Semantic decoding using ai / ML model

Abstract

The present disclosure describes a novel method of incorporating semantic awareness at the early stages of the decoding process in wireless mobile communication system including base station (e.g., gNB, TRP, TN, NTN) and mobile station (e.g., UE). The system enhances efficiency and effectiveness of semantic data transmission by selecting appropriate decoding. With semantic data transmission applied to radio access network, it is necessary to adapt transmission strategies based on the actual needs of the receiving application. Therefore, semantic data operation can be set up between network and UE by providing adaptive AI / ML model selection.

Description

[0001] 202407117

[0002] - 1 -

[0003] TITLE

[0004] Method of advanced adaptive processing for radio access network

[0005] TECHNNICAL FIELD

[0006] The present disclosure relates to the field of semantic data communication with adaptive AI / ML model selection for semantic pre-decoding, where techniques for incorporating semantic awareness at the early stages of the decoding process applicable to radio access network are presented.

[0007] BACKGROUND

[0008] In 3GPP (Third Generation Partnership Project), one of the selected study items as the approved Release 18 package of 5G Advanced is AI / ML (artificial intelligence / machine learning) as described in the related document (RP-213599) addressed in 3GPP TSG (Technical Specification Group) RAN (Radio Access Network) meeting #94e. The official title of AI / ML study item is “Study on AI / ML for NR Air Interface”. The goal of this study item is to identify a common AI / ML framework and areas of obtaining gains using AI / ML based techniques with use cases. According to 3GPP, the main objective of this study item is to study AI / ML framework for air-interface with target use cases by considering performance, complexity, and potential specification impact. In particular, AI / ML model, terminology and description to identify common and specific characteristics for framework are included as one of key work scopes. Regarding AI / ML framework, various aspects are under consideration for investigation and one of key items is about lifecycle management of AI / ML model where multiple stages are included as mandatory for model training, model deployment, model inference, model monitoring, model updating etc. Also in 3GPP, two-sided (AI / ML) model is defined as a paired AI / ML model(s) over which joint inference is performed, where joint inference comprises AI / ML Inference whose inference is performed jointly across the UE and the network. Also for one-sided (AI / ML) model, UE-side (AI / ML) model is defined as an AI / ML model whose inference is performed entirely at the UE and network-side (AI / ML) model is defined as an AI / ML model whose inference is performed entirely at the network. Currently, AI / ML specification work is at the stage of work item discussion for Release 19. Earlier, in 3GPP TR 37.817 for Release 17, titled as Study on 202407117

[0009] - 2 - enhancement for Data Collection for NR and EN-DC, UE (user equipment) mobility was also considered as one of AI / ML use cases and one of scenarios for model training / inference is that both functions are located within RAN node. Followingly, in Release 18 the new work item of “Artificial Intelligence (Al) / Machine Learning (ML) for NG-RAN” was initiated to specify data collection enhancements and signaling support within existing NG-RAN interfaces and architecture. L1 / L2 signaling refers to the control information carried on physical layer and MAC (medium access control) layer, respectively, while RRC (radio resource control) signaling is used for higher- level control of the radio connection. For the above active standardization works, RAN-based AI / ML model is considered very significant for both network and UE to meet any desired model operations (e.g., model training, inference, selection, switching, update, monitoring, etc.). Model information can be signaled to pair both network-side and UE-side models for various lifecycle management (LCM) operations.

[0010] However, signaling overhead indicating model information can be very high especially when model based LCM is processed between base station (BS / gNB) and multiple UEs. In LCM, model training is one of the most important parts for model deployment and currently there is no specification defined for signaling methods and network-UE behaviors so as to identify the required dataset when model updating / re- training as any activated model can be also impacted due to model / data drift. When ML condition changes, the enabled AI / ML model(s) can be impacted for model performance due to data / model drift. In this case, model re-training / updating can be executed. Along with 5G Advanced AI / ML standardization activities as described above, 6G study is scheduled to start in the upcoming release stage as the evolution of wireless communication technologies towards 6G has led to an unprecedented increase in data traffic and diversity of applications across RANs. Traditional communication paradigms, which focus on transmitting raw data without considering its semantic content or importance, are becoming increasingly inefficient in meeting the demands of emerging applications such as autonomous vehicles, extended reality (XR), and industrial Internet of Things (loT). As 6G networks aim to support these diverse applications with stringent requirements, there is a growing need for more efficient and context-aware communication systems. Semantic communication 202407117

[0011] - 3 - has emerged as a promising approach to address these challenges by focusing on the meaning and importance of information rather than raw bit transmission. This paradigm shift aims to improve spectral efficiency, reduce latency, and enhance overall system performance. Along with the recent advances in 3GPP, AI / ML technologies have notably expanded device intelligence, fostering federation and cooperation among distributed AI / ML entities. These advancements impose new requirements on future 6G mobile network architectures, necessitating the integration of communication, computation, control, and intelligence. The concept of Al-native 6G systems, specifically tailored for semantic and goal-oriented communications, has gained traction. These systems aim to go beyond the established limits of current sense-compute-connect-control models and transition toward semantic communication-based Al architectures, protocols, and services.

