WO2025233095A1

WO2025233095A1 - Ue capability signaling

Info

Publication number: WO2025233095A1
Application number: PCT/EP2025/060367
Authority: WO
Inventors: Malgorzata Tomala; Amaanat ALI; Sakira HASSAN
Original assignee: Nokia Technologies Oy
Current assignee: Nokia Technologies Oy
Priority date: 2024-05-10
Filing date: 2025-04-15
Publication date: 2025-11-13
Anticipated expiration: 2026-11-10
Also published as: GB202406612D0; GB2640958A

Abstract

Methods, apparatus and systems for UE capability signaling are disclosed.

Description

UE capability signaling

TECHNOLOGICAL FIELD

The present disclosure is related to but not limited to communication networks as defined by the 3GPP standard, such as the 5G and/or 6G standard. The disclosure particularly relates to artificial intelligence and/or machine learning, AIML, in particular to a signaling of capabilities of a user equipment, UE, regarding AIML.

BACKGROUND

Artificial intelligence and/or machine learning, AIML, will play a role in mobile communication networks. For instance, an AIML functionality may refer to an AIML-enabled feature or an AIML- enabled feature group, FG. A feature or feature group may be enabled (e.g., at a user equipment, UE] by at least one configuration (e.g., obtained from a network node], A supported and/or available AIML- enabled feature or feature group may be provided by the UE to the network, e.g., by reporting UE capabilities.

SUMMARY OF SOME EXEMPLARY EMBODIMENTS

A network node may be informed by the UE about supported and/or available UE capabilities, wherein the UE capabilities at least partially may relate to AIML.

It has been recognized that a static signaling of UE capabilities which may for instance indicate (e.g., AIML-] features or feature groups (e.g., supported by the UE] may be insufficient for at least some AIML use cases. For instance, diverse and/or various applicable conditions affecting at least one of an AIML-feature, an AIML-functionality or an AIML model supported by a UE may render static UE capabilities insufficient. For instance, a subset of (e.g., all] (e.g., identified] AIML models (e.g., supported by a given UE] may become (e.g., in-]applicable, e.g., after a first signaling to a network entity. For instance, at least one internal condition of a UE, e.g., such as memory, battery status or combinations thereof, may cause (e.g., in-]applicability of at least one AIML model. In another example, an (un-]availability of an AIML model, e.g., due to a configuration or an activation for performing inference for a given period of time may cause (in-]applicability. It has been recognized that at least one legacy operation which may be based on static UE capabilities may need adoption. This may in particular be the case due to dynamically changing conditions, e.g., which affect (e.g., availability of] AIML functions or function groups (e.g., providable by and/or] of the UE. A challenge may be that a network-side entity (e.g., network node] may be responsible to deduce which AIML components (e.g., models, functions, function groups] may be providable by a given UE. E.g., different AIML UE components may be indicated in an (e.g., regular and/or static] UE capability list, yet it may not be certain (e.g., to a network entity] which of these may be applicable, e.g., currently, in the future, for a given set of conditions, for a certain AIML use case and/or for a certain AIML functionality.

A further problem which may arise is that due to a changing environment (e.g., of a given UE], UE capabilities may need additional information, e.g., in order for an AIML-enabled functionality to work properly and/or in order for a network entity to appropriately configure an AIML-enabled functionality. Additional information, for instance additional conditions and/or applicability-related information (e.g., provided by a UE, e.g., to a network node] which may specify, e.g., under which conditions a model and/or functionality is applicable and/or suitable. Yet, an unstructured provision of such additional information may lead to ambiguous, if not contradictive conclusions (e.g., by the network node]. Even when assuming that various levels of collaboration between a network node (e.g., gNB] and a UE may be identified and considered (e.g., leaving such considerations to implementation], additional (e.g., applicability] information may not lead to an adequate assessment of a suitability of a UE to (e.g., currently] get involved in an (e.g., a given and/or required, e.g., by the network node] AIML operation.

It has been recognized that certain UE capabilities may relate to a (e.g., theoretical, given the right (e.g. hypothetical] conditions] supported UE capabilities. For instance, a UE may be manufactured with a given AIML capability, e.g., a certain model, a certain training mode, a certain network size, and/or combinations thereof. It has been recognized, that such (e.g., at least essentially invariable UE capabilities] may be signaled (e.g., by a UE, e.g., to a network node] as a first type of UE capabilities, which may be referred to as leading (e.g., AIML] UE capabilities.

At a given point in time (e.g., depending on at least one condition], an (e.g., invariable, supported] UE capability (i.e., despite being supported] may not be available (e.g., due to conditions such as for instance at least one of restricted available memory, models may only be initialized up to a given size, unavailable connectivity to further devices, unavailability of sensors, battery constraints and/or combinations thereof]. Thus, while supported UE capabilities may stay unchanged over time, availability of such supported UE capabilities may change over time (e.g., depending on conditions]. It has been recognized that it may be advantageous to enable a provision of complementing UE capability information, which may in particular relate to an availability (e.g., applicability] of a given (e.g., supported] AIML function. Additionally or alternatively, additional information such as conditions, parameters and/or combinations thereof may be provided as complementary UE capabilities. It has been recognized that a first potential solution, which may for instance reduce standardization impact, may be to consider a dynamic change of UE capability. For instance, an assessment on an additional and/or applicable condition may be left to the UE implementation. A UE may be allowed to send (e.g., different and/or various] UE capability information (e.g., different than previously]. For instance, according to a (e.g., standardized] legacy framework, provided (e.g., by UE to network node] UE capabilities may be stored at the network side (e.g., in radio access network, RAN, and/or core network (e.g., in UE Context]]. For instance, UE capabilities may be maintained in an UE registration area, e.g., to limit signaling overhead. An underlying principle may be that upon (e.g., fresh] UE connection to a network (e.g., to a network node], the UE may provide a (e.g., complete] set of UE capabilities, e.g., for a given RAT and/or for a given frequency band, e.g., based on a network enquiry (e.g., UE capability enquiry]. Combining a legacy method with an approach which may enable dynamic UE capability changes may enable a same user to indicate different input on the complete set of its capabilities based on at least one device condition. However, it has further been recognized that the network may be confused regarding expectations with respect to UE performance and that the network may lose control of the user (e.g., UE], It has further been recognized that signaling may deviate (e.g., across UEs] and may hide actually supported (e.g., manufactured] UE capabilities of a given UE.

There is thus a need for a framework for providing complete UE capability information to the network (e.g., by a UE to a network node] which is capable of covering dynamic changes in UE capabilities without compromising control by the network and without providing contradicting and/ or inconsistent information and without creating excessive signal overhead, in order to enable compatible (e.g., compatible with UE capabilities] AIML activation from the network.

It is thus, inter alia, an object of the disclosure to overcome shortcomings of current approaches and better match the needs of future UE capability signaling.

The present disclosure proposes inter alia to group and/or signal AIML-based functionality capabilities based on an organization of AIML UE capabilities, e.g., into a hierarchical and/or branched structure of AIML UE capabilities. By means of such an organization (e.g., hierarchy, tree, and/or combinations thereof], a network entity (e.g., network node] can gradually collect AIML UE capabilities.

For instance, AIML UE capabilities may be defined, on the one hand, as leading (e.g., or unconditional] UE capabilities. A given leading (e.g., or unconditional] UE AIML capability may be further specified (e.g., completed] by complementary (e.g., conditional] UE AIML capabilities, e.g., upon network request. Complementary UE AIML capabilities may be provided (e.g., by a UE, e.g. to a network node] in a reactive manner (e.g., in a two-step approach, e.g., according to the first example aspect] or additionally or alternatively in a proactive manner (e.g., in a one-step approach, e.g., according to the second example aspect], A network entity (e.g., network node] may for instance control, which conditional AIML UE-capabilities (e.g., AIML-based functionality capabilities] are retrieved (e.g., by the network node from the UE] and may add, modify, suspend and/or release at least one complementary AIML-enabled functionality (e.g., according to an AIML UE-capability],

Both the leading (e.g., unconditional] and complementary (e.g., conditional] UE AIML capabilities may determine an applicability to enable a respective AIML functionality. Additionally or alternatively, a leading UE AIML capability may indicate a (e.g., theoretically, according to manufacturing and/or according to an (e.g., static and/or long-term] configuration of the UE] supported AIML functionality whereas a (e.g., corresponding to the leading UE AIML capability] complementary UE AIML capability may indicated an (e.g., current] applicable and/or available UE AIML capability, e.g., based on at least one (e.g., internal to the UE or external to the UE] condition.

Based on the amount, frequency (e.g., of a provision, e.g. from the UE to the network node and/or e.g., of a requesting, e.g., by the network node] and content type of these information (e.g., UE AIML capabilities, e.g., leading and/or complementary], a network entity (e.g., network node] may decide how to enable and trigger a respective AIML-enabled functionality. It may further be decided (e.g., by the network (e.g., network node]], whether a UE would change a mode of reporting. For instance, a UE may switch to (e.g., complete] configuration reporting. For instance, the UE may continue conditional reporting (e.g., reactive or proactive]. An AIML feature performance management may be further maintained by a negotiation and an alignment on conditions (e.g., conditions affecting AIML UE capability, e.g., signaled by conditional AIML UE capabilities] applicable for the particular user. Such synchronization may be enabled in a dynamic manner by relating the functionality to an ongoing activity of the UE and/or (e.g., UE] connection dynamics.

According to a first example aspect, a method is disclosed, comprising: obtaining (e.g., by radio resource control, RRC, signaling, medium access control control element, MAC-CE, downlink control information, DCI, non-access stratum, NAS, signaling and/or combinations thereof) an artificial intelligence (e.g., and/or machine learning), AIML, user equipment, UE, capability enquiry (e.g., from a network node), determining (e.g., based on the obtained AIML user equipment, UE, capability enquiry] at least one leading AIML UE capability, wherein the leading AIML UE capability belongs to a hierarchical AIML UE capability structure (e.g., known at the UE and/or network node], wherein the hierarchical (e.g., branched and/or step-wise] AIML UE capability structure defines an association between the at least one leading AIML UE capability and at least one complementary AIML UE capability (i.e., the complementary AIML UE capability is associated to (e.g., dependent on, appended by, belongs to, configured to further specify] a respective (e.g., single one] of the at least one leading AIML UE capability according to the hierarchical AIML UE capability structure], and providing (e.g., by RRC signaling, MAC-CE, DCI, NAS signaling and/or combinations thereof] (e.g., in response to the obtaining of the UE capability enquiry; e.g., to the network node] an initial AIML UE capability response, wherein the initial AIML UE capability response indicates the at least one leading AIML UE capability.

According to a second example aspect, a method is disclosed, comprising: obtaining (e.g., by radio resource control, RRC, signaling, medium access control control element, MAC-CE, downlink control information, DCI, non-access stratum, NAS, signaling and/or combinations thereof] an artificial intelligence (e.g., and/or machine learning], AIML, user equipment, UE, capability enquiry (e.g., from a network node], determining (e.g., based on the obtained artificial intelligence/ machine learning, AIML, user equipment, UE, capability enquiry] at least one leading AIML UE capability and at least one complementary AIML UE capability, wherein the leading AIML UE capability and the complementary AIML UE capability belong to a hierarchical AIML UE capability structure (e.g., known at the UE and/or network node], wherein the hierarchical (e.g., branched and/or step- wise] AIML UE capability structure defines an association between the at least one leading AIML UE capability and at least one complementary AIML UE capability (i.e., the complementary AIML UE capability is associated to (e.g., dependent on, appended by, belongs to, configured to further specify] a respective (e.g., single one] of the at least one leading AIML UE capability according to the hierarchical AIML UE capability structure], and providing (e.g., by RRC signaling, MAC-CE, DCI, NAS signaling and/or combinations thereof] (e.g., in response to the obtaining of the UE capability enquiry; e.g., to the network node] an initial AIML UE capability response, wherein the initial AIML UE capability response indicates the at least one leading AIML UE capability and the at least one complementary AIML UE capability.

These methods according to the first and/or second example aspect may for instance be performed and/or controlled by an apparatus, for instance a server. Alternatively, the respective method may be performed and/or controlled by more than one apparatus, for instance a server cloud comprising at least two servers. Alternatively, the respective method according to the first and/ or second example aspect may for instance be performed and/or controlled by an electronic device, e.g. a network node in a communication system and/or by a terminal device, e.g., a user equipment (UE], For instance, the method may be performed and/ or controlled by using at least one processor of the electronic device.

According to a further example aspect, a computer program is disclosed, the computer program when executed by a processor causing an apparatus, for instance a server, a network node or a terminal device, e.g., a UE, to perform and/ or control the actions of the method according to the first and/ or second example aspect. The computer program may be stored on computer-readable storage medium, in particular a tangible and/or non-transitory medium. The computer readable storage medium could for example be a disk or a memory or the like. The computer program could be stored in the computer readable storage medium in the form of instructions encoding the computer-readable storage medium. The computer readable storage medium may be intended for taking part in the operation of a device, like an internal or external memory, for instance a Read-(e.g., Only] Memory [ROM] or hard disk of a computer, or be intended for distribution of the program, like an optical disc.

According to a further example aspect, an apparatus is disclosed, configured to perform and/or control or comprising respective means for performing and/ or controlling the method according to the first and/or second example aspect.

The means of the apparatus can be implemented in hardware and/or software. They may comprise for instance at least one processor for executing computer program code for performing the required functions, at least one memory storing the program code, or both. Alternatively, they could comprise for instance circuitry that is designed to implement the required functions, for instance implemented in a chipset or a chip, like an integrated circuit. In general, the means may comprise for instance one or more processing means or processors.

The above-disclosed apparatus according to any aspect may be a module or a component for a device, for example a chip. Alternatively, the disclosed apparatus according to any aspect may be a device, for instance a server or server cloud. The disclosed apparatus according to any aspect may comprise (e.g., only] the disclosed components, for instance means, processor, memory, or may further comprise one or more additional components.

A terminal device, e.g., a user equipment (UE] may for instance correspond to a mobile device such as for example a mobile phone, tablet, smartwatch, a laptop, a Personal Digital Assistant [PDA] device, a wearable, an Internet-of-Things (IOT] device, an IIOT (Industrial IOT] device, a vehicle and/or combinations thereof. Such a user equipment may also be referred to as user device.

A network node may correspond to a component of a communication network such as for instance a Base Transceiver Station [BTS], a nodeB, an evolved node B [eNB], a Next Generation NodeB [gNB], a distributed unit [DU], a central unit [CU] and/or combinations thereof.

The method according to the first and second example aspect comprises obtaining an AIML UE capability enquiry. The AIML UE capability enquiry may for instance be received by an apparatus, for instance a UE, performing and/or controlling the method according to the first and/or second example aspect. The AIML UE capability enquiry may be received from a network entity, for instance a network node. The AIML UE capability enquiry may indicate a request for at least one (e.g., at least one leading and/ or at least one complementary] AIML UE capability, e.g. by the network node, e.g. directed to the apparatus performing and/for controlling the method according to the first and/or second example aspect.

The AIML UE capability enquiry may correspond to and/or comprise a UE Capability Enquiry message. The UE Capability Enquiry message may be a downlink message sent over a signaling radio bearer from a network node to a UE (e.g., apparatus performing and/ or controlling the method according to the first and/or second example aspect] and is used to request at least one or more UE capabilities. The network (e.g., network node] may initiate a procedure (e.g., UE capability procedure] to a (e.g., to the] UE in a connected state of the UE to the network node (e.g., over a dedicated channel], e.g., when the network node requires UE radio access capability information. In particular, in a context of the present disclosure, the requested and/or required information may contain at least one UE capability information and may additionally comprise additional information, e.g., on a UE ability to perform at least one AI/ML-enabled operation, e.g. an applicability information to perform a certain AI/ML- enabled operation.

The AIML UE capability enquiry may for instance be obtained (e.g., received from a network entity such as a network node] by a radio resource control, RRC, signaling, a medium access control control element, MAC CE, a downlink control information, DCI, a non-access stratum, NAS, signaling and/or combinations thereof.

The AIML UE capability enquiry may at least partially relate to an AIML functionality, in particular a capability of the apparatus performing and/ or controlling the method according to the first and/ or second example aspect to provide a respective AIML functionality. The AIML UE capability enquiry may in particular fully relate to at least one or more AIML functionalities, i.e. may be unrelated to an (e.g., any] non-AIML UE capability. Additionally or alternatively, the AIML UE capability enquiry may at least partially relate to at least one or more non-AIML UE capabilities.

