WO2025091370A1

WO2025091370A1 - Type b model identification procedure for model id based lcm

Info

Publication number: WO2025091370A1
Application number: PCT/CN2023/129290
Authority: WO
Inventors: Peng Cheng; Huaning Niu; Wei Zeng; Weidong Yang; Dawei Zhang; Oghenekome Oteri
Original assignee: Apple Inc
Current assignee: Apple Inc
Priority date: 2023-11-02
Filing date: 2023-11-02
Publication date: 2025-05-08
Anticipated expiration: 2026-05-02

Abstract

Systems and methods may use type B model identification procedures for model ID based lifecycle management (LCM). Some embodiments may include training a new model for wireless communication. A user equipment (UE) or network node may send an indication comprising a new model identification (ID) associated with the new model. The UE or network node may also send the new model together with the new model ID.

Description

TYPE B MODEL IDENTIF ICATION PROCEDURE FOR MODEL ID BASED LCM

TECHNICAL FIELD

This application relates generally to wireless communication systems, including model identification via over-the-air signaling.

BACKGROUND

Wireless mobile communication technology uses various standards and protocols to transmit data between a base station and a wireless communication device. Wireless communication system standards and protocols can include, for example, 3rd Generation Partnership Project (3GPP) Long Term Evolution (LTE) (e.g., 4G) , 3GPP New Radio (NR) (e.g., 5G) , and Institute of Electrical and Electronics Engineers (IEEE) 802.11 standard for Wireless Local Area Networks (WLAN) (commonly known to industry groups as ) .

As contemplated by the 3GPP, different wireless communication systems' standards and protocols can use various radio access networks (RANs) for communicating between a base station of the RAN (which may also sometimes be referred to generally as a RAN node, a network node, or simply a node) and a wireless communication device known as a user equipment (UE) . 3GPP RANs can include, for example, Global System for Mobile communications (GSM) , Enhanced Data Rates for GSM Evolution (EDGE) RAN (GERAN) , Universal Terrestrial Radio Access Network (UTRAN) , Evolved Universal Terrestrial Radio Access Network (E-UTRAN) , and/or Next-Generation Radio Access Network (NG-RAN) .

Each RAN may use one or more radio access technologies (RATs) to perform communication between the base station and the UE. For example, the GERAN implements GSM and/or EDGE RAT, the UTRAN implements Universal Mobile Telecommunication System (UMTS) RAT or other 3GPP RAT, the E-UTRAN implements LTE RAT (sometimes simply referred to as LTE) , and NG-RAN implements NR RAT (sometimes referred to herein as 5G RAT, 5G NR RAT, or simply NR) . In certain deployments, the E-UTRAN may also implement NR RAT. In certain deployments, NG-RAN may also implement LTE RAT.

A base station used by a RAN may correspond to that RAN. One example of an E-UTRAN base station is an Evolved Universal Terrestrial Radio Access Network (E- UTRAN) Node B (also commonly denoted as evolved Node B, enhanced Node B, eNodeB, or eNB) . One example of an NG-RAN base station is a next generation Node B (also sometimes referred to as a g Node B or gNB) .

A RAN provides its communication services with external entities through its connection to a core network (CN) . For example, E-UTRAN may utilize an Evolved Packet Core (EPC) while NG-RAN may utilize a 5G Core Network (5GC) .

BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS

To easily identify the discussion of any particular element or act, the most significant digit or digits in a reference number refer to the figure number in which that element is first introduced.

FIG. 1 illustrates a table of cases for model delivery/transfer in accordance with some embodiments.

FIG. 2 illustrates an example signal flow diagram for a model transfer with a known model structure for type B1 in accordance with some embodiments.

FIG. 3 illustrates an example signal flow diagram for a model transfer with a known model structure for type B2 in accordance with some embodiments.

FIG. 4 illustrates an example signal flow diagram for a model transfer with an unknown model structure for type B1 in accordance with some embodiments.

FIG. 5 illustrates an example signal flow diagram for a model transfer with an unknown model structure for type B2 in accordance with some embodiments.

FIG. 6 illustrates an example signal flow diagram for a model update for type B1 in accordance with some embodiments.

FIG. 7 illustrates an example signal flow diagram for a model update for type B2 in accordance with some embodiments.

FIG. 8 illustrates a sub-model used for model transfer with fine-tuning in accordance with some embodiments.

FIG. 9 illustrates a signal flow diagram for model identification for a network side additional condition in accordance with some embodiments.

FIG. 10 illustrates a signal flow diagram for model identification for a UE side additional condition in accordance with some embodiments.

FIG. 11 illustrates a method for a UE, according to embodiments herein.

FIG. 12 illustrates a method for a network node, according to embodiments herein.

FIG. 13 illustrates an example architecture of a wireless communication system, according to embodiments disclosed herein.

FIG. 14 illustrates a system for performing signaling between a wireless device and a network device, according to embodiments disclosed herein.

DETAILED DESCRIPTION

Various embodiments are described with regard to a UE. However, reference to a UE is merely provided for illustrative purposes. The example embodiments may be utilized with any electronic component that may establish a connection to a network and is configured with the hardware, software, and/or firmware to exchange information and data with the network. Therefore, the UE as described herein is used to represent any appropriate electronic component.

Goals of wireless communication systems include providing reliable, efficient, and secure communication between UEs and network nodes. Artificial intelligence (AI) /machine learning (ML) may be used to assist in meeting these goals. For example, AI technologies can be used in several ways within a wireless system, including network optimization. For instance, AI can be used to optimize the performance of the network by analyzing data, predicting future traffic patterns, and identifying areas of congestion. This can help to improve network efficiency and reduce downtime. Further, AI can be used to automate network operations, such as provisioning, configuration, and optimization. This can help to reduce costs and improve operational efficiency.

