transmit the second semantic features based on the attention values ​​fed back from the terminal (3510). By setting a combination ratio for the second semantic features and combining them according to the set combination ratio, a synthetic semantic feature can be generated and transmitted. In this case, the terminal (3510) can perform a downstream task based on the synthetic semantic feature.

[0285] According to one embodiment, the base station (3520) comprises second semantic features can be transmitted to the terminal (3510) without combining the second semantic features. In this case, the terminal (3510) Set the combined weights for the received second semantic features After applying the combined weights, a synthetic semantic feature can be generated by combining the results. The terminal (3510) can perform a downstream task based on the synthetic semantic feature generated at the terminal.

[0286] In step S3519, the base station (3520) and the terminal (3510) can control the combining ratio. That is, the base station (3520) can receive combined weight information for each of the second semantic features from the terminal (3510) and perform re-segmentation of the background knowledge of the base station (3520) based on the fed-back combined weight information. Specifically, the base station (3520) can control the combining ratio for the second semantic features calculated by the base station (3520). and the combined weights for the second semantic features obtained from the terminal (3510). The difference can be calculated, and based on the calculated difference, the joint ratio and / or joint weight can be maintained, or re-partitioning can be performed on the background knowledge. For example, the joint ratio and combined weights If the difference is less than or equal to the threshold, the base station (3520) sets the combining ratio and combined weights Semantic communication with the terminal (3510) is performed based on the coupling ratio. On the other hand, and combined weights If the difference is greater than the threshold, the base station (3520) can perform an appropriate combination ratio control operation for the downstream task of the terminal (3510) by redistributing the background knowledge and re-performing step S3503 and subsequent steps.

[0287] In the above-described disclosure, the base station operates as a source and the terminal operates as a destination. However, embodiments of the present disclosure are not limited to this. For example, a terminal may operate as a source and another terminal communicating with that terminal may operate as a destination. Alternatively, a terminal may operate as a source and the base station may operate as a destination.

[0288] The proposed methods described above can be implemented independently, but they can also be implemented as a combination (or merge) of some of the proposed methods. Rules can be defined so that the base station notifies the terminal of the applicability of the proposed methods (or information about the rules of the proposed methods) through a predefined signal (e.g., a physical layer signal or a higher layer signal).

[0289] The present disclosure may be embodied in other specific forms without departing from the technical ideas and essential features described herein. Therefore, the above detailed description should not be construed as limiting in all respects but rather as illustrative. The scope of the present disclosure should be determined by a reasonable interpretation of the appended claims, and all modifications within the equivalent scope of the present disclosure are intended to be included within the scope of the present disclosure. Furthermore, claims that are not explicitly cited in the claims may be combined to form an embodiment or incorporated into a new claim through a post-filing amendment.

[0290] Embodiments of the present disclosure can be applied to various wireless access systems. Examples of various wireless access systems include the 3rd Generation Partnership Project (3GPP) or 3GPP2 systems.

[0291] The embodiments of the present disclosure can be applied not only to the various wireless access systems described above, but also to all technical fields that utilize these various wireless access systems. Furthermore, the proposed method can also be applied to mmWave and THz communication systems utilizing ultra-high frequency bands.

[0292] Additionally, embodiments of the present disclosure can be applied to various applications such as autonomous vehicles and drones.

Claims

1. A method performed by a terminal in a wireless communication system, A step of receiving first control information from a base station; A step of receiving a first data signal based on first control information from the base station; A step of transmitting feedback information for a first data signal to the base station; A step of receiving second control information from the base station; and A step of receiving a second data signal based on second control information from the base station, The above feedback information includes information used to determine similarities between each of the partial background knowledge and the background knowledge possessed by the terminal, based on a plurality of semantic features generated based on partial background knowledge derived from the background knowledge information of the base station. A method wherein the second data signal includes a synthetic semantic feature that combines a plurality of semantic features based on the feedback information.

2. In claim 1, A method wherein the first data signal comprises a plurality of semantic features generated based on partial background knowledge derived from background knowledge information of the base station.

3. In claim 2, A step of determining whether there is an intersection between each of the partial background knowledge used to generate the plurality of semantic features and the background knowledge possessed by the terminal; and A method further comprising the step of requesting a change to at least one of the partial background knowledges if the above intersection does not exist.

4. In claim 3, A method of transmitting, as feedback information, information to be used to reduce each of the partial background knowledges to an intersection portion with the background knowledge possessed by the terminal when the above intersection exists.

