Instruction-triggered transmission of known data for training artificial neural networks

Abstract

A method of wireless communication performed by a receiving device includes communicating with a sending device about the contents of a known payload. The method also includes requesting a known payload from the sending device for training an artificial neural network. The method also receives a known payload in response to the request. The method also performs online training of the artificial neural network using the known payload. A method of wireless communication performed by a sending device includes communicating with a receiving device about the contents of a known payload and then sending an indication to the receiving device that the known payload will be sent. The sending device unicasts the known payload to the receiving device for online training of the neural network.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims the benefit of U.S. Patent Application No. 17 / 176,738, filed on February 16, 2021, entitled “INDICATION TRIGGERING TRANSMISSION OF KNOWN DATA FOR TRAINING ARTIFICIAL NEURAL NETWORKS,” which claims the benefit of U.S. Provisional Patent Application No. 62 / 980,066, filed on February 21, 2020, entitled “INDICATION TRIGGERING TRANSMISSION OF KNOWN DATA FOR TRAINING ARTIFICIAL NEURAL NETWORKS,” the entire contents of the foregoing U.S. patent applications are expressly incorporated herein by reference. Technical Field

[0003] Aspects of the present disclosure relate generally to wireless communications, and more particularly, to techniques and apparatus for triggering future known payloads to be used to train artificial neural networks. Background Art

[0004] Wireless communication systems are widely deployed to provide various telecommunication services such as telephony, video, data, messaging, and broadcasting. Typical wireless communication systems may employ multiple access technologies that can support communication with multiple users by sharing available system resources (e.g., bandwidth, transmit power, etc.). Examples of such multiple access technologies include code division multiple access (CDMA) systems, time division multiple access (TDMA) systems, frequency division multiple access (FDMA) systems, orthogonal frequency division multiple access (OFDMA) systems, single carrier frequency division multiple access (SC-FDMA) systems, time division synchronous code division multiple access (TD-SCDMA) systems, and long term evolution (LTE) systems. LTE / LTE-Advanced is a collection of enhancements to the Universal Mobile Telecommunications System (UMTS) mobile standard promulgated by the Third Generation Partnership Project (3GPP).

[0005] A wireless communication network may include multiple base stations (BSs), which may support multiple user equipments (UEs) for communication. User equipments (UEs) may communicate with base stations (BSs) via downlinks and uplinks. Downlinks (or forward links) refer to the communication link from BSs to UEs, and uplinks (or reverse links) refer to the communication link from UEs to BSs. As will be described in more detail, BSs may be referred to as Node Bs, gNBs, Access Points (APs), Radio Heads, Transmit Receive Points (TRPs), New Radio (NR) BSs, Fifth Generation (5G) Node Bs, etc.

[0006] The above-mentioned multiple access technologies have been adopted in various telecommunication standards to provide a common protocol that enables different user devices to communicate at municipal, national, regional and even global levels. New Radio (NR), which may also be referred to as 5G, is a collection of enhancements to the LTE mobile standard promulgated by the Third Generation Partnership Project (3GPP). NR is designed to better support mobile broadband Internet access by improving spectrum efficiency, reducing costs, improving services, utilizing new spectrum, and using orthogonal frequency division multiplexing (OFDM) (CP-OFDM) with cyclic prefix (CP) on the downlink (DL), using CP-OFDM and / or SC-FDM (e.g., also known as discrete Fourier transform spread OFDM (DFT-s-OFDM)) on the uplink (UL), and supporting beamforming, multiple input multiple output (MIMO) antenna technology and carrier aggregation and better integration with other open standards. However, as the demand for mobile broadband access continues to increase, further improvements to NR and LTE technologies are needed. Preferably, these improvements should be applicable to other multiple access technologies and telecommunication standards that adopt these technologies.

[0007] An artificial neural network may include interconnected groups of artificial neurons (e.g., neuron models). An artificial neural network may be a computing device or represented as a method to be performed by a computing device. A convolutional neural network (such as a deep convolutional neural network) is a feedforward artificial neural network. A convolutional neural network may include layers of neurons that may be configured in a tiled receptive field. It is desirable to apply neural network processing to wireless communications to achieve greater efficiency. Summary of the invention

[0008] According to one aspect of the present disclosure, a method of wireless communication performed by a receiving device includes communicating with a sending device about the contents of a known payload. The method also includes requesting the known payload for training an artificial neural network from the sending device. The method also includes receiving the known payload in response to the request. The method also includes performing online training of the artificial neural network using the known payload.

[0009] In another aspect, a method of wireless communication performed by a transmitting device includes communicating with a receiving device about the contents of a known payload. The method also includes sending an indication to the receiving device notifying the known payload that the known payload will be sent. The method also includes unicasting the known payload to the receiving device for online training of a neural network.

[0010] According to another aspect of the present disclosure, a receiving device has a memory and at least one processor, the at least one processor is coupled to the memory and is configured to communicate with a sending device about the content of a known payload. The receiving device is also configured to request a known payload for training an artificial neural network from the sending device. The receiving device is configured to receive a known payload in response to the request. The receiving device is also configured to perform online training of the artificial neural network using the known payload.

[0011] According to one aspect of the present disclosure, a sending device has a memory and at least one processor, the at least one processor is coupled to the memory and is configured to communicate with a receiving device about the content of a known payload. The sending device is also configured to send an indication to notify the receiving device that the known payload will be sent. The sending device is also configured to unicast the known payload to the receiving device for online training of a neural network.

[0012] Aspects generally include methods, apparatus, systems, computer program products, non-transitory computer-readable media, user equipment, base stations, wireless communication devices, and processing systems as generally described with reference to and as described in the accompanying drawings and the specification.

[0013] The features and technical advantages of the examples according to the present disclosure have been summarized quite extensively above so that the following detailed description may be better understood. Additional features and advantages will be described below. The disclosed concepts and specific examples may be easily used as a basis for modifying or designing other structures for achieving the same purpose of the present disclosure. Such equivalent structures do not depart from the scope of the appended claims. The features of the concepts disclosed herein, their organization and methods of operation, and associated advantages will be better understood from the following description when considered in conjunction with the accompanying drawings. Each of the accompanying drawings is provided for the purpose of illustration and description and is not provided as a limitation of the claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to be able to understand the above-mentioned features of the present disclosure in detail, a more specific description briefly summarized above can be obtained by reference to some aspects, some of which are shown in the accompanying drawings. However, it should be noted that the accompanying drawings only illustrate certain typical aspects of the present disclosure and should not be considered as limiting the scope of the present disclosure, as the specification may allow for other equally effective aspects. The same reference numerals in different drawings may identify the same or similar elements.

[0015] Figure 1 is a block diagram conceptually illustrating an example of a wireless communication network in accordance with various aspects of the present disclosure.

[0016] Figure 2is a block diagram conceptually illustrating an example of a base station communicating with a user equipment (UE) in a wireless communication network according to various aspects of the present disclosure.

[0017] Figure 3 An example implementation of designing a neural network using a system on a chip (SOC) including a general-purpose processor according to certain aspects of the present disclosure is shown.

[0018] Figure 4A , Figure 4B and Figure 4C is a schematic diagram illustrating a neural network according to aspects of the present disclosure.

[0019] Figure 4D is a schematic diagram illustrating an exemplary deep convolutional network (DCN) according to aspects of the present disclosure.

[0020] Figure 5 is a block diagram illustrating an exemplary deep convolutional network (DCN) according to aspects of the present disclosure.

[0021] Fig. 6A and Figure 6B is a timing diagram illustrating triggering future known payloads for training an artificial neural network in accordance with various aspects of the present disclosure.

[0022] Figure 7 is a schematic diagram illustrating an example process, for example, performed by a receiving device, according to various aspects of the present disclosure.

[0023] Figure 8 is a schematic diagram illustrating an example process, for example, performed by a sending device, according to various aspects of the present disclosure. DETAILED DESCRIPTION

[0024] Various aspects of the present disclosure are described more fully below with reference to the accompanying drawings. However, the present disclosure can be embodied in a variety of different forms and should not be construed as being limited to any specific structure or function presented throughout the present disclosure. Rather, these aspects are provided to make the present disclosure in-depth and complete, and to fully convey the scope of the present disclosure to those skilled in the art. Based on the teachings, it should be understood by those skilled in the art that the scope of the present disclosure is intended to cover any aspect of the present disclosure disclosed, whether implemented independently of any other aspect of the present disclosure or implemented in combination with any other aspect of the present disclosure. For example, any number of aspects set forth can be used to implement a device or practice method. In addition, the scope of the present disclosure is intended to cover a device or method practiced using a structure, function, or structure and function other than or different from the various aspects of the present disclosure set forth. It should be understood that any aspect of the disclosed disclosure can be implemented by one or more elements of the claims.