Life cycle management (LCM) may refer to the automated management and optimization of various aspects throughout the life cycle of network components, services, or applications using AI techniques. The life cycle of a network component typically includes phases such as planning, deployment, operation, and maintenance. LCM may leverage AI capabilities to streamline and automate tasks related to these phases, ensuring efficient management and optimization of the network. With AI-powered LCM, network operators can benefit from advanced analytic s, machine learning algorithms, and automation to plan and optimize network resources. AI algorithms can analyze network performance data, predict traffic patterns, and optimize resource allocation to improve efficiency and quality of service. LCM may also be used to automate configuration and deployment. For example, AI can automate the configuration and deployment of network elements, reducing manual intervention and minimizing human errors. Further, AI-enabled monitoring and diagnostics tools can proactively detect anomalies and potential issues, enabling quick problem resolution and reducing network downtime. Further, by analyzing network data, AI algorithms can predict potential failures or degradation in network components, allowing operators to schedule maintenance activities in advance, optimizing performance and reducing costs. Accordingly, LCM may utilize artificial intelligence techniques to automate and optimize network management tasks throughout the life cycle of network components, leading to improved efficiency, enhanced performance, and better user experience.

A device can leverage multiple AI/ML models for different conditions through a process called model selection or model switching. This approach allows the device to dynamically choose the most appropriate AI model based on the prevailing conditions or requirements. The models may be used to enhance different aspects of wireless communication in different scenarios. Embodiments herein provide apparatuses, systems, and methods of transferring and updating models based on a model identification (ID) .

In some embodiments, LCM may be based on a model ID. Such model ID based LSM may rely on accurate model identification. Some embodiments herein provide model identification procedure for model-ID-based LCM. For model identification of UE-side (one sided model) or UE-part of two-sided models, embodiments may categorize model identification types as follows. Type A may be the case where a model is identified to the network (NW) (if applicable) and UE (if applicable) without over-the-air signaling. The model may be assigned with a model ID during the model identification, which may be referred/used in over-the-air signaling after model identification.

Type B may be the case where a model is identified via over-the-air signaling. There may be two different types of type B models. Type B1 may refer to model identification initiated by the UE, and where the NW assists the remaining steps (if any) of the model identification. The model may be assigned with a model ID during the model identification. Type B2 may refer to model identification initiated by the NW, and where the UE responds (if applicable) for the remaining steps (if any) of the model identification. Similar to Type B1, in Type B2 the model may be assigned with a model ID during the model identification.

In some embodiments, when a model of a known structure at UE (e.g., Case z4) is transferred from NW, the new model being identified (e.g., via Type B2) may have the same structure as a previously identified model at the Network and UE. In some embodiments, model ID may or may not be globally unique, and different types of model IDs may be created for a single model for various LCM purposes.

FIG. 1 illustrates a table 102 of cases for model delivery/transfer. Embodiments herein may be used for at least the illustrated cases for model delivery/transfer to UE, training location, and model delivery/transfer format combinations for UE-side models and UE-part of two-sided models.

For inference for UE-side models, to ensure consistency between training and inference regarding NW-side additional conditions (if identified) , the following options can be taken as potential approaches (when feasible and necessary) : model identification to achieve alignment on the NW-side additional condition between NW-side and UE-side; model training at NW and transfer to UE, where the model has been trained under the additional condition; information and/or indication on NW-side additional conditions is provided to UE; consistency assisted by monitoring (by UE and/or NW, the performance of UE-side candidate models/functionalities to select a model/functionality) . Other approaches are not precluded. This may not exclude the possibility that different approaches can achieve the same function.

Some embodiments herein focus on model identification type B, which is an over the air signaling procedure. Model identification type A is an offline approach. In some embodiments, type 2 model identification may be used for model transfer. Some solutions provide for model identification with a known model structure (model update) . Some solutions provide for model identification with unknown model structure (new model) . In some embodiments, type 2 model identification may be used for two sided/one sided model update. Some solutions provide for type B1 for NW side model identification. Some solutions provide for type B2 for UE side model identification. In some embodiments, type B model identification may be used to achieve alignment between NW side condition and UE side condition.

FIG. 2 illustrates an example signal flow diagram 202 for a model transfer with a known model structure for type B1 in accordance with some embodiments. The illustrated embodiment shown is a type B1 model transfer that is initiated by the network (e.g., network node 204) . A UE capability report may indicate to the NW that the UE 206 supports a model parameter update with a known model structure. If the UE 206 does not indicate that it supports the model parameter update in the UE capability report, the NW may avoid transferring the model.

The NW may train a new model 220. The network node 204 can indicate through Radio Resource Control (RRC) signaling a new model ID 208 and/or meta-information. The RRC signaling can be cell specific signaling (e.g., System Information Block (SIB) ) or UE specific signaling. When the new model ID 208 and/or meta-information is sent via UE specific signaling, the RRC message may be sent based on UE capability report.

The model ID may be indicated in various ways depending on the model ID design. In some embodiments, if the model ID is defined with a version number of part of the ID, then the updated model may be signaled with version ID only. The other parts of the model ID may be the same to indicate that the model is a known model. The model structure may be implied by the model ID itself which is previously identified. In some embodiments, the version number may be associated with a timestamp. In some embodiments, if the model ID is not defined with a version number, a new model ID may be associated with a previously identified model ID in the signaling. The previously identified model ID may implicitly indicate the known model structure. In some embodiments, a value tag can be added to a previously identified model ID. The value tag may indicate that it is an updated version with the same model structure.

The UE 206 may check 210 whether the new model ID can be supported. Meta information may provide more information on the updated model for the UE to check whether it supports and knows model structure. For example, the meta-information may describe what the model is for (e.g., the model is for CSI compression) . The meta-information may also describe use scenarios. For example, the meta-information may describe that the model is for indoors, outdoors, and/or a certain antenna configuration. The described meta-information are provided as examples other information regarding the model may be included in the meta-information.

The UE 206 may send the network node 204 a request 212 to download the model if the UE 206 capability supports the model parameter update and the indicated known model ID is already supported at the UE. The UE 206 may send the request 212 via Medium Access Control Control Element (MAC CE) or RRC message.

