Method and system for context-aware reporting of AIML model inapplicability
The method and system for context-aware reporting of AIML model inapplicability by UE using a hierarchical bitmap addresses the lack of granular reporting, enabling network entities to optimize resource allocation and reduce wastage by providing detailed information on model inapplicability.
Patent Information
- Application Number
- PCT/EP2026/063215
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2025-05-08
- Filing Date
- 2026-05-07
- Publication Date
- 2026-11-12
AI Technical Summary
Existing wireless communication systems lack a mechanism for user equipment (UE) to efficiently and granularly report the root cause of Artificial Intelligence/Machine Learning (AIML) model inapplicability due to dynamic changes, leading to resource wastage when the network remains unaware of the reason for inapplicability.
A method and system for context-aware reporting of AIML model inapplicability by user equipment (UE) using a hierarchical bitmap to provide scope, entity side, and cause indicators, allowing network entities to dynamically manage resources based on granular information.
Enables network entities to optimize resource allocation by differentiating between temporary and persistent limitations, reducing unnecessary resource allocation, and enhancing system efficiency and responsiveness.
Smart Images

Figure EP2026063215_12112026_PF_FP_ABST
Abstract
Description
202502593DESCRIPTIONTITLEMethod and System for Context-Aware Reporting of AIML Model InapplicabilityTECHNICAL FIELD
[0001] The present disclosure relates generally to wireless communication systems. More specifically, the disclosure pertains to methods and systems for managing resources for Artificial Intelligence / Machine Learning (AIML) models, and in particular, to techniques for a user equipment to report the cause of AIML model inapplicability to a network entity.BACKGROUND OF THE INVENTION
[0002] Wireless communication systems, such as those based on 3rd Generation Partnership Project (3GPP) standards, are increasingly employing Artificial Intelligence and Machine Learning (AIML) to optimize network performance and manage resources efficiently. In many scenarios, AIML models are deployed on a User Equipment (UE) to perform tasks such as mobility prediction or radio resource management.
[0003] Several approaches for configuring and managing such models are known in the art. For example, US 2022 / 0400373 A1 discloses a method where a user equipment transmits its machine learning (ML) capabilities to a base station and receives ML configuration information in return, indicating a neural network function and a corresponding ML model. As another example, US 2022 / 0337487 A1 discloses a network entity that determines model parameters for digitally analyzing input data by executing a training algorithm, searching a local database, or requesting the parameters from another network entity. Furthermore, WO 2022 / 161624 A1 discloses a method for identifying and selecting a candidate ML model based on a request that specifies output features and input data types.
[0004] While these and other documents address the initial configuration, selection, and management of AIML models, they do not solve the specific technical problem that202502593arises when a model, previously reported as applicable, becomes inapplicable due to dynamic changes at the UE. Specifically, the prior art lacks a mechanism for the UE to efficiently and granularly report the root cause of such inapplicability to the network.
[0005] In a typical operation, a UE may report to the network that a specific Al ML model is available and applicable for use. Based on this report, the network allocates resources and expects the UE to perform model inference. However, after this initial report, the state of the UE can change. The AIML model may become inapplicable due to various UE-side conditions, such as low battery, a full data buffer preventing new inputs, or an unavailable computation core that is occupied by other processes.
[0006] Current mechanisms for reporting model applicability, as discussed within 3GPP, generally only allow for a binary indication — that is, whether a model is applicable or not. This lack of granularity provides no visibility into the root cause of the inapplicability. As a result, when a model on the UE side becomes inapplicable, the network remains unaware of the reason. The network will continue to expect model inference from the UE and will keep the corresponding network resources allocated, leading to significant resource wastage.
[0007] Therefore, there exists a need for an improved method and system that enables a UE to provide the network with context-aware information regarding the reasons for AIML model inapplicability, allowing the network to optimize its resource allocation dynamically.SUMMARY OF THE INVENTION
[0008] The present disclosure provides a solution to the aforementioned problems by introducing methods, a system, a user equipment, a network entity, and a computer program product for context-aware reporting of AIML model inapplicability with increased granularity and contextual information. The object is to enable a network entity to dynamically and efficiently manage resources based on granular information provided by a user equipment. The object is achieved by a method, a system, a user equipment, a network entity, and a computer program product as defined in the appended claims.202502593
[0009] According to a first aspect of the disclosure, a method for reporting the inapplicability of an Artificial Intelligence / Machine Learning, AIML, model in a wireless communication network is provided. The method is performed by a user equipment, UE, and comprises providing information to a network entity. The information indicates a cause for an inapplicability of the AIML model and comprises at least two of a scope indicator for indicating a duration of the inapplicability, an entity side indicator for indicating an origin of the inapplicability, and a cause indicator for indicating a cause for the inapplicability.