[0025] Several aspects of telecommunication systems will now be presented with reference to various devices and techniques. These devices and techniques will be described in the following detailed description and illustrated in the accompanying drawings by various blocks, modules, components, circuits, steps, processes, algorithms and / or the like (collectively referred to as "elements"). These elements can be implemented using hardware, software, or a combination thereof. Whether these elements are implemented as hardware or software depends on the specific application and the design constraints imposed on the overall system.

[0026] It should be noted that although various aspects may be described using terminology typically associated with 5G and beyond wireless technologies, various aspects of the present disclosure may be applicable to communication systems based on other generations, such as and including 3G and / or 4G technologies.

[0027] Artificial intelligence (AI) / machine learning (ML) algorithms can improve wireless communications. The AI / ML modules can run at the UE, the base station, or in the case of a distributed algorithm, jointly across the UE and the base station. Although the present disclosure may involve AI / ML modules on the UE side, the present disclosure explicitly considers the autoencoder scenario. In the autoencoder scenario, joint training occurs.

[0028] In order to obtain useful results, the neural network executing the AI / ML algorithm needs to be trained. Offline training of the neural network can derive the weights and biases of the neural network from the training data. However, offline training does not take into account the dynamic real-world environment of wireless communication. Online training takes into account the dynamic environment. Online training can be done using known air transmissions (or signaling) that reflect the wireless environment. For the online training process, instead of training from scratch, the training can start with the values ​​of weights and biases derived from the offline training phase, and further fine-tune the neural network for the wireless channel, noise, and / or other environmental characteristics.

[0029] Reference signals, such as demodulation reference signal (DMRS), channel state information reference signal (CSI-RS), etc., are known transmissions. However, reference signals may not provide the appropriate type of information for certain types of machine learning models. Data transmission may be desirable for specific machine learning models, such as for log-likelihood ratio (LLR) calculations, MIMO demapping, etc. Known data transmissions may include, for example, known payloads of a physical downlink control channel (PDCCH), a physical uplink control channel (PUCCH), a physical downlink shared channel (PDSCH), and / or a physical uplink shared channel (PUSCH).

[0030] According to various aspects of the present disclosure, the base station and the UE communicate in advance (know / agree) on the payload / content of specific physical channel transmissions. These known transmissions can be immediately used as ground truth labels in the online training of the neural network. For the purpose of online training, the signaling framework can enable the UE to train their respective neural networks. "Known payloads" known to both parties of the transmission may be beneficial in a variety of contexts because the receiver does not need to fully decode the payload to find labels for training. Based on which type of neural network the UE or base station will train, the receiving device may require different types of data. Therefore, a receiver-specific known payload can be generated. A known payload may include data for a specific receiver (e.g., modulation and coding scheme (MCS), rank, beam pair, etc.).

[0031] Figure 1 1 is a schematic diagram showing a network 100 in which various aspects of the present disclosure may be practiced. The network 100 may be a 5G or NR network, or some other wireless network, such as an LTE network. The wireless network 100 may include a plurality of BSs 110 (shown as BSs 110a, BSs 110b, BSs 110c, and BSs 110d) and other network entities. A BS is an entity that communicates with a user equipment (UE), and may also be referred to as a base station, NR BS, Node B, gNB, 5G Node B (NB), access point, transmit receive point (TRP), etc. Each BS may provide communication coverage for a particular geographic area. In 3GPP, the term "cell" may refer to a coverage area of ​​a BS and / or a BS subsystem serving the coverage area, depending on the context in which the term is used.

[0032] A BS may provide communication coverage for macro cells, pico cells, femto cells, and / or other types of cells. A macro cell may cover a relatively large geographic area (e.g., several kilometers in radius) and may allow unrestricted access by UEs with service subscriptions. A pico cell may cover a relatively small geographic area and may allow unrestricted access by UEs with service subscriptions. A femto cell may cover a relatively small geographic area (e.g., a home) and may allow restricted access by UEs associated with the femto cell (e.g., UEs in a closed subscriber group (CSG)). The BS of a macro cell may be referred to as a macro BS. The BS of a pico cell may be referred to as a pico BS. The BS of a femto cell may be referred to as a femto BS or a home BS. In Figure 1In the example shown, BS 110a may be a macro BS for macro cell 102a, BS 110b may be a pico BS for pico cell 102b, and BS 110c may be a femto BS for femto cell 102c. A BS may support one or more (e.g., three) cells. The terms "eNB", "base station", "NR BS", "gNB", "TRP", "AP", "Node B", "5G NB", and "cell" may be used interchangeably herein.

[0033] In some aspects, the cells may not necessarily be stationary, and the geographic area of ​​the cells may move depending on the location of the mobile BS. In some aspects, the BSs may be interconnected with each other and / or with one or more other BSs or network nodes (not shown) in the wireless network 100 through various types of backhaul interfaces (such as direct physical connections, virtual networks, etc.) using any suitable transport network.

[0034] The wireless network 100 may also include a relay station. A relay station is an entity that can receive data transmissions from an upstream station (e.g., a BS or a UE) and send data transmissions to a downstream station (e.g., a UE or a BS). A relay station may also be a UE that can relay transmissions for other UEs. Figure 1 In the example shown, a relay station 110d may communicate with a macro BS 110a and a UE 120d to facilitate communication between the BS 110a and the UE 120d. A relay station may also be referred to as a relay BS, a relay base station, a relay, or the like.

[0035] The wireless network 100 may be a heterogeneous network including different types of BSs (e.g., macro BSs, pico BSs, femto BSs, relay BSs, etc.). These different types of BSs may have different transmit power levels, different coverage areas, and different effects on interference in the wireless network 100. For example, a macro BS may have a high transmit power level (e.g., 5 to 40 watts), while a pico BS, a femto BS, and a relay BS may have a lower transmit power level (e.g., 0.1 to 2 watts).

[0036] A network controller 130 may be coupled to a set of BSs and may provide coordination and control for these BSs. The network controller 130 may communicate with the BSs via a backhaul. The BSs may also communicate with each other directly or indirectly, for example, via a wireless or wired backhaul.

[0037] UE 120 (e.g., 120a, 120b, 120c) can be dispersed throughout the wireless network 100, and each UE can be stationary or mobile. UE can also be referred to as an access terminal, terminal, mobile station, subscriber unit, station, etc. UE can be a cellular phone (e.g., a smart phone), a personal digital assistant (PDA), a wireless modem, a wireless communication device, a handheld device, a laptop computer, a cordless phone, a wireless local loop (WLL) station, a tablet, a camera, a gaming device, a netbook, a smartbook, an ultrabook, a medical device or equipment, a biometric sensor / device, a wearable device (smart watch, smart clothing, smart glasses, smart wristband, smart jewelry (e.g., smart ring, smart bracelet)), an entertainment device (e.g., a music or video device or a satellite radio), a vehicle component or sensor, a smart meter / sensor, an industrial manufacturing device, a global positioning system device, or any other suitable device configured to communicate via a wireless or wired medium.

[0038] Some UEs may be considered to be machine type communication (MTC) or evolved or enhanced machine type communication (eMTC) UEs. MTC and eMTC UEs include, for example, robots, drones, remote devices, sensors, meters, monitors, location tags, etc., which may communicate with a base station, another device (e.g., a remote device), or some other entity. For example, a wireless node may provide a connection to a network (e.g., a wide area network such as the Internet or a cellular network) or provide a connection to a network via a wired or wireless communication link. Some UEs may be considered to be Internet of Things (IoT) devices, and / or may be implemented as NB-IoT (narrowband Internet of Things) devices. Some UEs may be considered to be customer premises equipment (CPE). UE 120 may be included in a housing that houses components of UE 120 (such as a processor component, a memory component, etc.).

[0039] Generally, any number of wireless networks can be deployed in a given geographic area. Each wireless network can support a specific RAT and can operate on one or more frequencies. RAT can also be referred to as radio technology, air interface, etc. Frequency can also be referred to as carrier, channel, etc. In order to avoid interference between wireless networks of different RATs, each frequency can support a single RAT in a given geographic area. In some cases, NR or 5G RAT networks can be deployed.

[0040] In some aspects, two or more UEs 120 (e.g., shown as UE 120a and UE 120e) may communicate directly (e.g., without using base station 110 as an intermediary to communicate with each other) using one or more sidelink channels. For example, UE 120 may use peer-to-peer (P2P) communication, device-to-device (D2D) communication, vehicle-to-everything (V2X) protocol (e.g., which may include vehicle-to-vehicle (V2V) protocol, vehicle-to-infrastructure (V2I) protocol, etc.), mesh network, etc. In this case, UE 120 may perform scheduling operations, resource selection operations, and / or other operations described elsewhere as being performed by base station 110.