The network node 204 may send the model file 214 together with the new model ID to the UE 206. The UE 206 may need to compile the downloaded model, or request UE server to compile the model. The compilation may happen by implementation and may take some time to finish. When the UE 206 is ready to run inferencing (e.g., after compilation of the model) , the UE 206 may send an uplink message 216 to the network node 204, indicating the model is ready to run and the UE 206 is ready to perform inferencing. The uplink message 216 may be a UE assisted information (UAI) , an uplink MAC CE for model activation request, or UE capability signaling. The network node 204 may send activation/deactivation messages 218 to the UE 206 to activate or deactivate the model. The UE 206 and the network node 204 may continue the remaining life cycle management procedure (e.g., activating, monitoring, deactivating, or switching the model) .

FIG. 3 illustrates an example signal flow diagram 302 for a model transfer with a known model structure for type B2 in accordance with some embodiments. The illustrated embodiment shown is a type B2 model transfer that is initiated by the UE 306. UE or UE side offline training may produce a new model 308 with updated model parameters. For example, the UE 306 may use newly collected data to update model parameters.

The UE 306 can indicate to the network node 304 a new model ID 310 and/or meta-information. The UE may send the indication through a MAC CE or RRC message. In some embodiments, the indication of a new model ID 310 and/or meta-information may be sent via a UE capability report.

The network node 304 may check 312 whether the new model ID can be supported. Meta information may provide more information on the updated model for the network node 304 to check whether it supports and knows model structure. For example, the meta-information may describe what the model is for (e.g., the model is for CSI compression) . The meta-information may also describe use scenarios. For example, the meta-information may describe that the model is for indoors, outdoors, and/or a certain antenna configuration. The described meta-information are provided as examples other information regarding the model may be included in the meta-information.

The network node 304 may send uplink grant for model transfer (e.g., request 314) , if network node 304 supports model parameter updates. The UE 304 may send the model file 316 together with the new model ID to the network node 304 in Physical Uplink Shared Channel (PUSCH) . Uplink (UL) DCI, MAC CE, or RRC message can be used to trigger the UL model transfer.

The UE 206 may send the network node 204 a request 212 to download the model if the UE 206 capability supports the model parameter update and the indicated known model ID is already supported at the UE. The UE 206 may send the request 212 via MAC CE or RRC message. When the network node 304 is ready to inference, the network node 304 may indicate activation 318 of the model ID. The UE 306 and the network node 304 may continue the remaining life cycle management procedure (e.g., activating, monitoring, deactivating, or switching the model) .

FIG. 4 illustrates an example signal flow diagram 402 for a model transfer with an unknown model structure for type B1 in accordance with some embodiments. The illustrated embodiment shown is a type B1 model transfer that is initiated by the network (e.g., network node 404) . A UE capability report may indicate to the NW that the UE 406 supports model transfer type z5 (e.g., model transfer of an unknown model) .

If the UE 406 has the capability to support model transfer type z5, NW can indicate through RRC signaling the new model ID (e.g., new model ID announcement 408) . There may be no need to link to a previous model ID. The UE 406 may check 410 whether the new model ID is supported.

The UE 406 may send a request 412 to download the model to the network node 404. The request 412 may be sent via a MAC CE or a RRC message. The network node 404 may send the model file 414 together with the new model ID to the UE 406.

The UE 406 may need to compile the downloaded model, or request UE server to compile the model. The compilation may happen by implementation and may take some time to finish. When the UE 406 is ready to run inferencing (e.g., after compilation of the model) , the UE 406 may send an uplink message 416 to the network node 404, indicating the model is ready to run and the UE 406 is ready to perform inferencing. The uplink message 416 may be an UAI, an UL MAC CE for model activation request, or UE capability signaling. The network node 404 may send activation/deactivation messages 418 to the UE 406 to activate or deactivate the model. The UE 406 and the network node 404 may continue the remaining life cycle management procedure (e.g., activating, monitoring, deactivating, or switching the model) .

FIG. 5 illustrates an example signal flow diagram 502 for a model transfer with an unknown model structure for type B2 in accordance with some embodiments. The illustrated embodiment shown is a type B2 model transfer that is initiated by the UE 506. UE or UE side offline training may produce a new model 508.

The UE 506 can indicate 510 to the network node 504 the availability of a new model. The UE may send the indication through a MAC CE or RRC message. The UE may indicate this is a z5 model transfer without previously identified structure.

The network node 504 may check 516 whether the NW supports unknown structure (z5) . If unknown model structure is not supported the network node 504 may end the procedure. If the unknown model structure is supported, the network node 504 may send an UL grant for model transfer (e.g., request UL model transfer 512) . The UE 506 may send the model file together with the new model ID in PUSCH (e.g., UL model transfer 514) . In some embodiments, the model file with the new model ID may be sent in an RRC message. Some embodiments may use fragmentation to fit the new model file and new model ID in the RRC message.

When the network node 504 is ready to inference, the network node 504 may indicate activation 518 of the model ID. The UE 506 and the network node 504 may continue the remaining life cycle management procedure (e.g., activating, monitoring, deactivating, or switching the model) .

FIGS. 2 and 3 relate to the transfer of known models, whereas FIGS. 4 and 5 relate to the transfer of unknown models. Some differences between these approaches is the UE capability of supporting z5 (unknown model) , and whether a previous known model ID is linked with the new model ID. In some embodiments, it may possible to combine the two solution as a common procedure.

FIGS. 2-5 illustrate model transfer procedures. Model transfer may be used for both transferring a new model and providing an update to a model. Some embodiments may use the procedures in FIGS. 6 and 7 to provide a model update. In some embodiments, when training collaboration type 3 are used model update procedures may be used. Training collaboration type 1 may use model transfer.

FIG. 6 illustrates an example signal flow diagram 602 for a model update for type B1 in accordance with some embodiments. The illustrated embodiment shown is a type B1 model update that is initiated by the network (e.g., network node 604) . For NW first training, when the network node 604 may train an updated model 608, and transfers new training dataset 610 to the UE 606.