[0012] However, current 3GPP specifications lack mechanisms to differentiate between data types based on their semantic importance, leading to inefficient use of network resources. Moreover, existing error correction and retransmission schemes do not consider the semantic nature of the data, resulting in unnecessary retransmissions and increased latency. To address these challenges, novel approaches such as goal- oriented semantic communication frameworks are being explored. These frameworks incorporate both semantic and effectiveness levels for various tasks with diverse data types, aiming to facilitate information exchange between intelligent agents in a more relevant, effective, and timely manner. The development of semantic communication protocols for 6G involves various levels of sophistication, from task-oriented neural protocols to language-oriented semantic protocols harnessing large language models (LLMs) and generative models. This evolution in protocol design aims to address the non-stationary tasks expected in 6G systems and offer the ability to tailor signaling messages for specific tasks. In aspect of these advancements and challenges, there is a significant need for innovative methods and systems that can enable efficient semantic data transmission in 6G wireless networks. These solutions should incorporate AI / ML techniques to enhance semantic understanding, optimize resource allocation, and improve overall system performance while meeting the diverse requirements of emerging applications. 202407117

[0013] - 4 -

[0014] With the evolution of wireless systems toward 5G-Advanced and 6G, semantic communication has emerged as a key paradigm to improve efficiency by transmitting meaning rather than raw data. Current 3GPP specifications for NR do not yet define mechanisms for semantic-aware processing or AI / ML-driven decoder selection. At the same time, 3GPP Release 18 and 19 have introduced AI / ML frameworks for RAN optimization, including model identification and lifecycle management, which are expected to become foundational for Al-native 6G architectures. However, existing solutions lack a standardized approach to dynamically select AI / ML models for semantic decoding based on context and importance of information. This gap creates challenges in achieving low latency, interoperability, and efficient resource allocation. The present invention addresses these challenges by introducing semantic quotientbased adaptive model selection, mapping relation tables for standardized signaling, and integration with established 3GPP procedures, thereby enabling intent-aware and context-driven radio operation.

[0015] US2023216932A1 relates to a method for filtering data traffic based on user information and areas of interest by optimizing how data is handled in communication systems by focusing on user preferences and specific data needs, ensuring that only relevant traffic is processed or transmitted.

[0016] US2023199746A1 describes a method aimed at enhancing communication systems, particularly in data processing and transmission by optimizing the flow of data across networks, ensuring more efficient handling of complex data interactions.

[0017] US2023291497A1 describes a method or system related to optimizing data processing and transmission in communication networks by improving how devices, particularly in wireless communication environments, handle complex data flows.

[0018] US2023412709A1 involves a method focused on optimizing data transmission and communication efficiency by improving how devices manage, transmit, and process data, potentially using techniques like enhanced encoding or optimized data flow to handle complex data interactions. 202407117

[0019] - 5 -

[0020] WO20231 13302A1 focuses on advancements in semantic communication systems designed to enhance transmission efficiency for transmitting semantic-related information instead of raw data, reducing the amount of data sent over communication networks.

[0021] WO2024038926A1 focuses on a method and device for optimizing wireless communication transmission, efficiently transmitting data in communication systems by improving the way signals are structured and sent between devices.

[0022] BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 is an exemplary table of mapping relation information about semantic decoding.

[0024] Figure 2 is an exemplary flow chart of enabling semantic pre-decoding at network side.

[0025] Figure 3 is an exemplary flow chart of processing semantic pre-decoding at UE side.

[0026] Figure 4 is an exemplary flow chart of the overall processing for semantic predecoding.

[0027] Figure 5 is an exemplary signaling flow of semantic pre-decoding.

[0028] DETAILED DESCRIPTION

[0029] The detailed description set forth below, with reference to annexed drawings, is intended as a description of various configurations and is not intended to represent the only configurations in which the concepts described herein may be practiced. The detailed description includes specific details for the purpose of providing a thorough understanding of the various concepts. However, it will be apparent to those skilled in 202407117

[0030] - 6 - the art that these concepts may be practiced without these specific details. In particular, although terminology from 3GPP 5G NR may be used in this disclosure to exemplify embodiments herein, this should not be seen as limiting the scope of the invention.

[0031] Some of the embodiments contemplated herein will now be described more fully with reference to the accompanying drawings. Other embodiments, however, are contained within the scope of the subject matter disclosed herein, the disclosed subject matter should not be construed as limited to only the embodiments set forth herein; rather, these embodiments are provided by way of example to convey the scope of the subject matter to those skilled in the art.

[0032] Generally, all terms used herein are to be interpreted according to their ordinary meaning in the relevant technical field, unless a different meaning is clearly given and / or is implied from the context in which it is used. All references to a / an / the element, apparatus, component, means, step, etc. are to be interpreted openly as referring to at least one instance of the element, apparatus, component, means, step, etc., unless explicitly stated otherwise. The steps of any methods disclosed herein do not have to be performed in the exact order disclosed, unless a step is explicitly described as following or preceding another step and / or where it is implicit that a step must follow or precede another step. Any feature of any of the embodiments disclosed herein may be applied to any other embodiment, wherever appropriate. Likewise, any advantage of any of the embodiments may apply to any other embodiments, and vice versa. Other objectives, features and advantages of the enclosed embodiments will be apparent from the following description.

[0033] In some embodiments, a more general term “network node” may be used and may correspond to any type of radio network node or any network node, which communicates with a UE (directly or via another node) and / or with another network node. Examples of network nodes are NodeB, MeNB, ENB, a network node belonging to MCG or SCG, base station (BS), multi-standard radio (MSR) radio node such as MSR BS, eNodeB, gNodeB, network controller, radio network controller (RNC), base station controller (BSC), relay, donor node controlling relay, base 202407117

[0034] - 7 - transceiver station (BTS), access point (AP), transmission points, transmission nodes, RRU, RRH, nodes in distributed antenna system (DAS), core network node (e.g. Mobile Switching Center (MSC), Mobility Management Entity (MME), etc.), Operations & Maintenance (O&M), Operations Support System (OSS), Self Optimized Network (SON), positioning node (e.g. Evolved- Serving Mobile Location Centre (E-SMLC)), Minimization of Drive Tests (MDT), test equipment (physical node or software), etc.