[0041] As described above, Figure 1 Provided as an example only. Other examples can be found with Figure 1 Described differently.

[0042] Figure 2 A block diagram of a design 200 of a base station 110 and a UE 120 is shown, which may be Figure 1 Base station 110 may be equipped with T antennas 234a through 234t, and UE 120 may be equipped with R antennas 252a through 252r, where in general T≥1 and R≥1.

[0043] At the base station 110, the transmit processor 220 may receive data for one or more UEs from the data source 212, select one or more modulation and coding schemes (MCS) for each UE based at least in part on a channel quality indication (CQI) received from the UE, process (e.g., decode and modulate) the data for each UE based at least in part on the MCS selected for the UE, and provide data symbols for all UEs. The transmit processor 220 may also process system information (e.g., for semi-static resource partitioning information (SRPI), etc.) and control information (e.g., CQI requests, grants, upper layer signaling, etc.) and provide overhead symbols and control symbols. The transmit processor 220 may also generate reference symbols for reference signals (e.g., cell-specific reference signals (CRS)) and synchronization signals (e.g., primary synchronization signals (PSS) and secondary synchronization signals (SSS)). The transmit (TX) multiple-input multiple-output (MIMO) processor 230 may perform spatial processing (e.g., precoding) on ​​data symbols, control symbols, overhead symbols, and / or reference symbols (if applicable), and may provide T output symbol streams to T modulators (MOD) 232a to 232t. Each modulator 232 may process a corresponding output symbol stream (e.g., for OFDM, etc.) to obtain an output sample stream. Each modulator 232 may also process (e.g., convert to analog, amplify, filter, and up-convert) the output sample stream to obtain a downlink signal. The T downlink signals from modulators 232a to 232t may be transmitted via T antennas 234a to 234t, respectively. According to various aspects described in more detail below, position coding may be used to generate synchronization signals to convey additional information.

[0044] At the UE 120, antennas 252a to 252r may receive downlink signals from the base station 110 and / or other base stations, and may provide received signals to demodulators (DEMODs) 254a to 254r, respectively. Each demodulator 254 may condition (e.g., filter, amplify, downconvert, and digitize) the received signal to obtain input samples. Each demodulator 254 may also process the input samples (e.g., for OFDM, etc.) to obtain received symbols. The MIMO detector 256 may obtain received symbols from all R demodulators 254a to 254r, perform MIMO detection on the received symbols (if applicable), and provide detected symbols. The receive processor 258 may process (e.g., demodulate and decode) the detected symbols, provide decoded data for the UE 120 to the data sink 260, and provide decoded control information and system information to the controller / processor 280. The channel processor may determine reference signal received power (RSRP), received signal strength indicator (RSSI), reference signal received quality (RSRQ), channel quality indicator (CQI), etc. In some aspects, one or more components of UE 120 may be included in a housing.

[0045] On the uplink, at the UE 120, a transmit processor 264 may receive and process data from a data source 262 and control information from a controller / processor 280 (e.g., for reports including RSRP, RSSI, RSRQ, CQI, etc.). The transmit processor 264 may also generate reference symbols for one or more reference signals. The symbols from the transmit processor 264 may be precoded by a TX MIMO processor 266 (if applicable), further processed by modulators 254a through 254r (e.g., for DFT-s-OFDM, CP-OFDM, etc.), and transmitted to the base station 110. At the base station 110, uplink signals from the UE 120 and other UEs may be received by the antenna 234, processed by the demodulator 232, detected by the MIMO detector 236 (if applicable), and further processed by the receive processor 238 to obtain decoded data and control information transmitted by the UE 120. The receive processor 238 may provide decoded data to a data sink 239 and decoded control information to a controller / processor 240. The base station 110 may include a communication unit 244 and communicate with the network controller 130 via the communication unit 244. The network controller 130 may include a communication unit 294, a controller / processor 290, and a memory 292.

[0046] The controller / processor 240 of the base station 110, the controller / processor 280 of the UE 120, and / or Figure 2Any other component of the system may perform one or more techniques associated with triggering a known payload for neural network training, as described in more detail elsewhere. For example, the controller / processor 240 of the base station 110, the controller / processor 280 of the UE 120, and / or Figure 2 Any other component of the can perform or direct e.g. Figure 7-Figure 8 The operations of processes 700, 800 and / or other processes described herein may be performed. Memories 242 and 282 may store data and program codes for base station 110 and UE 120, respectively. Scheduler 246 may schedule UEs for data transmission on the downlink and / or uplink.

[0047] In some aspects, UE 120 may include means for communicating, means for receiving, means for unicasting, means for sending, means for requesting, and means for training. These means may include combining Figure 2 One or more components of UE 120 or base station 110 are described.

[0048] As mentioned above, Figure 2 Provided as an example only. Other examples may differ from the reference Figure 2 The content described.

[0049] In some cases, different types of devices supporting different types of applications and / or services may coexist in one cell. Examples of different types of devices include UE handheld devices, customer premises equipment (CPE), vehicles, Internet of Things (IoT) devices, etc. Examples of different types of applications include ultra-reliable low-latency communications (URLLC) applications, massive machine type communications (mMTC) applications, enhanced mobile broadband (eMBB) applications, vehicle-to-everything (V2X) applications, etc. In addition, in some cases, a single device may support different applications or services simultaneously.

[0050] Figure 3An example implementation of a system on chip (SOC) 300 is shown, which may include a central processing unit (CPU) 302 or a multi-core CPU configured to trigger a known payload for neural network training according to certain aspects of the present disclosure. The SOC 300 may be included in a base station 110 or a UE 120. Variables (e.g., neural signals and synaptic weights), system parameters associated with a computing device (e.g., a neural network with weights), delays, frequency bin information, and task information may be stored in a memory block associated with a neural processing unit (NPU) 108, a memory block associated with a CPU 302, a memory block associated with a graphics processing unit (GPU) 304, a memory block associated with a digital signal processor (DSP) 306, a memory block 318, or may be distributed in multiple blocks. Instructions executed at the CPU 302 may be loaded from a program memory associated with the CPU 302, or may be loaded from the memory block 318.

[0051] The SOC 300 may also include additional processing modules customized for specific functions, such as a GPU 304, a DSP 306, a connection block 310 (which may include a fifth generation (5G) connection, a fourth generation long term evolution (4G LTE) connection, a Wi-Fi connection, a USB connection, a Bluetooth connection, etc.), and a multimedia processor 312 that can detect and recognize gestures, for example. In one implementation, the NPU is implemented in the CPU, DSP, and / or GPU. The SOC 300 may also include a sensor processor 314, an image signal processor (ISP) 316, and / or a navigation module 320 (which may include a global positioning system).

[0052] SOC 300 may be based on the ARM instruction set. In one aspect of the present disclosure, the instructions loaded into the general processor 302 may include code for communicating with a sending device about the contents of a known payload. The instructions may also include code for requesting a known payload for training an artificial neural network from a sending device. The instructions may also include code for receiving a known payload in response to the request, and code for performing online training of an artificial neural network using a known payload. The instructions loaded into the general processor 302 may include code for communicating with a sending device about the contents of a known payload, and code for sending an indication to notify a receiving device that a known payload will be sent. The instructions may also include code for unicasting a known payload to a receiving device to train a neural network.

[0053] Deep learning architectures can perform object recognition tasks by learning to represent inputs at successively higher levels of abstraction in each layer, thereby building useful feature representations of the input data. In this way, deep learning solves the main bottleneck of traditional machine learning. Before the advent of deep learning, machine learning methods for object recognition problems may rely heavily on human-engineered features, possibly combined with shallow classifiers. For example, a shallow classifier can be a two-class linear classifier, in which the weighted sum of the feature vector components can be compared to a threshold to predict which class the input belongs to. Human-engineered features can be templates or kernels customized for a specific problem domain by engineers with domain expertise. In contrast, deep learning architectures can learn to represent features similar to those that human engineers might design, but through training. In addition, deep networks can learn to represent and recognize new types of features that humans may not have considered.

[0054] Deep learning architectures can learn hierarchies of features. For example, if presented with visual data, the first layer can learn to recognize relatively simple features in the input stream, such as edges. In another example, if presented with auditory data, the first layer can learn to recognize the spectral power of specific frequencies. Using the output of the first layer as input, the second layer can learn to recognize combinations of features, such as simple shapes for visual data or combinations of sounds for auditory data. For example, higher layers can learn to represent complex shapes in visual data or words in auditory data. Even higher layers can learn to recognize common visual objects or spoken phrases.

[0055] Deep learning architectures can perform particularly well when applied to problems that have a natural hierarchical structure. For example, the classification of motor vehicles can benefit from first learning to recognize wheels, windshields, and other features. These features can be combined in different ways at higher levels to recognize cars, trucks, and airplanes.