The UE 606 may use the new training dataset 610 to train an encoder 612 offline. The UE 606 may indicate 614 to the network node 604 that the updated model ID is supported. Further, the UE 606 may indicate to the network node 604 a desire to update model ID/dataset ID that the UE 606 supports. This indication may be done through a UE capability report, or UAI can indicate through RRC signaling the new model ID/dataset ID.

The network node 604 may send activation/deactivation messages 616 to the UE 606 to activate or deactivate the model. The UE 606 and the network node 604 may continue the remaining life cycle management procedure (e.g., activating, monitoring, deactivating, or switching the model) .

FIG. 7 illustrates an example signal flow diagram 702 for a model update for type B2 in accordance with some embodiments. The illustrated embodiment shown is a type B2 model update that is initiated by the UE 706. As an example, for UE first training, the UE 706 may train an updated model 708, and transfer a new training dataset 710 to the network node 704.

The network node 704 may use the new training dataset 710 to train an encoder 712 offline. The network node 704 may indicate 714 to the UE 706 that the updated model ID is supported. The network node 704 may indicate to the UE 706 a desire to update model ID/dataset ID that the network node 704 supports. This indication may be done through RRC configuration/Downlink Control Information (DCI) and/or MAC CE.

The network node 704 may send activation/deactivation messages 716 to the UE 706 to activate or deactivate the model. The UE 706 and the network node 704 may continue the remaining life cycle management procedure (e.g., activating, monitoring, deactivating, or switching the model) .

In some embodiments, sub-models may be used with mode transfer and model update procedures. For direct AI/ML positioning, evaluation results show that fine-tuning/re-training a previous model with a dataset of a new deployment scenario improves the model performance for the new deployment scenario. After fine-tuning/re-training a previous model with the dataset of the new deployment scenario, the performance of the updated model may degrade for the previous deployment scenario (e.g., previous clutter parameter setting) that the previous model was trained for.

Accordingly, in some embodiments, a wireless communication system may keep the original model in case there may be a need for a switch back to the model. To do so, a sub-model/child-model framework may be created. The framework may use the original model ID as the parent model IDs and the fine-tuned/re-trained models as the sub-models/child models associated with the parent model. A UE may indicate the number of child models supportable as a capability

In the embodiments described with relation to FIGS. 1-6, the UE/network may indicate a new sub-model ID during the transfer. In some embodiments, the parent model may be kept by default. In some embodiments, associated signaling may be used to indicate if the parent model is kept or not.

FIG. 8 illustrates a sub-model 802 used for model transfer with fine-tuning in accordance with some embodiments. In model fine-tuning, layers that are close to the input of the model may be frozen and layers close to the output of the model may be updated. For example, model 806 may have 6 layers and a model ID set to X. A sub-model 802 may be created by freezing the first four layers 804 closer to the input and changing the two layers 808 closer to the output of the model.

In some embodiments, a model transfer or model update may indicate which layers to change and how they are changed. Sub-model IDs may be defined that indicate changes to specific layers. For example, sub-ID set to Y may indicate the sub-model 802 with the two layers 808 closest to the output changed. Model identification and transfer may involve identification/transfer of only the identified/configured layers. For instance, a model transfer or update may provide a map indicating the layers that have changed.

Currently, certain conditions that the UE side applies are not provided to the network side, and certain conditions that the network side applies are not provided to the UE side. Some of these conditions may be useful for more accurately training models. Accordingly, FIGS. 9 and 10 illustrate how additional conditions may be included in model identification.

Specifically, FIG. 9 illustrates a signal flow diagram 902 for model identification for a network side additional condition in accordance with some embodiments. A network side additional condition may refer to a network side implementation that is transparent to the UE. For example, a network antenna to port virtualization for CSI compression use case may be a network side additional condition. Further, network side beam pattern for Synchronization Signal Block (SSB) and Channel State Information Reference Signal (CSI-RS) transmission for beam management may be a network side additional condition.

The signal flow diagram 902 illustrates type B1 that is initiated by the network node 904. The network node 904 may determine a model ID 906 corresponding to its own additional condition. The model ID may be a logical ID to categorize the dataset. The network node 904 may create new model IDs to use as identifiers for different condition configurations. For example, if the network node 904 changes a side beam pattern, the network node may associate the new side beam pattern with a new model ID. The network node 904 may trigger 908 a data collection procedure with the model ID.

The UE 904 may collect data 912 and train the model offline. The UE 904 may collect data with the same ID, and offline train the model. In some embodiments, the UE can train the model with generalized performance working with multiple model IDs/datasets.

In some embodiments, the network node 904 can send a configuration 914 of the use case with the model IDs to the UE 910 for inferencing. The UE 910 may indicate 916 to the network node 904 whether the model ID is supported or not. In other embodiments, the UE 910 can report the supported model IDs to the network node 904 in either UE capability or UAI, and the network node 904 can configure/activate the model ID for inferencing.

The network node 904 may send activation/deactivation messages 918 to the UE 910 to activate or deactivate the model. The UE 910 and the network node 904 may continue the remaining life cycle management procedure (e.g., activating, monitoring, deactivating, or switching the model) .

FIG. 10 illustrates a signal flow diagram 1002 for model identification for a UE side additional condition in accordance with some embodiments. A UE side additional condition may refer to a UE side implementation that is transparent to the network. For example, a UE antenna placement on the phone may be a UE side additional condition. Further, UE side beam pattern for beam management may be a UE side additional condition.

The signal flow diagram 1002 illustrates type B2 that is initiated by the UE 1004. For network side data collection, the network node 1006 may trigger the data collection procedure 1008. The UE 1004 may decide 1010 the UE side additional condition and corresponding model ID. In a data report 1012 from the UE 1004 to the network node 1006, the UE 1004 may send the model ID corresponding to the UE side additional condition. The network node 1006 may collect data with the same ID, and offline train 1014 the model. The network node 1006 can train the model with generalized performance working with multiple model IDs/datasets.

In some embodiments, the network node 1006 can configure 1016 the model IDs to the UE’s confirmation which model matches with UE side additional condition. The UE may indicate 1018 the corresponding model ID, and the network can perform inferencing.