[0035] In some embodiments, the non-limiting term user equipment (UE) or wireless device may be used and may refer to any type of wireless device communicating with a network node and / or with another UE in a cellular or mobile communication system. Examples of UE are target device, device to device (D2D) UE, machine type UE or UE capable of machine to machine (M2M) communication, PDA, PAD, Tablet, mobile terminals, smart phone, laptop embedded equipped (LEE), laptop mounted equipment (LME), USB dongles, UE category Ml, UE category M2, ProSe UE, V2V UE, V2X UE, etc.

[0036] Additionally, terminologies such as base station / gNodeB and UE should be considered non-limiting and do in particular not imply a certain hierarchical relation between the two; in general, “gNodeB” could be considered as device 1 and “UE” could be considered as device 2 and these two devices communicate with each other over some radio channel. And in the following the transmitter or receiver could be either gNodeB (gNB), or UE.

[0037] As will be appreciated by one skilled in the art, aspects of the embodiments may be embodied as a system, apparatus, method, or program product. Accordingly, embodiments may take the form of an entirely hardware embodiment, an entirely software embodiment (including firmware, resident software, micro-code, etc.) or an embodiment combining software and hardware aspects.

[0038] For example, the disclosed embodiments may be implemented as a hardware circuit comprising custom very-large-scale integration (“VLSI”) circuits or gate arrays, off- the-shelf semiconductors such as logic chips, transistors, or other discrete 202407117

[0039] - 8 - components. The disclosed embodiments may also be implemented in programmable hardware devices such as field programmable gate arrays, programmable array logic, programmable logic devices, or the like. As another example, the disclosed embodiments may include one or more physical or logical blocks of executable code which may, for instance, be organized as an object, procedure, or function.

[0040] Furthermore, embodiments may take the form of a program product embodied in one or more computer readable storage devices storing machine readable code, computer readable code, and / or program code, referred hereafter as code. The storage devices may be tangible, non- transitory, and / or non-transmission. The storage devices may not embody signals. In a certain embodiment, the storage devices only employ signals for accessing code.

[0041] Any combination of one or more computer readable medium may be utilized. The computer readable medium may be a computer readable storage medium. The computer readable storage medium may be a storage device storing the code. The storage device may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, holographic, micromechanical, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing.

[0042] More specific examples (a non-exhaustive list) of the storage device would include the following: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random-access memory (“RAM”), a read-only memory (“ROM”), an erasable programmable read-only memory (“EPROM” or Flash memory), a portable compact disc readonly memory (“CD-ROM”), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a computer readable storage medium may be any tangible medium that can contain or store a program for use by or in connection with an instruction execution system, apparatus, or device.

[0043] Code for carrying out operations for embodiments may be any number of lines and may be written in any combination of one or more programming languages including 202407117

[0044] - 9 - an object- oriented programming language such as Python, Ruby, Java, Smalltalk, C++, or the like, and conventional procedural programming languages, such as the “C” programming language, or the like, and / or machine languages such as assembly languages. The code may execute entirely on the user’s computer, partly on the user’s computer, as a stand-alone software package, partly on the user’s computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer may be connected to the user’s computer through any type of network, including a local area network (“LAN”), wireless LAN (“WLAN”), or a wide area network (“WAN”), or the connection may be made to an external computer (for example, through the Internet using an Internet Service Provider (“ISP”)).

[0045] Furthermore, the described features, structures, or characteristics of the embodiments may be combined in any suitable manner. In the following description, numerous specific details are provided, such as examples of programming, software modules, user selections, network transactions, database queries, database structures, hardware modules, hardware circuits, hardware chips, etc., to provide a thorough understanding of embodiments. One skilled in the relevant art will recognize, however, that embodiments may be practiced without one or more of the specific details, or with other methods, components, materials, and so forth. In other instances, well-known structures, materials, or operations are not shown or described in detail to avoid obscuring aspects of an embodiment. Reference throughout this specification to “one embodiment,” “an embodiment,” or similar language means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment. Thus, appearances of the phrases “in one embodiment,” “in an embodiment,” and similar language throughout this specification may, but do not necessarily, all refer to the same embodiment, but mean “one or more but not all embodiments” unless expressly specified otherwise. The terms “including,” “comprising,” “having,” and variations thereof mean “including but not limited to,” unless expressly specified otherwise. An enumerated listing of items does not imply that any or all of the items are mutually exclusive, unless expressly specified otherwise. The terms “a,” “an,” and “the” also refer to “one or more” unless expressly specified otherwise. 202407117

[0046] - 10 -

[0047] Aspects of the embodiments are described below with reference to schematic flowchart diagrams and / or schematic block diagrams of methods, apparatuses, systems, and program products according to embodiments. It will be understood that each block of the schematic flowchart diagrams and / or schematic block diagrams, and combinations of blocks in the schematic flowchart diagrams and / or schematic block diagrams, can be implemented by code. This code may be provided to a processor of a general-purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions / acts specified in the flowchart diagrams and / or block diagrams.

[0048] The code may also be stored in a storage device that can direct a computer, other programmable data processing apparatus, or other devices to function in a particular manner, such that the instructions stored in the storage device produce an article of manufacture including instructions which implement the function / act specified in the flowchart diagrams and / or block diagrams.

[0049] The code may also be loaded onto a computer, other programmable data processing apparatus, or other devices to cause a series of operational steps to be performed on the computer, other programmable apparatus, or other devices to produce a computer implemented process such that the code which execute on the computer or other programmable apparatus provide processes for implementing the functions / acts specified in the flowchart diagrams and / or block diagrams.

