Capability reporting for multi-model artificial intelligence / machine learning user device features

Abstract

Apparatuses, methods, computer programs, computer program products, and computer readable media for capability reporting of multi-model AI / ML UE features are provided. The method includes generating user equipment capability information, including identifying at least one machine learning model available at the user equipment for a predetermined scene, assigning a unique identification to each of the at least one machine learning model, at least one machine learning model having a unique identification is associated with at least one of a parameter list, a data set, and a registration ID, and the generated user equipment capability information is reported to a network entity.

Description

Technical Field

[0001] Various example embodiments are directed to apparatus, methods, systems, computer programs, computer program products, and computer-readable media for capability reporting of multi-model AI / ML UE features.

[0002] abbreviation

[0003] 5G – Fifth Generation

[0004] gNB – 5G / NR base station

[0005] NR – New Radio

[0006] RAN – Radio Access Network

[0007] AI – Artificial Intelligence

[0008] ML – Machine Learning

[0009] UE – User Equipment

[0010] UL – Uplink

[0011] DL – Downlink

[0012] DCI – Downlink Control Information

[0013] MAC CE – Media Access Control Element

[0014] MIMO – Multiple Input Multiple Output

[0015] NN – Neural Network

[0016] CSI – Channel State Information

[0017] SSB – Synchronization Signal Block

[0018] RS – Reference Signal

[0019] TRP – Transmission Point

[0020] Tx – Transmitter

[0021] Rx – Receiver

[0022] RB – Residual Block

[0023] CNN – Convolutional Neural Network Background Art

[0024] Certain aspects of the present invention relate to the Rel-18 Study Item (SI) for Artificial Intelligence (AI) / Machine Learning (ML) for New Radio (NR) air interface (see 3GPP RP-213599).

[0025] The SI aims to explore the benefits of enhancing the air interface with capabilities that support AI / ML-based algorithms to improve performance and / or reduce complexity / overhead. The goal of such considerations is to lay the foundation for future air interface use cases that leverage AI / ML techniques. The initial set of use cases to be covered include CSI feedback enhancement (e.g., overhead reduction, improved accuracy, prediction), beam management (e.g., beam prediction in the time and / or spatial domain for overhead and latency reduction, improved beam selection accuracy), and positioning accuracy enhancement. For those use cases, the benefits should be evaluated (leveraging developed methodologies and defined KPIs) and the potential impact on the specification should be evaluated, both from the PHY layer perspective and from the protocol perspective.

[0026] A key expected outcome of such considerations is that “the AI / ML approaches used for selected sub-use cases need to be sufficiently diverse to support the various requirements for levels of gNB-UE collaboration”.

[0027] It is important to note that additional use cases may also be addressed during the “AI / ML for Air Interface” Work Item (WI) phase. Starting from Release 18, companies are likely to propose a large number of use cases and applications for ML in gNB and UE. The goal is to explore the advantages of enhancing the air interface with the support of implementing AI / ML based algorithms for enhanced performance and / or reduced complexity / overhead. The enhanced performance here depends on the use case considered and can be, for example, improved throughput, robustness, accuracy or reliability. The goal is that enough use cases will be considered to enable the identification of a common AI / ML framework, including functional requirements of the AI / ML architecture, which can be used for subsequent projects. The study should also identify areas where AI / ML can improve the performance of air interface functions. The regulatory impact will be evaluated in order to improve the overall understanding of what will be required to enable AI / ML technologies for air interfaces.

[0028] Generalizing AI / ML models to cover different scenarios / configurations is seen as a major challenge that RAN 1 should investigate, and there are different approaches to address this problem. It is unlikely that a single ML model will ultimately be necessary and sufficient for each and every deployment scenario. For example, developing a “universal” ML model for beam prediction in the spatial domain that performs well across different scenarios / configurations / settings / parameters may be difficult as its performance should be validated against data used for each and every deployment scenario. Additionally, having a universal model may have high training / inference complexity and may also have performance issues if the model does not perform well in all scenarios. Summary of the invention

[0029] Various example embodiments are directed to addressing at least some of the above-mentioned problems and / or difficulties and disadvantages.