An AIML UE capability may specify an ability of a UE to provide at least one aspect of a given AIML functionality.

The method according to the first and/or second example aspect further comprises determining at least one leading AIML UE capability. For instance, such determining may be based on the obtained AIML UE capability enquiry. E.g., obtaining the AIML UE capability enquiry may cause the apparatus performing and/or controlling the method according to the first and/or second example aspect to determine the at least one leading AIML UE capability.

The leading AIML UE capability belongs to a hierarchical AIML UE capability structure. The hierarchical AIML UE capability structure defines an association between the at least one leading AIML UE capability and at least one complementary AIML UE capability. For instance, hierarchical AIML UE capability structure may correspond to and/or enable a branched and/or stepwise definition of AIML UE capabilities. For instance, a UE may, e.g. in the first step, provide a leading AIML UE capability. The leading AIML UE capability may for instance specify at least one supported AIML functionality, e.g. supported by the UE. The UE may then, e.g. in a second step, further specify a complementary AIML UE capability in order to further specify its ability, and/or an applicability and/or an availability of an AIML functionality, for instance the supported AIML functionality and/or an AIML functionality related to the supported AIML functionality indicated by the leading AIML UE capability. In particular, based on the hierarchical AIML UE capability structure, a complementary AIML UE capability may always be associable to a leading AIML UE capability.

The hierarchical AIML UE capability structure may be visualized and/or structured as a branched tree, e.g., with higher-level leading AIML UE capabilities as nodes and lower-level complementary AIML UE capabilities as lower level nodes and/ or leaves of the tree. For instance, a given (e.g., any] leading AIML UE capability may be associated to at least one complementary AIML UE capability according to the tree. For instance, once a leading AIML UE capability is known, at least one associated complementary AIML UE capability may be derivable based on the hierarchical AIML UE capability structure. For instance, leading AIML UE capabilities of the hierarchical AIML UE capability structure may be represented as higher-level nodes of a tree (e.g., close to and/or connected to a root of the tree, e.g., by an edge] while complementary AIML UE capabilities may be represented as lower level (e.g., further away from the root of the tree, e.g., connected to the root via a node corresponding to a leading AIML UE capability] or by leaves of a the tree. Any lower-level node or leaf (i.e., complementary AIML UE capability] may be connected to a (e.g., single] higher-level node (i.e., leading AIML UE capability], e.g., by a branch or edge of the tree. Conditional to knowing at least one leading AIML UE capability, the corresponding (e.g., dependent] AIML UE capabilities may be derived as the lower-level nodes and/or leaves of the tree corresponding to (e.g., connected to] the higher-level node of the tree associated with the leading AIML UE capability. Other visualizations of the hierarchical AIML UE capability structure are possible. In general, the hierarchical AIML UE capability structure may be hierarchical in the sense that it defines respective dependencies of complementary AIML UE capabilities on leading AIML UE capabilities.

The hierarchical AIML UE capability structure may comprise at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 16, 32, 64 or more leading AIML UE capabilities. The hierarchical AIML UE capability structure may comprise at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 16, 32, 64 or more complementary AIML UE capabilities. The hierarchical AIML UE capability structure may comprise more complementary AIML UE capabilities than leading AIML UE capabilities. For instance, the hierarchical AIML UE capability structure may be fixed (e.g., pre-defined], alternatively the hierarchical AIML UE capability structure may be changeable, e.g., extendable. The hierarchical AIML UE capability structure may define, for any leading AIML UE capability at least one complementary. Additionally or alternatively, the hierarchical AIML UE capability structure may define additional information associated to leading and/or complementary AIML UE capabilities such as for instance conditions and/or parameters.

The hierarchical AIML UE capability structure may be known (e.g., stored, available, obtainable, retrievable and/or combinations thereof) to the apparatus (e.g. UE) performing and/or controlling the method according to the first and/or second example aspect and/or to a network entity, for instance a network node, in particular a network entity that provided the AIML UE capability enquiry.

Thus, the hierarchical AIML UE capability structure enables a separate signaling of a complementary AIML UE capability, e.g. without a corresponding leading AIML UE capability. The hierarchical AIML UE capability structure enables a receiver (e.g., network node)(e.g., of such separate signaling of an AIML UE capability) to make a connection between the complementarity AIML UE capability and a (nonsignaled and/or earlier signaled) leading AIML UE capability, based on the hierarchical AIML UE capability structure. Thereby, signaling overhead is reduced without compromising an ability of a UE (e.g. apparatus performing and/or controlling the method according to the first and/or second example aspect) to signal its (e.g. current, e.g. dynamically changing) AIML UE capabilities (e.g., as complementary AIML UE capabilities).

A complementary AIML UE capability may thus be associated to a respective at least one leading AIML UE capability. Such an association may be defined by the hierarchical AIML UE capability structure. For instance, a complementary AIML UE capability may be dependent on, belong to and/or (be configured to) further specify a respective leading AIML UE capability and/or a capability to provide respective AIML functionality corresponding to the leading and/or complementarity AIML capability. A complementary AIML UE capability may additionally or alternatively specify additional information, e.g., associated to a respective leading AIML UE capability, e.g., at least one condition and/or at least one parameter. The combination of leading and complementary AIML UE capability may enable a network node to configure an AIML functionality at the UE.

It has been recognized, that at least some aspects of AIML UE capabilities may be at least essentially invariable over time. For instance, a UE may be equipped with a chip which enables a given range of AIML functionalities. The UE may thus support AIML functionalities, which depends on the respective chip. As such invariable AIML UE capabilities do not change over time, there is limited need to signal such invariable AIML UE capabilities from the UE to the network. For instance, it may be sufficient, if the UE signals and invariable AIML UE capability ones, for instance in the form of a leading AIML UE capability. It has further been recognized, that at least some AIML UE capabilities are not invariable over time but may instead change, e.g., frequently. It has in particular been recognized, that an applicability of (e.g., in itself) invariable AIML UE capabilities (e.g., signaled by a respective leading AIML UE capability) may change over time. For instance, an AIML functionality which depends on an invariable AIML UE capability (e.g., a certain chip) may also depend on further factors, for instance availability of an (e.g., trained) AIML model using the chip. And availability of a model may in this case be considered an internal condition of the UE. Countless other conditions, such as for instance a current connection of the UE to another network device and/or two a positioning system and/or a positioning of the UE outside and/ or inside the building may influence currently available AIML UE capabilities and/or functionalities. It has been recognized that an efficient signaling of AIML UE capabilities may be achieved by differentiating into leading AIML UE capabilities, which may in particular correspond to invariable and/or unconditional AIML UE capabilities, and into complementary AIML UE capabilities, which may in particular correspond to variable and/ or conditional (e.g., dependent on internal and/or external conditions) AIML UE capabilities.

The method according to the second example aspect comprises determining, in addition to the at least one leading AIML UE capability, at least one complementary AIML UE capability (e.g., before providing the initial AIML UE capability response). Thus, according to the second example aspect, not only at least one leading AIML UE capability is determined, but also a complementary AIML UE capability. The complementary AIML UE capability is associated to the at least one leading AIML UE capability, based on the hierarchical AIML UE capability structure. The method according to the second example aspect may for instance relate to a proactive provision of AIML UE capabilities. Instead of first determining (e.g., and providing e.g. to a network node) a leading AIML UE capability and, in a later step, determining (e.g. and providing) a complementary AIML UE capability, both a leading and a corresponding complementarity AIML UE capability are determined (e.g. and provided).

The determining of both at least one leading and at least one complementary AIML UE capability may be based on the obtaining and/or on the obtained AIML UE capability enquiry. For instance the AIML UE capability enquiry may be configured to cause an apparatus performing and/ or controlling the method according to the first and/ or second example aspect to either determine (e.g., only) a leading AIML UE capability (e.g., thus corresponding to the first example aspect) or to instead determine both at least one leading AIML UE capability and at least one complementarity AIML UE capability (e.g., thus corresponding to the second example aspect).

The method according to the first and/or second example aspect comprises providing an initial AIML UE capability response. The initial AIML UE capability response may be provided, e.g. by the apparatus performing and/or controlling the method according to the first and/or second example aspect, to an entity of the network, in particular a network node, in particular the network node from which the AIML UE capability enquiry has been obtained.

The initial AIML UE response may for instance be provided (e.g., transmitted by a UE performing the method according to the first and/or second example aspect] by a radio resource control, RRC, signaling, a medium access control control element, MAC CE, a downlink control information, DCI, a non-access stratum, NAS, signaling and/or combinations thereof.

Providing the initial AIML UE response may be based on at least one of obtaining the AIML UE capability enquiry and/or on determining the at least one leading AIML UE capability (e.g., and/or determining at least one complementary AIML UE capability].

The AIML UE capability enquiry may be configured to cause an apparatus performing and/or controlling the method according to the first and/or second example aspect to perform the steps of determining at least one leading AIML UE capability (e.g., and/or a complementary AIML UE capability] and/ or providing the initial AIML UE capability response.

The initial AIML UE capability response indicates the (e.g., determined] at least one leading AIML UE capability. By indicating the at least one leading AIML UE capability, a receiving entity receiving the initial AIML UE capability may be informed of at least one AIML UE functionality supported by the apparatus performing and/for controlling the method according to the first and/or second example aspect.

According to the second example aspect, the initial AIML UE capability indicates, e.g. in addition to the at least one leading AIML UE capability, the (e.g., determined] at least one complementarity AIML UE capability. By enabling the UE to determine and/ or to provide and/ or by enabling the network node to enquire and/or obtain both an initial AIML UE capability and a complementary AIML UE capability, a number of required separate signaling steps between network node and UE may be reduced.

For instance, a network node may provide an initial AIML UE capability enquiry to a UE, for instance to a UE, which has previously become connected to the network node, e.g. a UE to which the network node has not yet sent her previous AIML UE capability enquiry. The initial AIML UE capability enquiry may be configured to enquire both at least one leading and at least one (e.g., corresponding to the leading] complementary AIML UE capability. The initial AIML UE capability enquiry may serve the purpose to gain knowledge about currently available AIML UE capabilities. As the network node does at this stage not have any information about supported AIML UE capabilities (e.g. corresponding to the leading AIML UE capability], the network node may (e.g., need to] request both leading and complementary AIML UE capabilities (e.g., as complementary AIML UE capabilities may not independently (e.g., fully] specify the UE’s capabilities to provide a given AIML functionality]. For instance the network node may, e.g., at a later time, requests complementary AIML UE capabilities only, i.e. corresponding to at least one initial or AIML UE capability, e.g. obtained as a response to the initial AIML UE capability enquiry.

In another example, a network node may currently be uninterested in requesting a particular AIML functionality at the UE, and may therefore not require any complementary AIML UE capability corresponding to such particular AIML functionality. A complementary AIML UE capability which the UE may provide at a given instance in time may be outdated at a later instance in time, e.g., an instance at which the network node may actually be interested in requesting the respective AIML functionality (e.g., corresponding to the complementary AIML UE capability]. In such a scenario, the reactive approach of the first example aspect provides the advantage of a reduced signaling volume when requesting the leading AIML UE capability (e.g., only, i.e.] without a complementary AIML UE capability. The network node is in this case still informed about the UE’s capability to (e.g., theoretically, e.g., conditionally on at least one condition being fulfilled] provide at least one AIML functionality. The network node is thus prepared, after having obtained the initial AIML UE capability response from the UE, to enquire at least one complementary AIML UE capability, based on the obtained at least one leading AIML UE capability, e.g., once it wishes to request a corresponding AIML functionality from the UE. Signaling of, e.g. unused (e.g. by the network node] complementary AIML UE capabilities may thus be prevented.

According to an embodiment of the second example aspect, the initial AIML UE capability response is provided as a single (e.g., coherent, uninterrupted] message (e.g., comprising a first bit indicating a leading AIML UE capability and a second bit (e.g., or multiple second bits] indicating a corresponding complementary AIML UE capability].

According to the second aspect, leading and complementary AIML UE capabilities are both enquired, determined and provided. In this case, the initial AIML UE capability response may, in particular in the second example aspect, be provided (by the apparatus performing answers or controlling the method according to the first and/or second example aspect] as a single message. The single message may for instance correspond to a (e.g., single] signaling, for instance at least one of an RRC, MAC CE, DCI, or NAS signaling. The single message may be separated in two at least two or more parts, for instance separated in time and/or frequency. The single message may be provided without further interaction between UE and network node. For instance, the initial AIML UE capability response may comprise at least one indication of the at least one leading AIML UE capability, for instance at least one bit (e.g., 1, 2, 3, 4, 5, 6, 7, or 8 bits] and at least one indication of the at least one complementary AIML UE capability, for instance at least one bit (e.g., 1, 2, 3, 4, 5, 6, 7, or 8 bits]. Multiple leading and/or complementary AIML UE capabilities may be transmitted in a single message, e.g., at least 2, 3, 4, 5, 6, 7, or 8 leading AIML UE capabilities and/or at least 2, 3, 4, 5, 6, 7, or 8 complementary AIML UE capabilities.

According to an embodiment of the first example aspect, the initial AIML UE capability response is unindicative (e.g., other than by the at least one leading AIML UE capability itself) of a (e.g., at least one and/or any) complementary AIML UE capability associated to (e.g., at least one of or any of) the at least one leading AIML UE capability indicated by the initial AIML UE capability response.

According to the first aspect, (e.g., only) at least one leading AIML UE capability is enquired, determined and provided. In this case, the initial AIML UE capability response may be on indicative of a complementary AIML UE capability. For instance, the initial AIML UE capability response may be indicative of at least one and/or any complementary AIML UE capability, in particular at least one and/ or any complementary AIML UE capability associated to the at least one leading AIML UE capability indicated by the initial AIML UE capability response. The initial AIML UE capability response may thus be free of any indication of a complementary AIML UE capability.

According to an embodiment of the first and second example aspect, the method further comprises obtaining (e.g., by RRC signaling, MAC-CE, DCI, NAS signaling and/or combinations thereof) a subsequent AIML UE capability enquiry (e.g., from a network node).

The method according to the first (e.g. and optionally, to the second) example aspect may further comprises obtaining at least one subsequent AIML UE capability enquiry. The subsequent AIML UE capability enquiry may be obtained from a network node, in particular the network node, from which the AIML UE capability enquiry has been obtained.

The subsequent AIML UE capability enquiry may for instance be obtained (e.g., received by the UE performing the method according to the first and/or second example aspect; e.g., from the network node) by a radio resource control, RRC, signaling, a medium access control control element, MAC CE, a downlink control information, DCI, a non-access stratum, NAS, signaling and/or combinations thereof.

For instance, the subsequent AIML UE capability enquiry may be targeted towards at least one complementary AIML UE capability. The subsequent AIML UE capability enquiry may be configured to enquire at least one complementary AIML UE capability, in particular at least one complementary AIML UE capability corresponding to at least one leading AIML UE capability, e.g. the at least one determined and/or provided (as part of the initial AIML UE capability response) leading AIML UE capability. For instance, the subsequent AIML UE capability enquiry may be obtained based on providing the initial AIML UE capability response. Additionally or alternatively, the subsequent AIML UE capability enquiry may be configured to request (e.g., enquire] the at least one complementary AIML UE capability (e.g., associated to at least one previously provided leading AIML UE capability].

The subsequent AIML UE capability enquiry may be obtained as part of a UE capability RRC procedure, e.g., by a UE capability RRC message. Additionally or alternatively, the subsequent AIML UE capability enquiry may be obtained by a another RRC procedure, for instance involving an RRC message such as at least one of User Information Request, User Information Response or UE Assistance Information.

According to an embodiment of the first example aspect, the method further comprises determining at least one complementary AIML UE capability.

For instance the at least one complementary AIML UE capability may be determined at least partially based on the subsequent AIML UE capability enquiry. Additionally and/or alternatively, the at least one complementary AIML UE capability may be determined at least partially based on at least one of the at least one (e.g., previously determined] leading AIML UE capability, e.g., and the hierarchical AIML UE capability structure.

According to an embodiment of the first and second example aspect, the method further comprises providing (e.g., by RRC signaling, MAC-CE, DCI, NAS signaling and/or combinations thereof] a subsequent AIML UE capability response, wherein the subsequent AIML UE capability response indicates (e.g., the] at least one determined complementary AIML UE capability (e.g., based on (e.g., obtaining of] the subsequent AIML UE capability enquiry and/or after determining the at least one complementary AIML UE capability].