[0050] The flowchart diagrams and / or block diagrams in the Figures illustrate the architecture, functionality, and operation of possible implementations of apparatuses, systems, methods, and program products according to various embodiments. In this regard, each block in the flowchart diagrams and / or block diagrams may represent a module, segment, or portion of code, which includes one or more executable instructions of the code for implementing the specified logical function(s). 202407117

[0051] - 11 -

[0052] It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the Figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. Other steps and methods may be conceived that are equivalent in function, logic, or effect to one or more blocks, or portions thereof, of the illustrated figures.

[0053] Although various arrow types and line types may be employed in the flowchart and / or block diagrams, they are understood not to limit the scope of the corresponding embodiments. Indeed, some arrows or other connectors may be used to indicate only the logical flow of the depicted embodiment. For instance, an arrow may indicate a waiting or monitoring period of unspecified duration between enumerated steps of the depicted embodiment. It will also be noted that each block of the block diagrams and / or flowchart diagrams, and combinations of blocks in the block diagrams and / or flowchart diagrams, can be implemented by special purpose hardware-based systems that perform the specified functions or acts, or combinations of special purpose hardware and code.

[0054] The description of elements in each figure may refer to elements of proceeding figures. Like numbers refer to like elements in all figures, including alternate embodiments of like elements.

[0055] The detailed description set forth below, with reference to the figures, is intended as a description of various configurations and is not intended to represent the only configurations in which the concepts described herein may be practiced. The detailed description includes specific details for the purpose of providing a thorough understanding of the various concepts. However, it will be apparent to those skilled in the art that these concepts may be practiced without these specific details. For instance, although 3GPP terminology, from e.g., 5G NR, may be used in this disclosure to exemplify embodiments herein, this should not be seen as limiting the scope of the present disclosure. 202407117

[0056] - 12 -

[0057] The disclosure is related to wireless communication system, which may be for example a 5G NR wireless communication system. More specifically, it represents a RAN of the wireless communication system, which is used exchange data with UEs via radio signals. For example, the RAN may send data to the UEs (downlink, DL), for instance data received from a core network (CN). The RAN may also receive data from the UEs (uplink, UL), which data may be forwarded to the CN.

[0058] In the examples illustrated, the RAN comprises one base station, BS. Of course, the RAN may comprise more than one BS to increase the coverage of the wireless communication system. Each of these BSs may be referred to as NB, eNodeB (or eNB), gNodeB (or gNB, in the case of a 5G NR wireless communication system), an access point or the like, depending on the wireless communication standard(s) implemented.

[0059] The UEs are located in a coverage of the BS. The coverage of the BS corresponds for example to the area in which UEs can decode a PDCCH transmitted by the BS.

[0060] An example of a wireless device suitable for implementing any method, discussed in the present disclosure, performed at a UE corresponds to an apparatus that provides wireless connectivity with the RAN of the wireless communication system, and that can be used to exchange data with said RAN. Such a wireless device may be included in a UE. The UE may for instance be a cellular phone, a wireless modem, a wireless communication device, a handheld device, a laptop computer, or the like. The UE may also be an Internet of Things (loT) equipment, like a wireless camera, a smart sensor, a smart meter, smart glasses, a vehicle (manned or unmanned), a global positioning system device, etc., or any other equipment that may run applications that need to exchange data with remote recipients, via the wireless device.

[0061] The wireless device comprises one or more processors and one or more memories. The one or more processors may include for instance a central processing unit (CPU), a digital signal processor (DSP), a field-programmable gate array (FPGA), an 202407117

[0062] - 13 - application specific integrated circuit (ASIC), etc. The one or more memories may include any type of computer readable volatile and non-volatile memories (magnetic hard disk, solid-state disk, optical disk, electronic memory, etc.). The one or more memories may store a computer program product, in the form of a set of programcode instructions to be executed by the one or more processors to implement all or part of the steps of a method for exchanging data, performed at a UE’s side, according to any one of the embodiments disclosed herein.

[0063] The wireless device can comprise also a main radio, MR, unit. The MR unit corresponds to a main wireless communication unit of the wireless device, used for exchanging data with BSs of the RAN using radio signals. The MR unit may implement one or more wireless communication protocols, and may for instance be a 3G, 4G, 5G, NR, WiFi, WiMax, etc. transceiver or the like. In preferred embodiments, the MR unit corresponds to a 5G NR wireless communication unit.

[0064] AI / ML Model is a data driven algorithm that applies AI / ML techniques to generate set of outputs based on set of inputs.

[0065] AI / ML model delivery is a generic term referring to delivery of an AI / ML model from one entity to another entity in any manner. Note is An entity could mean network node / function (e.g., gNB, LMF, etc.), UE, proprietary server, etc.

[0066] AI / ML model Inference is a process of using trained AI / ML model to produce set of outputs based on set of inputs.

[0067] AI / ML model testing is a subprocess of training, to evaluate the performance of final AI / ML model using dataset different from one used for model training and validation. Differently from AI / ML model validation, testing does not assume subsequent tuning of the model.

[0068] AI / ML model training is a process to train an AI / ML Model [by learning the input / output relationship] in data driven manner and obtain the trained AI / ML Model for inference. 202407117

[0069] - 14 -

[0070] AI / ML model transfer is a delivery of an AI / ML model over the air interface in manner that is not transparent to 3GPP signaling, either parameters of model structure known at the receiving end or new model with parameters. Delivery may contain full model or partial model.

