Enhanced RRC connection procedure

Abstract

There is provided a method, apparatus and computer program for causing the following to be performed: transmitting, to a first network node, a connection request, the connection request comprising a cause value, the cause value indicating that the connection request is related to a machine learning associated task and / or a machine learning model.

Description

ENHANCED RRC CONNECTION PROCEDURETECHNICAL FIELD

[0001] Various examples of this disclosure relate to a method, apparatus, system and computer program. In particular, but not exclusively, the present application relates to handling of machine learning associated tasks and / or machine learning models.BACKGROUND

[0002] A communication network can be seen as a facility that enables communications between two or more communication devices, or provides communication devices access to a data network. A mobile or wireless communication network is one example of a communication network. A communication device may be provided with a service by an application server.

[0003] Such communication networks operate in according with standards such as those provided by 3GPP (Third Generation Partnership Project) or ETSI (European Telecommunications Standards Institute). Examples of standards are the so-called 5G (5th Generation) standards and 6G (6th Generation) standards provided by 3GPP.SUMMARY

[0004] [Some example embodiments of this disclosure will be described with respect to certain aspects. These aspects are not intended to indicate key or essential features of the embodiments of this disclosure, nor are they intended to be used to limit the scope of thereof. Other features, aspects, and elements will be readily apparent to a person skilled in the art in view of this disclosure.

[0005] According to a first aspect, there is provided an apparatus comprising means for: transmitting, to a first network node, a connection request, the connection request comprising a cause value, the cause value indicating that the connection request is related to a machine learning associated task and / or a machine learning model.

[0006] According to a second aspect, there is provided an apparatus comprising: at least one processor; and at least one memory comprising code that, when executed by the at least one processor, causes the apparatus to perform: transmitting, to a first network node, a connection request, the connection request comprising a cause value,the cause value indicating that the connection request is related to a machine learning associated task and / or a machine learning model.

[0007] According to a third aspect, there is provided a method for an apparatus, the method comprising: transmitting, to a first network node, a connection request, the connection request comprising a cause value, the cause value indicating that the connection request is related to a machine learning associated task and / or a machine learning model.

[0008] According to a fourth aspect, there is provided an apparatus comprising: transmitting circuitry for transmitting, to a first network node, a connection request, the connection request comprising a cause value, the cause value indicating that the connection request is related to a machine learning associated task and / or a machine learning model.

[0009] The following may apply in respect of any (e.g., one or more, including all) of the above first to fourth aspects.

[0010] In some examples, the cause value may be a binary indication or an index value that maps a machine learning associated task and / or a machine learning model to a priority level.

[0011] In some examples, the machine learning associated task comprises at least one of: transfer of a machine learning model; transfer of one or more conditions of a machine learning model; transfer of machine learning model validation data; transfer of training data related to a machine learning model; or transfer of a machine learning data set.

[0012] The apparatus may be caused to perform: for receiving, from the first network node, a message indicating whether the connection request has been accepted or rejected by the first network node.

[0013] The apparatus may be caused to perform: when the message received from the first network node indicates that the connection request has been accepted, establishing a connection with the first network node and exchanging information to be used when executing a machine learning associated task and / or a machine learning model with the first network node over the established connection.

[0014] The apparatus may be caused to perform: when a load at the first network node is above or at a first threshold value, prioritizing a radio resource management relatedconnection over the established connection corresponding to the connection request related to a machine learning task and / or machine learning model.

[0015] The apparatus may be caused to perform: when the message received from the first network node indicates that the connection request has been rejected, initiating a backoff timer, and transmitting, to the first network node, a second connection request comprising the cause value when the backoff timer expires.

[0016] The apparatus may be caused to perform: when the message received from the first network node indicates that the connection request has been rejected, transmitting a second connection request comprising the cause value to another network node or cell.

[0017] The apparatus may be caused to perform: receiving, from a second network node, a connection release message prior to transmitting the connection request.

[0018] The apparatus may be caused to perform: based on the connection release message, releasing a connection with the second network node or switching from a connected state to another state.

[0019] In some examples, the another state may be an idle state or an inactive state.

[0020] In some examples, the another state is the idle state, the connection release message is a radio resource control release message, the connection request message is a radio resource control setup request message, a message indicating that the connection request has been accepted by the first network node is a radio resource control setup message and the cause value is an establishment cause value.

[0021] In some examples, the another state is the inactive state, the connection release message is a radio resource control release with a suspend configuration message, the connection request message is a radio resource control resume request message, a message indicating that the connection request has been accepted by the first network node is a radio resource control resume message and the cause value is a resume cause value.

[0022] The apparatus may be caused to perform: determining that signaling for a machine learning associated task and / or a machine learning model will be performed between the apparatus and a network node, wherein transmitting the connection request is performed based on said determination that signaling for a machine learning associated task and / or a machine learning model will be performed.

[0023] The apparatus may be caused to perform: determining that signaling for a machine learning associated task and / or a machine learning model will be performed autonomously by the apparatus, and transmitting, to the second network node, information related to an upcoming machine learning associated task and / or machine learning model.

[0024] The apparatus may be caused to perform: determining that signaling for a machine learning associated task and / or a machine learning model will be performed based on receiving, from the second network node, information related to an upcoming machine learning associated task and / or machine learning model.

[0025] In some examples, the connection release message received from the second network node comprises the information related to the upcoming machine learning associated task and / or machine learning model.

[0026] In some examples, the first network node is the second network node, or the first network node is different to the second network node.

[0027] In some examples, the first network node and / or second network node is one of: a network access node; a core network node; or an operations, administration and management function.

[0028] According to a fifth aspect, there is provided an apparatus comprising means for: receiving, from a user equipment, a connection request, the connection request comprising a cause value, the cause value indicating that the connection request is related to a machine learning associated task and / or a machine learning model; and determining, based on the cause value, whether to accept the connection request by configuring a connection to the user equipment.

[0029] According to a sixth aspect, there is provided an apparatus comprising: at least one processor, and at least one memory comprising code that, when executed by the at least one processor, causes the apparatus to perform: receiving, from a user equipment, a connection request, the connection request comprising a cause value, the cause value indicating that the connection request is related to a machine learning associated task and / or a machine learning model; and determining, based on the cause value, whether to accept the connection request by configuring a connection to the user equipment.

[0030] According to a seventh aspect, there is provided a method for an apparatus, the method comprising: receiving, from a user equipment, a connection request, theconnection request comprising a cause value, the cause value indicating that the connection request is related to a machine learning associated task and / or a machine learning model; and determining, based on the cause value, whether to accept the connection request by configuring a connection to the user equipment.

[0031] According to an eighth aspect, there is provided an apparatus comprising: receiving circuitry for receiving, from a user equipment, a connection request, the connection request comprising a cause value, the cause value indicating that the connection request is related to a machine learning associated task and / or a machine learning model; and determining circuitry for determining, based on the cause value, whether to accept the connection request by configuring a connection to the user equipment.

[0032] The following may apply in respect of any (e.g., one or more, including all) of the above fifth to eighth aspects.

[0033] In some examples, the cause value may be a binary indication or an index value that maps a machine learning associated task and / or a machine learning model to a priority level.

[0034] In some examples, the machine learning associated task comprises at least one of: transfer of a machine learning model; transfer of one or more conditions of a machine learning model; transfer of machine learning model validation data; transfer of training data related to a machine learning model; or transfer of a machine learning data set.

[0035] The apparatus may be caused to perform: transmitting, to the user equipment, a message indicating whether the connection request has been accepted or rejected by the apparatus, based on the determination of the apparatus.

[0036] The apparatus may be caused to perform: when determining to accept the connection request and after a connection is established with the user equipment, exchanging information to be used when executing a machine learning associated task and / or a machine learning model with the user equipment over the established connection.

[0037] The apparatus may be caused to perform: determining that a load at the apparatus is above or at a first threshold value, and prioritizing a radio resource management related connection over the established connection corresponding to theconnection request related to a machine learning associated task and / or a machine learning model.

[0038] The apparatus may be caused to perform: for determining that a load at the apparatus is above or at a first threshold value, and transmitting, to the user equipment, a command to release the established connection with the apparatus.

[0039] The apparatus may be caused to perform: when said determining determines to reject the connection request, transmitting a backoff timer to the user equipment.

[0040] The apparatus may be caused to perform: transmitting a connection release message to the user equipment, prior to receiving the connection request.

[0041] In some examples, the connection release message comprises an indication to release the connection with the apparatus, or to switch from a connected state to another state.

[0042] In some examples, the another state is an idle state or an inactive state.

[0043] In some examples, the another state is the idle state, the connection release message is a radio resource control release message, the connection request message is a radio resource control setup request message, a message indicating that the connection request has been accepted by the apparatus is a radio resource control setup message and the cause value is an establishment cause value.

[0044] In some examples, the another state is the inactive state, the connection release message is a radio resource control release with a suspend configuration message, the connection request message is a radio resource control resume request message, a message indicating that the connection request has been accepted by the apparatus is a radio resource control resume message and the cause value is a resume cause value.

[0045] The apparatus may be caused to perform: determining that signaling for a machine learning task and / or a machine learning model will be performed between the apparatus and the user equipment, wherein transmitting the connection release message is performed based on said determination that signaling for a machine learning associated task and / or a machine learning model will be performed.

[0046] The apparatus may be caused to perform: determining that signaling for a machine learning associated task and / or a machine learning model will be performed based on receiving, from the user equipment, information related to an upcoming machine learning associated task and / or machine learning model.

[0047] The apparatus may be caused to perform: determining that signaling for a machine learning associated task and / or a machine learning model will be performed autonomously by the apparatus, and transmitting, to the user equipment, information related to an upcoming machine learning associated task and / or machine learning model.