In some embodiments, the UE 1004 can report the supported model ID to the network node 1006 in either a UE capability report or UAI. The network node 1006 may use a model corresponding to the model ID for inferencing.

The network node 1006 may send activation/deactivation messages 1020 to the UE 1004 to activate or deactivate the model. The UE 1004 and the network node 1006 may continue the remaining life cycle management procedure (e.g., activating, monitoring, deactivating, or switching the model) .

FIG. 11 illustrates a method 1100 for a UE, according to embodiments herein. The illustrated method 1100 includes training 1102 a new model for wireless communication with a network node. The method 1100 further includes sending 1104 an indication to the network node comprising a new model ID associated with the new model. The method 1100 further includes receiving 1106, from the network node, an uplink grant comprising a request for a model transfer. The method 1100 further includes sending 1108, to the network node, the new model together with the new model ID. The method 1100 further includes receiving 1110, from the network node, a message to activate the new model.

In some embodiments of the method 1100, the new model comprises a model structure that is known to the network node. In some such embodiments, the new model ID comprises a version number, and wherein the new model is signaled with a version ID. In other such embodiments, the new model ID is associated with a previous model ID in the signaling, and wherein the previous model ID implicitly indicates the known model structure. Some such embodiments further comprise adding a value tag to a previous model ID to generate the new model ID.

In some embodiments, the method 1100 further comprises sending, to the network node, meta-information to provide more information on the new model.

In some embodiments of the method 1100, the model is an unknown model.

In some embodiments, the method 1100 further comprises training the new model to generate an updated model, sending, to the network node, a new training dataset for the updated model, and receiving, from the network node, an indication to update the new model ID or a dataset ID. In some such embodiments, when the new model is updated, the UE keeps both an original version of the new model and a sub-model comprising the updated model.

In some embodiments, the method 1100 further comprises associating a UE side additional condition with the new model ID.

In some embodiments, the method 1100 further comprises associating a network side additional condition with the new model ID, wherein the new model ID associated with the network side additional condition is part of a data collection procedure.

Embodiments contemplated herein include an apparatus comprising means to perform one or more elements of the method 1100. This apparatus may be, for example, an apparatus of a UE (such as a wireless device 1402 that is a UE, as described herein) .

Embodiments contemplated herein include one or more non-transitory computer-readable media comprising instructions to cause an electronic device, upon execution of the instructions by one or more processors of the electronic device, to perform one or more elements of the method 1100. This non-transitory computer-readable media may be, for example, a memory of a UE (such as a memory 1406 of a wireless device 1402 that is a UE, as described herein) .

Embodiments contemplated herein include an apparatus comprising logic, modules, or circuitry to perform one or more elements of the method 1100. This apparatus may be, for example, an apparatus of a UE (such as a wireless device 1402 that is a UE, as described herein) .

Embodiments contemplated herein include an apparatus comprising: one or more processors and one or more computer-readable media comprising instructions that, when executed by the one or more processors, cause the one or more processors to perform one or more elements of the method 1100. This apparatus may be, for example, an apparatus of a UE (such as a wireless device 1402 that is a UE, as described herein) .

Embodiments contemplated herein include a signal as described in or related to one or more elements of the method 1100.

Embodiments contemplated herein include a computer program or computer program product comprising instructions, wherein execution of the program by a processor is to cause the processor to carry out one or more elements of the method 1100. The processor may be a processor of a UE (such as a processor (s) 1404 of a wireless device 1402 that is a UE, as described herein) . These instructions may be, for example, located in the processor and/or on a memory of the UE (such as a memory 1406 of a wireless device 1402 that is a UE, as described herein) .

FIG. 12 illustrates a method 1200 for a network node, according to embodiments herein. The illustrated method 1200 includes training 1202 a new model for wireless communication with a UE. The method 1200 further includes sending 1204 an announcement to the UE comprising a new model ID associated with the new model. The method 1200 further includes receiving 1206, from the UE, a request to download the new model. The method 1200 further includes sending 1208, to the UE, the new model together with the new model ID. The method 1200 further includes receiving 1210, from the UE, an indication that the UE is ready to use the new model for inferencing. The method 1200 further includes sending 1212, to the UE, a message to activate the new model.

In some embodiments of the method 1200, the new model comprises a model structure that is known to the network node. In some such embodiments, the new model ID comprises a version number, and wherein the new model is signaled with a version ID. In other such embodiments, the new model ID is associated with a previous model ID in the signaling, and wherein the previous model ID implicitly indicates the known model structure. Some such embodiments further comprise adding a value tag to a previous model ID to generate the new model ID.

In some embodiments, the method 1200 further comprises sending, to the UE, meta-information to provide more information on the new model.

In some embodiments of the method 1200, the model is an unknown model.

In some embodiments, the method 1200 further comprises training the new model to generate an updated model, sending, to the UE, a new training dataset for the updated model, and receiving, from the UE, an indication that the updated model is supported. In some such embodiments, when the new model is updated, storing both an original version of the new model and a sub-model comprising the updated model.

In some embodiments, the method 1200 further comprises associating a network side additional condition with the new model ID.

In some embodiments, the method 1200 further comprises associating a UE side additional condition with the new model ID.

Embodiments contemplated herein include an apparatus comprising means to perform one or more elements of the method 1200. This apparatus may be, for example, an apparatus of a base station (such as a network device 1418 that is a base station, as described herein) .

Embodiments contemplated herein include one or more non-transitory computer-readable media comprising instructions to cause an electronic device, upon execution of the instructions by one or more processors of the electronic device, to perform one or more elements of the method 1200. This non-transitory computer-readable media may be, for example, a memory of a base station (such as a memory 1422 of a network device 1418 that is a base station, as described herein) .

Embodiments contemplated herein include an apparatus comprising logic, modules, or circuitry to perform one or more elements of the method 1200. This apparatus may be, for example, an apparatus of a base station (such as a network device 1418 that is a base station, as described herein) .