[0030] It is an object of various example embodiments to provide apparatus, methods, systems, computer programs, computer program products and computer readable media for capability reporting for multi-model AI / ML UE features.

[0031] According to one aspect of various example embodiments, there is provided a method for use in a user equipment, comprising:

[0032] Generate user equipment capability information, including

[0033] Identifying at least one machine learning model that can be used for a predetermined scenario at a user device;

[0034] assigning a unique identifier to each of the at least one machine learning model;

[0035] Associating at least one machine learning model having the unique identifier with at least one of a parameter list, a data set, and a registration ID; and

[0036] The generated user equipment capability information is reported to a network entity.

[0037] According to another aspect of various example embodiments, there is provided an apparatus for use in a user equipment, comprising:

[0038] The component used to generate user equipment capability information includes:

[0039] means for identifying at least one machine learning model that can be used for a predetermined scenario at a user device,

[0040] means for assigning a unique identifier to each of the at least one machine learning model,

[0041] A component for associating at least one machine learning model having the unique identifier with at least one of a parameter list, a data set, and a registration ID, and a component for reporting the generated user equipment capability information to a network entity.

[0042] According to another aspect of the invention, there is provided a computer program product comprising code means adapted to produce the steps of any of the methods described above when loaded into a memory of a computer.

[0043] According to another aspect of the invention, there is provided a computer program product as defined above, wherein the computer program product comprises a computer readable medium on which the software code is partly stored.

[0044] According to another aspect of the invention, there is provided a computer program product as defined above, wherein the program is directly loadable into an internal memory of a processing device.

[0045] According to one aspect of various exemplary embodiments, there is provided a computer readable medium storing a computer program as described above.

[0046] According to an exemplary aspect, a computer program product is provided, comprising a computer executable computer program code, which, when the program is executed on a computer (e.g., a computer of an apparatus according to any one of the above-mentioned apparatus-related exemplary aspects of the present disclosure), is configured to cause the computer to execute a method according to any one of the above-mentioned method-related exemplary aspects of the present disclosure.

[0047] Such a computer program product may include (or be embodied as) a (tangible) computer-readable (storage) medium storing computer-executable computer program code, and / or the program may be directly loaded into the internal memory of a computer or its processor.

[0048] Further aspects and features of the invention are set out in the dependent claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] These and other objects, features, details and advantages will become more apparent from the following detailed description of various aspects / embodiments which will be described in conjunction with the accompanying drawings, in which:

[0050] Figure 1 is a sequence diagram illustrating an example of a signaling procedure according to certain aspects of the present invention;

[0051] Figure 2 is a flow chart illustrating an example of a method according to certain aspects of the present invention.

[0052] Figure 3 is a block diagram illustrating another example of an apparatus according to certain aspects of the present invention.

[0053] Figure 4 is a block diagram illustrating another example of an apparatus according to certain aspects of the present invention. DETAILED DESCRIPTION

[0054] The present disclosure is described herein with reference to specific non-limiting examples and embodiments that are presently considered to be conceivable. It will be appreciated by those skilled in the art that the present disclosure is by no means limited to these examples and can be applied more broadly.

[0055] It should be noted that the following description of the present disclosure and its embodiments mainly relates to specifications used as non-limiting examples for certain exemplary network configurations and deployments. That is, the present disclosure and its embodiments are mainly described with respect to 3GPP specifications used as non-limiting examples for certain exemplary network configurations and deployments. Therefore, the description of the example embodiments given herein particularly relates to terms directly related thereto. Such terms are used only in the context of the non-limiting examples presented and naturally do not limit the present disclosure in any way. On the contrary, any other communication or communication-related system deployment, etc., may also be utilized as long as it meets the features described herein.

[0056] In the following, various embodiments and implementations of the present disclosure and aspects or embodiments thereof are described using several variants and / or alternatives. It should generally be noted that, depending on certain requirements and constraints, all described variants and / or alternatives may be provided individually or in any possible combination (also including combinations of individual features of various variants and / or alternatives).