[0048] The apparatus may be caused to perform: determining that signaling for a machine learning associated task and / or a machine learning model will be performed, based on receiving, from a core network, information related to an upcoming machine learning associated task and / or machine learning model.

[0049] The apparatus may be caused to perform: receiving a release command from the core network, the release command triggering the apparatus to provide the connection release message to the user equipment.

[0050] In some examples, the connection release message comprises the information related to the upcoming machine learning associated task and / or machine learning model.

[0051] The apparatus may be caused to perform: pre-empting machine learning related connections when it is determined that a load at the apparatus is above or at a first threshold value.

[0052] In some examples, the apparatus is one of: a network access node; a core network node; or an operations, administration and management function.

[0053] According to an aspect, there is provided a non-transitory computer readable medium comprising program instructions that, when executed by an apparatus, cause the apparatus to perform at least the method according to any of the preceding aspects.

[0054] In the above, many different embodiments have been described. It should be appreciated that further embodiments may be provided by the combination of any two or more of the embodiments described above.DESCRIPTION OF FIGURES

[0055] Some examples will now be described, by way of non-limiting and illustrative example only, with reference to the accompanying Figures in which:

[0056] Figures 1A and 1 B shows a representation of a communication system;

[0057] Figure 2 shows a representation of an apparatus for the communication system of Figures 1 A and 1 B according to some examples;

[0058] Figure 3 shows a representation of an apparatus according to some examples;

[0059] Figure 4 shows an example RRC connection establishment and / or resumption procedure;

[0060] Figure 5 shows an example RRC connection release procedure;

[0061] Figure 6 shows an example signalling diagram;

[0062] Figure 7 shows an example signalling diagram;

[0063] Figure 8 shows an example signalling diagram;

[0064] Figure 9 shows an example signalling diagram;

[0065] Figure 10 shows an example signalling diagram;

[0066] Figure 11A shows an example signalling diagram;

[0067] Figure 11 B shows an example signalling diagram;

[0068] Figure 12 shows an example signalling diagram;

[0069] Figure 13 shows an example signalling diagram;

[0070] Figure 14 shows an example signalling diagram;

[0071] Figure 15 illustrates an example method;

[0072] Figure 16 illustrates an example method;

[0073] Figure 17 illustrates an example method;

[0074] Figure 18 illustrates an example method;

[0075] Figure 19 shows a representation of an apparatus according to some examples.DETAILED DESCRIPTION

[0076] The following provides an example of a communication environment in which the presently described techniques may be deployed. It is understood that this communication environment is not limiting, and is merely being used to provide at least one example for describing where such techniques may be deployed.

[0077] Stated differently, in the following various examples are explained with reference to communication devices capable of communication with a communication system. Before explaining in detail the examples of the methods and apparatuses of the present disclosure, a 5thgeneration communication system (5GS), an access network and a core network (5GC) thereof, and communication devices are briefly explained with reference to Figures 1A, 1 B, 2 and 3.

[0078] Figure 1A shows a schematic representation of a 5G communication system (5GS). The 5GS may comprise a user equipment (UE), an access network such as a 5G radio access network (5G-RAN) or next generation radio access network (NG- RAN), a 5G core network (5GC), and one or more application functions. An application function may be deployed in the 5GS as trusted application function or may be deployed or host on one or more application servers of the data network. Such application functions are untrusted application functions. The 5GS connects the UE to a data network the access network and the 5GC (e.g., a UPF of the 5GC).

[0079] The 5G-RAN may comprise one or more radio access nodes, such as gNodeB (gNB). A gNB may comprise one or more gNodeB distributed units connected to one or more gNodeB centralized units. The 5G-RAN may be as illustrated below in Figure 1 B.

[0080] The 5GC may comprise the following network functions: Network Slice Selection Function (NSSF); Network Exposure Function; Network Repository Function (NRF); Policy Control Function (PCF); Unified Data Management (UDM); Application Function (AF); Authentication Server Function (AUSF); an Access and Mobility Management Function (AMF); and Session Management Function (SMF), and a user plane function (UPF). Figure 1A also shows the various interfaces (N1 , N2 etc.) that may be implemented between the various elements of the system.

[0081] The term “terminal device” refers to any end device that may be capable of wireless communication. By way of example rather than limitation, a terminal device may also be referred to as a communication device, user equipment (UE), a Subscriber Station (SS), a Portable Subscriber Station, a mobile device, a Mobile Station (MS), or an Access Terminal (AT). The terminal device may comprise, but not limited to, a mobile phone, a cellular phone, a smart phone, voice over IP (VoIP) phones, wireless local loop phones, a tablet, a wearable terminal device, a personal digital assistant (PDA), portable computers, desktop computer, image capture terminal devices such as digital cameras, gaming terminal devices, music storage and playback appliances, vehicle-mounted wireless terminal devices, wireless endpoints, mobile stations, laptop-embedded equipment (LEE), laptop-mounted equipment (LME), USB dongles, smart devices, wireless customer-premises equipment (CPE), a machine-type communications (MTC) device, an Internet of Things (loT) device, a watch or other wearable, a head-mounted display (HMD), a vehicle, a drone, a medicaldevice and applications (e.g., remote surgery), an industrial device and applications (e.g., a robot and / or other wireless devices operating in an industrial and / or an automated processing chain contexts), a data consumer electronics device, a device operating on commercial and / or industrial wireless networks, and the like. The terminal device may also correspond to a Mobile Termination (MT) part of an Integrated Access and Backhaul (IAB) node (e.g., a relay node). In the following description, the terms “terminal device”, “communication device”, “terminal”, “user device”, “user equipment” and “UE” may be used interchangeably.

[0082] As used herein, the term “network device” is used interchangeably with “network access node” or “cell” and refers to a node in a communication network via which a terminal device accesses the network and receives services therefrom. The network device may refer to a base station (BS) or an access point (AP), for example, a node B (NodeB or NB), an evolved NodeB (eNodeB or eNB), an NR NB (also referred to as a gNB), a Remote Radio Unit (RRU), a radio header (RH), a remote radio head (RRH), a relay, an Integrated Access and Backhaul node, a low power node such as a femto, a pico, a non-terrestrial network (NTN) or non-ground network device such as a satellite network device, a low earth orbit (LEO) satellite and a geosynchronous earth orbit (GEO) satellite, an aircraft network device, and so forth, depending on the applied terminology and technology. In some examples, radio access network (RAN) split architecture comprises a Centralized Unit (CU) and a Distributed Unit (DU) at an IAB donor node. An IAB node comprises a Mobile Terminal (IAB-MT) part that behaves like a UE toward the parent node, and a DU part of an IAB node behaves like a base station toward the next-hop IAB node.

[0083] In some examples, a link from the network device 120 to the user device 110 or 115 is referred to as a DL. In some examples, a link from the user device 110 or 115 to the network device 120 is referred to as a UL. Links are also referred to herein as “channels”. In DL, the network device 120 is a Tx device (or a transmitter), and the user device 110 or 115 is a Rx device (or a receiver). In UL, the user device 110 or 115 is a Tx device (or a transmitter), and the network device 120 is a Rx device (or a receiver). A link between the user device 110 and another user device (not shown) is referred to as a sidelink (SL). In SL, one of the user devices is a Tx device (or a transmitter), and the other of the user devices is a Rx device (or a receiver).

[0084] Communications in the communication environment 100 may be implemented according to any proper communication protocol(s), comprising, but not limited to, cellular communication protocols of the first generation (1 G), the second generation (2G), the third generation (3G), the fourth generation (4G), the fifth generation (5G), the sixth generation (6G), and the like, wireless local network communication protocols such as Institute for Electrical and Electronics Engineers (IEEE) 802.11 and the like, and / or any other protocols currently known or to be developed in the future. Moreover, the communication may utilize any proper wireless communication technology, comprising but not limited to: Code Division Multiple Access (CDMA), Frequency Division Multiple Access (FDMA), Time Division Multiple Access (TDMA), Frequency Division Duplex (FDD), Time Division Duplex (TDD), Multiple-Input Multiple-Output (MIMO), Orthogonal Frequency Division Multiple (OFDM), Discrete Fourier Transform spread OFDM (DFT-s-OFDM) and / or any other technologies currently known or to be developed in the future.

[0085] As described below, signaling between apparatus described in connection with Figure 1 B may be performed using service bearers and / or data bearers. Signalling Radio Bearers (SRBs) are Radio Bearers (RBs) that carry signalling information between a UE and a network node. SRBs may be considered to be RBs that are used for the transmission of radio resource control (RRC) messages associated with machine learning (ML) tasks and / or ML / AI models. 3GPP has described various SRBs, each of which being usable for different signalling exchange configurations: SRB0; SRB1 ; SRB2; SRB3; SRB4; and SRB5. In some examples, Data Radio Bearers (DRBs) may be considered to be Radio Bearers (RBs) that are used for carrying user data between a UE and network node. The DRBs may be used to transmit ML or Al related data between the UE and network.

[0086] Figure 2 illustrates an example of a control apparatus 200 for controlling a function of the access network (e.g., a 5G-RAN or the NG-RAN illustrated in Figures 1A and 1 B) illustrated on Figures 1A and 1 B. The control apparatus 200 may comprise at least one random access memory (RAM) 211 a, at least on read only memory (ROM) 211 b, at least one processor 212, 213 and a network interface 214. The at least one processor 212, 213 may be coupled to the RAM 211 a and the ROM 211 b. The at least one processor 212, 213 may be configured to execute an appropriate software code 215. Execution of the software code 215 may for example may cause the apparatusto perform operations for controlling a function of the access network. The software code 215 may be stored in the ROM 211 b. The control apparatus 200 may be interconnected with another control apparatus 200 for controlling another function of the 5G-RAN or the NG-RAN. In some examples, each function of the 5G-RAN or the NG-RAN is deployed or hosted on a control apparatus 200. In alternative examples, two or more functions of the 5G-RAN or the NG-RAN may share a control apparatus.