[0071] AI / ML model validation is a subprocess of training, to evaluate the quality of an AI / ML model using dataset different from one used for model training, that helps selecting model parameters that generalize beyond the dataset used for model training.

[0072] Data collection is a process of collecting data by the network nodes, management entity, or UE for the purpose of AI / ML model training, data analytics and inference.

[0073] Federated learning I federated training is a machine learning technique that trains an AI / ML model across multiple decentralized edge nodes e.g., UEs, gNBs each performing local model training using local data samples. The technique requires multiple interactions of the model, but no exchange of local data samples.

[0074] Functionality identification is a process / method of identifying an AI / ML functionality for the common understanding between the NW and the UE. Note is Information regarding the AI / ML functionality may be shared during functionality identification. Where AI / ML functionality resides depends on the specific use cases and sub use cases.

[0075] Model activation means enable an AI / ML model for specific AI / ML-enabled feature.

[0076] Model deactivation means disable an AI / ML model for specific AI / ML-enabled feature.

[0077] Model download means Model transfer from the network to UE.

[0078] Model identification is A process / method of identifying an AI / ML model for the common understanding between the NW and the UE. The process / method of model 202407117

[0079] - 15 - identification may or may not be applicable and regarding the AI / ML model may be shared during model identification.

[0080] Model monitoring is A procedure that monitors the inference performance of the AI / ML model.

[0081] Model parameter update is Process of updating the model parameters of model. Model selection is the process of selecting an AI / ML model for activation among multiple models for the same AI / ML enabled feature. Model selection may or may not be carried out simultaneously with model activation.

[0082] Model switching is deactivating currently active AI / ML model and activating different AI / ML model for specific AI / ML-enabled feature.

[0083] Model update is Process of updating the model parameters and / or model structure of model.

[0084] Model upload is Model transfer from UE to the network.

[0085] Network-side (AI / ML) model is an AI / ML Model whose inference is performed entirely at the network.

[0086] Offline field data is the data collected from field and used for offline training of the AI / ML model.

[0087] Offline training is an AI / ML training process where the model is trained based on collected dataset, and where the trained model is later used or delivered for inference. Note is This definition only serves as guidance. There may be cases that may not exactly conform to this definition but could still be categorized as offline training by commonly accepted conventions.

[0088] Online field data is the data collected from field and used for online training of the AI / ML model. 202407117

[0089] - 16 -

[0090] Online training is an AI / ML training process where the model being used for inference) is (typically continuously) trained in (near) real-time with the arrival of new training samples. Note is the notion of (near) real-time vs. non real-time is context- dependent and is relative to the inference time-scale. This definition only serves as guidance.

[0091] There may be cases that may not exactly conform to this definition but could still be categorized as online training by commonly accepted conventions. Note is Fine- tuning / re-training may be done via online or offline training. This note could be removed when we define the term fine-tuning.

[0092] Reinforcement Learning (RL) is a process of training an AI / ML model from input (a.k.a. state) and feedback signal (a.k.a. reward) resulting from the model’s output (a.k.a. action) in an environment the model is interacting with.

[0093] Semi-supervised learning is a process of training model with mix of labelled data and unlabeled data.

[0094] Supervised learning is a process of training model from input and its corresponding labels.

[0095] Two-sided (AI / ML) model is a paired AI / ML Model(s) over which joint inference is performed, where joint inference comprises AI / ML Inference whose inference is performed jointly across the UE and the network, i.e, the first part of inference is firstly performed by UE and then the remaining part is performed by gNB, or vice versa.

[0096] UE-side (AI / ML) model is an AI / ML Model whose inference is performed entirely at the UE.

[0097] Unsupervised learning is a process of training model without labelled data. 202407117

[0098] - 17 -

[0099] Proprietary-format models is ML models of vendor-Zdevice-specific proprietary format, from 3GPP perspective. They are not mutually recognizable across vendors and hide model design information from other vendors when shared.

[0100] Open-format models is ML models of specified format that are mutually recognizable across vendors and allow interoperability, from 3GPP perspective. They are mutually recognizable between vendors and do not hide model design information from other vendors when shared.

[0101] The present invention relates to a novel method and system for semantic-aware data transmission in 6G networks, specifically focusing on transmission data types and retransmission request types. AI / ML based techniques are currently applied to many different applications and 3GPP also started to work on its technical investigation to apply to multiple use cases based on the observed potential gains. AI / ML lifecycle can be split into several stages such as data collection / pre-processing, model training, model testing / validation, model deployment / update, model monitoring etc., where each stage is equally important to achieve target performance with any specific model(s). In applying AI / ML model for any use case or application, one of the challenging issues is to manage the lifecycle of AI / ML model. It is mainly because the data / model drift occurs during model deployment / inference and it results in performance degradation of AI / ML model. Fundamentally, the dataset statistical changes occur after model is deployed and model inference capability is also impacted with unseen data as input. In a similar aspect, the statistical property of dataset and the relationship between input and output for the trained model can be changed with drift occurrence.

[0102] In this context, model training or re-training is one of key issues for model performance maintenance as model performance such as inferencing and / or training is dependent on different model execution environment with varying configuration parameters. To handle this issue, collaboration between UE and gNB is highly important to track model performance and re-configure model corresponding to different environments. AI / ML model needs model monitoring after deployment because model performance cannot be maintained continuously due to drift and 202407117

[0103] - 18 - update feedback is then provided to re-train / update the model or select alternative model. When AI / ML model enabled wireless communication network is deployed, it is then important to consider how to handle AI / ML model in activation with reconfiguration for wireless devices under operations such as model training, inference, updating, etc.