[0010] In some embodiments of the first aspect, the information is provided in the form of a hierarchical bitmap. This bitmap may be dynamically updated based on real-time conditions of the UE. The scope indicator may be represented by a 1 -bit value, the entity side indicator by a 2-bit value, and the cause indicator by a 4-bit value.
[0011] In a further embodiment, the scope indicator indicating the duration of the inapplicability as one of temporary or permanent.
[0012] In another embodiment, the entity side indicator indicating the origin of the inapplicability as being selected from a group comprising: a UE-side issue, a networkside issue, and a metadata or configuration issue.
[0013] In yet another embodiment, the cause indicator provides a cause for inapplicability selected from a group comprising: low battery, full data buffer, unavailable computation core, model not provisioned, and model expired.
[0014] In some embodiments, the step of providing the information may comprise signaling the information using Layer 1 (L1 ), Layer 2 (L2), or Layer 3 (L3) of the wireless communication protocol stack. Layer 1 refers to the Physical Layer (PHY), used for fast, time-critical control information. Layer 2 refers to the Data Link Layer, which includes the Medium Access Control (MAC) sublayer, and offers a balance of speed and information capacity. Layer 3 refers to the Network Layer, primarily associated with the Radio Resource Control (RRC) protocol, and is used for more complex configuration procedures.202502593
[0015] According to a second aspect of the disclosure, a method for managing resources for an Artificial Intelligence / Machine Learning, AIML, model in a wireless communication network is provided. The method is performed by a network entity and comprises receiving information from a user equipment, UE, wherein the information indicates an inapplicability of the AIML model and comprises at least two of a scope indicator, an entity side indicator, and a cause indicator as described above. The method further comprises adjusting, by the network entity, at least one parameter associated with the AIML model based on the received information. The proposed solution enables: improved network-side decision making, reduced unnecessary resource allocation, improved responsiveness to UE conditions, enhanced overall system efficiency.
[0016] According to a second aspect, a network entity receives the information and adapts at least one network operation based on the reported inapplicability. In some embodiments of the second aspect, adjusting the at least one parameter comprises disabling or reconfiguring one or more UE-side model parameters to optimize resource allocation.
[0017] The method may further comprise storing, by the network entity, the received information as historical data to predict future inapplicability conditions. In a preferred embodiment, the network entity is a next-generation NodeB, gNB.
[0018] According to a third aspect of the disclosure, a system for managing resources for an AIML model in a wireless communication network is provided. The system comprises a user equipment, UE, configured to perform the method according to the first aspect, and a network entity configured to perform the method according to the second aspect.
[0019] According to further aspects, the disclosure provides a user equipment, UE, comprising a processor and memory configured to perform the method of the first aspect, and a network entity comprising a processor and memory configured to perform the method of the second aspect.202502593
[0020] A still further aspect of the disclosure is a computer program product comprising instructions which, when the program is executed by a computer, cause the computer to perform the method according to the first and / or second aspects.BRIEF DESCRIPTION OF THE DRAWINGS
[0021] Further features and advantages of the present disclosure will become apparent from the following detailed description, taken in combination with the appended drawings, in which:
[0022] FIG. 1 is a schematic block diagram illustrating a wireless communication system according to an embodiment of the present disclosure.
[0023] FIG. 2 is a flowchart illustrating a method performed by a user equipment for reporting AIML model inapplicability, according to an embodiment of the present disclosure.
[0024] FIG. 3 is a flowchart illustrating a method performed by a network entity for managing resources based on the reported inapplicability, according to an embodiment of the present disclosure.DETAILED DESCRIPTION OF THE DRAWINGS
[0025] The following detailed description is provided to enable any person skilled in the art to make and use the disclosed subject matter. The description sets forth various embodiments, examples, and specific details to provide a thorough understanding of the disclosed concepts. However, it will be apparent to those skilled in the art that the disclosed subject matter may be practiced without some or all of these specific details, or with variations thereof.