[0056] Neural networks can be designed with a variety of connection patterns. In a feedforward network, information is passed from a lower layer to a higher layer, where each neuron in a given layer communicates with a neuron in a higher layer. As described above, hierarchical representations can be established in successive layers of a feedforward network. Neural networks can also have loops or feedback (also known as top-down) connections. In a loop connection, the output from a neuron in a given layer can be communicated to another neuron in the same layer. The loop architecture can help identify more than one pattern across a chunk of input data that is sequentially passed to the neural network. The connection from a neuron in a given layer to a neuron in a lower layer is called a feedback (or top-down) connection. A network with many feedback connections can be helpful when the recognition of high-level concepts can help distinguish specific low-level features of the input.

[0057] The connections between layers of a neural network can be either fully connected or partially connected. Figure 4A An example of a fully connected neural network 402 is shown. In the fully connected neural network 402, neurons in a first layer may communicate their outputs to every neuron in a second layer, such that every neuron in the second layer will receive input from every neuron in the first layer. Figure 4B An example of a locally connected neural network 404 is shown. In the locally connected neural network 404, neurons in a first layer may be connected to a limited number of neurons in a second layer. More generally, the locally connected layers of the locally connected neural network 404 may be configured such that each neuron in the layer will have the same or similar connection pattern, but with connection strengths (e.g., 410, 412, 414, and 416) that may have different values. The locally connected connection patterns may produce spatially different receptive fields in higher layers because higher layer neurons in a given area may receive inputs that are tuned through training to properties of a limited portion of the total input to the network.

[0058] An example of a locally connected neural network is a convolutional neural network. Figure 4C An example of a convolutional neural network 406 is shown. The convolutional neural network 406 can be configured such that the connection strengths associated with the inputs of each neuron in the second layer are shared (e.g., 408). Convolutional neural networks may be well suited for problems where the spatial location of the inputs is meaningful.

[0059] One type of convolutional neural network is a deep convolutional network (DCN). Figure 4D A detailed example of a DCN 400 designed to recognize visual features from an image 426 input from an image capture device 430 (such as a vehicle-mounted camera) is shown. The DCN 400 of the current example can be trained to recognize traffic signs and numbers provided on traffic signs. Of course, the DCN 400 can be trained for other tasks, such as recognizing lane markings or recognizing traffic lights.

[0060] The DCN 400 may be trained using supervised learning. During training, an image (such as an image 426 of a speed limit sign) may be presented to the DCN 400, and then a forward pass may be calculated to produce an output 422. The DCN 400 may include a feature extraction portion and a classification portion. Upon receiving the image 426, a convolution layer 432 may apply a convolution kernel (not shown) to the image 426 to generate a first feature atlas 418. For example, the convolution kernel of the convolution layer 432 may be a 5×5 kernel that generates a 28×28 feature map. In this example, because four different feature maps are generated in the first feature atlas 418, four different convolution kernels are applied to the image 426 at the convolution layer 432. A convolution kernel may also be referred to as a filter or a convolution filter.

[0061] The first feature atlas 418 may be subsampled by a maximum pooling layer (not shown) to generate a second feature atlas 420. The maximum pooling layer reduces the size of the first feature atlas 418. That is, the size of the second feature atlas 420 (such as 14×14) is smaller than the size of the first feature atlas 418 (such as 28×28). The reduced size provides similar information to subsequent layers while reducing memory consumption. The second feature atlas 420 may be further convolved via one or more subsequent convolutional layers (not shown) to generate one or more subsequent feature atlases (not shown).

[0062] exist Figure 4D In the example of , the second feature atlas 420 is convolved to generate a first feature vector 424. In addition, the first feature vector 424 is further convolved to generate a second feature vector 428. Each feature of the second feature vector 428 may include a number corresponding to a possible feature of the image 426, such as "sign", "60", and "100". A softmax function (not shown) can convert the numbers in the second feature vector 428 into probabilities. In this way, the output 422 of the DCN 400 is the probability that the image 426 includes one or more features.

[0063] In this example, the probability of "logo" and "60" in output 422 is higher than the probability of other contents (such as "30", "40", "50", "70", "80", "90", and "100") in output 422. Before training, the output 422 generated by DCN 400 may be incorrect. Therefore, the error between output 422 and the target output can be calculated. The target output is the ground truth value of image 426 (e.g., "logo" and "60"). The weights of DCN 400 can then be adjusted so that the output 422 of DCN 400 is closer to the target output.

[0064] To adjust the weights, the learning algorithm may calculate a gradient vector for the weights. The gradient may indicate the amount by which the error will increase or decrease if the weights are adjusted. At the top layer, the gradient may correspond directly to the value of the weights connecting the activated neurons in the penultimate layer and the neurons in the output layer. In lower layers, the gradient may depend on the value of the weights and the calculated error gradients of higher layers. The weights may then be adjusted to reduce the error. This way of adjusting weights may be referred to as "back propagation" because it involves a "backward pass" through the neural network.

[0065] In practice, the error gradient of the weights may be calculated over a small number of examples so that the calculated gradient is close to the true error gradient. This approximation method may be referred to as stochastic gradient descent. Stochastic gradient descent may be repeated until the achievable error rate of the entire system stops decreasing or until the error rate reaches a target level. After learning, a new image (e.g., a speed limit sign of image 426) may be presented to the DCN, and a forward pass through the network may produce an output 422 that may be considered an inference or prediction of the DCN.

[0066] A Deep Belief Network (DBN) is a probabilistic model that includes multiple layers of hidden nodes. A DBN can be used to extract a hierarchical representation of a training data set. A DBN can be obtained by stacking multiple layers of restricted Boltzmann machines (RBMs). RBMs are artificial neural networks that can learn probability distributions of input sets. Because RBMs can learn probability distributions without information about the class to which each input should be classified, RBMs are often used for unsupervised learning. Using a hybrid unsupervised and supervised paradigm, the bottom RBM of a DBN can be trained in an unsupervised manner and can be used as a feature extractor, while the top RBM can be trained in a supervised manner (based on the joint distribution of inputs from the previous layer and the target class) and can be used as a classifier.

[0067] A deep convolutional network (DCN) is a network of convolutional networks configured with additional pooling and normalization layers. DCNs have achieved state-of-the-art performance on a variety of tasks. DCNs can be trained using supervised learning, in which both the input and output targets are known for multiple samples and used to modify the weights of the network by using a gradient descent method.

[0068] The DCN can be a feed-forward network. In addition, as described above, the connections from neurons in the first layer of the DCN to a set of neurons in the next higher layer are shared between the neurons in the first layer. The feed-forward and shared connections of the DCN can be used for fast processing. For example, the computational burden of the DCN can be much smaller than that of a neural network of similar size that includes loops or feedback connections.

[0069] The processing of each layer of the convolutional network can be viewed as a spatially invariant template or basis projection. If the input is first decomposed into multiple channels (such as the red, green, and blue channels of a color image), then the convolutional network trained on this input can be considered to be three-dimensional, with two spatial dimensions along the axes of the image and the third dimension capturing color information. The output of the convolutional connection can be considered to form a feature map in the subsequent layer, with each element of the feature map (e.g., 220) receiving input from a range of neurons in the previous layer (e.g., feature map 218) and from each of the multiple channels. The values ​​in the feature map can be further processed with nonlinearities (such as rectification, max(0, x)). The values ​​from neighboring neurons can be further pooled (which corresponds to downsampling) and can provide additional local invariance and dimensionality reduction. Normalization corresponding to whitening can also be applied by lateral inhibition between neurons in the feature map.

[0070] The performance of deep learning architectures can improve as more labeled data points become available or as computing power increases. Modern deep neural networks are often trained using thousands of times the computing resources available to a typical researcher just 15 years ago. New architectures and training paradigms can further improve the performance of deep learning. Rectified linear units can reduce the training problem known as vanishing gradients. New training techniques can reduce overfitting, allowing larger models to achieve better generalization. Encapsulation techniques can extract data within a given receptive field and further improve overall performance.

[0071] Figure 5 is a block diagram showing a deep convolutional network 550. Based on connections and weight sharing, the deep convolutional network 550 may include multiple different types of layers. Figure 5 As shown, the deep convolutional network 550 includes convolution blocks 554A and 554B. Each of the convolution blocks 554A and 554B may be configured with a convolution layer (CONV) 356, a normalization layer (LNorm) 558, and a maximum pooling layer (MAXPOOL) 560.