[0087] Figure 3 illustrates an example of a communication device 300, such as the UE illustrated on Figures 1A and 1 B. The communication device 300 may be provided by any device capable of sending and receiving radio signals. Non-limiting examples of a communication device 300 comprise a user equipment, a mobile station (MS) or mobile device such as a mobile phone or what is known as a ’smart phone’, a computer provided with a wireless interface card or other wireless interface facility (e.g., USB dongle), a personal data assistant (PDA) or a tablet provided with wireless communication capabilities, a machine-type communications (MTC) device, an Internet of things (loT) type communication device or any combinations of these or the like. The communication device 300 may comprise a transceiver for transmitting and / or receiving, for example, wireless signals carrying communications, for example radio signals. The communications may be one or more of voice, electronic mail (email), text messages, multimedia data, machine data and so on.

[0088] The communication device 300 may receive wireless signals (e.g., radio signals) over an air or radio interface 307 via appropriate apparatus for receiving and may transmit wireless signals via appropriate apparatus for transmitting radio signals. In Figure 3 transceiver is designated schematically by block 306. The transceiver 306 may comprise, for example, a radio part and associated antenna arrangement. The antenna arrangement may be arranged internally or externally to the mobile device and may comprise one or more antenna elements. The antenna arrangement may be a multi-input multi-output (MIMO) antenna.

[0089] The communication device 300 may be provided with at least one processor 301 , memory 302 comprising at least one memory ROM 302a, at least one RAM 302b and other possible components 303 for use in software and hardware aided execution of tasks it is designed to perform, including control of access to and communications with access networks (e.g., the 5G-RAN or NG-RAN illustrated in Figures 1A and 1 B) and other communication devices. The at least one processor 301 is coupled to theRAM 302b and the ROM 302a. The at least one processor 301 may be configured to execute an appropriate software code 308. The software code 308 may for example allow to perform one or more operations of the communication device. The software code 308 may be stored in the ROM 302a.

[0090] The processor, the ROM, and the RAM, the transceiver and other circuitry of the communication device (e.g., a modem) can be provided on a circuit board, in chipsets, or in a system on chip. The circuit board, chipsets or system on chip is denoted by reference 304. The communication device 300 may optionally have a user interface such as keypad 305, touch sensitive screen or pad, combinations thereof or the like. Optionally one or more of a display, a speaker and a microphone may be provided depending on the type of communication device.

[0091] The terms “sending”, “transmitting” and “providing” used herein may be used interchangeably.

[0092] The terms “receiving” and “obtaining” used herein may be used interchangeably.

[0093] The present disclosure considers information that may be provided during a radio resource control (RRC) connection establishment procedure and / or an RRC connection release procedure. To provide greater context, the examples of Figures 4 and 5 illustrate general 3GPP methods for such procedures.

[0094] Figure 4 shows an example RRC connection establishment and / or resumption procedure between a UE 400 (e.g., such as a UE described in connection with Figures 1A and 3) and a network node 402 (e.g., such as a network node described in connection with Figures 1A, and 2). The network node 402 may be an eNB or a gNB. It is understood that resuming a connection between a UE and network node includes establishing a new connection between the UE and network node.

[0095] In this example, at S401 , the UE 400 may transmit a connection request message to the network node 402. For example, the UE 400 may provide an RRC setup request message (e.g., RRCSetupRequest) to the network node 402. The connection setup request message may request the establishment of an RRC connection. Alternatively, the UE 400 may provide an RRC resume request message (e.g., RRCResumeRequest) to the network node 402. The connection resume request message may request to resume an existing RRC connection.

[0096] At S402, the network node 402 may transmit a connection message to the UE 400 to indicate that the connection request has been accepted by the network node 402. For example, the network node 402 may provide an RRC setup message (e.g., RRCSetup) to the UE 400. The connection setup message may comprise configuration information for the UE 400 to access network resources. Alternatively, the network node 402 may provide an RRC resume message (e.g., RRCResume) to the UE 400. The connection resume message may comprise configuration information for the UE 400 to access network resources.

[0097] At S403, the UE 400 may transmit, to the network node 402, an establishment complete message. For example, the UE 400 may provide an RRC setup complete message (e.g., RRCSetupComplete) to the network node 402. The setup complete message may indicate that the UE 400 has successfully applied the configuration information of the connection setup message and established a connection with the network node 402. Alternatively, the UE 400 may provide an RRC resume complete message (e.g., RRCResumeComplete) to the network node 402. The resume complete message may indicate that the UE 400 has successfully applied the configuration information of the connection resume message and has resumed a connection and / or established a new connection with the network node 402.

[0098] Figure 5 shows an example RRC connection release procedure between a UE 500 and a network node 502. In this example, at S501 , the network node 502 transmits, to the UE, a connection release message. For example, the network node 502 sends the UE an RRCRelease message. The connection release message may indicate to the UE 500 that the UE 500 should release an established RRC connection.

[0099] In some examples, when a UE establishes or resumes an RRC Connection (e.g., such as during the example of Figure 4), the UE may include an establishment or resume cause in the connection request message. The establishment or resume cause may be considered to be an indication that indicates a reason relating to why the UE is requesting a connection establishment or resumption with the network node. For example, the establishment or resume cause may indicate a service and / or properties defining a service that the UE intends to receive via the connection. The establishment or resume cause may enable the network (e.g., the RAN) to keep track of the number of connection attempts and / or success and / or failures per establishment or resume cause. The network may also calculate and monitor Accessibility KeyPerformance Indicators (KPIs) based on the establishment or resume cause. In some examples, the KPIs may comprise RRC Connection attempts and / or success and / or failures.

[0100] Legacy establishment or resume causes may comprise at least one of: emergency; highPriorityAccess; mobile terminated (mt)-Access; mobile originated (mo)-Signalling; mo-Data; mo-VoiceCall; mo-VideoCall; mo-SMS; mps- PriorityAccess; and mcs-PriorityAccess.

[0101] In an example, when the UE sets up an RRC connection for the purpose of a voice call, the UE may send an RRC connection request message with an establishment cause set to mo-VoiceCall. The establishment cause may only indicate the initial purpose for connection establishment of the UE.

[0102] In some examples, once the RRC connection is set up, the established connection can be used for other services in parallel with the initial purpose for connection establishment.

[0103] For example, when an RRC connection is established for a voice call, the UE can also access other services, such as short message services (SMS), using the same RRC connection.

[0104] Analogously, when an RRC connection has been established, the RRC signalling over the RRC connection can be used to carry an artificial intelligence (Al) and / or a machine learning (ML) model-related payload in addition to a non-AI and / or non-ML model-related payload. In some examples, the Al and / or the machine learning (ML) model-related payload information related to ML associated tasks and / or ML / AI models.

[0105] In some examples, information related to ML associated tasks and / or ML / AI models may be exchanged over an RRC established or resumed connection (for example, using an SRB or a DRB). Table 1 below shows various example ML- associated tasks that may result in an ML-associated data transfer between the network and UE(s).Table 1 : ML associated tasks intended for data transfer between a UE and network.

[0106] In some examples, the exchange of information related to ML associated tasks and / or ML / AI models may occur at any time depending on different triggers (relating to, for example, ML Model download, ML Data transfer etc.).

[0107] As discussed above, in some examples, the exchange of information between a UE and a network related to ML associated tasks and / or ML / AI models may be sent in parallel with non-AI / ML related RRM signalling or traffic over a single RRC connection after the RRC Connection is established. For example, when a first trigger condition is met for transferring ML / AI data to an RRC-Connected UE, an SRB and / or DRB may be established for exchanging information related to ML associated tasks and / or ML / AI models over an RRC connection, wherein the UE already has established SRBs or DRBs for radio resource management (RRM) signalling over the same RRC connection.

[0108] In this way, there may be a mix of ML-associated and non-ML-associated signalling and traffic in a network. The ratio may vary depending on multiple factors, including for example, busy hours and the fulfilment of any ML-associated data transfer trigger conditions.

[0109] The following recognises that it may be useful to treat Al and / or ML related RRC signalling differently to non-AI / ML related RRC signalling. For example, AI / ML modelRRC signalling may be configured such that it does not prevent non-AI / ML RRC signalling being signalled in accordance with minimum quality requirements associated with information transmitted for the non-AI / ML RRC signalling.

[0110] A two stage RRC signalling format has been proposed for signalling information related to AI / ML related RRC signalling and non-AI / ML related RRC signalling Al. In this example, a configuration of an RRC connection is modified based on the type of traffic being transmitted over it. For example, the two stage RRC signalling format may comprise a first RRC reconfiguration message for reconfiguring the RRC connection for carrying non-AL / ML related information. The first RRC reconfiguration message may comprise information and / or an indication scheduling a second RRC reconfiguration message, wherein the second RRC reconfiguration message is for reconfiguring the RRC connection for carrying AI / ML-related information.

[0111] In some examples, the two stage RRC signalling format is deployed via a single type of RRC connection. Stated differently, a single type of RRC connection may be used to transfer both non-AI / ML related information and AI / ML related information using the same SRB or DRB. However, messages cannot be prioritized within a given SRB because a given SRB is associated with a single logical channel. Therefore, all the RRC messages using the same SRB will have the same priority.

[0112] This can result in one or more issues.