The method according to the first (e.g. and second] example aspect may further comprise providing a subsequent AIML UE capability response. The subsequent AIML UE capability response may be provided to a network node, in particular the network node, from which the subsequent AIML UE capability enquiry has been obtained.

The subsequent AIML UE capability response may for instance be provided (e.g., transmitted by the UE performing the method according to the first and/or second example aspect; e.g., to the network node] by a radio resource control, RRC, signaling, a medium access control control element, MAC CE, a downlink control information, DCI, a non-access stratum, NAS, signaling and/or combinations thereof.

Providing the subsequent AIML UE capability response may be based on (e.g., obtaining of) the subsequent AIML UE capability enquiry. For instance, the subsequent AIML UE capability enquiry may be configured to cause a determining of at least one complementary AIML UE capability and/or to cause a providing of the at least one subsequent AIML UE capability response.

The subsequent AIML UE capability response indicates (e.g., the determined] at least one complementary AIML UE capability.

For instance, the UE performing and/or controlling the method according to the first (e.g. and/or the second] example aspect may first obtain an AIML UE capability enquiry, determine a leading AIML UE capability (e.g., based on the AIML UE capability enquiry], provide an indication of the determined leading AIML UE capability to the network node, then obtain a subsequent AIML UE capability enquiry from the network node (e.g. based on the provided, determined leading AIML UE capability], and then (e.g. based on the subsequent AIML UE capability enquiry] determine a complementary AIML UE capability (e.g., based on the leading AIML UE capability and/or the hierarchical AIML UE capability structure] and then provide the determined complementary AIML UE capability to the network node as part of the subsequent AIML UE capability response.

According to an embodiment, a method may comprise obtaining a subsequent AIML UE capability enquiry from the network node (e.g. based on a previously provided, determined leading AIML UE capability], and then (e.g. based on the subsequent AIML UE capability enquiry] determine a complementary AIML UE capability (e.g., based on the leading AIML UE capability and/or the hierarchical AIML UE capability structure] and then provide the determined complementary AIML UE capability to the network node as part of the subsequent AIML UE capability response.

According to an embodiment of the first and second example aspect, the method further comprises obtaining (e.g., by RRC signaling, MAC-CE, DCI, NAS signaling and/or combinations thereof) an AIML configuration (e.g., RRCReconfiguration signaling) (e.g., based on the provided leading AIML UE capability and/or complementary AIML UE capability) (e.g., from a (e.g., the) network node).

The method according to the first and/ or second example aspect may further comprise obtaining an AIML configuration. The AIML configuration may for instance be obtained from a network node, in particular from the network node, from which the AIML capability enquiry and/or the subsequent AIML capability enquiry have been obtained and/or to which the initial AIML UE capability response and/or the subsequent AIML capability response have been provided.

The AIML configuration may for instance be obtained (e.g., received by the UE performing the method according to the first and/or second example aspect; e.g., from the network node) by a radio resource control, RRC, signaling, a medium access control control element, MAC CE, a downlink control information, DCI, a non-access stratum, NAS, signaling and/or combinations thereof. In particular, the AIML configuration may be obtained as an RRCReconfiguration signaling.

The AIML configuration may be (e.g., obtained] based on (e.g., the providing of] the leading AIML UE capability and/or the complementary AIML UE capability. The AIML configuration may further at least partially be based on the hierarchical AIML UE capability structure.

The AIML configuration may be configured to specify at least one AIML functionality to be provided by the apparatus performing and/ or controlling the method according to the first and/ or second example aspect.

By providing the leading and associated complementary AIML UE capability to network node, the UE enables network node to select and/ or configure at least one AIML functionality.

According to an embodiment of the first and second example aspect, at least one of the AIML UE capability enquiry, the initial AIML UE capability response, the leading AIML UE capability and/or the complementary AIML UE capability relate to at least one of an AIML functionality, an AIML use case or an AIML life cycle management phase (e.g., a phase such as e.g., training, inference, performance monitoring, data collection, update, fine-tuning and/or combinations thereof] (e.g., and may additionally be partly unrelated to an AIML functionality or an AIML use case],

A given AIML UE capability (e.g., leading and/or complementary] may relate to at least one AIML functionality. An AIML functionality may for instance correspond to a signal processing, data processing, filtering, selection, restructuring, classification, segmentation, resource management, and/or optimization functionality which is at least partially assisted and/or enabled by artificial intelligence and/or machine learning.

Then AIML use case may for instance correspond to a task to be fulfilled by an entity (e.g. UE and/or network node] of a communication network. For instance, such a use case may correspond to channel state information, CSI, feedback enhancement, beam management, and/or positioning enhancements. For instance, functionalities within a given use case may partially be fulfilled, supported and/or enabled by an AIML functionality. For instance, usage of an AIML functionality may be optional. For instance, the network node may decide whether an AIML functionality is used within a given use case, for instance depending on the AIML UE capabilities, which the UE may signal, for instance as part of a (e.g., initial or subsequent] AIML UE capability response. The AIML UE capability enquiry, the initial AIML UE capability response, the leading AIML UE capability and/or the complementary AIML UE capability may further relate to an AIML lifecycle management, LCM, phase. Example AIML LCM phases may correspond to training, inference, performance monitoring, data collection, (e.g., AIML model] update, fine-tuning and/or combinations thereof. It has been recognized, that AIML functionalities carry an increased complexity compares to a majority of previously used approaches for implementing network functionalities. As AIML functionalities are typically trained on data, as opposed to for instance operator-defined heuristics and/or rules, AIML functionalities may need to be trained, adapters, and continuously maintained in order to perform correctly within any given use case. The UE and/or the network node may be required to fulfill managing roads in terms of the lifecycle management phases of at least one or more AIML functionalities. Thus, both the AIML UE capability enquiries of the network node and/or the AIML UE capability responses of the UE may at least partially depend on a given (e.g. current or desired] AIML life cycle management phase.

The AIML UE capability enquiry, the initial AIML UE capability response, the leading AIML UE capability and/or the complementary AIML UE capability may exclusively relate to AIML are may, additionally or alternatively at least partially relate to non-AIML functionalities, use-cases and life cycle management.

For instance, a (e.g., initial] AIML UE capability enquiry may be unspecific to AIML. In other words, the AIML UE capability enquiry may correspond to a legacy UE capability enquiry (e.g., compatible with . For instance, the AIML capability enquiry may relate to a use case. The UE may, for instance in its initial AIML UE capability response, indicate its (e.g., leading AIML UE] capability to provide at least one AIML functionality relevant to the use case indicated by the at least one AIML UE capability enquiry.

According to an embodiment of the first and second example aspect, the AIML UE capability enquiry is at least one of indicative of a request for (e.g., at least one] complementary AIML UE capability (e.g., by a respective tag], indicative of a request for (e.g., at least one] leading AIML UE capability (e.g., by a respective, different tag], unindicative of a request for (e.g., at least one] complementary AIML UE capability (e.g., by a respective tag], or unindicative of a request for (e.g., at least one] leading AIML UE capability (e.g., by a respective, different tag].

The AIML UE capability enquiry and/or the subsequent AIML UE capability enquiry may indicate, what kind of AIML UE capability is enquired. For instance, the (e.g., subsequent] AIML UE capability enquiry may be indicative of a request for at least one complementary AIML UE capability. For instance, this may be the case within the method according to the second exemplary aspect.

For instance, the (e.g., subsequent] AIML UE capability enquiry may be indicative of a request for at least one leading AIML UE capability. For instance, this may be the case within the methods according to the first and second exemplary aspect.

For instance, the (e.g., subsequent] AIML UE capability enquiry may be unindicative of a request for at least one complementary AIML UE capability. For instance, this may be the case within the method according to the first exemplary aspect.

For instance, the (e.g., subsequent] AIML UE capability enquiry may be unindicative of a request for at least one (e.g., or any] leading AIML UE capability. For instance, this may be the case within the method according to the first or second exemplary aspect. For instance, the UE may provide any applicable leading AIML UE capability in response. Alternatively, the AIML UE capability enquiry may be indicative of at least one complementary AIML UE capability and the respective method may further comprise determining at least one leading AIML UE capability, e.g., based on the indicated complementary AIML UE capability and/or the hierarchical AIML UE capability structure. Additionally or alternatively, an AIML UE capability enquiry devoid of an indication of a leading AIML UE capability may be provided by the network node and/ or obtained by the UE, in case the UE already provided a respective leading AIML UE capability to the network node.

An indication of a request may for instance, in any of the disclosed cases, correspond to a respective tag, information element and/ or a bit within the respective enquiry.

According to an embodiment of the first and second example aspect, the leading AIML UE capability is (e.g., determined to be] at least one of indicative of a support of at least one (e.g., high-level, generic] AIML-functionality (e.g., as opposed to availability and/or applicability e.g., for a given (e.g., current] AIML use case; e.g., without enabling the AIML-functionality by itself, i.e., without an AIML UE complementary capability; e.g., wherein a set of supported AIML functionalities are invariable, e.g., known from the time of manufacturing], comprises one bit (e.g., per AIML functionality; e.g., indicating support (e.g., as true or false]], static (e.g., invariable] over time (e.g., or variable over time], independent of at least one (e.g., or any] external (e.g., scenario, dataset, indoor or outdoor location of the apparatus] or internal (e.g., UEs (e.g., currently installed] AIML models, training assumptions, training sets, computational resources, memory, battery] condition (e.g., of the apparatus, of another apparatus to which the apparatus is connected (e.g., a network node], or of a network the apparatus is connected to (e.g., overload condition]], provided in an invariable way (e.g., UE-signaling type does not change], indicative of a (e.g. supported] AIML use case (e.g., generic and/or higher level, not specific functionality] (e.g., leading UE capability may indicate that (e.g., in principle] at least one AIML functionality is supported under a given (e.g., current] use case], comprises at least one parameter of an AIML functionality (e.g., in addition to indicating support of at least one AIML functionality; e.g., at least partially non-AIML-related], configured to enable a network node to provide a (e.g., one of multiple possible per AIML use case and/or per AIML functionality] reference (e.g., and/or basic] AIML configuration (e.g., based on the at least one parameter; e.g., but not more; e.g., not a full, functional configuration; e.g., wherein the reference AIML configuration insufficient for AIML function implementation and/or requiring at least one complementary AIML UE capability in order to construct a (e.g., complete] AIML configuration], associated to at least two (e.g., or more, e.g., at least 3, 4, 5, 6, 7, 8, 16, 32] complementary AIML UE capabilities, insufficient (e.g., without at least one (e.g., associated] complementary AIML UE capability] for an implementation of at least one AIML-functionality (e.g., a functionality associated with the leading AIML UE capability].

The leading AIML UE capability may indicate a support (e.g. by the UE] of at least one AIML functionality. In particular, the supported AIML functionality may correspond to a high level, generic, and/or theoretic AIML functionality. The leading AIML UE capability may correspond to a coarsegrained, unspecific, categoric indication, that, e.g. in principle, certain AIML functionality and/or a certain range of AIML functionalities is providable by the apparatus performing and/or controlling the method according to the first and/or second example aspect.

Providing a given leading AIML UE capability to a network node may in particular be on indicative of an actual availability and/or applicability of the respective AIML functionality indicated by the leading AIML UE capability.

A leading AIML UE functionality may be indicated and/ or comprise, one bit. For instance, one bits may be provided, e.g. by the UE, per UE ML capability and/or functionality. A leading AIML UE functionality may be indicated as an indication of support, for instance as a bit, which may indicate that a given AIML functionality is supported or not. For instance, at least one, multiple, e.g. a set of, AIML functionalities corresponding indicated by at least one or more leading AIML UE capabilities may be invariable over time. For instance, the UE may be equipped, e.g. from the time of manufacturing, with at least one or more AIML functionalities indicated by respective leading AIML UE capabilities.

In particular, at least one and/or any leading AIML UE capability may be independent of at least one (e.g. or any] condition. A condition may for instance be an external condition, such as for instance at least one or a given scenario, e.g. regarding connectivity of the apparatus performing and/or controlling the method according to the first and/or second example aspect, a data set, for instance and availability of data set, for instance data sets not residing within a memory of the apparatus (e.g., but in a different, remote device], or an indoor and/or outdoor location of the apparatus. A condition may additionally or alternatively be internal, i.e. related to the apparatus performing and/or controlling the method according to the first and/or second example aspect. For instance, an internal condition may relate to at least one of an AIML model (e.g. whether a given AIML model is (e.g. currently] installed, at least one assumption on which training of an AIML model was based, a training set, at least one computational resource such as for instance an available random access memory, or a (e.g. current] battery state, for instance a state of charge.

The leading AIML UE capability may be provided in an invariable way. For instance, the type of signaling used, e.g., RRC, MAC CE, DCI and/ or an NAS signaling may stay invariable.

The leading AIML UE capability may be indicative of at least one use case. For instance, the leading AIML UE capability may indicate that a (e.g. supported] AIML functionality may be available for a given use case.

The leading AIML UE capability may comprise at least one parameter of a given AIML functionality. For instance, the network node obtaining an indication of at least one leading AIML UE capability may be enabled to generate and/ or construct an AIML configuration. Such an AIML configuration may for instance be based (e.g. only] on the leading AIML UE capability. In this case, the AIML configuration may be considered a baseline, reference and/or root AIML configuration.

For instance, AIML configurations may be differentiated into functional (e.g. full] AIML configurations and reference AIML configurations. For instance, a leading AIML UE capability may (e.g. only] enable a network node to construct a reference AIML configuration. A reference AIML configuration may be insufficient to fully specify an AIML functionality at the UE. In order to construct a full and/ or functional network node may require, e.g. in addition to a leading AIML configuration, at least one complementary AIML UE capability, e.g., corresponding to the leading AIML UE capability. For instance, a leading AIML UE capability may be associated to one (e.g. single] complementary AIML UE capability. Additionally or alternatively, a leading AIML UE capability may be associated to more than one complementary AIML UE capabilities, for instance at least 2, 3, 4, 5, 6, 7, 8, 16, 32 or more complementary AIML UE capabilities.

According to an embodiment of the first and second example aspect, at least one of the subsequent AIML UE capability enquiry is configured to request at least one (e.g., or multiple] complementary AIML UE capability (e.g., associated (e.g., respectively] with the at least one leading AIML UE capability; e.g., according to the hierarchical AIML UE capability structure].

According to an embodiment of the first and second example aspect, at least one of the at least one determined complementary AIML UE capability is determined based on the subsequent AIML UE capability enquiry, the at least one determined complementary AIML UE capability is determined based on the hierarchical AIML UE capability structure, the at least one determined complementary AIML UE capability is associated to (e.g., specific to] the at least one leading AIML UE capability, the at least one determined complementary AIML UE capability is determined based on at least one condition (e.g., external or internal], or the determining of at least one complementary AIML UE capability comprises analyzing at least one (e.g., current] AIML UE capability (e.g., of the apparatus; e.g., for a (e.g., the current] use case](e.g., based on at least one condition (e.g., external and/or internal]].

Receiving the subsequent AIML UE capability enquiry may cause the apparatus performing and/or controlling the method according to the first and/or second example aspect to determine at least one complementary AIML UE capability. The determining may at least partially be based on the hierarchical AIML UE capability structure. For instance, the subsequent AIML UE capability enquiry may indicate a leading AIML UE capability, to which at least one complementary AIML UE capability is required. Based on the hierarchical AIML UE capability structure, the method may be able to deter line the applicable complementary AIML UE capabilities, e.g., which correspond to the leading AIML UE capability.

The complementary AIML UE capability may be associated to the (e.g., previously deter her mind and/or provided] leading AIML UE capability.

For instance, the at least one determines complementary AIML UE capability be determined based on at least one condition, for instance a condition external and/or internal to an apparatus performing and/or controlling the method according to the first and/or second example aspect. Determining all at least one complementary AIML UE capability may comprise analyzing at least one AIML UE capability. For instance, the apparatus performing and/or controlling the method according to the first and/ or second example aspect may analyze, whether a certain (e.g. requested and/ or enquired, e.g., based on the (e.g., subsequent] AIML UE capability enquiry] complementary AIML UE capability is currently providable by the apparatus.