[0104] For determining information about model identification (e.g., model ID), frequent model ID assignment / re-assignment processes might occur due to model drift related to model performance variation and / or model applicable condition change. In addition, if all identified models at UE side are to be assessed and monitored, the associated signaling overhead and computing power demand can increase significantly. And UEs often have a list of active / inactive ML models, each assigned a unique identifier. And the network and UEs may have to take frequent signaling exchanges to determine the related configuration. This can lead to increased signaling overhead and reduced radio resource efficiency. For example, more than one ML models are in active between gNB and UE where multiple UEs have their own on-device models in operation. In addition, each models might be in different LCM phases such as data collection, training, inferencing or monitoring, resulting in high signaling overhead for model-related resource transfers (e.g., datasets, model parameters, architectures, hyperparameters) via L1 / L2 or RRC signaling where L1 / L2 signaling refers to the control information carried on physical layer and MAC (medium access control) layer such as MAC CE (control element), respectively, while RRC (radio resource control) signaling is used for higher-level control of the radio connection. This invention aims to enhance the efficiency, reliability, and adaptability of data transmission in next-generation wireless communication systems by providing a comprehensive framework for semantic-aware data transmission, addressing the challenges of efficient resource utilization, reduced latency, and improved accuracy in semantic data communication.

[0105] In this method, adaptive AI / ML model selection for semantic pre-decoding is configured such that key semantic elements from the encoded semantic data before full decoding are identified and different semantic quotients are assigned to those key 202407117

[0106] - 19 - semantic elements where semantic quotient represents the degree to which a piece of semantic information aligns with and impacts the semantic decoding processes.

[0107] For example, a high semantic quotient indicates semantic information that is highly relevant, easily understood, and significantly impactful to the decoding process. The most appropriate AI / ML model is then selected to process semantic decoding of each key semantic elements or group of semantic elements. For example, semantic elements with high semantic quotient are routed to more complex, full semantic decoders. And semantic elements with low semantic quotient may be processed by simpler, faster decoders. Different semantic elements based on their quotients are allowed to be decoded simultaneously if applicable. Two or more AI / ML models for semantic decoding are also allowed to be activated selectively so as to maintain a buffer of recent semantic context to aid in the interpretation of new semantic data. Feedback message based on semantic decoding output is used to refine the reordering of quotients of key semantic elements as well as to identify the updated key semantic elements. The above overall flow of semantic data processing is defined as semantic pre-decoding. In addition, mapping relation information having semantic identifiers is configured to support adaptive AI / ML model selection for semantic predecoding. Specifically, the initial configuration of the mapping relation information is delivered to the UE using RRC signaling. The RRC reconfiguration message is then used to send the full mapping relation information or updates to the UE. For dynamic updates or real-time selection of AI / ML models based on mapping relation information, L1 / L2 signaling is used such that a new MAC CE is defined to indicate which AI / ML model to use for semantic decoding. In implementation aspect, the base mapping relation information is pre-configured in the UE as part of the device firmware or software. Updates to the mapping relation information (e.g., based on network conditions and semantic communication requirements) is configured at the network side and delivered to the UE using L1 / L2 or RRC signaling. The elements of mapping relation information (e.g., look-up table) include AI / ML model ID, semantic application domain, semantic feature and semantic quotient. Specifically, AI / ML model ID is a unique identifier for each model available at the UE side. Semantic application domain is the data type of specific application the model is designed to process. Semantic feature is a component of the meaning associated with semantic 202407117

[0108] - 20 - concept. Semantic quotient is the degree to which a piece of semantic information aligns with and impacts the semantic decoding processes.

[0109] The elements included in mapping relation table are allowed to be one-to-one or one- to-many / many-to-one or many-to-many mapping depending on its implementation scenarios. For example, the same AI / ML model ID is mapped onto two or more semantic application domain identifiers and / or semantic features. Mapping relation information consists of the varying combinations of identifiers for AI / ML model ID (e.g., ML_Model_ID), semantic application domain (e.g., Sem_AppDomain_ID), semantic feature (e.g., Sem_Feature_ID) and semantic quotient (e.g., Sem_Quot_ID) where the size of mapping relation table and the associated each specific identifier values of mapping relation elements need to be pre-configured according to different use cases and / or deployment scenarios. As the key semantic elements or features are identified at UE side, it can quickly look up the pre-configured mapping relation table(s). The table includes application domain to allow for context-specific quotient assignment, as the same feature might have different quotients in different applications. The quotient from the table is used to classify the semantic elements quickly and the semantic elements are routed to the most relevant decoder based on the associated AI / ML model ID indicated in the table. The table can be updated based on feedback from the semantic decoder output, allowing the system to learn and adapt over time. Regarding a new signaling or message to support adaptive AI / ML model selection for semantic pre-decoding, a new IE (information element) to indicate semantic communication capabilities is added to UE capability information message as a new RRC signaling where this new IE then contains supportability of semantic pre-decoding, semantic quotient, list of semantic decoders or AI / ML model IDs (e.g., with different complexities).

[0110] A new IE to configure semantic communication parameters for network is added to RRC reconfiguration message as a new RRC signaling for network where this new IE then contains (de-)activation of semantic pre-decoding, semantic quotient threshold, semantic mapping relation table identifier. A new MAC CE is added for dynamic semantic quotient indication that contains semantic quotient. Regarding key benefits of incorporating a mapping relation information (e.g., via look-up table), the pre- 202407117

[0111] - 21 - defined mappings speed up the process of identifying key semantic elements to be associated with the most appropriate AI / ML model of semantic decoding and the overall computational burden to execute the semantic pre-decoding is reduced by serving as a quick reference guide for the semantic pre-decoder, allowing it to make rapid decisions about semantic quotient and appropriate decoding paths, which is crucial for the low-latency requirements.