[0113] For example, when the non-ML-associated traffic is high and / or overloaded (e.g., the maximum RRC connection capacity is reached or is closed to being reached), the network may want to de-prioritise the ML-associated traffic / users, for example, by pre-empting existing RRC Connections related to ML associated tasks and / or ML / AI models. Pre-empting existing RRC Connections related to ML associated tasks and / or ML / AI models may comprise releasing said connection when the network becomes overloaded. Moreover, in some examples, the network may want to de-prioritise RRC Connections related to ML associated tasks and / or ML / AI models as RRC Connections related to ML associated tasks and / or ML / AI models are of low priority.

[0114] However, pre-empting existing connections alone may not be sufficient to handle high and / or overloaded non-ML-associated traffic.

[0115] As another example, it may be useful for the network to stop admitting UEs that initiate new connections for the purpose of an ML associated task when the network is overloaded.

[0116] This is illustrated in the following example in which a network node has a maximum capacity of 10 RRC Connections is considered. In this example, the network node may have a configured overload condition. The term “overload condition” may be used interchangeably with the term “overload trigger” herein. The overload condition may be a threshold value related to the load of the network node. For example, the configured threshold value may indicate a value at which the network node is deemed to be overloaded or close to being overloaded. For example, the threshold value may indicate that the network node is at or close to its maximum capacity of RRC Connections. For example, the configured threshold value may be a threshold at which low priority connections are to be rejected. The overload condition may be configured in terms of a percentage of the total number of established RRC connections compared to the maximum capacity of RRC connections.

[0117] In this example, if 90% is configured as the overload condition, the network node may determine that the overload condition is triggered when 9 UEs are RRC- Connected. For example, the network node may determine that the load at the network node is at or above the threshold value when 9 UEs are RRC-Connected.

[0118] Using current methods, when a 10thUE initiates a low priority RRC Connection with the network node (e.g., for the purpose of exchanging information related to ML associated tasks and / or AI / ML models), the UE may send a connection request to the network node with a legacy establishment cause. In this case, the network node may accept the low priority connection for the purpose of exchanging information related to ML associated tasks and / or AI / ML models, even though the overload condition has been triggered. At this point, if another UE initiates a higher priority RRC Connection (e.g., for the purpose of exchanging information related to a non-ML / AI related RRM task), the another UE may be rejected because the maximum RRC connection capacity has been reached.

[0119] In this way, connections related to ML associated tasks and / or AI / ML models (i.e., low priority connections) may not be identified before a connection related configured for ML associated tasks and / or AI / ML models is established. For example,connections related to ML associated tasks and / or AI / ML models may not be deprioritised or pre-empted before said connection is established.

[0120] Establishing an ML / AI-related connection (e.g., when the overload condition is triggered) and subsequently releasing said connection when the network node identifies said connection as an ML / AI-related connection may introduce unnecessary signalling overhead and unnecessary processing overhead in the network node. In some examples, if the maximum RRC connection capacity is reached or the overload condition is triggered, but UEs are still attempting to connect to the network during this period, the UEs may be rejected, which may degrade the Accessibility KPIs.

[0121] Therefore, in the current state of the art, when AI / ML related signalling and non- AI / ML related signalling is exchanged over a same established connection, the network node is not able to identify when a connection request is for establishing a connection for the purpose of exchanging information related to ML associated tasks and / or AI / ML models. In this way, there may be a negative impact on exchanging information related to non-AI / ML related RRM tasks.

[0122] Examples herein aim to address one or more of the problems indicated above.

[0123] Various examples herein provide a method for creating an RRC Connection associated with ML associated tasks and / or ML models. In particular, certain examples herein provide a method for the network node to identify that an RRC Connection request is for an ML associated task and / or model.

[0124] In more detail, various examples herein provide a set of new establishment and / or resume cause values that are usable to distinguish when a connection is being established (or resumed) for AI / ML-related information exchange (e.g., as opposed to non-AI / ML-related information exchange). In some examples, each of the newly introduced cause values is mapped with one or more ML associated tasks and / or models.

[0125] In examples described below, the UE provides a network node or a cell of a network node with a connection request comprising at least one of the new cause values, wherein the cause value indicates that the connection request is related to an ML associated task and / or an ML model. In this way, the network node may identify that a connection request is for ML related signalling / data exchange, and treat the connection request accordingly. For example, the network node may identify that a connection request is for exchanging information to be used when executing an MLassociated task and / or ML model. The network node may determine whether to grant or refuse the connection request based on signalling load currently experienced (or expected to be experienced by) the network node for non-AI / ML associated information signalling. For example, when the network node is currently experiencing a high load, the network node may cause the connection request comprising an AI / ML- related establishment cause value to be refused.

[0126] In some examples, the UE may determine that signaling for an ML associated task and / or an ML model is to be performed between the UE and a network node. In some examples, the UE may autonomously determine that signaling for an ML associated task and / or an ML model is to be performed. For example, the UE may determine that signaling for an ML associated task and / or an ML model is to be performed without receiving triggering signaling from the network node. In some examples, the UE transmits an indication to the network node indicating that signaling for an ML associated task and / or an ML model is to be performed. In some examples, the UE may transmit the indication indicating that signalling for an ML associated task and / or ML model is to be performed, when an RRC connection between the network node and the UE is released.

[0127] In some other examples, the network node may autonomously determine that signaling for an ML associated task and / or an ML model is to be performed. For example, the network may determine that signaling for an ML associated task and / or an ML model is to be performed without receiving triggering signaling from the UE. In some examples, the network transmits an indication to the UE indicating that signaling for an ML associated task and / or an ML model is to be performed. In some examples, the network transmits the indication to the UE indicating that signaling for an ML associated task and / or an ML model is to be performed, when an RRC connection between the network node and UE is released.

[0128] In some other examples, a core network may determine that signaling for an ML associated task and / or an ML model is to be performed. In some examples, the core network is notified that signaling for an ML associated task and / or an ML model is to be performed from an over the top (OTT) Server (e.g., when the UE camps on the network node and is RRC connected). In some examples, the core network may transmit an indication to the network node indicating that signaling for an ML associated task and / or an ML model is to be performed. For example, a core networknode may transmit an indication to the network node indicating that signaling for an ML associated task and / or an ML model is to be performed. In some examples, the core network node may be a LMF or an AMF. In some examples an operations, administration and management (OAM) function may transmit an indication to the network node indicating that signaling for an ML associated task and / or an ML model is to be performed. In some examples, after receiving said indication, the network node may subsequently provide an indication to the UE indicating that signaling for an ML associated task and / or an ML model is to be performed.

[0129] The above-mentioned principles are illustrated with respect to Figures 6 to 12. In particular, Figures 6 to 12 illustrate principles that are later illustrated together in the examples of any of Figures 13 to 18.

[0130] It is understood that the terms “network node” and “cell” used herein may be used interchangeably. It is understood that the term “network node” may be refer to a network access node (e.g., a RAN) or a core network node or an OAM function.

[0131] Figure 6 shows an example signalling diagram in which a network node 602 indicates to a UE 600 that signalling for an ML associated task and / or an ML model is to be performed. The signaling example as described in Figure 6 may be used in connection with S1302 of Figure 13, as described below (e.g., when the network node autonomously determines that signaling for an ML associated task and / or ML model is to be performed).

[0132] In this example, the network node 602 may determine that signaling for an ML associated task and / or an ML model is to be performed. For example, as shown at S601 , the network node 602 may provide an ML Pending Task Notification to the UE 600. The ML Pending Task Notification may notify the UE 600 that an ML related RRC connection is intended to be set up, for example during or after the current RRC Connection is released.

[0133] In some examples, the ML Pending Task Notification comprises an ML Pending Task Notification information element (IE). In some examples, the ML Pending Task Notification IE comprises a binary value indicating whether an RRC Connection related to an ML associated task and / or an ML model is required. In some examples, the binary value may be indicated by “True” or “False”. In some examples, the ML Pending Task Notification comprises an ML associated task ID IE and / or a priority level of the ML Pending Task. In some examples, this procedure may be initiated any time duringthe current RRC Connection when criteria and / or conditions for setting up an RRC Connection related to an ML associated task and / or an ML model are met.

[0134] Figure 7 shows an example signalling diagram. As shown at S701 , a network node 702 provides an RRC connection release message (e.g., RRCRelease) to a UE 700, wherein the connection release message comprises the indication that signalling for an ML associated task and / or an ML model is to be performed. The signaling example as described in Figure 7 may be used in connection with S1303 of Figure 13, as described below.

[0135] For example, the connection release message may comprise the ML Pending Task Notification IE and / or the ML associated task ID IE. For example, the connection release message may comprise an indication that there is an upcoming ML associated task and / or ML model.

[0136] Figure 8 shows an example signalling diagram. As shown at S801 , a network node 802 provides an RRC reconfiguration message to a UE 800, wherein the RRC reconfiguration message comprises the indication that signalling for an ML associated task and / or an ML model is to be performed. The signaling example as described in Figure 8 may be used in connection with S1404 of Figure 14, as described below.

[0137] For example, the reconfiguration message may comprise the ML Pending Task Notification IE and / or the ML associated task ID IE. For example, the reconfiguration message may comprise an indication that there is an upcoming ML associated task and / or ML model. In some examples, the reconfiguration message comprising the ML Pending Task Notification indicates to the UE 800 that an RRC Connection related to an ML associated task and / or an ML model is intended to be set up after the current RRC Connection is released. As shown at S802, upon receipt of the RRC reconfiguration message, the UE 800 may transmit, to the network node 802, an RRC reconfiguration complete message.

[0138] Figure 9 shows an example signalling diagram. As shown at S901 , a core network node 902 transmits a UE context release command message to a network access node 900. It is understood that the core network node 902 may be replaced with an 0AM function in this example. The signaling example as described in Figure 9 may be used in connection with S1302 of Figure 13, as described below (i.e. when a core network node or an 0AM function determines that signalling for an ML associated task is to be performed).