[0026] The embodiments described herein are to be considered illustrative and not restrictive. Accordingly, the scope of protection is not limited to the disclosed embodiments, but is defined by the appended claims and their equivalents.202502593
[0027] In the drawings, like reference numerals refer to like elements throughout. The drawings are schematic and not necessarily to scale, and are intended to illustrate the general principles of the disclosed subject matter rather than depict actual implementations.
[0028] As used herein, the terms “comprising,” “including,” “having,” and similar terms are intended to be open-ended, meaning “including but not limited to,” unless expressly stated otherwise. The use of singular terms such as “a,” “an,” and “the” includes plural references unless the context clearly dictates otherwise.
[0029] Where a feature is described in connection with one embodiment, it may also be used in combination with other embodiments, unless explicitly stated otherwise or technically infeasible.
[0030] The order of steps in any method disclosed herein is not intended to be limiting unless explicitly stated. Steps may be performed in a different order or concurrently, where technically appropriate.
[0031] Furthermore, the terms “first,” “second,” and similar designations are used for distinction only and do not imply any particular order or importance.
[0032] The present specification discloses a system or an apparatus for performing the operations of the methods. Such system or apparatus may be specially constructed for the required purposes. It may also comprise a computer or other device selectively activated or reconfigured by a computer program stored in the computer. The algorithms and methods
[0033] presented herein are not inherently related to any particular computer or other apparatus. Various machines may be used with programs in accordance with the teachings herein. Alternatively, the construction of more specialized apparatus to perform the required method steps may be appropriate.
[0034] In addition, the present specification also implicitly discloses a computer program. It would be apparent to a person skilled in the art that the individual steps of the method described herein may be put into effect by computer code. The computer program is not intended to be limited to any particular programming language and implementation thereof. A variety of programming languages and coding thereof may be202502593used to implement the teachings of the disclosure contained herein. The computer program is also not intended to be limited to any particular control flow. There are many other variants of the computer program which can use different control flows without departing from the spirit or scope of the disclosure.
[0035] Furthermore, one or more of the steps of the computer program may be performed in parallel rather than sequentially. Such a computer program may be stored on any computer-readable medium. The computer-readable medium may include storage devices such as magnetic or optical disks, memory chips, or other storage devices suitable for interfacing with a computer. The computer-readable medium may also include a hard-wired medium, such as exemplified in the Internet system, or a wireless medium, such as exemplified in the mobile telephone system. The computer program, when loaded and executed on such a computer, effectively results in an apparatus that implements the steps of the preferred method.
[0036] The present disclosure generally contemplates the facilitation and optimization of a network and / or user equipment UE efficiency. This can be, for example, in association with 3GPP based standards for energy efficiency or power saving. The present disclosure specifically contemplates the application of Artificial Intelligence / Machine Learning AIML in wireless communication. It may be applied to multiple cases for the Physical PHY layer, Media Access Control MAC layer, and higher layers.
[0037] The foregoing and other aspects of the disclosure will be discussed in further detail with reference to FIG. 1 through FIG. 3 hereinafter.
[0038] FIG. 1 is a schematic block diagram illustrating a system 100 for context-aware reporting of AIML model inapplicability. The system 100 operates within a wireless communication network 106. The system 100 includes at least one user equipment, UE, 102. The system 100 also includes at least one network entity 104. The UE 102 and the network entity 104 are configured to communicate with each other through the wireless communication network 106.
[0039] The UE 102 is a wireless device. Examples of a UE 102 are a target device, a device-to-device D2D UE, a machine-type UE or a UE capable of machine-to-machine M2M communication. Other examples include a PDA, a PAD, a tablet, a mobile202502593terminal, a smart phone, a laptop with embedded equipment LEE, or laptop with mounted equipment LME. A UE 102 can also be a USB dongle, a ProSe UE, or a V2X UE. The UE 102 contains at least one AIML model. The UE 102 is configured to perform the method steps described in relation to FIG. 2.