5. In claim 4, A method in which the second data signal is generated based on partial background knowledge reduced to the intersection portion based on the feedback information.

6. In claim 1, A method in which the above feedback information is transmitted according to a cycle set based on a performance metric for the performance result of a downstream task performed based on the above synthetic semantic feature.

7. A method performed by a base station in a wireless communication system, A step of transmitting first control information to a terminal; A step of transmitting a first data signal to the terminal based on the first control information; A step of receiving feedback information for the first data signal from the terminal; A step of transmitting second control information to the terminal; and A step of transmitting a second data signal based on the second control information from the terminal, The above feedback information includes information used to determine similarities between each of the partial background knowledge and the background knowledge possessed by the terminal, based on a plurality of semantic features generated based on partial background knowledge derived from the background knowledge information of the base station. A method wherein the second data signal includes a synthetic semantic feature that combines a plurality of semantic features based on the feedback information.

8. In claim 7, A method further comprising the step of generating a first data signal including the plurality of semantic features by using each of the plurality of encoders that utilize each of the partial background knowledges as attention.

9. In claim 7, A method further comprising a step of changing at least one of the partial background knowledge based on the feedback information.

10. In claim 7, Based on the above feedback information, a step of reducing each of the partial background knowledges to an intersection portion with the background knowledge possessed by the terminal; and A method further comprising a step of generating the synthetic semantic feature by using the partial background knowledge reduced to the intersection portion.

11. In claim 7, Further comprising a step of determining a combination ratio for the plurality of semantic features based on the feedback information; A method in which the above-mentioned synthetic semantic feature is generated by combining the plurality of semantic features based on the combining ratio.

12. In a terminal in a wireless communication system, Transmitter and receiver; and comprising a processor connected to the above transceiver, The above processor, Receive first control information from the base station, Receive a first data signal based on first control information from the above base station, Transmit feedback information for the first data signal to the base station, Receive second control information from the above base station, Control to receive a second data signal based on second control information from the above base station, The above feedback information includes information used to determine similarities between each of the partial background knowledge and the background knowledge possessed by the terminal, based on a plurality of semantic features generated based on partial background knowledge derived from the background knowledge information of the base station. The second data signal is a terminal including a synthetic semantic feature that combines a plurality of semantic features based on the feedback information.

13. In a base station in a wireless communication system, Transmitter and receiver; and comprising a processor connected to the above transceiver, The above processor, Transmit first control information to the terminal, Transmitting a first data signal to the terminal based on the first control information, Receive feedback information for the first data signal from the terminal, Transmit second control information to the above terminal, Controls the above terminal to transmit a second data signal based on the above second control information, The above feedback information includes information used to determine similarities between each of the partial background knowledge and the background knowledge possessed by the terminal, based on a plurality of semantic features generated based on partial background knowledge derived from the background knowledge information of the base station. A base station wherein the second data signal includes a synthetic semantic feature that combines a plurality of semantic features based on the feedback information.

14. In communication devices, At least one processor; At least one computer memory coupled to said at least one processor and storing instructions that direct operations when executed by said at least one processor, The above actions are, A step of receiving first control information from a base station; A step of receiving a first data signal based on first control information from the base station; A step of transmitting feedback information for a first data signal to the base station; A step of receiving second control information from the base station; and A step of receiving a second data signal based on second control information from the base station, The above feedback information includes information used to determine similarities between each of the partial background knowledge and the background knowledge possessed by the communication device, based on a plurality of semantic features generated based on partial background knowledge derived from the background knowledge information of the base station. A communication device wherein the second data signal includes a synthetic semantic feature that combines a plurality of semantic features based on the feedback information.

15. In a non-transitory computer-readable medium storing at least one instruction, comprising at least one instruction executable by the processor; At least one of the above commands causes the device to: Receive a first data signal based on first control information from the above base station, Transmit feedback information for the first data signal to the base station, Receive second control information from the above base station, Control to receive a second data signal based on second control information from the above base station, The above feedback information includes information used to determine similarities between each of the partial background knowledge and the background knowledge possessed by the terminal, based on a plurality of semantic features generated based on partial background knowledge derived from the background knowledge information of the base station. A non-transitory computer-readable medium wherein the second data signal comprises a synthetic semantic feature that combines a plurality of semantic features based on the feedback information.

Images