[0072] The convolution layer 556 may include one or more convolution filters that can be applied to the input data to generate a feature map. Although only two of the convolution blocks 554A, 554B are shown, the present disclosure is not limited thereto, and on the contrary, any number of convolution blocks 554A, 554B may be included in the deep convolution network 550, depending on design preferences. The normalization layer 558 may normalize the output of the convolution filter. For example, the normalization layer 558 may provide whitening or lateral suppression. The maximum pooling layer 560 may provide spatial downsampling aggregation to obtain local invariance and dimensionality reduction.

[0073] For example, the parallel filter bank of the deep convolutional network may be loaded onto the CPU 302 or GPU 304 of the SOC 300 to achieve high performance and low power consumption. In an alternative embodiment, the parallel filter bank may be loaded onto the DSP 306 or ISP 316 of the SOC 300. In addition, the deep convolutional network 550 may access other processing blocks that may be present on the SOC 300, such as the sensor processor 314 and the navigation module 320 dedicated to sensors and navigation, respectively.

[0074] The deep convolutional network 550 may also include one or more fully connected layers 562 (FC1 and FC2). The deep convolutional network 550 may also include a logistic regression (LR) layer 564. Between each layer 556, 558, 560, 562, 564 of the deep convolutional network 550 are weights (not shown) to be updated. The output of each of the layers (e.g., 556, 558, 560, 562, 564) may be used as an input to a subsequent one of the layers (e.g., 556, 558, 560, 562, 564) in the deep convolutional network 550 to learn a hierarchical feature representation from the input data 552 (e.g., image, audio, video, sensor data, and / or other input data) provided at the first one in the convolutional block 554A. The output of the deep convolutional network 550 is a classification score 566 for the input data 552. The classification score 566 may be a set of probabilities, where each probability is a probability of the input data, including a feature from a feature set.

[0075] As mentioned above, Figure 3-Figure 5 Provided as an example. Other examples may differ from the reference Figure 3-Figure 5 The content described.

[0076] As described above, artificial intelligence (AI) / machine learning (ML) algorithms can improve wireless communications. The AI / ML module can run on the UE, the base station, or in the case of a distributed algorithm, jointly on the UE and the base station. Although the present disclosure relates to the AI / ML module on the UE side, the present disclosure explicitly contemplates the autoencoder scenario and the scenario in which the AI ​​module runs on the network side. In the autoencoder scenario, joint training occurs.

[0077] In order to obtain useful results, the neural network executing the AI / ML algorithm needs to be trained. Offline training of the neural network can derive the weights and biases of the neural network from the training data. However, offline training does not take into account the dynamic real-world environment of wireless communication. Online training takes into account the dynamic environment, including conditions for which the neural network is not trained. Online training can be accomplished using known air transmissions (or signaling) that reflect the wireless environment. For the online training process, instead of training from scratch, training can start with the values ​​of weights and biases derived from the offline training phase. The online training process then further fine-tunes the neural network for the wireless channel, noise, and / or other environmental characteristics.

[0078] Reference signals, such as demodulation reference signals (DMRS), channel state information reference signals (CSI-RS), etc., are known transmissions. However, reference signals may not provide the appropriate type of information for certain types of machine learning models. Data transmission may be required for specific machine learning models, such as for log-likelihood ratio (LLR) calculations, MIMO demapping, etc. Known data transmissions may include, for example, known payloads of a physical downlink control channel (PDCCH), a physical uplink control channel (PUCCH), a physical downlink shared channel (PDSCH), and / or a physical uplink shared channel (PUSCH).

[0079] A generalized machine learning (ML) model for a MIMO demapper can be represented, for example, as y=Hx+n, where y is the received vector, x is the transmitted symbol vector, n is the noise vector, and H is the channel matrix. A neural network (NN) can be trained offline to determine (estimated symbols), and can benefit from fine-tuning using online training. The input to the neural network is the received observations (y) and the estimated channel matrix (H), and the output is the detected transmitted symbols, e.g. In this case, the ground truth labels are known payloads used to train the neural network. Therefore, the device (e.g., UE or gNB) can perform further online training without having to decode y to use the decoded Serves as ground truth labels for the neural network. Because the payload is known, the machine learning model can use the training data to learn how to interpret the symbols in the presence of channel characteristics, noise characteristics, etc.

[0080] This is just an illustrative example of how known data transmission is useful for the purpose of online training. The above example has application in the context of unicast transmission, where the base station sends training data exclusively to a given UE (given multiple-input multiple-output (MIMO) rank, modulation and coding scheme (MCS), etc.) and the UE uses the training data to train its neural network. Furthermore, the use case is not limited to MIMO demappers, the neural network can be any type of neural network.

[0081] Regular data transmissions on these channels can be considered known once decoded. Regular data transmissions include system information blocks (SIBs) that are known to be repeated periodically between boundaries when the system information (SI) changes. Regular data transmissions may also include unicast transmissions once a cyclic redundancy check (CRC) passes. However, decoding regular data transmissions for training artificial neural networks may use excessive memory and computational overhead, resulting in high latency. For example, received modulation symbols are stored until decoding is completed, and only after decoding is completed can the decoded symbols be used for training purposes.

[0082] Therefore, a mode in which both the base station and the UE know and / or agree in advance on the payload content of a particular physical channel transmission may be desirable. These known transmissions can be immediately used as ground truth labels in online training of the neural network.

[0083] For the purpose of online training, the signaling framework can enable UEs to train their respective neural networks. Sending a "known payload" known to both parties can be beneficial in a variety of contexts because the receiver does not need to fully decode the payload to find the label for training.

[0084] The base station can unicast the aforementioned "known data" so that the UE can use the data to train its neural network. These transmissions are UE-specific and can be customized for a specific topic (e.g., for a specific neural network). Similarly, the UE can unicast known data to the base station so that the base station can train its neural network. In some configurations, the known data is transmitted point-to-point because other receiving devices are unlikely to find the training data useful. The training data can be specifically customized for a specific receiving device.

[0085] Based on which type of neural network the UE or base station will train, the receiving device may require different types of data. Therefore, a receiver-specific known payload can be generated. The known payload can include data for a specific receiver (e.g., MCS, rank, beam pair, etc.).

[0086] In some aspects of the present disclosure, a known payload may be generated in a manner that enables the UE and the base station to independently generate the known payload. For example, the known payload may be based on a scrambling seed configured by radio resource control (RRC), or may be any sequence known to both the UE and the base station.

[0087] The transmission of known data in the PDSCH for downlink training can be triggered by a UE request. The transmission of known data in the PUSCH for uplink training can be triggered by a base station request. The UE and base station requests can include specific characteristics, such as MCS, beam pairs, rank, etc. The training of a specific neural network will be based on these requested characteristics.

[0088] The transmission of known data in the PDSCH for downlink training may also be initiated by a base station indication. The indication may be in the PDCCH or the medium access control (MAC) layer control element (MAC-CE). The transmission of known data in the PUSCH for uplink training may be initiated by a UE indication. The indication may be through the PUCCH or the MAC-CE.

[0089] According to aspects of the present disclosure, the location of the known payload (e.g., on which physical channel) and the exact time-frequency resources, periodicity (aperiodic, semi-persistent, periodic, etc.), duration and / or aggregation level (for PDCCH), payload size, etc. of the known payload can be explicitly signaled separately from the payload. For example, the base station can send information about the known payload. The explicit signaling can be via RRC messages, MAC control elements (CEs), or via PDCCH downlink control information (DCI).

[0090] Some non-limiting examples of known data locations include PDSCH, PUSCH, PUCCH, and PDCCH. For PDCCH, the known payload can be paired with a known payload in PDSCH or PUSCH. As previously described, for PDCCH and PUCCH, the known payload can be periodic, aperiodic, or semi-persistent, where the parameters are signaled separately. Parameters can include where the known payload is sent and how long it is sent. In the case of periodic transmissions, a semi-persistent scheduling or configuration grant (CG) can be defined onto which the known payload is mapped.

[0091] When a known payload is sent on the PUCCH, parameters related to the known payload may be signaled separately. The signaling may be via an RRC message, MAC-CE, or downlink control information (DCI). The parameters signaled separately may include the PUCCH format, uplink control information (UCI) multiplexed on the PUSCH, their payload sizes, etc. For example, the UE may be signaled the type of UCI multiplexed on the PUSCH, such as whether HARQ-ACK is multiplexed on the PUSCH, or whether HARQ-ACK and the first part of CSI are multiplexed on the PUSCH.

[0092] Fig. 6A is a timing diagram illustrating triggering future known payloads for training an artificial neural network in accordance with various aspects of the present disclosure. Fig. 6A In the example shown, base station 110 is shown as sending a known payload to UE 120 for online training of the UE's neural network. This scenario is for illustration purposes only, as UE 120 may also send a known payload to base station 110, or first UE 120 may send a known payload to another UE (not shown), for example, via sidelink communication.