According to an embodiment of the first and second example aspect, at least one complementary AIML UE capability associated to the at least one leading AIML UE capability is (e.g., determined to be] at least one of indicative of an (e.g., current] availability (and/or applicability] of at least one AIML functionality (e.g., for a given use case, e.g., may provide information required for using a given AIML functionality (e.g., for the given use case]](e.g., low-level, detailed] AIML-functionality (e.g., as opposed to (e.g., mere] support (e.g., indicated by a leading AIML UE capability e.g., for a given (e.g., current] AIML use case; e.g., while enabling the AIML-functionality (e.g., by itself], e.g., in combination with an associated AIML UE lead capability; e.g., availability and/or applicability vary over time], comprising (e.g., only] one bit (e.g., or at least 2, 3, 4, 5, 6, 7, 8, 16, or 32 bits], larger in information (e.g., is represented by more bits] than a leading AIML UE capability (e.g., the leading UE capability to which the complementary UE capability is associated], variable over time, dependent of at least one external (e.g., scenario, dataset, indoor or outdoor location of the apparatus] or internal (e.g., UEs (e.g., currently installed] AIML models, training assumptions, training sets, computational resources, memory, battery] condition, indicative of a non-AIML UE capability (e.g., may be related inter alia, but not only to AIML], indicative of at least one additional information, wherein the additional information comprises at least one of at least one parameter of an AIML functionality (e.g., required to make use of the functionality; e.g., enabling a (e.g., full and/or functional] AIML configuration (e.g., at the networkside], e.g., not (e.g., only] a reference AIML configuration], a group of features (e.g., wherein the group is at least one of a list, a sequence or a set of parameters], or a list of (e.g., AIML (e.g., functionality]] parameters, configured to enable a network node to provide a (e.g., full and/ or functional] AIML configuration (e.g., based on the at least one parameter], sufficient (e.g., jointly with a respective associated leading AIML UE capability] for implementation of at least one AIML-functionality (e.g., associated with the leading AIML UE capability], configured to be provided by a same or by different means than the leading AIML UE capability, or associated to a single (e.g., or more than one] leading AIML UE capability. The complementary AIML UE capability may complement the leading AIML unique ability. For instance, the complementary AIML UE capability may be indicative of an (e.g. current] availability and/or applicability of at least one AIML functionality. The complementary AIML UE capability may thus indicate, whether the apparatus performing and/or controlling the method according to the first and/or second example aspect is (e.g. currently] able to provide a given AIML functionality. The complementary AIML UE capability may thus (e.g. in combination with the corresponding leading AIML UE capability] enable, e.g. to the network node, to provide a complete, functional and/or affordable AIML configuration and thus fully configure a given AIML functionality in accordance with the (e.g., current] ability of the UE.

The complementary AIML UE capability may be valid for a limited amount of time. For instance the apparatus performing and/or controlling the method according to the first and/or second example aspect may provide, e.g. along with the complementary AIML UE capability, an indication of an exploration time, after which the complementary AIML UE capability may not be up-to-date anymore. Additionally or alternatively, the network side, for example the network node obtaining the at least one complementary AIML UE capability may disregard the complementary AIML UE capability after a predefined timespan.

The complementary AIML UE capability may comprise and/or be represented by (e.g. only] one bit. For instance, the complementary AIML UE capability may be represented by more than one bit, for instance at least 2, 3, 4, 5, 6, 7, 8, 16, 32, 64, 128 or more bits.

In particular, the complementary AIML UE capability may be larger in information content than a (e.g. corresponding] leading AIML capability. For instance, the leading AIML UE capability may comprise a single bit only, which indicates whether a given AIML functionality is supported by the apparatus or not. On the other hand, a complementary AIML unique ability may further indicate details, such as for instance parameters for the corresponding AIML functionality, and thus may be larger in information content than the leading AIML UE capability.

The complementary AIML UE capability may be variable over time. For instance, the complementary AIML UE capability may be dependent on at least one condition, for instance external or internal to the apparatus performing and schedule controlling the method according to the first and/or second example aspect. Determining the complementary AIML UE capability may involve evaluating at least one condition.

The complementary AIML UE capability may be indicative of at least one non-AIML UE capability. The complementary AIML UE capability may be indicative of at least one additional information. For instance, such additional information may relate to AIML functionality parameters, AIML functionalities and/or AIML functionality groups.

The complementary AIML UE capability may be configured to enable a network node to provide a (e.g., full and/or functional] AIML configuration. For instance, at least one AIML functionality parameter provided with the complementary AIML UE capability may be used by a receiving network node to construct a full and/or functional AIML configuration.

The complementary AIML UE capability may be sufficient (e.g., jointly with a respective associated leading AIML UE capability and/or at least partially based on the hierarchical AIML UE capability structure] for an implementation and/ or configuration of at least one AIML-functionality (e.g., associated with the leading AIML UE capability].

The complementary AIML UE capability may be configured to be provided by a same or by different means than the leading AIML UE capability. For instance, the leading AIML UE capability may be provided semi-statically, e.g., by RRC, while a complementary AIML UE capability may be provided in a dynamic fashion, e.g. DCI or uplink control information, UCI, or by a MAC CE.

A given complementary AIML UE capability may for instance be associated to (e.g., only] a single (e.g., or more than one] leading AIML UE capability.

According to an embodiment of the first and second example aspect, at least one of the leading AIML UE capability or the complementary AIML UE capability is grouped depending on a type of a mobile device (e.g., UE; e.g., extended reality, XR, virtual reality, VR, RedCap, enhanced mobile broadband, eMBB, massive Machine Type Communication, mMTC , or combinations thereof).

Signaling off AIML capabilities may be unified by grouping based on the type of war device.

According to an embodiment of the first and second example aspect, an AIML use case comprises at least one of beam management,

(e.g., UE) positioning, channel state information, CSI, compression,

CSI prediction, or

UE mobility. According to an embodiment of the first and second example aspect, an AIML functionality is a new radio, NR,/6G radio feature at least one of assisted or realized based on an AIML.

According to a third example aspect, a method is disclosed, comprising: providing (e.g., by RRC signaling, MAC-CE, DCI, NAS signaling and/or combinations thereof] (e.g., to a UE] an AIML UE capability enquiry, obtaining (e.g., by RRC signaling, MAC-CE, DCI, NAS signaling and/or combinations thereof] (e.g., from the UE] an initial AIML UE capability response, wherein the initial AIML UE capability response indicates at least one leading AIML UE capability, wherein the leading AIML UE capability belongs to a hierarchical AIML UE capability structure (e.g., known at the UE and/or network node], wherein the hierarchical (e.g., branched and/or step-wise] AIML UE capability structure defines an association between the at least one leading AIML UE capability and at least one complementary AIML UE capability (i.e., the complementary AIML UE capability is associated to (e.g., dependent on, appended by, belongs to, configured to further specify] a respective (e.g., single one] of the at least one leading AIML UE capability according to the hierarchical AIML UE capability structure].

According to a fourth example aspect, a method is disclosed, comprising: means for providing (e.g., by RRC signaling, MAC-CE, DCI, NAS signaling and/or combinations thereof) (e.g., to a UE) an AIML UE capability enquiry, and means for obtaining (e.g., by RRC signaling, MAC-CE, DCI, NAS signaling and/or combinations thereof] (e.g., from the UE) an initial AIML UE capability response, wherein the initial AIML UE capability response indicates at least one leading AIML UE capability and at least one complementary AIML UE capability, wherein the leading AIML UE capability and the complementary AIML UE capability belong to a hierarchical AIML UE capability structure (e.g., known at the UE and/or network node), wherein the hierarchical (e.g., branched and/or step- wise) AIML UE capability structure defines an association between the at least one leading AIML UE capability and at least one complementary AIML UE capability (i.e., the complementary AIML UE capability is associated to (e.g., dependent on, appended by, belongs to, configured to further specify) a respective (e.g., single one) of the at least one leading AIML UE capability according to the hierarchical AIML UE capability structure).

These methods according to the third and/ or fourth example aspect may for instance be performed and/or controlled by an apparatus, for instance a server. Alternatively, the respective method may be performed and/or controlled by more than one apparatus, for instance a server cloud comprising at least two servers. Alternatively, the respective method according to the third and/ or fourth example aspect may for instance be performed and/or controlled by an electronic device, e.g. a node in a communication system and/or by a terminal device, e.g., a user equipment (UE], For instance, the method may be performed and/ or controlled by using at least one processor of the electronic device.

According to a further example aspect, a computer program is disclosed, the computer program when executed by a processor causing an apparatus, for instance a server, a network node or a terminal device, e.g., a UE, to perform and/or control the actions of the method according to the third and/or fourth example aspect.

The computer program may be stored on computer-readable storage medium, in particular a tangible and/or non-transitory medium. The computer readable storage medium could for example be a disk or a memory or the like. The computer program could be stored in the computer readable storage medium in the form of instructions encoding the computer-readable storage medium. The computer readable storage medium may be intended for taking part in the operation of a device, like an internal or external memory, for instance a Read-(e.g., Only] Memory [ROM] or hard disk of a computer, or be intended for distribution of the program, like an optical disc.

A terminal device, e.g., a user equipment [UE] may for instance correspond to a mobile device such as for example a mobile phone, tablet, smartwatch, a laptop, a Personal Digital Assistant [PDA] device, a wearable, an Internet-of-Things (IOT) device, an IIOT (Industrial IOT] device, a vehicle and/or combinations thereof. Such a user equipment may also be referred to as user device. A network node may correspond to a component of a communication network such as for instance a Base Transceiver Station (BTS], a nodeB, an evolved node B (eNB], a Next Generation NodeB (gNB], a distributed unit [DU], a central unit (CU] and/or combinations thereof.

According to an embodiment of the fourth example aspect the initial AIML UE capability response is provided as a single (e.g., coherent, uninterrupted] message (e.g., comprising a first bit indicating a leading AIML UE capability and a second bit (e.g., or multiple second bits] indicating a corresponding complementary AIML UE capability].

According to an embodiment of the third example aspect, the initial AIML UE capability response is unindicative (e.g., other than by the at least one leading AIML UE capability itself] of a (e.g., at least one and/or any] complementary AIML UE capability associated to (e.g., at least one of or any of] the at least one leading AIML UE capability indicated by the initial AIML UE capability response.

According to an embodiment of the third and fourth example aspect, the method further comprises providing (e.g., by RRC signaling, MAC-CE, DCI, NAS signaling and/or combinations thereof] (e.g., if the initial AIML UE capability response is unindicative of any complementary AIML UE capability] a subsequent AIML UE capability enquiry (e.g., to the UE] based on the initial AIML UE capability response.

According to an embodiment of the third and fourth example aspect, the method further comprises obtaining (e.g., by RRC signaling, MAC-CE, DCI, NAS signaling and/or combinations thereof] (e.g., in response to the subsequent AIML UE capability enquiry] (e.g., from the UE] a subsequent AIML UE capability response, wherein the subsequent AIML UE capability response indicates at least one (e.g., determined] complementary AIML UE capability (e.g., based on the subsequent AIML UE capability enquiry].

According to an embodiment of the third and fourth example aspect, the method further comprises providing (e.g., by RRC signaling, MAC-CE, DCI, NAS signaling and/or combinations thereof] (e.g., to the UE] an AIML configuration (e.g., RRCReconfiguration signaling] based on the obtained leading AIML UE capability and/or complementary AIML UE capability.

According to an embodiment of the third and fourth example aspect, at least one of the AIML UE capability enquiry, the initial AIML UE capability response, the leading AIML UE capability and/or the complementary AIML UE capability relate to at least one of an AIML functionality, an AIML use case or an AIML life cycle management phase (e.g., a phase such as e.g., training, inference, performance monitoring, data collection, update, fine-tuning and/or combinations thereof) (e.g., and may additionally be partly unrelated to an AIML functionality or an AIML use case).

According to an embodiment of the third and fourth example aspect, wherein the AIML UE capability enquiry is at least one of indicative of a request for (e.g., at least one) complementary AIML UE capability (e.g., by a respective tag), indicative of a request for (e.g., at least one) leading AIML UE capability (e.g., by a respective, different tag), unindicative of a request for (e.g., at least one) complementary AIML UE capability (e.g., by a respective tag), or unindicative of a request for (e.g., at least one) leading AIML UE capability (e.g., by a respective, different tag).

According to an embodiment of the third and fourth example aspect, wherein the leading AIML UE capability is (e.g., determined to be) at least one of indicative of a support of at least one (e.g., high-level, generic) AIML-functionality (e.g., as opposed to availability and/or applicability e.g., for a given (e.g., current) AIML use case; e.g., without enabling the AIML-functionality by itself, i.e., without an AIML UE complementary capability; e.g., wherein a set of supported AIML functionalities are invariable, e.g., known from the time of manufacturing), comprises one bit (e.g., per AIML functionality; e.g., indicating support (e.g., as true or false)), static (e.g., invariable) over time (e.g., or variable over time), independent of at least one (e.g., or any) external (e.g., scenario, dataset, indoor or outdoor location of the apparatus) or internal (e.g., UEs (e.g., currently installed) AIML models, training assumptions, training sets, computational resources, memory, battery) condition (e.g., of the apparatus, of another apparatus to which the apparatus is connected (e.g., a network node), or of a network the apparatus is connected to (e.g., overload condition)), provided in an invariable way (e.g., UE-signaling type does not change), indicative of a (e.g. supported) AIML use case (e.g., generic and/or higher level, not specific functionality) (e.g., leading UE capability may indicate that (e.g., in principle) at least one AIML functionality is supported under a given (e.g., current) use case), comprises at least one parameter of an AIML functionality (e.g., in addition to indicating support of at least one AIML functionality; e.g., at least partially non-AIML-related), configured to enable a network node to provide a (e.g., one of multiple possible per AIML use case and/or per AIML functionality) reference (e.g., and/or basic) AIML configuration (e.g., based on the at least one parameter; e.g., but not more; e.g., not a full, functional configuration; e.g., wherein the reference AIML configuration insufficient for AIML function implementation and/or requiring at least one complementary AIML UE capability in order to construct a (e.g., complete] AIML configuration], associated to at least two (e.g., or more, e.g., at least 3, 4, 5, 6, 7, 8, 16, 32] complementary AIML UE capabilities, insufficient (e.g., without at least one (e.g., associated] complementary AIML UE capability] for an implementation of at least one AIML-functionality (e.g., a functionality associated with the leading AIML UE capability].

According to an embodiment of the third and fourth example aspect, wherein at least one of the subsequent AIML UE capability enquiry is configured to request at least one (e.g., or multiple] complementary AIML UE capability (e.g., associated (e.g., respectively] with the at least one leading AIML UE capability].

According to an embodiment of the third and fourth example aspect, wherein at least one of the at least one indicated complementary AIML UE capability is determined based on the subsequent AIML UE capability enquiry, the at least one indicated complementary AIML UE capability is determined based on the hierarchical AIML UE capability structure, the at least one obtained complementary AIML UE capability is associated to (e.g., specific to] the at least one leading AIML UE capability, the at least one indicated complementary AIML UE capability is determined based on at least one condition (e.g., external or internal], or a determining of at least one complementary AIML UE capability comprises analyzing at least one (e.g., current] AIML UE capability (e.g., of the apparatus; e.g., for a (e.g., the current] use case](e.g., based on at least one condition (e.g., external and/or internal]].

According to an embodiment of the third and fourth example aspect, wherein at least one complementary AIML UE capability associated to the at least one leading AIML UE capability is (e.g., determined to be] at least one of indicative of an (e.g., current] availability (and/or applicability] of at least one AIML functionality (e.g., for a given use case, e.g., may provide information required for using a given AIML functionality (e.g., for the given use case]] (e.g., low-level, detailed] AIML-functionality (e.g., as opposed to (e.g., mere] support (e.g., indicated by a leading AIML UE capability e.g., for a given (e.g., current] AIML use case; e.g., while enabling the AIML-functionality (e.g., by itself], e.g., in combination with an associated AIML UE lead capability; e.g., availability and/or applicability vary over time], comprising (e.g., only] one bit (e.g., or at least 2, 3, 4, 5, 6, 7, 8, 16, or 32 bits], larger in information (e.g., is represented by more bits] than a leading AIML UE capability (e.g., the leading UE capability to which the complementary UE capability is associated], variable over time, dependent of at least one external (e.g., scenario, dataset, indoor or outdoor location of the apparatus] or internal (e.g., UEs (e.g., currently installed] AIML models, training assumptions, training sets, computational resources, memory, battery] condition, indicative of a non-AIML UE capability (e.g., may be related inter alia, but not only to AIML], indicative of at least one additional information, wherein the additional information comprises at least one of at least one parameter of an AIML functionality (e.g., required to make use of the functionality; e.g., enabling a (e.g., full and/or functional] AIML configuration (e.g., at the networkside], e.g., not (e.g., only] a reference AIML configuration], a group of features (e.g., wherein the group is at least one of a list, a sequence or a set of parameters], or a list of (e.g., AIML (e.g., functionality]] parameters, configured to enable a network node to provide a (e.g., full and/ or functional] AIML configuration (e.g., based on the at least one parameter], sufficient (e.g., jointly with a respective associated leading AIML UE capability] for implementation of at least one AIML-functionality (e.g., associated with the leading AIML UE capability], configured to be provided by a same or by different means than the leading AIML UE capability, or associated to a single (e.g., or more than one] leading AIML UE capability.