[0040] The network entity 104 is a node within the wireless communication network 106. The network entity 104 can correspond to any type of radio network node or any network node that communicates with the UE 102. Examples of a network entity 104 include a NodeB, an eNodeB, or a gNodeB. It can also be a base station BS, a network controller, or an access point AP. In a preferred embodiment, the network entity 104 is a next-generation NodeB gNB. The network entity 104 is configured to perform the method steps described in relation to FIG. 3.
[0041] FIG. 2 is a flowchart illustrating a method performed by the user equipment 102. FIG. 3 is a flowchart illustrating a method performed by the network entity 104. These methods address a specific technical problem. A UE 102 may report to the network that an AIML model is applicable. However, at a later stage, the model at the UE 102 may become inapplicable. This can be due to reasons like low battery, a full data buffer, or an unavailable computation core. Current 3GPP mechanisms only allow for a binary indication of applicability. This lack of granularity means the network does not know the root cause of the inapplicability. As a result, the network may waste resources by repeatedly pushing unusable models. The UE 102 also lacks a standardized framework to describe these inapplicability conditions. The invention solves this problem by enabling the network to optimize resources when a UE model becomes inapplicable after it had earlier reported applicability.
[0042] The main idea of the invention is that the UE 102 provides structured, multi-level information to the network entity 104 to report the cause of model inapplicability. The provision of cause-aware inapplicability information enables the network entity to differentiate between temporary and persistent limitations, and between UE-side and network-side constraints. This allows the network entity to selectively suspend, modify, or avoid model-dependent procedures, thereby improving resource utilization, reducing signaling overhead, and enhancing service continuity.202502593
[0043] As shown in FIG. 2, the method at the UE 102 comprises a step of providing 200 this information. The information indicates an inapplicability of the AIML model. The information is structured to comprise at least two and preferably three distinct components. First, it comprises a scope indicator. The scope indicator indicates a duration of the inapplicability. Second, it comprises an entity side indicator. The entity side indicator indicates an origin of the inapplicability. Third, it comprises a cause indicator. The cause indicator indicates a cause for the inapplicability. The process at the UE 102 then ends 202. The provision of this information is a dynamic process. The content of the hierarchical bitmap is updated based on the real-time conditions of the UE 102. The UE 102 provides the updated information whenever it determines that the applicability status of the AIML model has changed.
[0044] In some embodiments, the information is provided in the form of a hierarchical bitmap. This provides a compact and efficient format for the information. In alternative embodiment, the information may be represented using different encoding formats, including but not limited to enumerated values, structured fields, or variable-length information elements. The hierarchical bitmap represents one efficient implementation but is not limiting.
[0045] In a further specific embodiment, this hierarchical bitmap has a particular structure. The scope indicator is represented by a 1 -bit value. The entity side indicator is represented by a 2-bit value. The cause indicator is represented by a 4-bit value.
[0046] The scope indicator may be used to indicate the duration of the inapplicability as one of temporary or permanent. This allows the network entity 104 to decide whether to de-allocate resources for a short time or for a longer period.
[0047] The entity side indicator may be used to indicate the origin of the inapplicability. It can be selected from a group comprising a UE-side issue, a network-side issue, or a metadata or configuration issue. This helps the network entity 104 to diagnose the problem correctly.
[0048] The cause indicator may indicate a cause for the inapplicability. This cause can be selected from a group that includes, but is not limited to, low battery, full data buffer,202502593or an unavailable computation core. Other reasons can include that a model is not provisioned or that a model has expired.
[0049] The step of providing 200 the information can be performed by signaling on different layers of the communication protocol stack. This may be Layer 1 L1 , Layer 2 L2, or Layer 3 L3. Layer 1 is the Physical Layer PHY. Layer 2 is the Data Link Layer, which includes the Medium Access Control MAC sublayer. Layer 3 is the Network Layer, often associated with the Radio Resource Control RRC protocol. Using any of these layers provides flexibility for implementation in a standard.
[0050] As shown in FIG. 3, the method at the network entity 104 begins by receiving this information from the UE 102. The network entity 104 then performs a step of adjusting 201 at least one parameter associated with the AIML model. This adjustment is based on the received information. The process at the network entity 104 then ends 202.