[0093] At time t1, UE 120 and base station 110 agree on a known payload. In some aspects of the present disclosure, both base station 110 and UE 120 know / agree on the payload content of a particular physical channel transmission in advance. For example, UE 120 and base station 110 communicate about the content of the known payload. These known transmissions can be immediately used as ground truth labels in online training of neural networks. Sending a "known payload" that is known to both parties is beneficial in a variety of contexts because the receiver does not need to fully decode the payload to find the label for training. In some aspects, base station 110 configures UE 120 with a known payload, for example via RRC signaling. For example, a known payload can be generated using a scrambling seed shared with the UE via RRC configuration.

[0094] At time t2, UE 120 requests a known payload from base station 110. The UE request may include specific characteristics, such as MCS, beam pairs, rank, etc. The training of a specific neural network will be based on these requested characteristics. The request from UE 120 may be triggered by a performance degradation of the UE-side neural network. For example, if the performance of the neural network does not meet certain criteria, the neural network may be outdated. Therefore, fine-tuning or updating of the network can be performed using online training.

[0095] At time t3, base station 110 transmits the aforementioned known data, for example via unicast, so that UE 120 can use the data to train its neural network. The transmission is UE-specific and can be customized for a specific subject (e.g., for a specific neural network). The training data can be specifically customized for a specific receiving device.

[0096] At time t4, UE 120 performs online training of its neural network based on the received known payload. Because the payload is known, the machine learning model can use the training data to learn how to interpret the symbols in the presence of channel characteristics, noise characteristics, etc.

[0097] Figure 6B is a timing diagram illustrating triggering future known payloads to be used for training an artificial neural network in accordance with various aspects of the present disclosure. Figure 6B In the illustrated example, UE 120 is shown sending a known payload to base station 110 for online training of the base station's neural network.

[0098] At time t1, UE 120 and base station 110 agree on a known payload. In some aspects of the present disclosure, both base station 110 and UE 120 know / agree in advance on the payload content of a particular physical channel transmission. For example, UE 120 and base station 110 communicate about the content of the known payload. In some aspects, base station 110 configures UE 120 with the known payload, for example, via RRC signaling. For example, the known payload can be generated using a scrambling seed shared with the UE via RRC configuration.

[0099] At time t2, the base station 110 requests a known payload from the UE 120. The request may include specific characteristics, such as MCS, beam pairs, rank, etc. The training of a specific neural network will be based on these requested characteristics. The request from the base station 110 may be triggered by a performance degradation of the neural network on the UE side. For example, if the performance of the neural network does not meet certain criteria, the neural network may be outdated. Therefore, fine-tuning or updating of the network can be performed using online training.

[0100] At time t3, UE 120 transmits the aforementioned known data so that base station 110 can use the data to train its neural network. At time t4, base station 110 performs online training of its neural network based on the received known payload.

[0101] Figure 7 7 is a diagram illustrating an example process 700, for example, performed by a receiving device, according to various aspects of the present disclosure. The example process 700 is an example of triggering a known payload, for example, for training an artificial neural network.

[0102] like Figure 7 As shown, in some aspects, process 700 may include communicating with a transmitting device about the contents of a known payload (block 705). For example, a UE (e.g., using antenna 252, DEMOD 254, TX MIMO processor 266, transmit processor 264, MIMO detector 256, receive processor 258, controller / processor 280, etc.) or a base station (e.g., using antenna 234, DEMOD 232, TX MIMO processor 230, transmit processor 220, MIMO detector 236, receive processor 238, controller / processor 240, etc.) may communicate about the contents.

[0103] Process 700 may include requesting a known payload for training an artificial neural network (block 710). For example, a UE (e.g., using antenna 252, DEMOD 254, TX MIMO processor 266, transmit processor 264, controller / processor 280, etc.) or a base station (e.g., using antenna 234, DEMOD 232, TX MIMO processor 230, transmit processor 220, controller / processor 240, etc.) may request a known payload for training an artificial neural network.

[0104] like Figure 7 As shown, in some aspects, process 700 may include receiving a known payload in response to the request (block 715). For example, a UE (e.g., using antenna 252, DEMOD 254, MIMO detector 256, receive processor 258, controller / processor 280, etc.) or a base station (e.g., using antenna 234, DEMOD 232, MIMO detector 236, receive processor 238, controller / processor 240, etc.) may receive the known payload in response to the request.

[0105] like Figure 7 As shown, in some aspects, process 700 may include online training of an artificial neural network using a known payload (block 720). For example, a UE (e.g., using controller / processor 280, memory 282, etc.) and a base station (e.g., using controller / processor 240, memory 242, etc.) may train an artificial neural network using a known payload. The known payload may be used to generate ground truth labels for training the neural network.

[0106] Figure 8 8 is a diagram illustrating an example process 800, for example, performed by a sending device, according to various aspects of the present disclosure. The example process 800 is an example of unicasting known data for training a neural network.

[0107] like Figure 8As shown, in some aspects, process 800 may include communicating with a receiving device about the contents of a known payload (block 805). For example, a UE (e.g., using antenna 252, DEMOD 254, TX MIMO processor 266, transmit processor 264, MIMO detector 256, receive processor 258, controller / processor 280, etc.) or a base station (e.g., using antenna 234, DEMOD 232, TX MIMO processor 230, transmit processor 220, MIMO detector 236, receive processor 238, controller / processor 240, etc.) may communicate about the contents.

[0108] like Figure 8 As shown, in some aspects, process 800 may include sending an indication notifying a receiving device that a known payload will be sent (block 810). For example, a UE (e.g., using antenna 252, DEMOD 254, TX MIMO processor 266, transmit processor 264, controller / processor 280, memory 282, etc.) or a base station (e.g., using antenna 234, DEMOD 232, TX MIMO processor 230, transmit processor 220, controller / processor 240, memory 242, etc.) may send an indication notifying a receiving device of a known payload.

[0109] like Figure 8 As shown, in some aspects, process 800 may include unicasting a known payload to a receiving device for online training of a neural network (block 815). For example, a UE (e.g., using antenna 252, DEMOD 254, TX MIMO processor 266, transmit processor 264, controller / processor 280, memory 282, etc.) or a base station (e.g., using antenna 234, DEMOD 232, TX MIMO processor 230, transmit processor 220, controller / processor 240, memory 242, etc.) may unicast a known payload to a receiving device.

[0110] Examples of implementations are described in the following numbered clauses:

[0111] 1. A method for wireless communication by a receiving device, comprising:

[0112] communicating with the sending device regarding the contents of the known payload;

[0113] requesting a known payload for training an artificial neural network from a sending device;

[0114] receiving a known payload in response to the request; and

[0115] Online training of the artificial neural network is performed using the known payload.

[0116] 2. A method according to clause 1, wherein the known payload is received in at least one of a physical downlink shared channel (PDSCH) or a physical uplink shared channel (PUSCH).

[0117] 3. A method according to any of the preceding clauses, wherein the request for a known payload originates from a user equipment (UE) for downlink training or from a base station for uplink training.

[0118] 4. A method according to any of the preceding clauses, wherein the request describes a characteristic of a known payload, the characteristic comprising at least one of a modulation and coding scheme (MCS), a beam pair or a rank, and the training is based on the characteristic.

[0119] 5. The method according to any of the preceding clauses further includes receiving information including periodicity, time and frequency resources of a known payload, and a payload size, the information being carried by at least one of radio resource control (RRC) signaling, a medium access control layer (MAC) control element, or PDCCH downlink control information (DCI).

[0120] 6. A method according to clause 1, wherein the known payload is received in at least one of a physical downlink control channel (PDCCH) or a physical uplink control channel (PUCCH).

[0121] 7. A method according to clause 6, wherein the known payload is received in a PDCCH and is paired with another known payload in one of a physical downlink shared channel (PDSCH) or a physical uplink shared channel (PUSCH).

[0122] 8. A method according to any of the preceding clauses, wherein the known payload is based on a radio resource control (RRC) scrambling seed.

[0123] 9. A method according to any of the preceding clauses, wherein, for a UE, the request occurs via a physical uplink control channel (PUCCH) or a medium access control layer (MAC) control element (CE), or for a base station, the request occurs via a physical downlink control channel (PDCCH) or MAC-CE.

[0124] 10. A method according to any of the preceding clauses, wherein the known payload is received in a physical uplink control channel (PUCCH) with relevant parameters separately signaled via radio resource control (RRC) signaling, downlink control information (DCI) or media access control layer (MAC) control element (CE), and the relevant parameters include at least one of the PUCCH format, multiplexed uplink control information (UCI) type or payload size.

[0125] 11. A method of wireless communication performed by a transmitting device, comprising:

[0126] communicating with a receiving device the contents of a known payload;

[0127] sending an indication notifying a receiving device that a known payload will be sent; and

[0128] Unicast a known payload to receiving devices for online training of neural networks.