According to an embodiment of the third and fourth example aspect, wherein at least one of the leading AIML UE capability or the complementary AIML UE capability is grouped depending on a type of a mobile device (e.g., UE; e.g., extended reality, XR, virtual reality, VR, RedCap, enhanced mobile broadband, eMBB, massive Machine Type Communication, mMTC , or combinations thereof).

According to an embodiment of the third and fourth example aspect, an AIML use case comprises at least one of beam management, positioning, channel state information, CSI, compression,

CSI prediction, or

UE mobility.

According to an embodiment of the third and fourth example aspect, an AIML functionality is a new radio, NR,/6G radio feature at least one of assisted or realized based on an AIML.

According to a fifth example aspect, a method is disclosed, comprising: by a first apparatus (e.g., a network node], providing (e.g., by RRC signaling, MAC-CE, DCI, NAS signaling and/or combinations thereof] (e.g., to a second apparatus] an AIML UE capability enquiry, obtaining (e.g., by RRC signaling, MAC-CE, DCI, NAS signaling and/or combinations thereof] (e.g., from the second apparatus] an initial AIML UE capability response, wherein the initial AIML UE capability response indicates at least one leading AIML UE capability, wherein the leading AIML UE capability belongs to a hierarchical AIML UE capability structure (e.g., known at the UE and/or network node], wherein the hierarchical (e.g., branched and/or step-wise] AIML UE capability structure defines an association between the at least one leading AIML UE capability and at least one complementary AIML UE capability (i.e., the complementary AIML UE capability is associated to (e.g., dependent on, appended by, belongs to, configured to further specify] a respective (e.g., single one] of the at least one leading AIML UE capability according to the hierarchical AIML UE capability structure] and by a second apparatus (e.g., a terminal device], obtaining (e.g., by radio resource control, RRC, signaling, medium access control control element, MAC-CE, downlink control information, DCI, non-access stratum, NAS, signaling and/or combinations thereof] the artificial intelligence (e.g., and/or machine learning], AIML, user equipment, UE, capability enquiry (e.g., from the first apparatus], determining (e.g., based on the obtained AIML user equipment, UE, capability enquiry] the at least one leading AIML UE capability, and providing (e.g., by RRC signaling, MAC-CE, DCI, NAS signaling and/or combinations thereof] (e.g., in response to the obtaining of the UE capability enquiry; e.g., to the first apparatus] the initial AIML UE capability response, wherein the initial AIML UE capability response indicates the at least one leading AIML UE capability.

According to a sixth example aspect, a method is disclosed, comprising: by a first apparatus (e.g., a network node], means for providing (e.g., by RRC signaling, MAC-CE, DCI, NAS signaling and/or combinations thereof] (e.g., to a second apparatus] an AIML UE capability enquiry, and means for obtaining (e.g., by RRC signaling, MAC-CE, DCI, NAS signaling and/or combinations thereof] (e.g., from the second apparatus] an initial AIML UE capability response, wherein the initial AIML UE capability response indicates at least one leading AIML UE capability and at least one complementary AIML UE capability, wherein the leading AIML UE capability and the complementary AIML UE capability belong to a hierarchical AIML UE capability structure (e.g., known at the UE and/ or network node], wherein the hierarchical (e.g., branched and/or step-wise] AIML UE capability structure defines an association between the at least one leading AIML UE capability and at least one complementary AIML UE capability (i.e., the complementary AIML UE capability is associated to (e.g., dependent on, appended by, belongs to, configured to further specify] a respective (e.g., single one] of the at least one leading AIML UE capability according to the hierarchical AIML UE capability structure], and by a second apparatus (e.g., a UE] means for obtaining (e.g., by radio resource control, RRC, signaling, medium access control control element, MAC-CE, downlink control information, DCI, non-access stratum, NAS, signaling and/or combinations thereof) the artificial intelligence (e.g., and/or machine learning), AIML, user equipment, UE, capability enquiry (e.g., from the first apparatus), means for determining (e.g., based on the obtained artificial intelligence/ machine learning, AIML, user equipment, UE, capability enquiry) the at least one leading AIML UE capability and at least one complementary AIML UE capability, and means for providing (e.g., by RRC signaling, MAC-CE, DCI, NAS signaling and/or combinations thereof] (e.g., in response to the obtaining of the UE capability enquiry; e.g., to the first apparatus) the initial AIML UE capability response, wherein the initial AIML UE capability response indicates the at least one leading AIML UE capability and the at least one complementary AIML UE capability.

BRIEF DESCRIPTION OF THE FIGURES

Fig. la,b shows a flow chart illustrating example embodiments of the present disclosure according to all example aspects;

Fig. 2 shows a signaling diagram in which example aspects of the disclosure are illustrated;

Fig. 3 shows a signaling diagram in which example aspects of the disclosure are illustrated;

Fig. 4 shows a signaling diagram in which example aspects of the disclosure are illustrated;

Fig. 5 shows a signaling diagram in which example aspects of the disclosure are illustrated;

Fig. 6 shows a flow chart illustrating an embodiment according to the first example aspect of the disclosure;

Fig. 7 shows a flow chart illustrating an embodiment according to the second example aspect of the disclosure;

Fig. 8 shows a flow chart illustrating an embodiment according to the third example aspect of the disclosure;

Fig. 9 shows a flow chart illustrating an embodiment according to the fourth example aspect of the disclosure;

Fig. 10 shows a block diagram illustrating an embodiment according to the first example aspect of the disclosure;

Fig. 11 shows a block diagram illustrating an embodiment according to the second example aspect of the disclosure;

Fig. 12 shows a block diagram illustrating an embodiment according to the third example aspect of the disclosure; Fig. 13 shows a block diagram illustrating an embodiment according to the fourth example aspect of the disclosure;

Fig. 14 shows a schematic illustration of examples of tangible and non-transitory computer- readable storage media.

DETAILED DESCRIPTION OF THE FIGURES

The following description serves to deepen the understanding of the present disclosure and shall be understood to complement and be read together with the description of example embodiments of the present disclosure as provided in the above SUMMARY section of this specification.

A method according to the example aspects may enable and/or trigger at least one AIML functionality. It is proposed to group AIML UE capability into two categories or (e.g., main] branches.

In a first category, leading (e.g., unconditional] AIML UE capabilities may be defined. An unconditional AIML UE capability information element may for instance represent (e.g., a category of] at least one static UE hardware characteristics. Such leading AIML UE capabilities may convey simple and at least in some instances limited information on whether the UE is able to provide an AIML-enabled feature and/or functionality.

At least in some cases, a leading AIML UE capability may relate to (e.g., an enabling of] an AIML specific use case or purpose. For instance, a leading AIML UE capability may be single bit indication (e.g., per use case]. For instance, a leading AIML UE capability may be a single bit for at least one of an AIML enabled beam management, an AIML enabled positioning accuracy enhancement, or an AIML-enabled CSI feedback enhancement (CSI reporting]. For instance, a leading AIML UE capability information element may correspond to a sequence or list for one or more (e.g., AIML UE capability] use cases. A respective element in the list may be defined based a collective set of grouped components that may determine a particular use case support. An element itself may be a (e.g., deterministic] indicator for further components and/or conditional AIML capabilities. For instance, a leading AIML UE capability may comprise a list of three values which relate to beam management optimization, positioning accuracy enhancement, and channel state information, CSI, feedback enhancement. The leading AIML UE capability may be set to a positive value for beam management optimization (e.g., value "true” or "supported”; e.g., bit-value 1 instead of 0], This bit may imply that AIML-BeamManagement is an AIML UE capability use case, e.g., and that the UE is capable to support at least one conditional AIML UE capability for this use case.

A second branch and/or category of AIML UE capabilities are complementary (e.g., conditional] AIML UE capabilities. A conditional AIML UE capability may relate to a (e.g., category of] leading (e.g., unconditional] AIML UE capability. Conditional AIML UE capability may convey semi-static information (e.g., information elements]. Such information elements] may take varying (e.g., alternating] values which may change dynamically over a life-time of a UE. This may be in contrast to leading AIML UE capabilities which may stay constant across a life-time of a UE.

A complementary AIML UE capability may represent a e.g., sequence of components that list and/ or a list of applicable conditions to determine to what extend an AIML feature (and/or functionality] may be configured (e.g., is configurable] and may be activated (e.g., is activatable], e.g., at a given point of time (e.g., now or in the future].

The branch of complementary AIML UE capabilities may again be divided into two sub-branches. One sub-branch (e.g., the complementary AIML UE capabilities in this sub-branch] may be dependent on generic AI/ML conditions (e.g. related to life cycle management of AI/ML model, corresponding to AI/ML model or training, monitoring or inference], the while the other sub-branch of complementary AIML UE capabilities may consider (e.g., different] use cases.

The first sub-branch of the complementary AIML UE capabilities may relate to an AIML purpose that may differentiate between generic applicability conditions for at least one or more AIML functionalities. For instance, one bit or multiple bits (e.g., group bits] may indicate at least one AIML functionality condition of e.g., inference, monitoring, training, maximum number of supported AIML- enabled features, maximum number of AI/ML-enabled functionalities, model-ID based LCM AI/ML support, UE-sided model support and/or combinations thereof. This sub-branch may change over time, e.g., depending on a user preference and/or at least one (e.g., internal or external] condition. If a user (e.g., of an apparatus performing and/or controlling the first and/or second method] is already involved in a maximum number of supported ML functionalities for one specific use case (e.g. Beam Management], they (e.g., their device, e.g., UE] may decide to indicate a conditional AIML capability information (e.g., bit] for another use case (e.g., enhanced positioning] and for a functionality (e.g., inference] to false. Even though at some other occasion a same user (e.g., a device of the user, e.g., a UE] may indicate its readiness to perform the previously deactivated functionality for a use case (e.g., inference for Positioning Enhancements], This sub-branch may enable a UE to signal different values of parameters when a corresponding (e.g., corresponding to a functionality to which the parameter applies] conditional AIML UE capability changes, gets invalid or becomes valid.

Another sub-branch of complementary AIML UE capabilities, may also correspond to (e.g., be signaled as] a group bits indication. This sub-branch may relate to at least one AIML function or functionalityspecific capability, wherein the capabilities may be raw UE capabilities. This sub-branch may be particularly compatible to a legacy approach wherein an (e.g., complete] set of multiple (e.g., various] individual UE capabilities is indicated. These UE capabilities may relate to an underlying (e.g., unconditional] use case. For instance, at least one or multiple groups of individual parameters such as: CSI-report framework, CSI-RS-Resource, codebook Parameters, beam management parameters, beam management SSB-CSI-RS, aperiodic- CSI-RS, beam report timing may refer to (e.g., raw] UE capabilities, e.g., per frequency band. A given (e.g., each] bit may individually not make a corresponding AIML feature working. However, (e.g., only] a (e.g., certain] set and (e.g., commonly] supported individual components may make an AIML-enabled feature functioning. This list may represent a sequence of groups of parameters which enable AIML feature optimization, e.g.: Ll-RSRP support, Beam Index support, CSI-RS support. Such group bits indication could be a consecutive set of supported or interlaced (e.g., parameters] depending on how the network (e.g., network node] may implement the corresponding functionalities.

For instance, at least one or more AIML-assisted beam management, BM, parameters may represent a sequence of components, e.g., wherein the (e.g., sequence of) components list(s) either an exhaustive or a less complete set of parameters which may determine to what extend an AIML-assisted feature (e.g., functionality) and/or an AIML-feature (e.g., -functionality) may be configured and/or activated. In case AIML-BM-Parameters comprise e.g., Measured DL Reference Signal (SSB, CSI-RS), Predicted Measurement Signal, Measurement periodicity, Measurement sets, this may imply that these are individual components and/or it may be determined that AIML-BeamManagement is a category (e.g., a current use case). For instance, a certain combination of parameters may determine at least one value of a leading (e.g., unconditional) AIML UE capability bit to be set as supported use case, functionality and/or category. If a UE is not capable of one component (e.g., (e.g., AIML-assisted) feature and/or functionality) which is (e.g., mandatorily) needed for a specific use case (e.g., category), this use case (e.g., category) in leading (e.g., unconditional) AIML Capability may be indicated as unsupported (e.g. lacking Beam Index Id implies the Beam Management optimization use case (e.g., category) under leading (e.g., unconditional) AIML capabilities may be signaled as unsupported).

For instance, a leading AIML UE capability may correspond to a static AIML capability whereas complementary AIML UE capability may correspond to a dynamic AIML capability. Where, leading AIML UE capability may be reported first. E.g., subsequently, complementary at least one AIML UE capability may be reported, e.g., based on a network request. For instance, complementary AIML UE capabilities may change over time and may additionally or alternatively include AIML feature (e.g., functionality) combinations with non-ML features, e.g., in terms of how both of (AIML and non-AIML) may be configured or can work together.

For instance, at least one leading AIML UE capability may be a higher level capability (e.g., compared to the complementary AIML UE capability). For instance, at least one (e.g., or all) leading AIML UE capabilities may correspond to use-case level information. For instance, complementary AIML UE capabilities may be a lower level and/or AIML functionality-level information, e.g., where leading information is reported first, followed by complementary information, e.g., based on network request. For instance, a supported use case may indicate that there is at least one ML functionality supported under the supported use case.

For instance, a leading AIML UE capability may indicate {support BM, support CSI, no support positioning}. E.g., the leading AIML UE capability may indicate parameters which may allow a basic/reference configuration to be provided to the UE. E.g., if requested by the network (e.g., a network node according to the third and/or fourth example aspect], detailed complementary AIML UE capabilities (e.g., capability parameters like CSI-report framework: xx, CSI-RS-Resource: yy, codebook Parameters: zz, ...] and/or possible combinations of the capability parameters may be requested from the UE (e.g., a UE according to the first and/or second aspect].

Complementary AIML UE capabilities may or may not change over time.

The following indications of leading and complementary AIML UE capability may adhere to a 2-step approach (e.g., according to the first and/or third example aspect; first leading capabilities are obtained and/or provided, then complementary capabilities] may adhere to a one-step approach (e.g., according to the second and/or fourth example aspect]. Leading AIML UE capabilities may correspond to (e.g., indicate] at least one or more supported functionalities. Complementary AIML UE capabilities may indicate (e.g., support and] availability (e.g., of a given AIML functionality]. Leading and complementary AIML UE capabilities may be reported together (e.g., according to the second and/or fourth example aspect], e.g., wherein the leading AIML UE capabilities may indicate a (e.g., general] capability support, i.e., the supported AIML functionality may not necessarily be (e.g., currently] available, e.g., but may be available for at least some conditions. Complementary AIML UE capabilities may indicate a (e.g., current] availability, i.e., whether the functionality is currently activatable. For instance, a first AIML functionality may correspond to {direct fingerprinting positioning up to 16 TRPs, CIR length of X, ...} leading (support]: True, complementary (availability]: False. For instance, such an indication may be provided and/or obtained if indoor direct positioning is supported but the UE is currently outdoors.

For instance, an AIML use case may correspond to a new radio, NR, and/or 6G radio feature that may for instance be chosen as a problem to be solved using AIML. An example of a use case may be beam management, positioning and CSI compression, CSI prediction and in Rel-19 UE mobility.

An AIML feature and/ or an AIML functionality may correspond to a NR/ 6G radio feature which may for instance be realized and/or be assisted using AIML. The radio feature has a corresponding AIML functionality that may be enabled by the network (e.g., network node], e.g., in response to the UE capabilities being known (using either with leading and complementary, or both]. An AIML component may correspond to an information element [IE] that indicates an AI/ML capability that may be common or specific to an AI/ML use case.