Embodiments contemplated herein include an apparatus comprising: one or more processors and one or more computer-readable media comprising instructions that, when executed by the one or more processors, cause the one or more processors to perform one or more elements of the method 1200. This apparatus may be, for example, an apparatus of a base station (such as a network device 1418 that is a base station, as described herein) .

Embodiments contemplated herein include a signal as described in or related to one or more elements of the method 1200.

Embodiments contemplated herein include a computer program or computer program product comprising instructions, wherein execution of the program by a processing element is to cause the processing element to carry out one or more elements of the method 1200. The processor may be a processor of a base station (such as a processor (s) 1420 of a network device 1418 that is a base station, as described herein) . These instructions may be, for example, located in the processor and/or on a memory of the base station (such as a memory 1422 of a network device 1418 that is a base station, as described herein) .

FIG. 13 illustrates an example architecture of a wireless communication system 1300, according to embodiments disclosed herein. The following description is provided for an example wireless communication system 1300 that operates in conjunction with the LTE system standards and/or 5G or NR system standards as provided by 3GPP technical specifications.

As shown by FIG. 13, the wireless communication system 1300 includes UE 1302 and UE 1304 (although any number of UEs may be used) . In this example, the UE 1302 and the UE 1304 are illustrated as smartphones (e.g., handheld touchscreen mobile computing devices connectable to one or more cellular networks) , but may also comprise any mobile or non-mobile computing device configured for wireless communication.

The UE 1302 and UE 1304 may be configured to communicatively couple with a RAN 1306. In embodiments, the RAN 1306 may be NG-RAN, E-UTRAN, etc. The UE 1302 and UE 1304 utilize connections (or channels) (shown as connection 1308 and connection 1310, respectively) with the RAN 1306, each of which comprises a physical communications interface. The RAN 1306 can include one or more base stations (such as base station 1312 and base station 1314) that enable the connection 1308 and connection 1310.

In this example, the connection 1308 and connection 1310 are air interfaces to enable such communicative coupling, and may be consistent with RAT (s) used by the RAN 1306, such as, for example, an LTE and/or NR.

In some embodiments, the UE 1302 and UE 1304 may also directly exchange communication data via a sidelink interface 1316. The UE 1304 is shown to be configured to access an access point (shown as AP 1318) via connection 1320. By way of example, the connection 1320 can comprise a local wireless connection, such as a connection consistent with any IEEE 802.11 protocol, wherein the AP 1318 may comprise a router. In this example, the AP 1318 may be connected to another network (for example, the Internet) without going through a CN 1324.

In embodiments, the UE 1302 and UE 1304 can be configured to communicate using orthogonal frequency division multiplexing (OFDM) communication signals with each other or with the base station 1312 and/or the base station 1314 over a multicarrier communication channel in accordance with various communication techniques, such as, but not limited to, an orthogonal frequency division multiple access (OFDMA) communication technique (e.g., for downlink communications) or a single carrier frequency division multiple access (SC-FDMA) communication technique (e.g., for uplink and ProSe or sidelink communications) , although the scope of the embodiments is not limited in this respect. The OFDM signals can comprise a plurality of orthogonal subcarriers.

In some embodiments, all or parts of the base station 1312 or base station 1314 may be implemented as one or more software entities running on server computers as part of a virtual network. In addition, or in other embodiments, the base station 1312 or base station 1314 may be configured to communicate with one another via interface 1322. In embodiments where the wireless communication system 1300 is an LTE system (e.g., when the CN 1324 is an EPC) , the interface 1322 may be an X2 interface. The X2 interface may be defined between two or more base stations (e.g., two or more eNBs and the like) that connect to an EPC, and/or between two eNBs connecting to the EPC. In embodiments where the wireless communication system 1300 is an NR system (e.g., when CN 1324 is a 5GC) , the interface 1322 may be an Xn interface. The Xn interface is defined between two or more base stations (e.g., two or more gNBs and the like) that connect to 5GC, between a base station 1312 (e.g., a gNB) connecting to 5GC and an eNB, and/or between two eNBs connecting to 5GC (e.g., CN 1324) .

The RAN 1306 is shown to be communicatively coupled to the CN 1324. The CN 1324 may comprise one or more network elements 1326, which are configured to offer various data and telecommunications services to customers/subscribers (e.g., users of UE 1302 and UE 1304) who are connected to the CN 1324 via the RAN 1306. The components of the CN 1324 may be implemented in one physical device or separate physical devices including components to read and execute instructions from a machine-readable or computer-readable medium (e.g., a non-transitory machine-readable storage medium) .

In embodiments, the CN 1324 may be an EPC, and the RAN 1306 may be connected with the CN 1324 via an S1 interface 1328. In embodiments, the S1 interface 1328 may be split into two parts, an S1 user plane (S1-U) interface, which carries traffic data between the base station 1312 or base station 1314 and a serving gateway (S-GW) , and the S1-MME interface, which is a signaling interface between the base station 1312 or base station 1314 and mobility management entities (MMEs) .

In embodiments, the CN 1324 may be a 5GC, and the RAN 1306 may be connected with the CN 1324 via an NG interface 1328. In embodiments, the NG interface 1328 may be split into two parts, an NG user plane (NG-U) interface, which carries traffic data between the base station 1312 or base station 1314 and a user plane function (UPF) , and the S1 control plane (NG-C) interface, which is a signaling interface between the base station 1312 or base station 1314 and access and mobility management functions (AMFs) .

Generally, an application server 1330 may be an element offering applications that use internet protocol (IP) bearer resources with the CN 1324 (e.g., packet switched data services) . The application server 1330 can also be configured to support one or more communication services (e.g., VoIP sessions, group communication sessions, etc. ) for the UE 1302 and UE 1304 via the CN 1324. The application server 1330 may communicate with the CN 1324 through an IP communications interface 1332.

FIG. 14 illustrates a system 1400 for performing signaling 1434 between a wireless device 1402 and a network device 1418, according to embodiments disclosed herein. The system 1400 may be a portion of a wireless communications system as herein described. The wireless device 1402 may be, for example, a UE of a wireless communication system. The network device 1418 may be, for example, a base station (e.g., an eNB or a gNB) of a wireless communication system.