[0057] According to certain aspects of the present invention, it is proposed to use multiple ML models for the same scenario (e.g., CSI compression, beam prediction, positioning, etc.), where the model can change depending on the situation / scenario / configuration. A scenario refers to at least one of a feature / sub-use case / use case. The above problems can be avoided by using multiple ML models for the same scenario. Therefore, it must be specified how to define multiple ML models for each specific scenario in the NR framework.

[0058] When more than one model can be used / selected for model inference on the UE side for a given scenario, the UE capability report (UE characteristics) shall take into account the following:

[0059] UE indicates the number of ML models (e.g., X=4, 6, 8) that the UE can support for the same functionality (e.g., CSI compression with a two-sided model, beam prediction in the spatial domain with a one-sided model, positioning estimation in NLOS scenarios with a one-sided model)

[0060] UE indicates the ML model ID (a number that uniquely identifies different models supporting the same function. For example, when X=4, ID=1, 2, 3, 4) and indicates the association between the ML model ID and the parameter list and / or data set and / or registration ID, where

[0061] ● Parameter lists defining radio conditions (deployment type, applicable scenarios, gNB / UE antenna configurations, clutter parameters, etc.) and / or ML specific details (reportable quantities, required / associated measurement configurations, required / associated auxiliary information, inputs / outputs and dimensions of the ML model) and / or limitations / conditions (inference latency, required warm-up time, fine-tuning requirements)

[0062] ● A dataset refers to a version of a dataset that is accessible to both the UE and the network through other means (operator-controlled servers, proprietary cloud), and

[0063] The network learns aspects related to the model based on the dataset associated with the model ID.

[0064] ● Registration ID refers to a unique version of an ML model with a unique identifier. Here, the trained model may be located in an operator-controlled server or a proprietary server, where the network may not have full access to the model. However, the network can understand the details / parameters associated with the model by referring to the registration ID.

[0065] UE indicates whether the model can be switched from one model to another

[0066] • If the switching is dynamic, the UE may also indicate any relevant delay considerations for switching from one model to another.

[0067] ●In one variant, model switching between certain model combinations may not be supported, while in certain other combinations, switching may be supported.

[0068] UE indicates whether the model can be disabled and the possibility to fall back to the parametric model

[0069] • If fallback is supported, the UE may also indicate any related delay considerations for switching and parameter lists associated with parameter models (e.g., for CSI compression, reporting based on Type I or Type II codebooks may be considered as parameter models).

[0070] • The UE indicates whether more than one model can be active at a given time.

[0071] ● If yes, the UE also reports the number of models supported in parallel, which model IDs can be supported in parallel, and any associated considerations / limitations for applying parallel operations of the ML models

[0072] ● In one example, for inter-cell beam prediction in the spatial domain, beam prediction may use multiple models in parallel, where each model may be a cell-specific model. For each of these cell-specific models, the UE measures the beam corresponding to that cell and uses these measurements at the input of the model, and the model output provides the best predicted beam for the same cell.

[0073] Based on the received UE capability information, the network configures the ML model parameters to decide /

[0074] Model switching is supported, or more than one model is considered to be activated at a given time, for a given feature related support towards the UE.

[0075] In the following, a specific example of the above aspect will be described by considering a beam management case (spatial domain beam prediction) as an example.

[0076] Specifically, according to the specific example,

[0077] • The UE indicates the number of ML models (e.g., X=4, 6, 8) that the UE can support for beam prediction in the spatial domain.

[0078] UE indicates the ML model ID (for X=4, ID=1, 2, 3, 4) and indicates the association between the ML model ID and the parameter list

[0079] ●Model ID 1 – Frequency Range 2 (FR2), Subcarrier Spacing (SCS) 120

[0080] kHz, dense urban scenarios, base station (BS) antenna configuration (one panel: (M, N, P, Mg, Ng) = (4, 8, 2, 1, 1), (dV, dH) = (0.5, 0.5)λ), set A dimension (64 beams, layout M1), set B dimension (4 or 8 or 16 beams, layout N1), auxiliary information (NULL), reportable quantity (beam ID, L1-RSRP), etc.