[0139] In some examples, the UE context release command message triggers the network access node 900 to provide a connection release message to the UE. In some examples, the UE context release command is sent to the network access node 900 by an AMF to request the release of a UE-associated logical connection over the next generation (NG) interface (e.g., over the interface between the network access node 900 and a UE).

[0140] In some examples, the UE context release command message comprises the indication that signalling for an ML associated task and / or an ML model is to be performed. For example, the UE context release command message may comprise a ML Pending Task Notification IE and / or the ML associated task ID IE. For example, the UE context release command message may comprise an indication that there is an upcoming ML associated task and / or ML model.

[0141] In some examples herein, the UE may transmit, to a network node, an RRC connection request message. For example, after the UE has determined that signaling for an ML associated task and / or an ML model is to be performed between the UE and the network, the UE may provide the network node with the connection request message. In some examples, the connection request message indicates the UE’s intent to set up or resume an RRC connection related to an ML associated task and / or an ML model. The connection request message may be provided during a new RRC connection setup or during an RRC connection resumption.

[0142] As shown at S902, upon receipt of the UE context release command message, the network access node 900 may transmit a UE context release complete message to the core network node 902.

[0143] Figure 10 shows an example signalling diagram. As shown at S1001 , a UE 1000 may transmit, to a network node 1002, an RRC connection request message. The connection request message may be a RRC connection setup request message (e.g., RRCSetupRequest). In other examples, the connection request message may be a RRC connection resume message (e.g, RRCResumeRequest). In some examples, the connection request message comprises an indication that the connection request message is related to an ML associated task and / or an ML model. For example, a connection setup request message may be a connection request for establishing a connection, with the network node 1002, for exchanging information to be used when executing an ML associated task and / or ML model. For example, aconnection resume request message may be a connection request for resuming an existing connection with the network node 1002, for exchanging information to be used when executing an ML associated task and / or ML model. In some examples, the connection request message is for establishing a connection and / or for configuring a connection with the network. For example, the connection setup request is for establishing a service bearer and / or establishing a configuration for a service bearer.

[0144] It is understood that the term “connection request” and “connection request message” may be used interchangeably herein. It is understood that the term “connection request message” may refer to a “connection setup request message” or a “connection resume request message”.

[0145] Figures 11A and 11 B illustrate different example forms of the newly disclosed cause values, and illustrate their use in the connection request of Figure 10. It is understood that other forms may be taken for the cause values. It is also understood that combinations of these forms may be utilised (e.g., a cause value may comprise the example cause values of Figures 11 A and 11 B). It is further understood that such cause values as described below may be comprised in other types of signalling, such as the connection request messages described in relation to Figures 4, 10, 12-18.

[0146] In some examples, the connection request comprises a binary indication indicating that that the connection request is related to an ML associated task and / or ML model. Figure 11A shows an example signalling procedure between a UE 1100A and a network node 1102A. As shown at S1101 A the UE 1100A transmits a connection request (e.g., RRCSetupRequest or RRCResumeRequest) to the network node 1102A, wherein the connection request comprises said binary indication. In this example, the connection request may comprise a Boolean information element (IE) setting an “ML associated task associated connection indication" to True and / or False. In this way, the connection request may identify whether the connection is ML related or non-ML related.

[0147] As shown at S1102A, upon receipt of the connection request, the network node 1102A may provide an RRC connection setup (e.g, RRCSetup) or resume message (e.g., RRCResume) to the UE 1100A. For example, the network node 1102A may provide a connection setup or resume message to the UE 1110A based on the received cause value comprised in the received connection request.

[0148] As shown at S1103A, upon receipt of the RRC connection setup or resume message, the UE 1100A may transmit to the network node 1102A, an RRC complete message. For example, the UE 1110A may transmit to the network node 1102A, a setup complete message (e.g., RRCSetupComplete) or a resume complete message (e.g., RRCResumeComplete).

[0149] ln this way, and as shown at S1104A, once the connection has been established or resumed between the UE 1100A and network node 1102A, ML related information may be exchanged between the network nodel 102A and UE 1100A.

[0150] In some other examples, the connection request may comprise a cause value indicating that the connection request is related to a machine learning associated task and / or a machine learning model where the cause value is based on a priority level of an ML associated task and / or ML model. For example, the cause value is an index value that maps an ML associated task and / or ML model to a priority level. This is illustrated in Figure 11 B.

[0151] Figure 11 B shows an example signalling procedure between a UE 1100B and a network node 1102B. As shown at S1101 B, the UE 1100B transmits a connection request (e.g., RRCSetupRequest or RRCResumeRequest) to the network node 1102B, wherein the connection request indicates a priority level of an ML associated task and / or ML model.

[0152] As shown at S1102B, upon receipt of the connection request, the network node 1102B may provide an RRC connection setup (e.g., RRCSetup) or resume message (e.g., RRCResume) to the UE 1100B. For example, the network node 1102B may provide a connection setup or resume message to the UE 1100B based on the received cause value comprised in the received connection request.

[0153] As shown at S1103B, upon receipt of the RRC connection setup or resume message, the UE 1100B may transmit to the network node 1102B, an RRC complete message. For example, the UE 1110B may transmit to the network node 1102B, a setup complete message (e.g., RRCSetupComplete) or a resume complete message (e.g., RRCResumeComplete).

[0154] ln this way, and as shown at S1104B, once the connection has been established or resumed between the UE 1100B and network node 1102B, ML related information may be exchanged between the network nodel 102B and UE 1100B.

[0155] In this example, the priority of the ML associated task and / or ML model is indicated using an establishment cause value or a resume cause value. In this way, the connection request may indicate the priority of the connection.

[0156] In some examples, one or more low priority establishment causes or resume causes (such as lowpriorityAccessl , lowpriorityAccess2) may be added in the list of legacy establishment causes supported by the UE and / or network. For example, an updated list of supported cause values may comprise: ENUMERATED {emergency, highPriorityAccess, mt-Access, mo-Signalling, mo-Data, mo-VoiceCall, mo-VideoCall, mo-SMS, mps-PriorityAccess, mcs-PriorityAccess, lowpriorityAccessl , lowpriorityAccess2, spare4, spare3, spare2, sparel }.

[0157] In some examples, the UE may map an ML associated task and / or ML model to a priority level (e.g., low priority). The UE may then map the priority level to an establishment cause value or a resume cause value. In this way, the cause value may be an index value that maps an ML associated task and / or ML model to a priority level.

[0158] In some examples, each ML associated task may be identified with a specific task-ID and relative priority, as shown below in Table 2. There may be various possible ML associated tasks based on different LCM phases (e.g., data collection, model training, inference and performance monitoring).Table 2: Priority and task-ID mapping to ML associated tasks

[0159] Figure 12 shows an example signalling diagram of connection priority identification at the network node 1202 using an ML related cause value.

[0160] At S1201 , the network node 1202 identifies an overload condition. In some examples, the overload condition is related to the maximum capacity of RRC Connections at the network node 1202. For example, the configured overloadcondition may be a threshold value at which low priority connections are to be rejected. For example, the overload condition may be triggered when 90% of the maximum capacity of RRC connections have been established.

[0161] At S1202, when the network node 1202 has identified the overload condition, the network node 1202 may be able to identify and pre-empt low priority RRC connections (e.g., RRC connections related to an ML associated task and / or an ML model). For example, the network node 1202 may pre-empt RRC Connections related to an ML associated task and / or an ML model when the overload condition is triggered. For example, the network node may pre-empt RRC Connections related to an ML associated task and / or an ML model when the load at the network node 1202 is at or above a threshold value. In this example, the load at the network node 1202 is related to the maximum capacity of RRC connections supported by the network node 1202.

[0162] In one example, a network node supports 100 RRC Connections of which 20 RRC Connections are related to an ML associated task and / or an ML model when at a given point in time. At that time instance, if there is a need to admit 10 RRM-task related RRC Connections, the network node may pre-empt an existing 10 ML-task associated RRC Connections (i.e., may pre-empt 10 existing RRC Connections related to an ML associated task and / or an ML model).

[0163] In some examples, the network node 1202 may enable connection identification of ML related connections. For example, the network node 1202 may activate a connection prioritisation. In some examples, the connection prioritisation may enable the network node 1202 to prioritize high priority connections (e.g., RRM Task associated RRC Connections) over low priority connections (e.g., RRC Connections related to an ML associated task and / or an ML model).

[0164] At S1203A, the UE 1200 may provide an RRC connection request (e.g., RRCSetupRequest or RRCResumeRequest) to the network node 1202. The RRC connection request may comprise a cause value related to an ML associated task and / or an ML model. For example, the cause value may comprise a binary value indicating that the connection is related to an ML associated task and / or an ML model. In other examples, the cause value may comprise an index value that maps an ML model and / or ML associated task to a priority level.

[0165] At S1204A, in a case where the network node 1202 has identified that the overload condition has been triggered, the network node 1202 may determine to rejectthe connection request related to an ML associated task and / or an ML model. For example, when the network node 1202 determines that an RRC connection load is at or above a configured threshold value, the network node 1202 may determine to reject the connection request related to an ML associated task and / or an ML model.

[0166] In some examples, if the overload condition is resolved (e.g., is no longer triggered) the network node 1202 may begin to accept RRC connection requests related to an ML associated task and / or an ML model. For example, when the network node 1202 determines that an RRC connection load is below a configured threshold value, the network node 1202 may determine to accept the connection request related to an ML associated task and / or an ML model.

[0167] At S1205A, when the network node 1202 determines to reject the ML related RRC connection request, the network node 1202 may transmit an RRC reject message to the UE 1200.