According to some examples, a method to enable and/or trigger at least one AI/ML functionality may at least essentially be based on classifying a given AIML UE capability into two main groups. Leading AIML UE Capability may be understood as a primary, dominant, static, and higher of importance capability, for which UE signaling does not change under any conditions. A complementary UE capability may be understood as secondary UE AIML capability that may depend on at least one or more conditionfs],

A leading AIML UE capabilities may correspond to a flag, e.g., a static UE capability information element. The leading AIML UE capability may represent an information on UE support for at least one or more AIML operations. The leading AIML UE capability may be a static, e.g., a UE hardware, characteristic. The leading AIML UE capability may convey limited information on whether the UE can current provide AIML-enabled features. The leading AIML UE capability may be dominant and may be invariable over time

A complementary AIML UE capability may relate to a (e.g., respective] leading AIML UE capability. The complementary AIML UE capability may convey semi-static information. The complementary AIML UE capability may for instance correspond to at least one or more information elementfs], e.g., which may take alternate values changing dynamically over UE life-time. For instance, during a life cycle management, LCM, monitoring operation on beam management, BM, (e.g., only] reported predicted Beam index may be low granularity reporting. The network (e.g., network node according to the third and/or fourth example aspect] could ask for a higher granularity (say predicted Ll-RSRP correspond to those beam index]. In a result, a handling conditional AIML UE capabilities may need at least one pre-requisite by recognizing leading information.

Examples of different parameters and values in Leading and Complementary UE capabilities: Leading info may correspond to FG capability bits, e.g., bits = 1. Complementary info may correspond to Model, training, additional info, e.g., Model=false, training=false, LCM purpose=true

The following shows an example of a realization of a signaling of AIML UE capabilities in a 3GPP standard, e.g., in TS 38.331.

— AIML Rel-18 extensions:

UE-NR-Capability-vl800 ::= SEQUENCE { Leading-MLCapability-rl8 SEQUENCE {

—Support of any parameter in the list indicates the UE is capable of AI/ML-enabled functionality indicated by the particular parameter and supports complete set of individual necessary components but functionality availability is conditioned by corresponding Complementary capabilities ML-BM-rl8 ENUMERATED {supported} OPTIONAL, ML-Positioning-rl8 ENUMERATED {supported} OPTIONAL, ML-CSI-rl8 ENUMERATED {supported}

}

—Below is example of Complementary UE Capabilities for Beam Management, but same structure can apply per each Category for Positioning and CSI and others

Conditional-MLCapability-rl8 SEQUENCE { ML-BM-Conditions-rl8 SEQUENCE { training-rl8 ENUMERATED {supported} OPTIONAL, monitoring-rl8 ENUMERATED {supported} OPTIONAL, inference-rl8 ENUMERATED {supported} OPTIONAL, ML-model-based-rl8 ENUMERATED {supported} OPTIONAL, Maximum-ML-BM-rl8 INTEGER [1..16] OPTIONAL,

}

—Below is example of raw UE Capabilities for individual components for Beam Management, same structure can apply per Category for Positioning and CSI beamManagement-rl8 SEQUENCE {

ExemplaryRRCIEEnablingBM-rl8 ENUMERATED {supported} OPTIONAL, ExemplaryRRCIEEnablingBM-rl8 ENUMERATED {supported} OPTIONAL,

}

For instance, leading AIML UE capabilities may correspond to low-granularity and/or functionalityspecific information. E.g., leading AIML UE capability may correspond to a flag that may be an AIML UE capability information element, IE. The IE may represent a category of information on whether the UE can support AIML-enabled feature for a given use case. The leading AIML UE capability may be dominant and a leading UE Capability on an AIML UE capability, represented by a flag, may not change over time. A leading AIML UE capability may be related to AIML-specific use case and/or may be a single bit indication per use case. E.g., a single bit may be provided for AIML-enabled Beam Management, or AIML-enabled positioning accuracy enhancement, and/or AIML-enabled CSI feedback enhancement (CSI reporting],

A UE capability information element may correspond to a sequence and/or list for one or more categories. E.g., a given (e.g., each] element in the list may be defined based on a collective set of grouped components that determine particular category support as a whole. The element itself may be a deterministic indicator for further components and complementary AIML capabilities. For instance, if an AIML leading UE capability consisting a list of three values: Beam Management , Positioning accuracy enhancement, CSI feedback enhancement, where the list has set bit to positive value for Beam Management (value "true” or "supported”], this may imply that AIML-BeamManagement is a Category and the UE is capable to support complementary AI/ML Capabilities for the Category.

Examples of different parameters and values in leading and complementary UE capabilities comprise the following.

Leading info (high level]: Beam Management. Complementary information (lower level]:

Feature: BM-Casel: Beam information (Beam ids, Ll-RSRP, beam pairs], Predicted Top Kbeam(s] among a set of beams (beam id, Ll-RSRP or both], Beam information on predicted Top K beam(s] among a set of beams and RSRP of predicted Top Kbeam(s] among a set of beams (beam ids, Ll-RSRP] Feature: BM-Case2: Beam information (Beam ids, Ll-RSRP, beam pairs], Predicted Top Kbeam(s] among a set of beams (beam id, Ll-RSRP or both]

Leading info (high level]: Positioning. Complementary information (lower level]:

Feature: AIML_direct_DL_CIR_UEside, Functionality 1-01 : N’t = 64 only, N_port = 2 only, N_TRP = 12 only; Functionality 1-02: N’t = 128 only, N_port = 2,4 only, N_TRP = 1,...,18 only

Feature: AIML_assisted_DL_CIR_LMFside: Functionality 2-01: intermediate_feature= ToA only, N’t = 128 only, N_port = 1 only, N_TRP = 15 only.

Examples of functionalities where positioning features are enabled by configuration of specific set of UE conditions:

F eatur e : Al M L_dir ect_D L_CI R_U Eside

Functionality 1-01 : N’t = 64 only, N_port = 2 only, N_TRP = 12 only

Functionality 1-02: N’t = 128 only, N_port = 2,4 only, N_TRP = 1,...,18 only

Functionality 1-03: ....

Functionality 1-04: ....

F eatur e : Al M L_assisted_D L_CI R_LM Fside

Functionality 2-01: intermediate_feature= ToA only, N’t = 128 only, N_port = 1 only, N_TRP = 15 only Functionality 2-02: intermediate_feature= LOS/NLOS indication only, N’t = 128 only, N_port = 2,4 only, N_TRP = 1,...,18 only Functionality 2-03: ....

Functionality 2-04: ....

Figure la shows a UE implementation for structuring leading (e.g., unconditional] and complementary (e.g., conditional] AIML UE capabilities, which may for instance relate to AIML-enabled features.

A device (e.g., apparatus performing and/or controlling the method according to the first and/or second example aspect] may support setting at least one leading (e.g., unconditional] AIML UE capability bit. Otherwise, the device may be considered a legacy device (e.g., which may be capable of setting at least one raw and/ or individual capability, which may for instance be static over UE lifetime. An AIML-capable device may, e.g., upon determining that the device is capable of at least one AIML- enabled functionality, set a leading (e.g., unconditional] AIML UE capability bit to "true”. E.g., if the device has a (e.g., complete] set of components for a given AIML feature, it may set a leading (e.g., unconditional] AIML UE capability bit to "true”. For instance, if the device has a capability to maintain at least one (e.g., semi-static] (e.g., AIML UE] capability, the device may monitor at least one of the at least one capabilities over time and/ or may monitor at least one or multiple (e.g., different] condition (e.g., internal and/or external to the device], and may additionally provide (e.g., AIML UE] capability information to the network (e.g., network node according to the third and/or fourth aspect] accordingly. The device (e.g., UE] may set at least one or more complementary (e.g., conditional] AIML UE capability bits, e.g., based on the monitoring.

The method shown in Fig. la may reveal benefits in particular if in addition, the network (e.g., network node according to the third and/or fourth aspect] is capable to apply a disclosed approach for enquiring (e.g., fetching] (e.g., AIML] UE capabilities. Instead of signaling an (e.g., entire] set of individual UE capabilities (e.g., multiple components that UE capabilities may consist of], the network (e.g., network node according to the third and/ or fourth aspect] may fetch the relevant AIML UE capabilities based on its own conditions and capabilities. The complete set of functionalities can be provided only upon NW request, saving signaling overhead and revealing NW implementation from deducing what is applicable for AI/ML operation. In order to provide a minimized impact of dynamic UE capabilities (and conditional changes], this structure would allow efficient procedure for integrating gradual information exchange between the user and the NW on UE status, only when the NW is interested and ready to make use of the additional conditions.

Fig. lb demonstrates such on-demand enquiring of UE capabilities by the network node (e.g., according to the third and/ or fourth example aspect]. Instead of providing an (e.g., complete] list of (e.g., AIML UE capability related] parameters, which may for instance not be needed for an actual UE operation, the UE (e.g., network node according to the first and/or second example aspect] may indicate (e.g., only] selected (e.g., single/ simple] UE capabilities, e.g., regarding its AI/ML abilities. Additionally or alternatively, a (e.g., complete] set of AIML functionalities and/or AIML UE capabilities may be provided (e.g., only] upon a corresponding network-side request (e.g., AIML UE capability enquiry]. In this way, signaling overhead may be saved. The network node may be freed of the need to implement a deducing of what AIML functionality maybe applicable, e.g., for a given AI/ML operation.

In order to provide a minimized impact of dynamic UE capabilities (and conditional changes], the present disclosure proposes ways by which an efficient procedure for integrating gradual information exchange between the user and the network on UE status are provided, in particular (e.g., only] when the NW is interested and ready to make use of at least one of the additional conditions and/or AIML functionalities.

For instance, in step S101 of Fig. lb, a network node (e.g., according to the third and/ or fourth example aspect] may provide a network request (e.g., AIML UE capability request] to a UE (e.g., according to the first and/or second example aspect]. Based on whether the UE supports a hierarchical AIML UE capability structure, it may either (no-branch] provide one or multiple individual AIML-supported feature components, e.g., as individual bits per AIML feature and/or use case. If, on the other hand, the hierarchical AIML UE capability structure is supported (yes-branch], the UE may signal (e.g., as an AIML UE capability response] at least one or more leading (e.g., unconditional] AIML UE capabilities in a first step. The network node may then decide whether it requires (e.g., for which use cases and/ or functionalities] complementary (e.g., conditional] AIML UE capabilities. In case it does, it may enquire (e.g., as a subsequent AIML UE capability request], at least one complementary (e.g., conditional] AIML UE capability which may in turn be signaled by the UE (e.g., in the form of a subsequent AIML UE capability response].

Any device (e.g., UE, e.g., according to the first and/or second example aspect] may support a setting of at least one leading AIML UE capability bit. For instance, (e.g., only] a UE is capable of at least one AIML-enabled functionality may set a leading capability bit to "true”. If the UE has a complete set of components for a given AIML feature group (category/functionality/use case], it set may set a ‘leading AIML UE capability” bit to "true”. For instance if it has the capability to maintain its semi-static capabilities, monitor them over time and different conditions, and provide capability information to the network accordingly, it can set "Complementary UE capabilities” bits.

Fig. 2 shows a basic signaling diagram involving a two-step (e.g., reactive] implementation of an embodiment of the present disclosure between a UE 100 (e.g. according to the first and/or second example aspect] and a network node, RAN 200, (e.g. according to the third and/or fourth example aspect]. The UE 100 and network node 200 may be part of a system according to the fifth and/or sixth example aspect.

In a first step S101 the network node 200 provides an AIML UE capability enquiry, e.g., obtained by the UE 100. The UE 100, and step S102 determines AIML enabled features. The UE 100 then provides, in step SI 03 an AIML UE capability response indicating at least one leading AIML UE capability. For instance, the leading AIML UE capability may be indicated as at least one bit and/or as a bit group. E.g., an individual bit may indicate that a corresponding use case may be supported with at least one AIML functionality. The network node 100 may decide to request additional information and may, in step S105, provide a subsequent AIML UE capability enquiry. The subsequent AIML UE capability enquiry may indicate at least one desired complementary AIML UE capability. In step SI 07 the UE provides at least one complementary AIML UE capability. For instance, the complementary AIML UE capability may relate to a use case indicated as supported by the leading AIML UE capability provided in step 103. The complementary AIML UE capability may comprise at least one parameter which may be used by the network node 200 in order to configure a supported AIML functionality, e.g. for a given use case.

Fig. 3 demonstrates a more detailed embodiment illustrating split into UE capability transfer for leading and for complementary AIML UE capability. Closely related Fig. 4 details this process for the example embodiment wherein the AIML-enabled functionality is beam management. Fig. 4 will be described in detail below.

A UE may indicate that that AIML is enabled (e.g., in principle], whereas a model is not available yet. In such a case, a training in a UE may not be possible and/ or only signaling and/ or reporting may be possible for inference, activation, deactivation, switching, and selection of functionality/model if Model is ‘true’ [available].

To activate an AIML-enabled functionality beam management, e.g., with an enhanced AIML UE capability transfer for leading and complementary AIML UE capability split, at least some of the following steps may be taken.

In step S101, the network node 200 (e.g., according to the third and/or fourth example aspect] initiates a UECapabilityEnquiry message. Thus, the network node provides an AIML UE capability enquiry to the UE.

In step S102, the UE 100 may differentiate, e.g. according to a hierarchical AIML UE capability structure, capabilities to branches. According to a first branch, leading AIML UE capability may be identified and, according to a second branch, complementary AIML UE capabilities may be identified. For instance, (e.g., complementary] AIML UE capabilities may correspond to an ability to perform AIML model training, ability to perform functionality based inference, UE capabilities for AI/ML-enabled feature category.

For instance for beam management, a list of individual AIML UE capabilities may comprise: for BM- Case-1: Top-K beam predictions (support of predicting best-K NZP CSI-RS resources based on SSB and/or CSI-RS based RSRP measurements, Set B conditions: Measured DL RS (Downlink Reference Signal], Measured set pattern], associated individual parameters for CSI reporting: CSI-report framework, CSI-RS-Resource, codebook Parameters, beam management parameters, beam management SSB-CSI-RS, aperiodic- CSI-RS, beam report timing can refer to individual UE capabilities per frequency band.

In step S103, the UE 100 may provide, e.g., in a response to the UECapabilityEnquiry (S101], AI/ML Leading UE Capability (e.g., only]. This may for instance correspond to a notification to the network node about a generic UE support for a particular AIML-enabled feature and/or feature category and/or use case. For instance, AI/ML Leading capability may correspond to (e.g., indicate] AI/ML-based Beam Management = supported, Positioning = non-supported, CSI-reporting= non-supported. For instance, the indication of leading AIML UE capability may be realized by a new (e.g., meaning of such] UE Capability information element. A field in the signaling does not impose device readiness (e.g., feature availability] to be configured with the feature. It may instead (e.g., only] notifies the network about at least one or more further nested related (e.g., complementary] capabilities which may for instance be fetched, e.g., in order to determine UE readiness (e.g., availability of a respective AIML-enabled feature and/or functionality].

Step S104: The network node 200 may fetch information and may determine whether the leading AIML UE capability is matching one or more network node deployment and/or network node capabilities. For instance, the network node 200 may determine whether it will implement an AIML- supported beam management and/or whether to allow the UE 100 to configure an AIML supported beam management based on correspond (e.g., obtained] AIML UE capability information.

Step S104: The network node 200 may decide that is will attempt to configure the UE with an AIML- supported functionality and/or use case, e.g., corresponding to the leading AIML UE capability obtained in step S103. The network node 200 thus prepares to collect further UE capability information the UE 100 for given and supported feature through a specific query for complementary AIML UE capability.

In step S105, network node 200 initiates an RRC message to enquire about complementary AIML UE capabilities. Such message may correspond to a subsequent AIML UE capability enquiry. For instance, the network node 200 may send the UECapabilityEnquiry with a specific flag which may indicate at least one complementary AIML UE capability, e.g., for a given category that was indicated in step SI 03 as a category (e.g., use case, AIML-enabled functionality] which the UE 100 supports. For instance, an indication of a given category and/or use case and/or AIML functionality may be brought by a new and/or by a new meaning of a UE capability information element. A conditional field in an UL signaling may not impose static availability (e.g., readiness] of the device to be configured with a given AIML feature. It may, however, notify the network node 200 about at least one or more applicability conditions which may for instance be valid at a given point of time but that can change over time. Optionally, or jointly the network node 200 sends the UECapabilityEnquiry with a specific flag for at least one or more AIML UE capabilities which may pertain to at least one individual AIML UE capability for a given category (e.g., a (e.g., complete] set of components that enable AIML-enabled beam management. Additionally or alternatively, a further (e.g., dedicated] RRC message may be used to request UE additional information.