[0081] ●Model ID 2–FR2, SCS 60kHz, indoor hotspot scenario, BS antenna configuration (one panel: (M, N, P, Mg, Ng) = (2, 2, 2, 1, 1), (dV, dH) = (0.5, 0.5)λ), set A dimension (32 beams, layout M2), set B dimension (8 or 16 beams, layout N2), auxiliary information (beam angle), reportable quantity (beam ID), etc.

[0082] UE indication model can switch from one model to another

[0083] ● Model switching is dynamic. Multiple combinations (combination 1, 2, ...) and related parameters for switching can be provided.

[0084] UE indication model can be disabled and fallback to parameter model can be performed

[0085] • The delay considerations and parameter list for handover associated with the parameter model may also be indicated by the UE.

[0086] • The UE indicates that no more than one model can be activated at a given time.

[0087] Based on the received UE capability information, the network configures the ML model related parameters to decide /

[0088] Model switching is supported, or allowing for more than one model to be activated at a given time for a given feature supported by the UE.

[0089] The above multiple ML models for the same scenario (e.g., use case feature group) can be indicated in the form of the following table:

[0090]

[0091] Table 1: Coexistence and switching of multiple model IDs

[0092] Table 1 discusses how the UE reports coexistence and handover when using multiple ML models for the same scenario. In the example above, the UE supports four such ML models, with ML model IDs 1 to 4 for different parameter lists (optionally providing a registration ID (global or vendor specific ID) and a dataset ID used for training). Combinations refer to potential combinations of ML models that are allowed to operate together. For example, combination 1 indicates that, for example, when ML models 1 to 3 are configured by the network in a specific band combination, they can operate together (model ID 4 in combination 1 cannot be configured to the UE). Table 2 below further describes how the network should be interpreted by the network for configuration purposes.

[0093]

[0094] Table 2: Description of each combination

[0095] Figure 1 is a sequence diagram illustrating an example of a signaling procedure according to certain aspects of the present invention.

[0096] exist Figure 1 In step S11, the network triggers the acquisition of ML model ID combinations for the same scenario (e.g., use case). That is, the network will request the UE to provide information about ML models that are available for a specific scenario at the UE and can be combined.

[0097] Therefore, the network sends a corresponding request to the user equipment in step S12, which request indicates, among other things, the available ML models and the scenarios for which the combination should be provided. Optionally, for example, a registration ID is requested to be provided by the user equipment.

[0098] In step S13, the UE identifies one or more machine learning models that are available at the user equipment for the predetermined scenario contained in the request received in step S12. The user equipment then assigns a unique identifier to each of the one or more machine learning models. In addition, the user equipment associates the one or more machine learning models with the unique identifier with at least one of the parameter list, the data set, and the optional registration ID. That is, the user equipment generates the user equipment capability information and compiles the ML model ID combination list.

[0099] Then, in step S14, the user equipment reports the generated user equipment capability information including available ML models and combinations to the network.

[0100] The network stores available ML models and combinations indicated by ML model IDs and configures ML model ID combinations.

[0101] The configuration is then sent to the user equipment in step S16.

[0102] The user equipment acknowledges receipt of the configured ML model ID combination in step S17 and starts to act according to the configured ML model ID combination received from the network.

[0103] In the following, a more general description of an example version of the invention is made with reference to Figures 2 to 4 conduct.

[0104] Figure 2 is a flow chart illustrating examples of methods according to some example versions of the invention.

[0105] According to an example version of the present invention, the method may be implemented in a user device or the like, or may be a part of a user device or the like. The method includes generating user device capability information in step S21. Generating user device capability information in step S21 includes identifying at least one machine learning model that can be used for a predetermined scenario at the user device, assigning a unique identifier to each of the at least one machine learning model, and associating at least one machine learning model having a unique identifier with at least one of a parameter list, a data set, and a registration ID.

[0106] In addition, the method includes reporting the generated user equipment capability information to a network entity in step S22. The network entity may be a base station, such as a gNB.