[0168] Alternatively, to S1203A to S1205A, at S1203B, the UE 1200 may transmit a RRM Task related RRC connection request (e.g., RRCSetupRequest or RRCResumeRequest) to the network node 1202. The RRC connection request may comprise a cause value that indicates that the connection request is related to a RRM task.

[0169] At S1204B, in the case where the network node 1202 has identified that the overload condition has been triggered, the network node 1202 may determine to accept the RRM Task connection request i.e. , because the cause value indicates that the RRM Task related connection request is of high priority.

[0170] At S1205B, if the network node 1202 determines to accept the connection request, the network node 1202 may transmit an RRCSetup or RRCResume message to the UE 1200.

[0171] At S1206B, the UE 1200 may provide the network node 1202 with an RRCSetupComplete or RRCResumeComplete message.

[0172] Figures 13 and 14, as described in further detail below, illustrate examples in which the examples of Figures 6-12 can be implemented.

[0173] Figure 13 provides an example signalling procedure according to the examples herein. Figure 13 shows an example signalling procedure between a UE 1300 and a network node 1302. In some examples, the network node 1302 is a network access node. In some examples, the network node 1302 is a core network node. For example,the core network node may be an AMF or an LMF. In some examples, the network node 1302 is an OAM function. In some examples, the network node 1302 is an RAN node.

[0174] At S1301 , the UE 1300 may be camped on a cell in the network node 1302 and may be RRC connected.

[0175] At S1302, the UE 1300 and / or network node 1302 determines that signalling for an ML associated task and / or ML model is intended and / or will be performed between the UE 1300 and a network node.

[0176] In some examples, the machine learning associated task comprises at least one of: transfer of a machine learning model; transfer of one or more conditions of a machine learning model; transfer of machine learning model validation data; transfer of training data related to a machine learning model; or transfer of a machine learning data set.

[0177] In some examples, the UE 1300 may autonomously determine that signalling for an ML associated task and / or ML model is intended and / or will be performed. In this case the UE 1300 may transmit, to the network node 1302, information related to signalling to be performed for an ML associated task and / or ML model. For example, the UE 1300 may transmit, to the network node 1302, the ML Pending Task Notification IE and / or the ML associated task ID IE. For example, the UE 1300 may transmit, to the network node 1302, information related to an upcoming ML associated task and / or ML model. In some examples, the UE 1300 may transmit the information related to signalling to be performed for an ML associated task and / or ML model to the network node 1302 when an RRC connection between the network node 1302 and the UE 1300 is released.

[0178] In some other examples, the network node 1302 may autonomously determine that signalling for an ML associated task and / or ML model is intended and / or will be performed. In this case, the network node 1302 may transmit, to the UE 1300, information related to signalling to be performed for an ML associated task and / or ML model. For example, the network node 1302 may send the UE 1300 the ML Pending Task Notification IE and / or the ML associated task ID IE and / or ML model ID. For example, the network node 1302 may transmit, to the UE 1300, information related to an upcoming ML associated task and / or ML model. In some examples, the network node 1302 may transmit the information related to signalling to be performed for anML associated task and / or ML model to the UE 1300 when an RRC connection between the network node 1302 and the UE 1300 is released.

[0179] In some other examples, when the network node 1302 is a network access node, a core network node or an 0AM function may determine that signalling for an ML associated task and / or ML model is intended and / or will be performed. In this case, the core network or 0AM function may transmit, to the network node 1302, a UE context release command message as described above with regards to Figure 9. The UE context release command message may trigger the network node 1302 to provide the UE 1300 with information related to signalling to be performed for an ML associated task and / or ML model.

[0180] In some examples, when the network node 1302 determines that a configured overload condition has been triggered, the network node 1302 may activate a connection prioritisation. The overload condition may be as described with respect to Figure 12. In some examples, the connection prioritisation may enable the network node 1302 to prioritize high priority connections (e.g., RRM Task associated RRC Connections or high priority ML related RRC Connections) over low priority connections (e.g., low priority ML related RRC Connections). The network node 1302 may add the UE 1300 to a list indicating that the network node 1302 is to transmit a ML Pending Task Notification to the UE 1300 when the RRC connection between the UE 1300 and network node 1302 has been released.

[0181] At S1303, the network node 1302 may transmit, to the UE 1300, a connection release message (e.g., RRCRelease).

[0182] In some examples, the connection release message comprises the information related to signalling to be performed for the ML associated task and / or ML model. For example, when the network node 1302 or the core network or the 0AM function determines that signalling for an ML associated task and / or ML model is intended and / or will be performed, the connection release message may comprise the information related to the signalling to be performed for the ML associated task and / or ML model. In some examples, the connection release message comprises an indication to the UE 1300 to release its current connection with the network node 1302 due to signalling that is to be performed for a ML associated task and / or ML model.

[0183] In some examples, the connection release message is provided to the UE 1300 based on the determination that signalling for an ML associated task and / or ML model is intended and / or will be performed.

[0184] In some examples, the connection release message comprises an indication to the UE 1300 to release the current connection with the network node 1302.

[0185] In some examples, it may be possible for the network node 1302 to create ML DRBs in the current RRC connection. However, when the network node 1302 is overloaded, and ML related tasks and / or ML models are deprioritised, the network node 1302 may transmit to, the UE 1300, a connection release message.

[0186] In some examples, the connection release message comprises an indication to the UE 1300 to switch to another state (e.g., an idle state or an inactive state).

[0187] For example, if the network node 1302 determines that the UE 1300 should be moved into an RRC idle state, the connection release message may be an RRC release message.

[0188] In some examples, the connection release message may indicate that the signalling for the ML associated task and / or ML model is to be initiated at a particular time. In this case, the UE 1300 may switch to an idle state.

[0189] In an example where the network node 1302 determines that the UE 1300 should be moved into an RRC inactive state, the connection release message may be an RRC release with a suspend configuration message.

[0190] At S1304, based on the connection release message received, the UE 1300 may release its connection with the network node 1302 and / or switch to an RRC idle state or switch to an RRC Inactive state.

[0191] At S1305, the UE 1300 determines to establish a connection for signalling that is to be performed, between the UE 1300 and a network node, for a ML associated task and / or ML model.

[0192] In some examples, the UE 1300 determines to establish a connection with a network node based on determined and / or received information related to the ML associated task and / or ML model that is to be signalled (e.g., the pending ML associated task and / or model). In some examples, the UE 1300 may determine to establish a connection with a network node based on the connection release message received from the network node 1302.

[0193] At S1306, the UE 1300 transmits a connection request message (e.g., RRCSetupRequest or RRCResumeRequest) to a network node. In some examples, the UE 1300 provides a connection request to network node 1302. In some other examples, the UE 1300 provides the connection request to another network node (e.g., a second network node). In some examples, the second network node is a network access node. In some examples, the second network node is a core network node. For example, the core network node may be an AMF or an LMF. In some examples, the second network node is an 0AM function.

[0194] In some examples, the connection request comprises a cause value. For example, the connection request comprises a cause value which indicates that the connection request is related to an ML associated task and / or ML model.

[0195] In some examples, the cause value corresponds to an ML associated task and / or an ML model. A cause value that corresponds to an ML associated task and / or an ML model may be referred to as an ML related cause value herein.

[0196] In some examples, the cause value is a binary indication. For example, the cause value indicates whether the connection request is related to an ML associated task and / or ML model via a binary indication (e.g., True / False). For example, the cause value is an ML associated task associated connection indication IE (which may be set to True if the connection is ML related and set to False if the connection is not ML related). This may be as described above in connection with Figure 11 A.

[0197] In some examples, the UE 1300 maps the ML associated task ID and / or ML model ID (associated with the ML associated task and / or ML model that is to be signalled) to a priority level. The UE 1300 may subsequently map the priority level (associated with the ML associated task ID and / or ML model ID) to an RRC establishment cause. For example, the UE 1300 maps the priority to the RRC Establishment cause “lowpriorityAccessl”. In this way, the cause value is an establishment cause value. In some examples, the UE 1300 may subsequently map the priority level (associated with the ML associated task ID and / or ML model ID) to an RRC resume cause. For example, the UE 1300 maps the priority to the RRC resume cause “lowpriorityAccessl”. In this way, the cause value is the resume cause value. This may be as described above with reference to Figures 11 B and 12.For example, the cause value is an index value that maps the ML associated task and / or ML model to a priority level.

[0198] Where the connection release message indicates switching to an idle state, the connection request message is an RRC setup request message and the cause value is an establishment cause value.

[0199] In a case where the connection release message indicates that the signalling for the ML associated task and / or ML model is to be initiated at a particular time, the UE 1300 may initiate the connection request message when the time instance notified / configured by the network node 1302 occurs.

[0200] In the case where the connection release message indicates switching to an inactive state, the connection request message is an RRC resume request message and the cause value is a resume cause value.

[0201] At S1307, the network node that has received the connection request determines whether to accept or reject the connection request. For example, if the connection request is received by network node 1302, network node 1302 may determine to accept or reject the connection request. In another example, if the connection request is received by a different network node to network node 1302 (i.e. , the second network node), the second / different network node may determine to accept or reject the connection request.

[0202] In some examples, network node 1302 or the second network node may transmit to the UE 1300 a message indicating whether the connection request has been accepted or rejected by the respective network node.

[0203] In some examples, the message indicating whether the connection request has been accepted is an RRC setup message. For example, where the connection release message indicates switching to an idle state, the message indicating whether the connection request has been accepted is an RRC setup message.

[0204] In some examples, the message indicating whether the connection request has been accepted is an RRC resume message. For example, where the connection release message indicates switching to an inactive state, the message indicating whether the connection request has been accepted is an RRC resume message.