[0129] 12. A method according to clause 11, wherein the indication comprises a base station indication in a physical downlink control channel (PDCCH) or a medium access control (MAC) layer control element (MAC-CE) in response to a known payload being transmitted in a physical downlink shared channel (PDSCH) for downlink training.

[0130] 13. A method according to any of the preceding clauses, wherein the indication comprises a user equipment (UE) indication in a physical uplink control channel (PUCCH) or a medium access control (MAC) layer control element (MAC-CE) in response to a known payload being transmitted in a physical uplink shared channel (PUSCH) for uplink training.

[0131] 14. The method according to any of the preceding clauses further includes sending information including the periodicity, time and frequency resources of a known payload, and a payload size, the information being carried by radio resource control (RRC) signaling, a medium access control layer (MAC) control element, or PDCCH downlink control information (DCI).

[0132] 15. A method according to any of the preceding clauses, wherein the known payload is based on a radio resource control (RRC) scrambling seed.

[0133] 16. A receiving device for wireless communication, comprising:

[0134] Memory; and

[0135] at least one processor coupled to the memory and configured to:

[0136] communicating with the sending device regarding the contents of the known payload;

[0137] requesting a known payload for training an artificial neural network from a sending device;

[0138] receiving a known payload in response to the request; and

[0139] Online training of the artificial neural network is performed using the known payload.

[0140] 17. A receiving device according to clause 16, wherein the known payload is received in at least one of a physical downlink shared channel (PDSCH) or a physical uplink shared channel (PUSCH).

[0141] 18. A receiving device according to any of the preceding clauses, wherein the request for a known payload originates from a user equipment (UE) for downlink training or from a base station for uplink training.

[0142] 19. A receiving device according to any of the preceding clauses, wherein the request describes characteristics of a known payload, the characteristics comprising at least one of a modulation and coding scheme (MCS), a beam pair and / or a rank, and the training is based on the characteristics.

[0143] 20. A receiving device according to clause 16, wherein the known payload is received in a physical downlink control channel (PDCCH) and is paired with another known payload in a physical downlink shared channel (PDSCH) or a physical uplink shared channel (PUSCH).

[0144] 21. A receiving device according to any of the preceding clauses, wherein the at least one processor is also configured to receive information including the periodicity, time / frequency and payload size of a known payload, which information is carried by at least one of radio resource control (RRC) signaling, medium access control layer (MAC) control elements or PDCCH downlink control information (DCI).

[0145] 22. A receiving device according to any of the preceding clauses, wherein the known payload is based on a radio resource control (RRC) scrambling seed.

[0146] 23. A receiving device according to any of the preceding clauses, wherein the at least one processor is also configured to make a request for a UE via a physical uplink control channel (PUCCH) or a medium access control layer (MAC) control element (CE), or for a base station via a physical downlink control channel (PDCCH) or MAC-CE.

[0147] 24. A receiving device according to clause 16, wherein the known payload is received in a physical uplink control channel (PUCCH) with relevant parameters separately signaled via radio resource control (RRC) signaling, downlink control information (DCI) or medium access control layer (MAC) control element (CE), and the relevant parameters include at least one of PUCCH format, multiplexed uplink control information (UCI) type or payload size.

[0148] 25. A transmitting device for wireless communication, comprising:

[0149] Memory; and

[0150] at least one processor coupled to the memory and configured to:

[0151] communicating with a receiving device the contents of a known payload;

[0152] sending an indication notifying a receiving device that a known payload will be sent; and

[0153] Unicast a known payload to receiving devices for online training of neural networks.

[0154] 26. A transmitting device according to clause 25, wherein, in response to a known payload being transmitted in a physical downlink shared channel (PDSCH) for downlink training, the indication comprises a base station indication in a physical downlink control channel (PDCCH) or a medium access control (MAC) layer control element (MAC-CE).

[0155] 27. A transmitting device according to clause 25, wherein the indication comprises a user equipment (UE) indication in a physical uplink control channel (PUCCH) or a medium access control (MAC) layer control element (MAC-CE) in response to a known payload being transmitted in a physical uplink shared channel (PUSCH) for uplink training.

[0156] 28. A transmitting device according to any of the preceding clauses, wherein the at least one processor is also configured to send information including the periodicity, time and frequency resources of a known payload, and a payload size, the information being carried by radio resource control (RRC) signaling, a medium access control layer (MAC) control element, or PDCCH downlink control information (DCI).

[0157] 29. A transmitting device according to any of the preceding clauses, wherein the known payload is based on a radio resource control (RRC) scrambling seed.

[0158] 30. A transmitting device according to any of the preceding clauses, wherein the at least one processor is also configured to make a request for a UE via a physical uplink control channel (PUCCH) or a medium access control layer (MAC) control element (CE), or for a base station via a physical downlink control channel (PDCCH) or MAC-CE.

[0159] The foregoing disclosure provides illustration and description, but is not intended to be exhaustive or to limit the aspects to the precise forms disclosed. Modifications and variations may be made in light of the above disclosure or may be acquired from practice of these aspects.

[0160] As used, the term "component" is intended to be broadly interpreted as hardware, firmware, and / or a combination of hardware and software. As used, a processor is implemented in hardware, firmware, and / or a combination of hardware and software.

[0161] Some aspects are described in conjunction with a threshold value. As used, satisfying a threshold value may refer to a value greater than a threshold value, greater than or equal to a threshold value, less than a threshold value, less than or equal to a threshold value, equal to a threshold value, not equal to a threshold value, etc., depending on the context.

[0162] It is clear that the described systems and / or methods can be implemented in various forms of hardware, firmware, and / or combinations of hardware and software. The actual dedicated control hardware or software code used to implement these systems and / or methods does not limit these aspects. Therefore, the operation and behavior of the systems and / or methods are described without reference to specific software code, and it should be understood that software and hardware can be designed to implement the systems and / or methods based at least in part on the description.

[0163] Although the specific combination of features is described in the claims and / or disclosed in the specification, these combinations are not intended to limit the disclosure of various aspects. In fact, a plurality of these features can be combined in a manner not specifically described in the claims and / or not specifically disclosed in the specification. Although each of the listed dependent claims may only directly quote a claim, the disclosure of various aspects includes the combination of each dependent claim and each other claim in the claims. The phrase pointing to "at least one" in a series of projects refers to any combination (including single members) of those projects. As an example, "at least one of a, b or c" is intended to cover a, b, c, ab, ac, bc and abc, and any combination (for example, aa, aaa, aab, aac, abb, acc, bb, bbb, bbc, cc and ccc) with the multiple of the same element.

[0164] Unless explicitly stated, any element, behavior or instruction used should not be interpreted as critical or essential. In addition, as used, the articles "a" and "an" are intended to include one or more items and can be used interchangeably with "one or more". In addition, as used, the terms "set" and "group" are intended to include one or more items (e.g., related items, unrelated items, combinations of related and unrelated items, etc.), and can be used interchangeably with "one or more". If only one item is intended to be used, the phrase "only one" or similar language is used. In addition, as used, the terms "have", "have", etc. are intended to be open terms. In addition, unless otherwise explicitly stated, the phrase "based on" is intended to mean "based at least in part on".

Claims

1. A method for wireless communication performed by a receiving device, comprising: communicating with the sending device regarding the contents of the known payload; requesting a known payload for training an artificial neural network from a transmitting device, wherein the request describes a characteristic for the known payload, the characteristic comprising at least one of a modulation and coding scheme (MCS), a beam pair, or a rank, the training being based on the characteristic; receiving a known payload in response to the request; and performing online training of an artificial neural network using the known payload, The method further comprises: Receive information including periodicity, time and frequency resources, and payload size of the known payload, the information being carried by at least one of radio resource control (RRC) signaling, a medium access control layer (MAC) control element, or PDCCH downlink control information (DCI), wherein the periodicity indicates whether the known payload is periodic, aperiodic, or semi-persistent.

2. The method according to claim 1, wherein: The known payload is received in at least one of a physical downlink shared channel (PDSCH) or a physical uplink shared channel (PUSCH).

3. The method according to claim 2, wherein: The request for a known payload originates from a user equipment (UE) for downlink training or from a base station for uplink training.

4. The method according to claim 1, wherein: The known payload is received in at least one of a physical downlink control channel (PDCCH) or a physical uplink control channel (PUCCH).

5. The method according to claim 4, wherein: The known payload is received in the PDCCH and is paired with another known payload in one of a physical downlink shared channel (PDSCH) or a physical uplink shared channel (PUSCH).

6. The method according to claim 1, wherein: The known payload is based on a radio resource control (RRC) scrambling seed.