The wireless device 1402 may include one or more processor (s) 1404. The processor (s) 1404 may execute instructions such that various operations of the wireless device 1402 are performed, as described herein. The processor (s) 1404 may include one or more baseband processors implemented using, for example, a central processing unit (CPU) , a digital signal processor (DSP) , an application specific integrated circuit (ASIC) , a controller, a field programmable gate array (FPGA) device, another hardware device, a firmware device, or any combination thereof configured to perform the operations described herein.

The wireless device 1402 may include a memory 1406. The memory 1406 may be a non-transitory computer-readable storage medium that stores instructions 1408 (which may include, for example, the instructions being executed by the processor (s) 1404) . The instructions 1408 may also be referred to as program code or a computer program. The memory 1406 may also store data used by, and results computed by, the processor (s) 1404.

The wireless device 1402 may include one or more transceiver (s) 1410 that may include radio frequency (RF) transmitter circuitry and/or receiver circuitry that use the antenna (s) 1412 of the wireless device 1402 to facilitate signaling (e.g., the signaling 1434) to and/or from the wireless device 1402 with other devices (e.g., the network device 1418) according to corresponding RATs.

The wireless device 1402 may include one or more antenna (s) 1412 (e.g., one, two, four, or more) . For embodiments with multiple antenna (s) 1412, the wireless device 1402 may leverage the spatial diversity of such multiple antenna (s) 1412 to send and/or receive multiple different data streams on the same time and frequency resources. This behavior may be referred to as, for example, multiple input multiple output (MIMO) behavior (referring to the multiple antennas used at each of a transmitting device and a receiving device that enable this aspect) . MIMO transmissions by the wireless device 1402 may be accomplished according to precoding (or digital beamforming) that is applied at the wireless device 1402 that multiplexes the data streams across the antenna (s) 1412 according to known or assumed channel characteristics such that each data stream is received with an appropriate signal strength relative to other streams and at a desired location in the spatial domain (e.g., the location of a receiver associated with that data stream) . Certain embodiments may use single user MIMO (SU-MIMO) methods (where the data streams are all directed to a single receiver) and/or multi user MIMO (MU-MIMO) methods (where individual data streams may be directed to individual (different) receivers in different locations in the spatial domain) .

In certain embodiments having multiple antennas, the wireless device 1402 may implement analog beamforming techniques, whereby phases of the signals sent by the antenna (s) 1412 are relatively adjusted such that the (joint) transmission of the antenna (s) 1412 can be directed (this is sometimes referred to as beam steering) .

The wireless device 1402 may include one or more interface (s) 1414. The interface (s) 1414 may be used to provide input to or output from the wireless device 1402. For example, a wireless device 1402 that is a UE may include interface (s) 1414 such as microphones, speakers, a touchscreen, buttons, and the like in order to allow for input and/or output to the UE by a user of the UE. Other interfaces of such a UE may be made up of transmitters, receivers, and other circuitry (e.g., other than the transceiver (s) 1410/antenna (s) 1412 already described) that allow for communication between the UE and other devices and may operate according to known protocols (e.g., and the like) .

The wireless device 1402 may include a model transfer and update module 1416. The model transfer and update module 1416 may be implemented via hardware, software, or combinations thereof. For example, the model transfer and update module 1416 may be implemented as a processor, circuit, and/or instructions 1408 stored in the memory 1406 and executed by the processor (s) 1404. In some examples, the model transfer and update module 1416 may be integrated within the processor (s) 1404 and/or the transceiver (s) 1410. For example, the model transfer and update module 1416 may be implemented by a combination of software components (e.g., executed by a DSP or a general processor) and hardware components (e.g., logic gates and circuitry) within the processor (s) 1404 or the transceiver (s) 1410.

The model transfer and update module 1416 may be used for various aspects of the present disclosure, for example, aspects of FIGS. 1-13. The model transfer and update module 1416 is configured to perform procedures to transfer and update an AI/ML model.

The network device 1418 may include one or more processor (s) 1420. The processor (s) 1420 may execute instructions such that various operations of the network device 1418 are performed, as described herein. The processor (s) 1420 may include one or more baseband processors implemented using, for example, a CPU, a DSP, an ASIC, a controller, an FPGA device, another hardware device, a firmware device, or any combination thereof configured to perform the operations described herein.

The network device 1418 may include a memory 1422. The memory 1422 may be a non-transitory computer-readable storage medium that stores instructions 1424 (which may include, for example, the instructions being executed by the processor (s) 1420) . The instructions 1424 may also be referred to as program code or a computer program. The memory 1422 may also store data used by, and results computed by, the processor (s) 1420.

The network device 1418 may include one or more transceiver (s) 1426 that may include RF transmitter circuitry and/or receiver circuitry that use the antenna (s) 1428 of the network device 1418 to facilitate signaling (e.g., the signaling 1434) to and/or from the network device 1418 with other devices (e.g., the wireless device 1402) according to corresponding RATs.

The network device 1418 may include one or more antenna (s) 1428 (e.g., one, two, four, or more) . In embodiments having multiple antenna (s) 1428, the network device 1418 may perform MIMO, digital beamforming, analog beamforming, beam steering, etc., as has been described.

The network device 1418 may include one or more interface (s) 1430. The interface (s) 1430 may be used to provide input to or output from the network device 1418. For example, a network device 1418 that is a base station may include interface (s) 1430 made up of transmitters, receivers, and other circuitry (e.g., other than the transceiver (s) 1426/antenna (s) 1428 already described) that enables the base station to communicate with other equipment in a core network, and/or that enables the base station to communicate with external networks, computers, databases, and the like for purposes of operations, administration, and maintenance of the base station or other equipment operably connected thereto.