[0112] If semantic pre-decoding is enabled (that can be either indicated by network side or decided by UE autonomously). Selection of AI / ML model for decoding is either decided by network side or UE side autonomously. If network side decides selection of AI / ML model for decoding, a specific AI / ML model ID is indicated by network or model itself is transferred to UE. If UE side decides selection of AI / ML model for decoding, the pre-configured mapping relation is referenced.

[0113] The proposed semantic pre-decoding control plane (model IDs, LCM-aware control, and cross-node collaboration) are consistent with 6G AI / ML for NG-RAN and the broader applications from Al-coordinated to Al-native architectures. By standardizing capability exposure and configuration for semantic decoding, the feature enables intent- / task-aware operation.

[0114] This invention is designed to integrate seamlessly with the 3GPP NR architecture and aligns with key specifications for RRC signaling and for MAC CE, as well as AI / ML frameworks introduced in Release 18 and 19. The proposed semantic adaptive AI / ML model selection enables low-latency semantic communication by classifying semantic elements before full decoding and routing them to appropriate AI / ML decoders. A mapping relation table, identified by a SemanticMappingTablelD, provides standardized associations between Model ID, semantic application domain, semantic feature, and semantic quotient, ensuring interoperability and lifecycle management across multi-vendor deployments. Configuration and capability exposure are achieved through new RRC information elements, while dynamic updates for time-critical adjustments are supported via a newly defined MAC CE. 202407117

[0115] - 22 -

[0116] Figure 1 shows an exemplary table of mapping relation information about semantic decoding. In this example, regarding key benefits of incorporating a mapping relation information (e.g., via look-up table), the pre-defined mappings speed up the process of identifying key semantic elements to be associated with the most appropriate AI / ML model of semantic decoding and the overall computational burden to execute the semantic pre-decoding is reduced by serving as a quick reference guide for the semantic pre-decoder, allowing it to make rapid decisions about semantic quotient and appropriate decoding paths, which is crucial for the low-latency requirements. Mapping relation information includes the key elements such as AI / ML model ID (e.g., a unique identifier for each model available at the UE side), semantic application domain (e.g., the data type of specific application the model is designed to process), semantic feature (e.g., a component of the meaning associated with semantic concept), semantic quotient (e.g., the degree to which a piece of semantic information aligns with and impacts the semantic decoding processes). The elements included in mapping relation table are allowed to be one-to-one or one-to- many / many-to-one or many-to-many mapping depending on its implementation scenarios. For example, the same AI / ML model ID is mapped onto two or more semantic application domain identifiers and / or semantic features. In the exemplary mapping relation table, mapping relation information consists of the varying combinations of identifiers for AI / ML model ID (ML_Model_ID), semantic application domain (Sem_AppDomain_ID), semantic feature (Sem_Feature_ID) and semantic quotient (Sem_Quot_ID) where the size of mapping relation table and the associated each specific identifier values of mapping relation elements need to be preconfigured according to different use cases and / or deployment scenarios. As the key semantic elements or features are identified at UE side, it can quickly look up the pre-configured mapping relation table. The table includes application domain to allow for context-specific quotient assignment, as the same feature might have different quotients in different applications. The quotient from the table is used to classify the semantic elements quickly. The semantic elements are routed to the most relevant decoder based on the associated AI / ML model ID indicated in the table. The table can be updated based on feedback from the semantic decoder output, allowing the system to learn and adapt over time. 202407117

[0117] - 23 -

[0118] Figure 2 shows an exemplary flow chart of enabling semantic pre-decoding at network side. In this example, configuration information to support semantic predecoding is generated at network side. Mapping relation information to support adaptive AI / ML model selection for semantic pre-decoding is also part of configuration information that is sent to UE. UE sends feedback information based on semantic decoding output so that feedback information is used to refine the reordering of quotients of key semantic elements as well as to identify the updated key semantic elements along with mapping relation information update.

[0119] Figure 3 shows an exemplary flow chart of processing semantic pre-decoding at UE side. In this example, based on the received configuration information to support semantic pre-decoding, UE activates semantic data processing with reception of semantic data for decoding. If semantic pre-decoding is enabled (that can be either indicated by network side or decided by UE autonomously), the pre-configured mapping relation information is used to select the appropriate AI / ML model(s) for decoding process along with identification of key semantic elements and the associated quotients. Selection of AI / ML model for decoding is either decided by network side or UE side autonomously. If network side decides selection of AI / ML model for decoding, a specific AI / ML model ID is indicated by network or model itself is transferred to UE. If UE side decides selection of AI / ML model for decoding, the pre-configured mapping relation is referenced. If semantic pre-decoding is not supported, then the baseline semantic decoder is enabled for processing.

[0120] Figure 4 shows an exemplary flow chart of the overall processing for semantic predecoding. In this example, key semantic elements from the incoming encoded data before full decoding are identified. Different semantic quotients are assigned to key semantic elements. The most appropriate AI / ML model is selected to process semantic decoding of each key semantic elements. The appropriate decoding path based on the pre-configured threshold is enabled with the associated AI / ML model. Based on feedback of semantic decoding output, the pre-configured mapping relation information is updated to refine the re-ordering of quotients of key semantic elements as well as to identify the updated key semantic elements. 202407117

[0121] - 24 -

[0122] Figure 5 shows an exemplary signaling flow of semantic pre-decoding. In this example, semantic communication capabilities is sent to network side (e.g., as part of UE capability information message) that contains supportability of semantic predecoding, semantic quotients, list of semantic decoders or AI / ML model IDs. Based on the received semantic communication capabilities from UE, semantic communication parameters is configured and sent to UE via RRC reconfiguration message where this configuration information contains (de-)activation of semantic pre-decoding, semantic quotient threshold, semantic mapping relation table identifier. A new MAC CE is allowed for dynamic semantic quotient indication that contains semantic quotient. Semantic pre-decoding process is activated at UE side. Based on feedback of semantic decoding output, the pre-configured mapping relation information is updated to refine the re-ordering of quotients of key semantic elements as well as to identify the updated key semantic elements. By applying the proposed method, faster processing of semantic data can be executed by classifying semantic quotients early and allocating computational resources more effectively. Therefore, the overall system latency can be reduced by routing semantic elements with low quotients through simpler decoders. Also semantic elements with high quotient receive more thorough processing potentially improve the accuracy of semantic interpretation.