7. The method according to claim 1, wherein: The request occurs via a physical uplink control channel (PUCCH) or a medium access control layer (MAC) control element (CE) for a UE, or via a physical downlink control channel (PDCCH) or MAC-CE for a base station.

8. The method according to claim 1, wherein: The known payload is received in a physical uplink control channel (PUCCH) with relevant parameters separately signaled via radio resource control (RRC) signaling, downlink control information (DCI) or medium access control layer (MAC) control element (CE), and the relevant parameters include at least one of a PUCCH format, a multiplexed uplink control information (UCI) type or a payload size.

9. A method of wireless communication performed by a transmitting device, comprising: communicating with a receiving device the contents of a known payload; sending an indication notifying a receiving device that a known payload will be sent; as well as unicasting a known payload to a receiving device for online training of a neural network, wherein the unicast known payload has characteristics specific to the receiving device, the characteristics including at least one of a modulation and coding scheme (MCS), a beam pair, or a rank, and the training is based on the characteristics, The method also includes sending information including the periodicity, time and frequency resources, and payload size of the known payload, the information being carried by radio resource control (RRC) signaling, a medium access control layer (MAC) control element, or PDCCH downlink control information (DCI), wherein the periodicity indicates whether the known payload is periodic, non-periodic, or semi-persistent.

10. The method according to claim 9, wherein: The indication includes a base station indication in a physical downlink control channel (PDCCH) or a medium access control (MAC) layer control element (MAC-CE) in response to a known payload being transmitted in a physical downlink shared channel (PDSCH) for downlink training.

11. The method according to claim 9, wherein: The indication includes a user equipment (UE) indication in a physical uplink control channel (PUCCH) or a medium access control (MAC) layer control element (MAC-CE) in response to a known payload being transmitted in a physical uplink shared channel (PUSCH) for uplink training.

12. The method according to claim 9, wherein: The known payload is based on a radio resource control (RRC) scrambling seed.

13. A receiving device for wireless communication, comprising: Memory; as well as at least one processor coupled to the memory and configured to: communicating with the sending device regarding the contents of the known payload; requesting a known payload for training an artificial neural network from a transmitting device, wherein the request describes a characteristic for the known payload, the characteristic comprising at least one of a modulation and coding scheme (MCS), a beam pair, or a rank, the training being based on the characteristic; receiving a known payload in response to the request; and performing online training of an artificial neural network using the known payload, The at least one processor is further configured to receive information including periodicity, time / frequency, and payload size of a known payload, the information being carried by at least one of radio resource control (RRC) signaling, a medium access control layer (MAC) control element, or PDCCH downlink control information (DCI), wherein the periodicity indicates whether the known payload is periodic, aperiodic, or semi-persistent.

14. The receiving device according to claim 13, wherein: The known payload is received in at least one of a physical downlink shared channel (PDSCH) or a physical uplink shared channel (PUSCH).

15. The receiving device according to claim 13, wherein: The request for the known payload originates from a user equipment (UE) for downlink training or from a base station for uplink training.

16. The receiving device according to claim 13, wherein: The known payload is received in a physical downlink control channel (PDCCH) and is paired with another known payload in a physical downlink shared channel (PDSCH) or a physical uplink shared channel (PUSCH).

17. The receiving device according to claim 13, wherein: The known payload is based on a radio resource control (RRC) scrambling seed.

18. The receiving device according to claim 13, wherein: The at least one processor is further configured to make a request via a physical uplink control channel (PUCCH) or a medium access control layer (MAC) control element (CE) for a UE, or via a physical downlink control channel (PDCCH) or MAC-CE for a base station.

19. The receiving device according to claim 13, wherein: The known payload is received in a physical uplink control channel (PUCCH) with relevant parameters separately signaled via radio resource control (RRC) signaling, downlink control information (DCI) or medium access control layer (MAC) control element (CE), and the relevant parameters include at least one of a PUCCH format, a multiplexed uplink control information (UCI) type or a payload size.

20. A transmitting device for wireless communication, comprising: Memory; as well as at least one processor coupled to the memory and configured to: communicating with a receiving device the contents of a known payload; sending an indication notifying a receiving device that a known payload will be sent; and unicasting a known payload to a receiving device for online training of a neural network, wherein the unicast known payload has characteristics specific to the receiving device, the characteristics including at least one of a modulation and coding scheme (MCS), a beam pair, or a rank, and the training is based on the characteristics, The at least one processor is further configured to send information including the periodicity, time and frequency resources, and payload size of the known payload, the information being carried by radio resource control (RRC) signaling, a medium access control layer (MAC) control element, or PDCCH downlink control information (DCI), wherein the periodicity indicates whether the known payload is periodic, non-periodic, or semi-persistent.

21. The transmitting device according to claim 20, wherein: The indication includes a base station indication in a physical downlink control channel (PDCCH) or a medium access control (MAC) layer control element (MAC-CE) in response to a known payload being transmitted in a physical downlink shared channel (PDSCH) for downlink training.

22. The transmitting device according to claim 20, wherein: The indication includes a user equipment (UE) indication in a physical uplink control channel (PUCCH) or a medium access control (MAC) layer control element (MAC-CE) in response to a known payload being transmitted in a physical uplink shared channel (PUSCH) for uplink training.

23. The transmitting device according to claim 20, wherein: The known payload is based on a radio resource control (RRC) scrambling seed.

24. The transmitting device according to claim 20, wherein: The at least one processor is further configured to make a request via a physical uplink control channel (PUCCH) or a medium access control layer (MAC) control element (CE) for a UE, or via a physical downlink control channel (PDCCH) or MAC-CE for a base station.

25. A receiving device for wireless communication, comprising: means for communicating with a sending device regarding the contents of a known payload; means for requesting a known payload for training an artificial neural network from a transmitting device, wherein the request describes characteristics for the known payload, the characteristics comprising at least one of a modulation and coding scheme (MCS), a beam pair, or a rank, the training being based on the characteristics; means for receiving a known payload in response to the request; and means for performing online training of an artificial neural network using said known payload, The receiving device further comprises: A component for receiving information including periodicity, time and frequency resources, and payload size of the known payload, the information being carried by at least one of radio resource control (RRC) signaling, a medium access control layer (MAC) control element, or PDCCH downlink control information (DCI), wherein the periodicity indicates whether the known payload is periodic, aperiodic, or semi-persistent.

26. A non-transitory computer-readable medium storing one or more instructions for wireless communication, the one or more instructions, when executed by one or more processors of a receiving device, causing the one or more processors to: communicating with the sending device regarding the contents of the known payload; Request a known payload from the sending device for training an artificial neural network, where The request describes a characteristic for the known payload, the characteristic comprising at least one of a modulation and coding scheme (MCS), a beam pair, or a rank, the training being based on the characteristic; receiving a known payload in response to the request; as well as Perform online training of an artificial neural network using the known payload, wherein the one or more instructions further cause the one or more processors to receive information including periodicity, time and frequency resources, and payload size of the known payload, the information being carried by at least one of radio resource control (RRC) signaling, medium access control layer (MAC) control elements, or PDCCH downlink control information (DCI), wherein the periodicity indicates whether the known payload is periodic, aperiodic, or semi-persistent.

27. A transmitting device for wireless communication, comprising: means for communicating with a receiving device that the contents of the known payload; means for sending an indication notifying a receiving device that a known payload will be sent; as well as means for unicasting a known payload to a receiving device for online training of a neural network, wherein the unicast known payload has characteristics specific to the receiving device, the characteristics including at least one of a modulation and coding scheme (MCS), a beam pair, or a rank, the training being based on the characteristics, Wherein, the sending device also includes: A component for sending information including the periodicity, time and frequency resources, and payload size of the known payload, the information being carried by radio resource control (RRC) signaling, a medium access control layer (MAC) control element, or PDCCH downlink control information (DCI), wherein the periodicity indicates whether the known payload is periodic, aperiodic, or semi-persistent.

28. A non-transitory computer-readable medium storing one or more instructions for wireless communication, which, when executed by one or more processors of a transmitting device, cause the one or more processors to: communicating with a receiving device the contents of a known payload; sending an indication notifying a receiving device that a known payload will be sent; and unicasting a known payload to a receiving device for online training of a neural network, wherein the unicast known payload has characteristics specific to the receiving device, the characteristics including at least one of a modulation and coding scheme (MCS), a beam pair, or a rank, and the training is based on the characteristics, in, The one or more instructions also cause the one or more processors to send information including the periodicity, time and frequency resources, and payload size of the known payload, the information being carried by radio resource control (RRC) signaling, a medium access control layer (MAC) control element, or PDCCH downlink control information (DCI), wherein the periodicity indicates whether the known payload is periodic, non-periodic, or semi-persistent.

Images