[0107] According to some example versions of the invention, the method also includes receiving a request from a network entity indicating a predetermined scenario for which the machine learning model is to be identified.

[0108] According to some example versions of the invention, the method also includes receiving, from a network entity, a configuration regarding a machine learning model to be used for a predetermined scenario based on the reported user equipment capability information, and acting according to the received configuration.

[0109] According to some example versions of the invention, the user device capability information includes an indication of whether the identified machine learning model can be switched from one model to another when applied to a predetermined scenario.

[0110] According to some example versions of the invention, the user device capability information includes: an indication of whether more than one of the identified machine learning models can be activated at a given time when applied to a predetermined scenario.

[0111] According to some example versions of the invention, the user device capability information includes an indication of whether the identified machine learning model can be disabled to switch to a parameter model when applied to a predetermined scenario.

[0112] According to some example versions of the invention, the parameter list defines at least one of the following items: radio conditions, including deployment type, applicable scenarios, base station / user equipment antenna configuration and at least one of clutter parameters, machine learning specific details, including at least one quantity that can be reported, required measurement configuration, required auxiliary information, inputs / outputs and dimensions of the machine learning model, and limitations / conditions, including at least one inference delay, required warm-up time, and fine-tuning requirements.

[0113] According to some example versions of the invention, a dataset refers to a version of a dataset that is accessible to both user devices and networks through other means including operator-controlled servers and / or private clouds, and the dataset specifies model-related aspects.

[0114] According to some example versions of the invention, a registration ID refers to a unique version of a machine learning model having a unique identifier.

[0115] Figure 3 is a block diagram illustrating another example of an apparatus according to some example versions of the invention.

[0116] exist Figure 3 In FIG. 1 , a circuit block diagram showing the configuration of a device 30 is shown, and the device 30 is configured to implement various aspects of the present invention described above. It should be noted that Figure 3The device 30 shown in the figure may include several additional elements or functions in addition to the elements or functions described below, which are omitted for simplicity because they are not essential for understanding the present invention. In addition, the device may also be another device with similar functions, such as a chipset, a chip, a module, etc., which may also be part of the device or attached to the device as a separate element, etc.

[0117] The apparatus 30 may include a processing function or processor 31, such as a CPU or the like, which executes instructions given by a program or the like. The processor 31 may include one or more processing parts dedicated to a specific process as described below, or the process may be run in a single processor. For example, the part for performing such a specific process may also be provided as a discrete element, or within one or more processors or processing parts, such as in one physical processor (such as a CPU) or in multiple physical entities. Reference symbol 32 represents a transceiver or input / output (I / O) unit (interface) connected to the processor 31. The I / O unit 32 may be used to communicate with one or more other network elements, entities, terminals, etc. The I / O unit 32 may be a combined unit including communication devices for multiple network elements, or may include a distributed structure with multiple different interfaces for different network elements. The apparatus 30 also includes at least one memory 33, which may be used to store data and programs to be executed by the processor 31 and / or as a working storage for the processor 31, for example.

[0118] The processor 31 is configured to perform processing related to the above-mentioned aspects.

[0119] Specifically, the apparatus 30 may be implemented in a user device or may be a part of a user device, and may be configured to perform the following steps in conjunction with Figure 2 Description of the process.

[0120] Therefore, according to some example versions of the present invention, there is provided an apparatus 30 for use in a user device, comprising at least one processor 31 and at least one memory 33 for storing instructions to be executed by the processor 31, wherein the at least one memory 33 and the instructions are configured, together with the at least one processor 31, so that the apparatus 30 at least performs the generation of user device capability information, including: identifying at least one machine learning model that can be used for a predetermined scenario at the user device, assigning a unique identifier to each of the at least one machine learning model, associating at least one machine learning model with a unique identifier with at least one of a parameter list, a data set, and a registration ID, and reporting the generated user device capability information to a network entity.

[0121] In addition, the present invention can also be implemented by an apparatus for a user equipment, including components for performing the above-mentioned processing, such as Figure 4 as shown in .