[0051] In some embodiments, the step of adjusting 201 the parameter comprises disabling or reconfiguring one or more UE-side model parameters. This action allows the network entity 104 to optimize resource allocation and prevent waste.
[0052] The method at the network entity 104 may further comprise storing the received information. This information is stored as historical data. The network entity 104 can then use this historical data to predict future inapplicability conditions.
[0053] The system 100 comprises a user equipment 102 and a network entity 104. The user equipment 102 is configured to perform the method of providing information about model inapplicability. The network entity 104 is configured to perform the method of receiving this information and adjusting resources accordingly.
[0054] A user equipment 102 may comprise a processor and a memory. The memory contains instructions which, when executed by the processor, cause the user equipment 102 to perform the method of providing the inapplicability information.
[0055] A network entity 104 may comprise a processor and a memory. The memory contains instructions which, when executed by the processor, cause the network entity202502593104 to perform the method of adjusting resources based on the inapplicability information.
[0056] A computer program product may comprise instructions. When the program is executed by a computer, these instructions cause the computer to perform the methods for reporting and managing AIML model inapplicability.
Claims
202502593CLAIMS1. A method for reporting an inapplicability of an Artificial Intelligence / Machine Learning, AIML, model in a wireless communication network (106), the method comprising:- providing (200), by a user equipment, UE, (102) information to a network entity (104), the information comprising at least two of:- a scope indicator for indicating a duration of the inapplicability; - an entity side indicator for indicating an origin of the inapplicability; and - a cause indicator for indicating a causecause for the inapplicability.
2. The method according to claim 1 , wherein the information is provided in the form of a hierarchical bitmap.
3. The method according to claim 2, wherein the hierarchical bitmap is dynamically updated based on real-time conditions of the UE (102).
4. The method according to claim 2 or 3, wherein the scope indicator is represented by a 1 -bit value, the entity side indicator is represented by a 2-bit value, and the cause indicator is represented by a 4-bit value.
5. The method according to any of the preceding claims, wherein the scope indicator indicates the duration of the inapplicability as one of temporary or permanent.
6. The method according to any of the preceding claims, wherein the entity side indicator indicates the origin of the inapplicability as being selected from a group comprising: a UE-side issue; a network-side issue; and a metadata or configuration issue.
7. The method according to any of the preceding claims, wherein the cause indicator indicates a causecause for inapplicability selected from a group comprising: low battery; full data buffer; unavailable computation core; model not provisioned; and model expired.2025025938. The method according to any of the preceding claims, wherein providing (200) the information comprises signaling the information using Layer 1, L1, Layer 2, L2, or Layer 3, L3.
9. A method for managing resources for an Artificial Intelligence / Machine Learning, AIML, model in a wireless communication network (106), the method comprising:- receiving, by a network entity (104) from a user equipment, UE, (102) information indicating an inapplicability of the AIML model, the information comprising at least two of:- a scope indicator for indicating a duration of the inapplicability; - an entity side indicator for indicating an origin of the inapplicability;and- a cause indicator for indicating a causecause for the inapplicability;and- adjusting (201 ), by the network entity (104), at least one parameter associated with the AIML model based on the received information.
10. The method according to claim 9, wherein adjusting (201) the at least one parameter comprises disabling or reconfiguring one or more UE-side model parameters to optimize resource allocation.
11. The method according to claim 9 or 10, further comprising:- storing, by the network entity (104), the received information as historical data to predict future inapplicability conditions.
12. The method according to any of claims 9 to 11 , wherein the network entity (104) is a next-generation NodeB, gNB.
13. A system (100) for managing resources for an Artificial Intelligence / Machine Learning, AIML, model in a wireless communication network (106), the system (100) comprising:- a user equipment, UE, (102) configured to perform the method of any one of claims 1 to 8; and- a network entity (104) configured to perform the method of any one of claims 9 to 12.20250259314. A user equipment, UE, (102) comprising a processor and a memory, the memory containing instructions which, when executed by the processor, cause the UE (102) to perform the method of any one of claims 1 to 8.
15. A network entity (104) comprising a processor and a memory, the memory containing instructions which, when executed by the processor, cause the network entity (104) to perform the method of any one of claims 9 to 12.
16. A computer program product comprising instructions which, when the program is executed by a computer, cause the computer to perform the method of any one of claims 1 to 12.