[0205] In some examples, when the message provided by network node 1302 or the second network node indicates that the connection request has been accepted, the UE 1300 establishes a connection with the network node 1302 or second network node. In some examples, when a connection is established between the UE 1300 and network node 1302 or the second network node, the UE 1300 and / or the network nodemay exchange information to be used when executing an ML associated task and / or an ML model with each other.

[0206] In some examples, network node 1302 or the second network node may accept the connection request when the overload condition is not triggered. For example, when network node 1302 or the second network node determines that the load is below a threshold value, network node 1302 or the second network node may accept the connection request.

[0207] In some examples, when network node 1302 or the second network node accepts the connection request, the network node 1302 or the second network node may prioritize RRM related connections the established connection. For example, when a load at the network node 1302 or the second network node is above or at a threshold value, the network node 1302 or the second network node may prioritize RRM related connections over ML related connections. For example, when a load at the network node 1302 or the second network node is above or at a threshold value, the network node 1302 or the second network node may prioritize RRM related connections over the established connection corresponding to the connection request related to an ML associated task and / or ML model.

[0208] In some examples, the UE 1300 and / or network node 1302 and / or the second network node may prioritize exchanging RRM related information over exchanging information to be used when executing an ML associated task and / or ML model. In some examples, prioritizing exchanging RRM related information may be achieved by configuring a higher priority SRB or DRB for exchanging the RRM related information than the bearer(s) configured to be used for exchanging the information to be used when executing an ML associated task and / or ML model. For example, when the overload condition is triggered, network node 1302 or the second network node may prioritize exchanging RRM related information over exchanging information to be used when executing an ML associated task and / or ML model. Information to be used when executing an ML associated task and / or ML model may be referred to as ML related information herein.

[0209] In some examples, the UE 1300 may initiate RRM actions (e.g., bearer setup) either after the ML associated task and / or ML model is completed or while the ML associated task and / or ML model is in progress.

[0210] In some examples, the network node 1302 or the second network node may reject the connection request when the overload condition is triggered. For example, when network node 1302 or the second network node determines that the load is above or at a threshold value, network node 1302 or the second network node may reject the connection request. In some examples, when network node 1302 or the second network node determines that the load is above or at a threshold value, network node 1302 or the second network node may pre-empt connections related to an ML associated task and / or ML model.

[0211] In some examples, if the network node 1302 or the second network node determines that the overload condition is triggered after accepting the connection request, network node 1302 or the second network node may transmit a command message to the UE 1300 to release the connection with the respective network node.

[0212] In some examples, when the network node 1302 or the second network node determines that the connection request has been rejected, the network node 1302 or second network node may transmit, to the UE 1300, a backoff timer.

[0213] In some examples, when the message obtained from the network node 1302 or second network node indicates that the connection request has been rejected, the UE 1300 may initiate a backoff timer. In some examples, the UE 1300 may initiate a backoff timer that has been configured at the UE 1300. In some examples, the UE 1300 may initiate the backoff timer received from the network node 1302 or second network node. In some examples, the UE 1300 may transmit a second connection request comprising the ML related cause value to network node 1302 or to the second network node when the backoff timer expires.

[0214] In some examples, when the message received from the network node 1302 or second network node indicates that the connection request has been rejected, the UE 1300 may transmit a second connection request comprising the ML related cause value to another network node (for example, a third network node). In some examples, the third network node is a network access node. In some examples, the third network node is a core network node. For example, the core network node may be an AMF or a LMF. In some examples, the third network node is an 0AM function.

[0215] Figure 14 shows another example that considers the use of the above- mentioned principles of Figures 6 to 12. In the example of Figure 14, a UE 1400 resumes a current ML related RRC connection with a network node 1402. In someexamples, the network node 1402 is a network access node. In some examples, the network node 1402 is a core network node. For example, the core network node may be an AMF or an LMF. In some examples, the network node 1402 is an OAM function.

[0216] In some examples, when the UE 1400 resumes its RRC Connection, the UE 1400 has one or more ML related DRBs that are suspended.

[0217] Figure 14 considers a scenario where the UE 1400 is configured with one DRB related ML associated task.

[0218] At S1401 , the UE 1400 may have setup or established an ML related connection with the network node 1402. For example, the UE 1400 may have an established connection related to an ML associated task and / or ML model.

[0219] At S1402, the network node 1402 may determine inactivity on the configured DRB. For example, the network node 1402 may determine that the UE 1400 is inactive.

[0220] At S1403, the network node 1402 may transmit, to the UE 1400, a connection release message (e.g., RRCRelease).

[0221] For example, when the network node 1402 determines inactivity on the configured DRB, the network node 1402 may transmit, to the UE 1400, an RRC release with suspend configuration message.

[0222] In some examples, when the UE 1400 receives the connection release message, the UE 1400 moves into an inactive state.

[0223] At S1404, the UE 1400 and / or network node 1402 determines that the connection related to an ML associated task and / or an ML model is to be resumed. For example, the UE 1400 and / or network node 1402 determines that signalling for an ML associated task and / or ML model is intended and / or will be performed between the UE 1400 and a network node. For example, when UL data is detected on the DRB, the UE 1400 may determine that the connection related to an ML associated task and / or an ML model is to be resumed. For example, when DL data is detected on the DRB, the network node 1402 may determine that the connection related to an ML associated task and / or an ML model is to be resumed. In this case, the network node 1402 may provide the UE 1400 with an indication to resume the connection related to an ML associated task and / or an ML model. For example, the network node 1402 may provide the UE 1400 with information of the DL data on the DRB.

[0224] At S1405, after determining and / or receiving an indication that the connection related to an ML associated task and / or an ML model is to be resumed, the UE 1400 provides a connection request (e.g., RRCResumeRequest) to the network node 1402.

[0225] In some examples, the connection request comprises a cause value. In some examples, the cause value indicates that the connection request is related to an ML associated task and / or an ML model.

[0226] In some examples, the cause value is based on determined and / or received UL or DL data on the DRB.

[0227] In some examples, the cause value is a binary indication. For example, the cause value indicates whether the connection request is related to an ML associated task and / or ML model via a binary indication (e.g., True / False). For example, the cause value is an ML associated task associated connection indication IE (which may be set to True if the connection is ML related and set to False if the connection is not ML related). This may be as described above in relation to Figure 11 A.

[0228] In some examples, the UE 1400 maps the ML associated task ID and / or ML model ID (associated with the ML associated task and / or ML model that is to be signalled) to a priority level. The UE 1400 may subsequently map the priority level (associated with the ML associated task ID and / or ML model ID) to an RRC resume cause. For example, the UE 1400 maps the priority to the RRC resume cause “lowpriorityAccessl ”. In this way, the cause value is the resume cause value. This may be as described above in connection with Figure 11 B and / or Figure 12.

[0229] For example, the cause value is an index value that maps the ML associated task and / or ML model to a priority level.

[0230] In some examples, the connection request is an RRC resume request message.

[0231] At S1406, the network node 1402 determines whether to accept or reject the connection request.

[0232] For example, the network node 1402 may determines whether to resume the connection related to an ML associated task and / or an ML model.

[0233] In some examples, the network node 1402 determines whether to resume the connection related to an ML associated task and / or an ML model based on the overload condition. From S1406, the method may proceed as discussed above from S1303 in Figure 13.

[0234] Some advantages of various examples herein comprise identifying and / or rejecting connections related to an ML associated task and / or an ML model (i.e. , low priority connections) based on the load of the network node. For example, connections related to an ML associated task and / or an ML model may be rejected before an RRC connection is established. In this way, latency, signalling overhead, processing overhead and / or KPI impact may be reduced.

[0235] An advantage of some examples herein comprises controlling RRC connections using separate thresholds for non-ML related signalling and ML-related signalling.

[0236] An advantage of some examples herein comprises eliminating the need for separate RRC messages.

[0237] Use cases of various examples of the present application may comprise at least one of: channel state information (CSI) feedback enhancement; Beam management; and Positioning accuracy enhancements.

[0238] CSI feedback enhancement may comprise spatial-frequency domain CSI compression using a two-sided ML / AI model and / or time domain CSI prediction using a UE sided ML / AI model.

[0239] Various examples herein provide a use case for AI / ML general frameworks for one-sided AI / ML models within the proposals studied in FS_NR_AIML_Air project [RAN2],

[0240] For example, AI / ML general frameworks for one-sided AI / ML models may comprise at least one of: signalling and protocol aspects of Life Cycle Management (LCM) enabling functionality and model selection, activation, deactivation, switching and / or fallback (wherein in some examples, identification related signalling may be required for signalling of LCM); signalling and / or mechanism(s) for LCM to facilitate model training, inference, performance monitoring and / or data collection for both UE- sided and NW-sided AI / ML models; and signalling mechanisms of applicable functionalities and / or models.

[0241] Beam management may comprise downlink (DL) transmission beam prediction for both UE-sided AI / ML models and NW-sided AI / ML models. For example, DL transmission beam prediction may comprise at least one of: spatial-domain DL transmission beam prediction for a set of beams (Set A) based on measurement results of another set of beams (Set B); temporal downlink beam prediction for a setof beams (Set A) based on the historic measurement results of another set of beams (Set B); specifying signalling and / or mechanism(s) to facilitate LCM operations specific to Beam Management use cases; and / or enabling method(s) to ensure consistency between training and inference stages regarding NW-side additional conditions for inference at the UE.

[0242] Positioning accuracy enhancements may comprise direct AI / ML positioning; AI / ML assisted positioning; specifying measurements / signalling / mechanism(s) to facilitate LCM operations specific to positioning accuracy enhancement use cases; investigating and specifying the signalling of measurement enhancements; and enabling method(s) to ensure consistency between training and inference stages regarding NW-side additional conditions for inference at UE for relevant positioning accuracy enhancement use cases.