[0122] That is, according to some example versions of the present invention, such as Figure 4 As shown in , the apparatus for use in a user equipment includes a component 41 for generating user equipment capability information. The component 41 for generating includes a component 411 for identifying at least one machine learning model that can be used for a predetermined scenario at the user equipment, a component 412 for assigning a unique identifier to each of the at least one machine learning model, and a component 413 for associating at least one machine learning model with a unique identifier with at least one of a parameter list, a data set, and a registration ID. In addition, the apparatus includes a component 42 for reporting the generated user equipment capability information to a network entity.

[0123] In addition, according to some example versions of the present invention, a computer program is provided, comprising instructions, which, when executed by an apparatus used in a user device, causes the apparatus to generate user device capability information, including identifying at least one machine learning model that can be used for a predetermined scenario at the user device, assigning a unique identifier to each of the at least one machine learning model, associating at least one machine learning model with the unique identifier with at least one of a parameter list, a data set, and a registration ID, and reporting the generated user device capability information report to a network entity.

[0124] The computer program product may comprise code means adapted to produce the steps of any of the methods described above when loaded into the memory of a computer.

[0125] According to some example versions of the invention, there is provided a computer program product as defined above, wherein the computer program product comprises a computer readable medium storing software code portions.

[0126] According to some example versions of the invention, there is provided a computer program product as defined above, wherein the program is directly loadable into an internal memory of a processing device / apparatus.

[0127] According to some example versions of the invention, there is provided a computer readable medium storing a computer program as described above.

[0128] According to some example versions of the present invention, a computer program product is provided, comprising a computer executable computer program code, which, when the program is run on a computer (e.g., a computer of a device according to any one of the above-mentioned device-related exemplary aspects of the present disclosure), is configured to cause the computer to execute a method according to any one of the above-mentioned method-related exemplary aspects of the present disclosure.

[0129] Such a computer program product may include (or be embodied as) a (tangible) computer-readable (storage) medium or the like on which a computer-executable computer program code is stored, and / or the program may be directly loaded into the internal memory of a computer or its processor.

[0130] Furthermore, the present invention may be implemented by an apparatus for use in a user equipment, the user equipment comprising a corresponding circuit system for performing the above-mentioned processing.

[0131] That is, according to some example versions of the present invention, there is provided an apparatus for use in a user device, comprising a generating circuit system for generating user device capability information. The generating circuit system comprises an identification circuit system for identifying at least one machine learning model that can be used for a predetermined scenario at the user device, an allocation circuit system for assigning a unique identifier to each of the at least one machine learning model, an association circuit system for associating at least one machine learning model having a unique identifier with at least one of a parameter list, a data set, and a registration ID, and a reporting circuit system for reporting the generated user device capability information to a network entity.

[0132] For more details about the functions of the apparatus and computer program, please refer to the above description of the method according to some example versions of the present invention, such as in conjunction with Figure 2 As described in.

[0133] In the exemplary description of the aforementioned device, only the units / components relevant to understanding the principles of the present invention have been described using functional blocks. The device may include other units / components necessary for its corresponding operation accordingly. However, the description of these units / components is omitted in this specification. The arrangement of the functional blocks of the device should not be interpreted as limiting the present invention, and the functions can be performed by one block or further split into sub-blocks.

[0134] When it is stated in the foregoing description that a device (or some other component) is configured to perform certain functions, this should be interpreted as equivalent to the following description: (i.e., at least one) processor or corresponding circuit system (possibly in cooperation with computer program code stored in the memory of the corresponding device) is configured to cause the device to perform at least the above functions. In addition, such functions should be interpreted as being equivalently implementable by specially configured circuit systems or components for performing the corresponding functions (i.e., the expression "the unit is configured to" should be interpreted as equivalent to expressions such as "components for").

[0135] As used in this application, the term "circuitry" may refer to one or more or all of the following:

[0136] (a) implemented solely in hardware circuitry (e.g., implemented solely in analog and / or digital circuitry) and

[0137] (b) a combination of hardware circuitry and software such as (where applicable):

[0138] (i) Combination of analog and / or digital hardware circuits and software / firmware

[0139] and

[0140] (ii) any portion of hardware processor(s) with software (including digital signal processor(s), software and memory(s) that work together to enable a device (such as a mobile phone or server) to perform various functions and

[0141] (c) Hardware circuits and / or processor(s), such as microprocessor(s) or portion(s) of microprocessor(s), require software (e.g., firmware) to operate, but the software may not be present when the software is not required for operation.