Claims

202407117- 25 -CLAIMS1. Method of advanced adaptive processing for radio access network by configuring adaptive AI / ML model selection for semantic pre-decoding, comprising:• Configuring new information elements and the associated message signaling for semantic data transmission;• Defining semantic identifiers for mapping relation information;• Generating mapping relation information to select AI / ML model adaptively;• Selecting AI / ML model for appropriate semantic decoding.

2. The method according to previous claim 1 , wherein key semantic elements from the encoded semantic data before full decoding are identified.

3. The method according to one of the previous claims, wherein different semantic quotients are assigned to each key semantic elements.

4. The method according to one of the previous claims, wherein semantic quotient represents the degree to which a piece of semantic information aligns with and impacts the semantic decoding processes.

5. The method according to one of the previous claims, wherein AI / ML model is selected to process semantic decoding of each key semantic elements or group of semantic elements such that semantic elements with high semantic quotient are routed to more complex semantic decoders and semantic elements with low semantic quotient may be processed by simpler decoders.

6. The method according to one of the previous claims, wherein different semantic elements based on their quotients are allowed to be decoded simultaneously.

7. The method according to one of the previous claims, wherein two or more AI / ML models for semantic decoding are allowed to be activated selectively.202407117- 26 -8. The method according to one of the previous claims, wherein feedback message based on semantic decoding output is used to refine the re-ordering of quotients of key semantic elements as well as to identify the updated key semantic elements.

9. The method according to one of the previous claims, wherein the initial configuration of the mapping relation information is delivered to the UE using RRC signaling.

10. The method according to one of the previous claims, wherein the RRC reconfiguration message is used to send the full mapping relation information or updates to the UE.

11. The method according to one of the previous claims, wherein L1 / L2 signaling for dynamic updates or real-time selection of AI / ML models is used such that MAC CE indicates which AI / ML model to use for semantic decoding.

12. The method according to one of the previous claims, wherein the base mapping relation information is pre-configured in the UE as part of the device firmware or software.

13. The method according to one of the previous claims, wherein updates to the mapping relation information (e.g., based on network conditions and semantic communication requirements) is configured at the network side and delivered to the UE using L1 / L2 or RRC signaling.

14. The method according to one of the previous claims, wherein the elements of mapping relation information (e.g., look-up table) include AI / ML model ID, semantic application domain, semantic feature and semantic quotient.

15. The method according to one of the previous claims, wherein the elements included in mapping relation table are allowed to be one-to-one or one-to- many / many-to-one or many-to-many mapping depending on its implementation scenarios.202407117- 27 -16. The method according to one of the previous claims, wherein the size of mapping relation table and the associated each specific identifier values of mapping relation elements are pre-configured according to different use cases and / or deployment scenarios.

17. The method according to one of the previous claims, wherein the mapping relation table includes application domain to allow for context-specific quotient assignment, as the same feature might have different quotients in different applications.

18. The method according to one of the previous claims, wherein the quotient from the table is used to classify the semantic elements quickly.

19. The method according to one of the previous claims, wherein a new IE to indicate semantic communication capabilities is added to UE capability information message as a new RRC signaling by containing supportability of semantic predecoding, semantic quotient, list of semantic decoders or AI / ML model IDs (e.g., with different complexities).

20. The method according to one of the previous claims, wherein a new IE to configure semantic communication parameters for network is added to RRC reconfiguration message as a new RRC signaling for network by containing (deactivation of semantic pre-decoding, semantic quotient threshold, semantic mapping relation table identifier.21 . The method according to one of the previous claims, wherein a new MAC CE is added for dynamic semantic quotient indication that contains semantic quotient.

22. The method according to one of the previous claims, wherein selection of AI / ML model for decoding is either decided by network side or UE side autonomously.202407117- 28 -23. The method according to one of the previous claims, wherein a specific AI / ML model ID is indicated by network or model itself is transferred to UE if network side decides selection of AI / ML model for decoding.

24. The method according to one of the previous claims, wherein the pre-configured mapping relation is referenced if UE side decides selection of AI / ML model for decoding.

25. Apparatus for advanced adaptive processing for radio access network by configuring adaptive AI / ML model selection for semantic pre-decoding the apparatus comprising a wireless transceiver, a processor coupled with a memory in which computer program instructions are stored, said instructions being configured to implement steps of the claims 1 to 24.

26. User Equipment comprising an apparatus according to claim 25.

27. gNB comprising an apparatus according to claim 25.

28. Wireless communication system for advanced adaptive processing for radio access network by configuring adaptive AI / ML model selection for semantic predecoding, wherein the wireless communication systems comprises user equipment according to claim 26, gNB according to claim 27, whereby the user equipment and the gNB each comprises a processor coupled with a memory in which computer program instructions are stored, said instructions being configured to implement steps of the claims 1 to 24.

Images