In step S106, the UE 100 receives the enquiry for at least one or more complementary AIML UE capability. E.g., the network node 200 may request for an AIML UE capability for a given category and/or for an AIML-enabled feature. The UE 100 may ensure that the complementary AIML UE capability is set according to the UE status and currently applicable conditions. A complementary AIML UE capability for a given AIML-enabled feature may be associated with at least one or more parameters that may for instance take different values at a given time. For instance, an AIML-enabled beam management may be available for training with a maximum number of possible configurations or training sessions in parallel. E.g., if at least one UE condition changes, the UE may indicate, e.g., in a complementary AIML UE capability, e.g., for a same AIML-enabled feature (e.g., beam management] that training may not be available. In addition, at least one or more codepoints for any of the conditional AIML UE capabilities may be extended to more values to enable semi-static updates of the device’s status: e.g., 'supported' or 'non-supported', 'activation', 'deactivation', 'suspended' or 'removed' etc.

In step S107, the UE 100 may invoke a procedure according to step S106. The UE 100 may in particular indicate complementary AIML UE capabilities for the use case of beam management which may comprise an ML model indication and/or a training indication, e.g., as individual bits.

In step S108, the network node 200 may decide to configure the UE 100 according to the received leading and/or complementary AIML UE capabilities and/or AI/ML-enabled features.

In step S109, the network node 200 may configure the UE 100 accordingly. For instance, the network node 200 may include an option that a respective AIML-enabled feature may be modified and/or released according to additional configurations. As an example, the RRC message (e.g. RRC Reconfiguration conveying configuration for AIML-enabled functionality] may provide at least one or more corresponding configuration parameters that may for instance match previously provided complementary capabilities. The additional configuration may for instance distinguish between complementary (e.g., conditional] AIML Configurations matching Complementary AIML-enabled capabilities.

An example of an amendment to the standard may be structured as follows.

RRCReconfiguration-vl900-IEs ::= SEQUENCE { nonCriticalExtension SEQUENCE {} OPTIONAL, conditionalMLReconfiguration-rl9 ConditionalMLReconfiguration-rl9,

}

ConditionalMLReconfiguration-rl9 :: = SEQUENCE { conditionalML-ToAddModList SEQUENCE (SIZE(l..max conditionalML]] OF conditionalML-Config OPTIONAL, - Need N conditionalML-ToReleaseList SEQUENCE (SIZEfl.. max conditionalML]] OF conditionalML-Config OPTIONAL, - Need N

}

According to further embodiment (e.g., in particular second and/or fourth example aspect], leading and complementary AIML UE capability may be provided in one step. An example is shown in Fig. 5.

The network side may be uncertainty about what the user (e.g., UE, e.g., according to the first and/or second example aspect] currently supports (e.g., can provide] at the first attempt of fetching AIML UE capabilities. Such first attempt may lead to obtaining leading AIML UE capabilities indicating support only but not yet complementary ones indicating availability. In an embodiment, it is thus proposed that information on dynamically changing capabilities (what the UE is capable of in the given time instance] can be also provided at the initial procedure.

For instance, instead of an indication (e.g., by leading AIML UE capability] of a support of an AIML functionality, e.g., as true/false, e.g., as one bit, the UE 100 may provide more information, for instance at least two bit fields. The first field (e.g., bit] may indicate support of an AIML functionality and/or use case as in one of true/ false. True may represent a "static” UE capability, e.g., leading info: supported functionalities and/or use cases. The second field (e.g., bit] may indicate an availability of an AIML functionality, e.g., true/false.. The "availability” may represent (for support=True] whether the respective AIML feature is currently activatable and may refer to complementary info (e.g., complementary AIML UE capability]. The joint support+availability information may be sent in a (e.g., proactive] one-step approach. For instance, in one step, the UE 100 may report: low battery, ML model degraded needs re-training before use, real-time reporting for inference is not preferred, and/ or combinations thereof.

This embodiment is illustrated in the signaling of Fig. 4. UE Capability Transfer in case Leading and Complementary ML enabled Capability are represented as two bits parameter (one static, the other reflecting availability]. The leading information may correspond to at least one supported AIML functionality. The complementary information may correspond to availability of the AIML functionality. The combination of support and availability information is sent in a one-step approach.

In Fig. 5, the additional step S102a at the UE 100 comprises assessing complementary AIML UE capabilities which are then provided in step S103. The network node 200 may then only require a single preparatory signaling S101 before providing the configuration in step S109.

The methods introduced are tailored procedures for AIML-enabled operations that may reuse some of the legacy components from the UE capabilities (e.g. CSI-reporting framework including Ll-RSRP report is legacy feature that can serve a purpose for AI/ML only if supported jointly with new UE capabilities]. While it builds on existing operations it reveals the NW from the overcomplex exercise to deduce what is applicable for AIML operations. Yet, it enables indicating to the NW changeable over time or per AIML category (e.g. use case] UE conditions that does not hide actual device capabilities.

Fig. 6 shows a flowchart of an example embodiment according to the first example aspect, for instance performed by a first apparatus (e.g., UE], The method comprises, in step Ml 00, obtaining an AIML UE capability enquiry, e.g., from a network node (e.g., performing and/or controlling a method according to the first (e.g., and/or second] example aspect]. The UE may then, in step M102 determine at least one leading AIML UE capability, e.g., based on the obtained AIML UE capability enquiry of step M100. In an initial AIML capability response, the UE may then in step M104 provide and/or indicate the at least one leading AIML UE capability, e.g. to the network node.

Fig. 7 shows a flowchart of an example embodiment according to the second example aspect, for instance performed by a first apparatus (e.g., UE], The method comprises, in step M200, obtaining an AIML UE capability enquiry, e.g., from a network node (e.g., performing and/or controlling a method according to the second (e.g., and/or first] example aspect]. The AIML UE enquiry may for instance be indicative of a desire to be informed about at least one AIML functionality, e.g., a currently available AIML functionality, e.g., about a currently available AIML-assisted use case. The UE may then, in step M202 determine at least one leading AIML UE capability and at least one complementary AIML UE capability, e.g., based on the obtained AIML UE capability enquiry of step M200. In an initial AIML capability response, the UE may then in step M204 provide and/or indicate the at least one leading AIML UE capability and the at least one complementary AIML UE capability, e.g., to the network node.

Fig. 8 shows a flowchart of an example embodiment according to the third example aspect, for instance performed by a second apparatus (e.g., a network node]. In a step M300, an AIML UE capability enquiry is provided, e.g., to a UE, e.g., to an apparatus performing a method according to the first (e.g., and/or second] example aspect. In Step M302, an initial AIML capability response is obtained (e.g., from the UE], The initial AIML capability response may indicate at least one leading AIML UE capability. For instance, the initial AIML capability response may be unindicative of a complementary AIML UE capability.

Fig. 9 shows a flowchart of an example embodiment according to the fourth example aspect, for instance performed by a second apparatus (e.g., a network node]. In a step M400, an AIML UE capability enquiry is provided, e.g., to a UE, e.g., to an apparatus performing a method according to the second (e.g., and/or first] example aspect. In Step M402, an initial AIML capability response is obtained (e.g., from the UE], The initial AIML capability response may indicate at least one leading AIML UE capability and at least one complementary AIML UE capability.

Fig. 10 shows an example block diagram of a first apparatus 100, for instance a UE. The UE 100 may perform a method according to the first example aspect. The UE 100 comprises a user interface A160, a program memory A110, a main memory A120, and a data memory A140. Further, it comprises a processor A130. The apparatus 100 may further comprise functional units A131, A132, A133 which correspond to the method steps shown in Fig. 6. A functional unit may for instance correspond to a code block within a memory A110, A120, A140. The AIML UE capability enquiry obtainer A131 and/or the initial AIML capability response provider A133 may for instance be connected to and/or control the communication interface Al 50.

Fig. 11 shows an example block diagram of a first apparatus 100, for instance a UE. The UE 100 may perform a method according to the second example aspect. The UE 100 comprises a user interface A160, a program memory A110, a main memory A120, and a data memory A140. Further, it comprises a processor A130. The apparatus 100 may further comprise functional units A131, A132, A133 which correspond to the method steps shown in Fig. 7. A functional unit may for instance correspond to a code block within a memory A110, A120, A140. The AIML UE capability enquiry obtainer A131 and/or the initial AIML capability response provider A133 may for instance be connected to and/or control the communication interface Al 50. Fig. 12 shows an example block diagram of a second apparatus 200, for instance a network node. The network node 200 may perform a method according to the third example aspect. The network node 200 comprises a user interface A260, a program memory A210, a main memory A220, and a data memory A240. Further, it comprises a processor A230. The apparatus 200 may further comprise functional units A231 and A232 which correspond to the method steps shown in Fig. 8. A functional unit may for instance correspond to a code block within a memory A210, A220, A240. The AIML UE capability enquiry provider A231 and/or the initial AIML capability response obtainer A232 may for instance be connected to and/or control the communication interface A250.

Fig. 13 shows an example block diagram of a second apparatus 200, for instance a network node. The network node 200 may perform a method according to the fourth example aspect. The network node 200 comprises a user interface A260, a program memory A210, a main memory A220, and a data memory A240. Further, it comprises a processor A230. The apparatus 200 may further comprise functional units A231 and A232 which correspond to the method steps shown in Fig. 9. A functional unit may for instance correspond to a code block within a memory A210, A220, A240. The AIML UE capability enquiry provider A231 and/or the initial AIML capability response obtainer A232 may for instance be connected to and/or control the communication interface A250.

Fig. 14 is a schematic illustration of examples of tangible and non-transitory computer-readable storage media according to the present invention that may for instance be used to implement program and/or main memory A110, A120, A140, A210, A220, A240 of the apparatus 100 and/or 200 of Fig. 9 to 13. Fig. 14 shows a flash memory 1400, which may for instance be soldered or bonded to a printed circuit board, a solid-state drive 1401 comprising a plurality of memory chips (e.g. Flash memory chips], a magnetic hard drive 1402, a Secure Digital [SD] card 1403, a Universal Serial Bus [USB] memory stick 1404, an optical storage medium 1405 (such as for instance a CD-ROM or DVD] and a magnetic storage medium 1406.

Some embodiments comprise:

Embodiment 1:

A first method comprising: obtaining an artificial intelligence, AIML, user equipment, UE, capability enquiry, determining at least one leading AIML UE capability, wherein the leading AIML UE capability belongs to a hierarchical AIML UE capability structure, wherein the hierarchical AIML UE capability structure defines an association between the at least one leading AIML UE capability and at least one complementary AIML UE capability, and providing an initial AIML UE capability response, wherein the initial AIML UE capability response indicates the at least one leading AIML UE capability. Embodiment 2:

A first method comprising: obtaining an artificial intelligence, AIML, user equipment, UE, capability enquiry, determining at least one leading AIML UE capability and at least one complementary AIML UE capability, wherein the leading AIML UE capability and the complementary AIML UE capability belong to a hierarchical AIML UE capability structure, wherein the hierarchical AIML UE capability structure defines an association between the at least one leading AIML UE capability and at least one complementary AIML UE capability, and providing an initial AIML UE capability response, wherein the initial AIML UE capability response indicates the at least one leading AIML UE capability and the at least one complementary AIML UE capability.

Embodiment 3:

The first method of embodiment 1 or 2, wherein the initial AIML UE capability response is provided as a single message.

Embodiment 4:

The first method of any of embodiments 1 to 3, wherein the initial AIML UE capability response is unindicative of a complementary AIML UE capability associated to the at least one leading AIML UE capability indicated by the initial AIML UE capability response.

Embodiment 5:

The first method of any of embodiments 1 to 4, further comprising obtaining a subsequent AIML UE capability enquiry.

Embodiment 6:

The first method of any of embodiments 1 to 5, further comprising determining at least one complementary AIML UE capability.

Embodiment 7:

The first method of any of embodiments 1 to 6, further comprising providing a subsequent AIML UE capability response, wherein the subsequent AIML UE capability response indicates at least one determined complementary AIML UE capability.

Embodiment 8:

The first method of any of embodiments 1 to 7, further comprising obtaining an AIML configuration.

Embodiment 9:

The first method of any of embodiments 1 to 8, wherein at least one of the AIML UE capability enquiry, the initial AIML UE capability response, the leading AIML UE capability and/or the complementary AIML UE capability relate to at least one of an AIML functionality, an AIML use case or an AIML life cycle management phase.

Embodiment 10:

The first method of any of embodiments 1 to 9, wherein the AIML UE capability enquiry is at least one of indicative of a request for complementary AIML UE capability, indicative of a request for AIML UE capability, unindicative of a request for complementary AIML UE capability, or unindicative of a request for AIML UE capability.

Embodiment 11:

The first method of any of embodiments 1 to 10, wherein the leading AIML UE capability at least one of indicative of a support of at least one AIML-functionality, comprises one bit, static over time, independent of at least one external or internal condition, provided in an invariable way, indicative of an AIML use case, comprises at least one parameter of an AIML functionality, configured to enable a network node to provide a reference AIML configuration, associated to at least two complementary AIML UE capabilities, insufficient for an implementation of at least one AIML-functionality.

Embodiment 12:

The first method of any of embodiments 1 to 11, wherein at least one of the subsequent AIML UE capability enquiry is configured to request at least one complementary AIML UE capability.

Embodiment 13:

The first method of any of embodiments 1 to 11, wherein at least one of the at least one determined complementary AIML UE capability is determined based on the subsequent AIML UE capability enquiry, the at least one determined complementary AIML UE capability is determined based on the hierarchical AIML UE capability structure, the at least one determined complementary AIML UE capability is associated to the at least one leading AIML UE capability, the at least one determined complementary AIML UE capability is determined based on at least one condition, or the determining of at least one complementary AIML UE capability comprises analyzing at least one AIML UE capability.

Embodiment 14:

The first method of any of embodiments 1 to 13, wherein at least one complementary AIML UE capability associated to the at least one leading AIML UE capability is at least one of indicative of an availability of at least one AIML functionality, comprising one bit, larger in information than a leading AIML UE capability, variable over time, dependent of at least one external or internal condition, indicative of a non-AIML UE capability, indicative of at least one additional information, wherein the additional information comprises at least one of at least one parameter of an AIML functionality, a group of features, or a list of parameters, configured to enable a network node to provide an AIML configuration, sufficient for implementation of at least one AIML-functionality, configured to be provided by a same or by different means than the leading AIML UE capability, or associated to a single leading AIML UE capability.

Embodiment 15:

The first method of any of embodiments 1 to 14, wherein at least one of the leading AIML UE capability or the complementary AIML UE capability is grouped depending on a type of a mobile device.

Embodiment 16:

The first method of any of embodiments 1 to 15, wherein an AIML use case comprises at least one of beam management, positioning, channel state information, CSI, compression, CSI prediction, or UE mobility. Embodiment 17:

The first method of any of embodiments 1 to 16, wherein an AIML functionality is a new radio, NR,/ 6G radio feature at least one of assisted or realized based on an AIML.

Embodiment 18:

A second method comprising: providing an AIML UE capability enquiry, obtaining an initial AIML UE capability response, wherein the initial AIML UE capability response indicates at least one leading AIML UE capability, wherein the leading AIML UE capability belongs to a hierarchical AIML UE capability structure, wherein the hierarchical AIML UE capability structure defines an association between the at least one leading AIML UE capability and at least one complementary AIML UE capability.

Embodiment 19:

A second method (e.g., according to the fourth example aspect] comprising: means for providing an AIML UE capability enquiry, and means for obtaining an initial AIML UE capability response, wherein the initial AIML UE capability response indicates at least one leading AIML UE capability and at least one complementary AIML UE capability, wherein the leading AIML UE capability and the complementary AIML UE capability belong to a hierarchical AIML UE capability structure, wherein the hierarchical AIML UE capability structure defines an association between the at least one leading AIML UE capability and at least one complementary AIML UE.

Embodiment 20:

The first method of embodiment 18 or 19, wherein the initial AIML UE capability response is provided as a single message.

Embodiment 21:

The first method of any of embodiments 18 to 20, wherein the initial AIML UE capability response is unindicative complementary AIML UE capability associated to the at least one leading AIML UE capability indicated by the initial AIML UE capability response.