[0142] This definition of circuitry applies to all uses of the term in this application, including in any claims. As a further example, as used in this application, the term "circuitry" also covers an implementation of only a hardware circuit or processor (or multiple processors) or a portion of a hardware circuit or processor and its (or their) accompanying software and / or firmware. The term "circuitry" also covers (for example, if applicable to a particular claim element) a baseband integrated circuit or processor integrated circuit for a mobile device or a similar integrated circuit in a server, cellular network device, or other computing or network device.

[0143] For the purpose of the present invention as described above, it should be noted that:

[0144] - the method steps which may be implemented as software code portions and executed using a processor at a device (as an example of an apparatus, a device and / or its modules, or as an example of an entity comprising a device and / or a module) are independent of the software code and may be specified using any known or future developed programming language as long as the functionality defined by the method steps is preserved;

[0145] - Generally, any method step is suitable for implementation in the form of software or hardware without changing the idea of ​​the aspect / embodiment and its modification in terms of the implemented functions;

[0146] - the method steps and / or devices, units or parts that may be implemented as hardware components at the above-defined device or any (multiple) modules thereof (e.g., devices that perform the functions of the device according to the above-mentioned aspects / embodiments) are hardware-independent and can be implemented using any known or future developed hardware technology or any hybrid technology of these technologies, such as MOS (metal oxide semiconductor), CMOS (complementary MOS), BiMOS (bipolar MOS), BiCMOS (bipolar CMOS), ECL (emitter coupled logic), TTL (transistor-transistor logic), etc., for example using ASIC (application specific IC (integrated circuit)) components, FPGA (field programmable gate array) components, CPLD (complex programmable logic device) components, APU (accelerated processor unit), GPU (graphics processor unit) or DSP (digital signal processor) components;

[0147] - devices, units or components (e.g., the means defined above, or any of their corresponding units / components) may be implemented as separate devices, units or components, but this does not exclude that they are implemented in a distributed manner throughout the system, as long as the functionality of the devices, units or components is preserved;

[0148] - a device may be represented by a semiconductor chip, a chipset or a (hardware) module comprising such a chip or chipset; however, this does not exclude the possibility that the functionality of the device or module is not implemented by hardware but as software in a (software) module, such as a computer program or computer program product comprising executable software code portions for execution / running on a processor; - a device may be considered as a device or as a component of more than one device, for example, whether they functionally cooperate with each other or are functionally independent of each other but located in the same device housing.

[0149] In general, it should be noted that the corresponding functional blocks or elements of the above aspects can be implemented in hardware and / or software form by any known components as long as they are suitable for performing the functions of the corresponding parts. The method steps can be implemented in separate functional blocks or by separate devices, or one or more method steps can be implemented in a single functional block or by a single device.

[0150] Generally, any method step is suitable for being implemented as software or by hardware without changing the idea of ​​the invention. Devices and components can be implemented as separate devices, but this does not exclude that they are implemented in a distributed manner in the entire system, as long as the functionality of the device is preserved. Such and similar principles should be considered to be known to those skilled in the art.

[0151] The software in this specification includes software code, such as a computer program or computer program product including code components or portions or for performing corresponding functions, and software (or a computer program or a computer program product) embodied on a tangible medium (such as a computer-readable (storage) medium) having a corresponding data structure or code components / portions stored thereon, or embodied in a signal or chip (possibly during its processing).

[0152] It should be noted that the above aspects / embodiments and general and specific examples are for illustrative purposes only and are by no means intended to limit the present invention thereto. On the contrary, it is intended to cover all changes and modifications falling within the scope of the appended claims.