The network device 1418 may include a model transfer and update module 1432. The model transfer and update module 1432 may be implemented via hardware, software, or combinations thereof. For example, the model transfer and update module 1432 may be implemented as a processor, circuit, and/or instructions 1424 stored in the memory 1422 and executed by the processor (s) 1420. In some examples, the model transfer and update module 1432 may be integrated within the processor (s) 1420 and/or the transceiver (s) 1426. For example, the model transfer and update module 1432 may be implemented by a combination of software components (e.g., executed by a DSP or a general processor) and hardware components (e.g., logic gates and circuitry) within the processor (s) 1420 or the transceiver (s) 1426.

The model transfer and update module 1432 may be used for various aspects of the present disclosure, for example, aspects of FIGS. 1-13. The model transfer and update module 1432 is configured to perform procedures to transfer and update an AI/ML model.

For one or more embodiments, at least one of the components set forth in one or more of the preceding figures may be configured to perform one or more operations, techniques, processes, and/or methods as set forth herein. For example, a baseband processor as described herein in connection with one or more of the preceding figures may be configured to operate in accordance with one or more of the examples set forth herein. For another example, circuitry associated with a UE, base station, network element, etc. as described above in connection with one or more of the preceding figures may be configured to operate in accordance with one or more of the examples set forth herein.

Any of the above described embodiments may be combined with any other embodiment (or combination of embodiments) , unless explicitly stated otherwise. The foregoing description of one or more implementations provides illustration and description, but is not intended to be exhaustive or to limit the scope of embodiments to the precise form disclosed. Modifications and variations are possible in light of the above teachings or may be acquired from practice of various embodiments.

Embodiments and implementations of the systems and methods described herein may include various operations, which may be embodied in machine-executable instructions to be executed by a computer system. A computer system may include one or more general-purpose or special-purpose computers (or other electronic devices) . The computer system may include hardware components that include specific logic for performing the operations or may include a combination of hardware, software, and/or firmware.

It should be recognized that the systems described herein include descriptions of specific embodiments. These embodiments can be combined into single systems, partially combined into other systems, split into multiple systems or divided or combined in other ways. In addition, it is contemplated that parameters, attributes, aspects, etc. of one embodiment can be used in another embodiment. The parameters, attributes, aspects, etc. are merely described in one or more embodiments for clarity, and it is recognized that the parameters, attributes, aspects, etc. can be combined with or substituted for parameters, attributes, aspects, etc. of another embodiment unless specifically disclaimed herein.

It is well understood that the use of personally identifiable information should follow privacy policies and practices that are generally recognized as meeting or exceeding industry or governmental requirements for maintaining the privacy of users. In particular, personally identifiable information data should be managed and handled so as to minimize risks of unintentional or unauthorized access or use, and the nature of authorized use should be clearly indicated to users.

Although the foregoing has been described in some detail for purposes of clarity, it will be apparent that certain changes and modifications may be made without departing from the principles thereof. It should be noted that there are many alternative ways of implementing both the processes and apparatuses described herein. Accordingly, the present embodiments are to be considered illustrative and not restrictive, and the description is not to be limited to the details given herein, but may be modified within the scope and equivalents of the appended claims.

Claims

A method for a user equipment (UE) , the method comprising:

training a new model for wireless communication with a network node;

sending an indication to the network node comprising a new model identification (ID) associated with the new model;

receiving, from the network node, an uplink grant comprising a request for a model transfer;

sending, to the network node, the new model together with the new model ID; and

receiving, from the network node, a message to activate the new model.
The method of claim 1, wherein the new model comprises a model structure that is known to the network node.
The method of claim 2, wherein the new model ID comprises a version number, and wherein the new model is signaled with a version ID.
The method of claim 2, wherein the new model ID is associated with a previous model ID in the signaling, and wherein the previous model ID implicitly indicates the known model structure.
The method of claim 2, further comprising adding a value tag to a previous model ID to generate the new model ID.
The method of claim 1, further comprising sending, to the network node, meta-information to provide more information on the new model.
The method of claim 1, wherein the new model is an unknown model.
The method of claim 1, further comprising:

training the new model to generate an updated model;

sending, to the network node, a new training dataset for the updated model; and

receiving, from the network node, a second indication to update the new model ID or a dataset ID.
The method of claim 8, wherein when the new model is updated, the UE keeps both an original version of the new model and a sub-model comprising the updated model.
The method of claim 1, further comprising associating a UE side additional condition with the new model ID.
The method of claim 1, further comprising associating a network side additional condition with the new model ID, wherein the new model ID associated with the network side additional condition is part of a data collection procedure.
A method for a network node, the method comprising:

training a new model for wireless communication with a user equipment (UE) ;

sending an announcement to the UE comprising a new model identification (ID) associated with the new model;

receiving, from the UE, a request to download the new model;

sending, to the UE, the new model together with the new model ID;

receiving, from the UE, an indication that the UE is ready to use the new model for inferencing; and

sending, to the UE, a message to activate the new model.
The method of claim 12, wherein the new model comprises a model structure that is known to the network node.
The method of claim 13, wherein the new model ID comprises a version number, and wherein the new model is signaled with a version ID.
The method of claim 13, wherein the new model ID is associated with a previous model ID in the signaling, and wherein the previous model ID implicitly indicates the known model structure.
The method of claim 13, further comprising adding a value tag to a previous model ID to generate the new model ID.
The method of claim 12, further comprising sending, to the UE, meta-information to provide more information on the new model.
The method of claim 12, wherein the new model is an unknown model.
The method of claim 12, further comprising:

training the new model to generate an updated model;

sending, to the UE, a new training dataset for the updated model; and

receiving, from the UE, an indication that the updated model is supported.
The method of claim 19, wherein when the new model is updated, storing both an original version of the new model and a sub-model comprising the updated model.
The method of claim 12, further comprising associating a network side additional condition with the new model ID.
The method of claim 12, further comprising associating a UE side additional condition with the new model ID.
An apparatus comprising means to perform the method of any of claim 1 to claim 21.
A computer-readable media comprising instructions to cause an electronic device, upon execution of the instructions by one or more processors of the electronic device, to perform the method of any of claim 1 to claim 21.
An apparatus comprising logic, modules, or circuitry to perform the method of any of claim 1 to claim 21.