Embodiment 22:

The first method of any of embodiments 18 to 21, further comprising providing a subsequent AIML UE capability enquiry based on the initial AIML UE capability response. Embodiment 23:

The first method of any of embodiments 18 to 22, further comprising obtaining a subsequent AIML UE capability response, wherein the subsequent AIML UE capability response indicates at least one complementary AIML UE capability.

Embodiment 24:

The first method of any of embodiments 18 to 23, further comprising providing an AIML configuration based on the obtained leading AIML UE capability and/or complementary AIML UE capability.

Embodiment 25:

The first method of any of embodiments 18 to 24, wherein at least one of the AIML UE capability enquiry, the initial AIML UE capability response, the leading AIML UE capability and/or the complementary AIML UE capability relate to at least one of an AIML functionality, an AIML use case or an AIML life cycle management phase.

Embodiment 26:

The first method of any of embodiments 18 to 25, wherein the AIML UE capability enquiry is at least one of indicative of a request for complementary AIML UE capability, indicative of a request for AIML UE capability, unindicative of a request for complementary AIML UE capability, or unindicative of a request for AIML UE capability.

Embodiment 27:

The first method of any of embodiments 18 to 26, wherein the leading AIML UE capability is at least one of indicative of a support of at least one AIML-functionality, comprises one bit, static over time, independent of at least one external or internal condition, provided in an invariable way, indicative of an AIML use case, comprises at least one parameter of an AIML functionality, configured to enable a network node to provide a reference AIML configuration, associated to at least two complementary AIML UE capabilities, insufficient for an implementation of at least one AIML-functionality. Embodiment 28:

The first method of any of embodiments 18 to 27, wherein at least one of the subsequent AIML UE capability enquiry is configured to request at least one complementary AIML UE capability.

Embodiment 29:

The first method of any of embodiments 18 to 28, wherein at least one of the at least one indicated complementary AIML UE capability is determined based on the subsequent AIML UE capability enquiry, the at least one indicated complementary AIML UE capability is determined based on the hierarchical AIML UE capability structure, the at least one obtained complementary AIML UE capability is associated to the at least one leading AIML UE capability, the at least one indicated complementary AIML UE capability is determined based on at least one condition, or a determining of at least one complementary AIML UE capability comprises analyzing at least one AIML UE capability.

Embodiment 30:

The first method of any of embodiments 18 to 29, wherein at least one complementary AIML UE capability associated to the at least one leading AIML UE capability is at least one of indicative of an availability of at least one AIML functionality, comprising one bit, larger in information than a leading AIML UE capability, variable over time, dependent of at least one external or internal condition, indicative of a non-AIML UE capability, indicative of at least one additional information, wherein the additional information comprises at least one of at least one parameter of an AIML functionality, a group of features, or a list of parameters, configured to enable a network node to provide an AIML configuration, sufficient for implementation of at least one AIML-functionality, configured to be provided by a same or by different means than the leading AIML UE capability, or associated to a single leading AIML UE capability.

Embodiment 31: The first method of any of embodiments 18 to 30, wherein at least one of the leading AIML UE capability or the complementary AIML UE capability is grouped depending on a type of a mobile device.

Embodiment 32:

The first method of any of embodiments 18 to 31, wherein an AIML use case comprises at least one of beam management, positioning, channel state information, CSI, compression, CSI prediction, or UE mobility.

Embodiment 33:

The first method of any of embodiments 18 to 32, wherein an AIML functionality is a new radio, NR,/6G radio feature at least one of assisted or realized based on an AIML.

Embodiment 34:

A first apparatus, e.g., a UE, comprising respective means for performing the method of any of Embodiments 1 to 17.

Embodiment 35:

An first apparatus, e.g., a UE, comprising at least one processor; and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to perform and/or control the method according any of embodiments 1 to 17.

Embodiment 36:

A second apparatus, e.g., a network node, comprising respective means for performing the method of any of Embodiments 18 to 33.

Embodiment 37:

An second apparatus, e.g., a network node, comprising at least one processor; and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to perform and/or control the method according any of embodiments 18 to 33.

Embodiment 38: A computer program, the computer program when executed by a processor causing an apparatus, e.g. the apparatus according to embodiment 34 or 35, to perform and/or control the actions and/or steps of the method of any of embodiments 1 to 17.

Embodiment 39:

A computer program product comprising a computer program according to embodiment 38.

Embodiment 40:

A computer program, the computer program when executed by a processor causing an apparatus, e.g. the apparatus according to embodiment 36 or 37, to perform and/or control the actions and/or steps of the method of any of embodiments 18 to 33.

Embodiment 41:

A computer program product comprising a computer program according to embodiment 40.

Embodiment 42:

A system comprising: at least one first apparatus according to any of the embodiments 34 or 35, and at least one second apparatus according to any of the embodiments 36 or 37.

Any presented connection in the described embodiments is to be understood in a way that the involved components are operationally coupled. Thus, the connections can be direct or indirect with any number or combination of intervening elements, and there may be merely a functional relationship between the components.

Further, as used in this text, the term ‘circuitry’ refers to any of the following:

(a] hardware-only circuit implementations (such as implementations in only analog and/or digital circuitry]

(b] combinations of circuits and software (and/or firmware], such as: (i] to a combination of processors] or (ii] to sections of processor^]/ software (including digital signal processor (s)J, software, and memory(ies] that work together to cause an apparatus, such as a mobile phone, to perform various functions] and

(c] to circuits, such as a microprocessor(s] or a section of a microprocessor(s], that re-quire software or firmware for operation, even if the software or firmware is not physically present.

This definition of ‘circuitry’ applies to all uses of this term in this text, including in any claims. As a further example, as used in this text, the term ‘circuitry’ also covers an implementation of merely a processor (or multiple processors] or section of a processor and its (or their] accompanying software and/or firmware. The term ‘circuitry’ also covers, for example, a baseband integrated circuit or applications processor integrated circuit for a mobile phone.

Any of the processors mentioned in this text, in particular but not limited to processors 130, 230, 330 of Figs. 9 to 15, could be a processor of any suitable type. Any processor may comprise but is not limited to one or more microprocessors, one or more processor(s] with accompanying digital signal processor(s], one or more processor(s] without accompanying digital signal processor(s], one or more special-purpose computer chips, one or more field-programmable gate arrays (FPGAS], one or more controllers, one or more application-specific integrated circuits (ASICS], or one or more computers]. The relevant structure/hardware has been programmed in such a way to carry out the described function.

Moreover, any of the actions or steps described or illustrated herein may be implemented using executable instructions in a general-purpose or special-purpose processor and stored on a computer- readable storage medium (e.g., disk, memory, or the like] to be executed by such a processor. References to ‘computer-readable storage medium’ should be understood to encompass specialized circuits such as FPGAs, ASICs, signal processing devices, and other devices.

Moreover, any of the actions described or illustrated herein may be implemented using executable instructions in a general-purpose or special-purpose processor and stored on a computer-readable storage medium (e.g., disk, memory, or the like] to be executed by such a processor. References to ‘computer-readable storage medium’ should be understood to encompass specialized circuits such as FPGAs, ASICs, signal processing devices, and other devices.

The wording "A, or B, or C, or a combination thereof’ or "at least one of A, B and C” may be understood to be not exhaustive and to include at least the following: (i] A, or (ii] B, or (iii] C, or (iv] A and B, or (v] A and C, or (vi] B and C, or (vii] A and B and C.

It will be understood that the embodiments disclosed herein are only exemplary, and that any feature presented for a particular exemplary embodiment and/or for a particular example aspect may be used with any (e.g., other] example aspect of the present disclosure on its own or in combination with any feature presented for the same or another particular exemplary embodiment and example aspects and/or in combination with any other feature not mentioned. It will further be understood that any feature presented for an example embodiment in a particular category may also be used in a corresponding manner in an example embodiment of any other category. LIST OF ABBREVIATIONS

Al Artificial Intelligence

BM Beam Management

CSI Channel State Information LCM Life Cycle Management

ML Machine Learning

UE User Equipment

RAN Radio Access Network

RRC Radio Resource Control

Claims

C l i m s

1. An apparatus comprising: means for obtaining an artificial intelligence, AIML, user equipment, UE, capability enquiry, means for determining at least one leading AIML UE capability, wherein the leading AIML UE capability belongs to a hierarchical AIML UE capability structure, wherein the hierarchical AIML UE capability structure defines an association between the at least one leading AIML UE capability and at least one complementary AIML UE capability, and means for providing an initial AIML UE capability response, wherein the initial AIML UE capability response indicates the at least one leading AIML UE capability.

2. The apparatus according to claim 1, wherein: the initial AIML UE capability response is unindicative of a complementary AIML UE capability associated to the at least one leading AIML UE capability indicated by the initial AIML UE capability response.

3. The apparatus according to claim 1 or 2, further comprising means for obtaining a subsequent AIML UE capability enquiry.

4. The apparatus according to any of claims 1 to 3, further comprising means for determining at least one complementary AIML UE capability.

5. The apparatus according to any of claims 1 to 4, further comprising means for providing a subsequent AIML UE capability response, wherein the subsequent AIML UE capability response indicates at least one determined complementary AIML UE capability.

6. The apparatus according to any of claims 1 to 5, further comprising means for obtaining an AIML configuration.

7. The apparatus according to any of claims 1 to 6, wherein: at least one of the AIML UE capability enquiry, the initial AIML UE capability response, the leading AIML UE capability and/or the complementary AIML UE capability relate to at least one of an AIML functionality, an AIML use case or an AIML life cycle management phase.

8. The apparatus according to any of claims 1 to 7, wherein the AIML UE capability enquiry is at least one of indicative of a request for complementary AIML UE capability, indicative of a request for leading AIML UE capability, unindicative of a request for complementary AIML UE capability, or unindicative of a request for leading AIML UE capability.

9. The apparatus according to any of claims 1 to 8, wherein the leading AIML UE capability is at least one of indicative of a support of at least one AIML-functionality, comprises one bit, static over time, independent of at least one external or internal condition, provided in an invariable way, indicative of a AIML use case, comprises at least one parameter of an AIML functionality, configured to enable a network node to provide a reference AIML configuration, associated to at least two complementary AIML UE capabilities, insufficient for an implementation of at least one AIML-functionality.

10. The apparatus according to any of claims 3 to 9 , wherein at least one of the subsequent AIML UE capability enquiry is configured to request at least one complementary AIML UE capability.

11. The apparatus according to any of claims 4 to 10, wherein at least one of the at least one determined complementary AIML UE capability is determined based on the subsequent AIML UE capability enquiry, the at least one determined complementary AIML UE capability is determined based on the hierarchical AIML UE capability structure, the at least one determined complementary AIML UE capability is associated to the at least one leading AIML UE capability, the at least one determined complementary AIML UE capability is determined based on at least one condition, or the determining of at least one complementary AIML UE capability comprises analyzing at least one AIML UE capability.

12. The apparatus according to any of claims 1 to 11, wherein at least one complementary AIML UE capability associated to the at least one leading AIML UE capability is at least one of indicative of an availability of at least one AIML functionality , comprising one bit, larger in information than a leading AIML UE capability, variable over time, dependent of at least one external or internal condition, indicative of a non-AIML UE capability, indicative of at least one additional information, wherein the additional information comprises at least one of at least one parameter of an AIML functionality, a group of features, or a list of parameters, configured to enable a network node to provide a (e.g., full and/or functional] AIML configuration, sufficient for implementation of at least one AIML-functionality, configured to be provided by a same or by different means than the leading AIML UE capability, or associated to a single leading AIML UE capability.

13. The apparatus according to any of claims 1 to 12, wherein at least one of the leading AIML UE capability or the complementary AIML UE capability is grouped depending on a type of a mobile device.

14. The apparatus according to any of claims 1 to 13, wherein an AIML use case comprises at least one of beam management, positioning, channel state information, CSI, compression,

CSI prediction, or

UE mobility.

15. The apparatus according to any of claims 1 to 14, wherein an AIML functionality is a new radio, NR,/6G radio feature at least one of assisted or realized based on an AIML.

16. An apparatus comprising means for providing a AIML UE capability enquiry, means for obtaining an initial AIML UE capability response, wherein the initial AIML UE capability response indicates at least one leading AIML UE capability, wherein the leading AIML UE capability belongs to a hierarchical AIML UE capability structure, wherein the hierarchical AIML UE capability structure defines an association between the at least one leading AIML UE capability and at least one complementary AIML UE capability.

17. The apparatus according to claim 16, wherein the initial AIML UE capability response is unindicative of a complementary AIML UE capability associated to the at least one leading AIML UE capability indicated by the initial AIML UE capability response.

18. The apparatus according to claim 16 or 17, further comprising means for providing a subsequent AIML UE capability enquiry based on the initial AIML UE capability response.

19. The apparatus according to any of claims 16 or 18, further comprising means for obtaining a subsequent AIML UE capability response, wherein the subsequent AIML UE capability response indicates at least one complementary AIML UE capability.

20. The apparatus according to any of claims 16 to 19, further comprising means for providing an AIML configuration based on the obtained leading AIML UE capability and/or complementary AIML UE capability.

21. The apparatus according to any of claims 16 to 20, wherein: at least one of the AIML UE capability enquiry, the initial AIML UE capability response, the leading AIML UE capability and/or the complementary AIML UE capability relate to at least one of an AIML functionality, an AIML use case or an AIML life cycle management phase.

22. The apparatus according to any of claims 16 to 21, wherein the AIML UE capability enquiry is at least one of indicative of a request for complementary AIML UE capability, indicative of a request for leading AIML UE capability, unindicative of a request for complementary AIML UE capability, or unindicative of a request for leading AIML UE capability.

23. The apparatus according to any of claims 16 to 22, wherein the leading AIML UE capability is at least one of indicative of a support of at least one AIML-functionality, comprises one bit, static over time, independent of at least one external or internal condition, provided in an invariable way, indicative of a AIML use case, comprises at least one parameter of an AIML functionality, configured to enable a network node to provide a reference AIML configuration, associated to at least two complementary AIML UE capabilities, insufficient for an implementation of at least one AIML-functionality.

24. The apparatus according to any of claims 18 to 23, wherein at least one of the subsequent AIML UE capability enquiry is configured to request at least one complementary AIML UE capability.

25. The apparatus according to any of claims 18 to 24, wherein at least one of the at least one indicated complementary AIML UE capability is determined based on the subsequent AIML UE capability enquiry, the at least one indicated complementary AIML UE capability is determined based on the hierarchical AIML UE capability structure, the at least one obtained complementary AIML UE capability is associated to the at least one leading AIML UE capability, the at least one indicated complementary AIML UE capability is determined based on at least one condition, or a determining of at least one complementary AIML UE capability comprises analyzing at least one AIML UE capability.

26. The apparatus according to any of claims 16 to 25, wherein at least one complementary AIML UE capability associated to the at least one leading AIML UE capability is at least one of indicative of an availability of at least one AIML functionality, comprising one bit, larger in information than a leading AIML UE capability, variable over time, dependent of at least one external or internal condition, indicative of a non-AIML UE capability, indicative of at least one additional information, wherein the additional information comprises at least one of at least one parameter of an AIML functionality, a group of features, or a list of parameters, configured to enable a network node to provide a (e.g., full and/or functional] AIML configuration, sufficient for implementation of at least one AIML-functionality, configured to be provided by a same or by different means than the leading AIML UE capability, or associated to a single leading AIML UE capability.

27. The apparatus according to any of claims 16 to 26, wherein at least one of the leading AIML UE capability or the complementary AIML UE capability is grouped depending on a type of a mobile device.

28. The apparatus according to any of claims 16 to 27, wherein an AIML use case comprises at least one of beam management, positioning, channel state information, CSI, compression,

CSI prediction, or

UE mobility.

29. The apparatus according to any of claims 16 to 28, wherein an AIML functionality is a new radio, NR,/6G radio feature at least one of assisted or realized based on an AIML.

30. A system comprising a first apparatus, comprising means for providing an AIML UE capability enquiry, means for obtaining an initial AIML UE capability response, wherein the initial AIML UE capability response indicates at least one leading AIML UE capability, wherein the leading AIML UE capability belongs to a hierarchical AIML UE capability structure, wherein the hierarchical AIML UE capability structure defines an association between the at least one leading AIML UE capability and at least one complementary AIML UE capability, and a second apparatus, comprising means for obtaining the artificial intelligence, AIML, user equipment, UE, capability enquiry, means for determining the at least one leading AIML UE capability, and means for providing the initial AIML UE capability response, wherein the initial AIML UE capability response indicates the at least one leading AIML UE capability.