Claims

1. A method for a user device, comprising: Generate user equipment capability information, including: identifying at least one machine learning model that can be used for a predetermined scenario at the user device, assigning a unique identifier to each of the at least one machine learning model, associating the at least one machine learning model having the unique identifier with at least one of: a parameter list, a data set, and a registration ID, and The generated user equipment capability information is reported to a network entity.

2. The method according to claim 1, further comprising: A request is received from the network entity indicating the predetermined scenario for which the machine learning model is to be identified.

3. The method according to claim 1 or 2, further comprising: receiving, from the network entity, configuration regarding: the machine learning model to be used for the predetermined scenario based on the reported user equipment capability information, and Act according to the received configuration.

4. The method according to any one of claims 1 to 3, wherein: The user device capability information includes: an indication of whether the identified machine learning model can be switched from one model to another when applied to the predetermined scenario.

5. The method according to any one of claims 1 to 4, wherein: The user device capability information includes: an indication of whether more than one of the identified machine learning models can be activated at a given time when applied to the predetermined scenario.

6. The method according to any one of claims 1 to 5, wherein: The user device capability information includes: an indication of whether the identified machine learning model can be disabled to switch to a parameter model when applied to the predetermined scenario.

7. The method according to any one of claims 1 to 6, wherein The parameter list defines at least one of the following: Radio conditions, including at least one of the following: deployment type, applicable scenario, base station / user equipment antenna configuration, and clutter parameters; Machine learning specific details, including: at least one reportable quantity, required measurement configuration, required auxiliary information, input / output and dimensionality of the machine learning model; as well as Constraints / conditions, including at least one inference latency, required warm-up time, and fine-tuning requirements.

8. The method according to any one of claims 1 to 7, wherein: The dataset refers to a version of the dataset, the dataset is accessible to both the user equipment and the network through other means including operator-controlled servers and / or a proprietary cloud, and the dataset specifies model-related aspects.

9. The method according to any one of claims 1 to 8, wherein: The registration ID refers to a unique version of the machine learning model with a unique identifier.

10. An apparatus comprising: The component used to generate user equipment capability information includes: means for identifying at least one machine learning model that can be used for a predetermined scenario at the user equipment, means for assigning a unique identifier to each of the at least one machine learning model, means for associating the at least one machine learning model having the unique identifier with at least one of: a parameter list, a data set, and a registration ID, and Means for reporting the generated user equipment capability information to a network entity.

11. The device according to claim 10, comprising means for: A request is received from the network entity indicating the predetermined scenario for which the machine learning model is to be identified.

12. The device according to claim 10 or 11, comprising means for: receiving, from the network entity, configuration regarding: the machine learning model to be used for the predetermined scenario based on the reported user equipment capability information, and Act according to the received configuration.

13. The device according to any one of claims 10 to 12, wherein: The user device capability information includes: an indication of whether the identified machine learning model can be switched from one model to another when applied to the predetermined scenario.

14. The device according to any one of claims 10 to 13, wherein: The user device capability information includes: an indication of whether more than one of the identified machine learning models can be activated at a given time when applied to the predetermined scenario.

15. The device according to any one of claims 10 to 14, wherein: The user device capability information includes: an indication of whether the identified machine learning model can be disabled to switch to a parameter model when applied to the predetermined scenario.

16. The device according to any one of claims 10 to 15, wherein: The parameter list defines at least one of the following: Radio conditions, including at least one of the following: deployment type, applicable scenario, base station / user equipment antenna configuration, and clutter parameters; Machine learning specific details, including: at least one reportable quantity, required measurement configuration, required auxiliary information, inputs / outputs and dimensions of the machine learning model; as well as Constraints / conditions, including at least one inference latency, required warm-up time, and fine-tuning requirements.

17. The device according to any one of claims 10 to 16, wherein The dataset refers to a version of the dataset, the dataset is accessible to both the user equipment and the network through other means including operator-controlled servers and / or a proprietary cloud, and the dataset specifies model-related aspects.

18. The device according to any one of claims 10 to 17, wherein: The registration ID refers to a unique version of the machine learning model with a unique identifier.

19. A computer program comprising instructions which, when executed by a device, cause the device to perform the method according to any one of claims 1 to 9.

Images