[0243] In some examples, direct AI / ML positioning may comprise: UE-based positioning with a UE-side model; UE-assisted / LMF-based positioning with a LMF-side model; and / or NG-RAN node assisted positioning with a LMF-side model. In some further examples, AI / ML assisted positioning may comprise: UE-assisted / LMF-based positioning with a UE-side model and / or NG-RAN node assisted positioning with a gNB-side model.

[0244] Figures 15 to 18 illustrate aspects of the above examples. It is therefore understood that the above description may provide features that have functional correspondence with features discussed below, and so may be used to better understand the presently described aspects.

[0245] Reference is made to Figure 15 which shows a method of some examples.

[0246] This method of Figure 15 may be performed by an apparatus. The apparatus may comprise or be a user equipment, such as described above in connection with Figure 3. Alternatively, or additionally, the apparatus may be such as discussed in relation to Figure 19.

[0247] The method may comprise, as referenced at S1501 , transmitting, to a first network node, a connection request, the connection request comprising a cause value, the cause value indicating that the connection request is related to a machine learning associated task and / or a machine learning model.

[0248] It may be understood that the connection request of S1501 relates to the connection request described for Figures 4, 10, 11 A, 11 B and 12-14.

[0249] The cause value of S1501 may comprise properties of a service for which a connection is being requested by the apparatus.

[0250] Stated differently the apparatus may request to establish or resume a connection with a first network node. The request to establish or resume the connection may comprise a cause value, wherein the cause value indicates that the request for establishing or resuming a connection with the first network node is for the purpose of exchanging an ML associated task and / or an ML model.

[0251] The cause value may be a binary indication. Stated differently, the cause value may indicate whether the connection request is related to an ML associated task and / or ML model using a binary indication. This may be as described above in connection with Figure 11A and 13-14.

[0252] The cause value may be an index value that maps a machine learning associated task and / or a machine learning model to a priority level. Stated differently, the apparatus may map the ML associated task and / or ML model to a priority level. For example, the apparatus may map a ML associated task ID and / or ML model ID to a priority level. The apparatus may subsequently map the priority level to a cause value. For example the apparatus may map a priority level (associated with a ML associated task and / or ML model) to a cause value. This may be as described above in connection with Figures 11 B, 12-14.

[0253] The ML associated task may comprise at least one of the ML associated tasks as described in relation to Table 1 . For example, the ML associated task may comprise at least one of: transfer of a machine learning model; transfer of one or more conditions of a machine learning model; transfer of machine learning model validation data; transfer of training data related to a machine learning model; or transfer of a machine learning data set.

[0254] The apparatus may receive, from a second network node, a connection release message, prior to transmitting the connection request of S1501 .

[0255] The connection release message may be as described above in connection with Figures 5, 7 and 13-14. For example, the connection release message may indicate to the apparatus to release a connection with the second network node or to switch from a connected state to another state. In this way, based on the connection release message, the apparatus may release a connection with the second network node or may switch from a connected state to another state.

[0256] In some examples, the another state is an idle state or an inactive state.

[0257] Stated differently, the connection release message received by the apparatus from the second network node, may indicate to the apparatus to switch from a connected to state to either an idle state or an inactive state. For example, the connection release message may indicate to the apparatus to release its connection with the second network node and enter into either an RRC idle state or an RRC inactive state.

[0258] In some examples, when the another state is the idle state, the connection release message is a radio resource control release message, the connection request message is a radio resource control setup request message, a message indicating that the connection request has been accepted by the first network node is a radio resource control setup message and the cause value is an establishment cause value.

[0259] Stated differently, when the connection release message indicates to the apparatus to switch to an idle state, the connection release message may be an RRC release message. The apparatus may subsequently transmit a connection request message (as described in S1501 ), to a first network node, wherein the connection request is a RRC setup request message. The connection setup request may comprise an establishment cause value. For example, the establishment cause value may indicate to the first network node that the apparatus is requesting to establish a connection with the first network node for the purpose of exchanging an ML associated task and / or an ML model. When the first network node accepts the connection request, the first network node may send a RRC setup message to the apparatus to indicate to the apparatus that the connection setup request has been accepted.

[0260] In some examples, when the another state is the inactive state, the connection release message is a radio resource control release with a suspend configuration message, the connection request message is a radio resource control resume request message, a message indicating that the connection request has been accepted by the first network node is a radio resource control resume message and the cause value is a resume cause value.

[0261] Stated differently, when the connection release message indicates to the apparatus to switch to an inactive state, the connection release message may be an RRC release message with a suspend configuration message. The apparatus may subsequently transmit a connection request message (as described in S1501 ), to afirst network node, wherein the connection request is a RRC resume request message. The connection resume request may comprise a resume cause value. For example, the resume cause value may indicate to the first or second network node that the apparatus is requesting to establish or resume a connection with the first or second network node for the purpose of exchanging an ML associated task and / or an ML model. When the first or second network node accepts the connection request, the first or second network node may send a RRC resume message to the apparatus to indicate to the apparatus that the connection resume request has been accepted.

[0262] Prior to S1501 , the apparatus may determine that signaling for a ML associated task and / or a ML model will be performed between the apparatus and a network node, wherein transmitting the connection request is performed based on said determination that signaling for a machine learning associated task and / or a machine learning model will be performed.

[0263] In some examples, the apparatus determines that signaling for a machine learning associated task and / or a machine learning model will be performed autonomously and transmits, to the second network node, information related to an upcoming machine learning associated task and / or machine learning model.

[0264] Stated differently, the apparatus may autonomously determine that signaling for an ML associated task and / or an ML model is to be performed. Autonomously determining may comprise the apparatus determining that signaling for an ML associated task and / or an ML model is to be performed without receiving triggering signaling from a network node. The apparatus may subsequently transmit information related to signaling to be performed for a ML associated task and / or ML model to the second network node. This may be as described above in connection with Figure 13 and 14.

[0265] In some examples, the apparatus determines that signaling for a machine learning associated task and / or a machine learning model will be performed based on receiving, from the second network node, information related to an upcoming machine learning associated task and / or machine learning model.

[0266] Stated differently, the apparatus may receive information related to signaling to be performed for a ML associated task and / or ML model from the second network node.

[0267] In this case, the connection release message received from the second network node may comprise the information related to the upcoming machine learning associated task and / or machine learning model.

[0268] In some examples, after the apparatus transmits the connection request of S1501 to the first network node, the apparatus may receive, from the first network node, a message indicating whether the connection request has been accepted or rejected by the first network node. This may be as described above in connection with Figure 13 and 14.

[0269] Stated differently, when the first network node has a load that is below a configured threshold value, the first network node may accept the connection request received from the apparatus. For example, when an overload condition is not triggered, the first network node may accept the connection request received from the apparatus. The overload condition may be as described above in connection with Figure 12.

[0270] In other examples, when the first network node has a load that is at or above a configured threshold value, the first network node may reject the connection request received from the apparatus. For example, when an overload condition is triggered, the first network node may reject the connection request received from the apparatus. The overload condition may be as described above in connection with Figure 12.

[0271] In some examples, when the message received from the first network node indicates that the connection request has been accepted, the apparatus may establish a connection with the first network node and may exchange information to be used when executing a machine learning associated task and / or a machine learning model with the first network node over the established connection.

[0272] It is to be understood that establishing a connection with the first network node may comprise establishing a new connection or resuming an existing connection.

[0273] In some examples, when a load at the first network node is above or at a first threshold value, the apparatus may prioritize an RRM related connection over an established connection corresponding to the connection request related to a machine learning task and / or machine learning model.

[0274] In other words, when the first network node accepts the connection request of S1501 , the UE may establish a connection with the first network node. When a load at the first network node becomes above or equal to a configured threshold, and thefirst network node receives a connection request for an RRM related task, the apparatus may receive an indication from the first network node to release the established connection for signalling information related to a ML associated task and / or ML model. The first network node may subsequently accept the connection request for the RRM related task.

[0275] In some examples, when the message received from the first network node indicates that the connection request has been rejected, the apparatus may initiate a backoff timer, and may transmit, to the first network node, a second connection request comprising the cause value when the backoff timer expires.

[0276] Stated differently, the apparatus may be configured with a backoff timer. For example, the apparatus may receive a backoff timer from the first network node. For example, the backoff timer may indicate how long the first network node will have a load at or above the configured threshold value. In some examples, after the time indicated by the backoff timer has expired, the apparatus may transmit a connection request message to the first network node, wherein the connection request message comprises a cause value as indicated in S1501.

[0277] In some examples, when the message received from the first network node indicates that the connection request has been rejected, the apparatus may transmit a second connection request comprising the cause value to another network node or cell.

[0278] Stated differently, when the first network node rejects the connection request received from the apparatus, the apparatus may transmit a connection request to a different network node or cell, wherein the connection request comprises the cause value as described in relation to S1501.

[0279] The first network node may be the second network node. For example, the first network node of S1501 may be the same entity as the second network node. In other examples, the first network node may be different to the second network node. For example, the first network node of S1501 is not the same entity as the second network node.

[0280] The first network node and / or second network node may be one of: a network access node; a core network node; or an operations, administration, and management function.

[0281] Reference is made to Figure 16 which shows a method of some examples.

[0282] This method of Figure 16 may be performed by an apparatus. The apparatus may comprise or be a network node, such as described above in connection with Figure 3.

[0283] Alternatively, or additionally, the apparatus may be such as discussed in relation to Figure 19.

[0284] The method may comprise, as referenced at S1601 , receiving, from a user equipment, a connection request, the connection request comprising a cause value, the cause value indicating that the connection request is related to a machine learning associated task and / or a machine learning model.

[0285] It may be understood that the connection request of S1601 relates to the connection request described for Figures 4, 10, 11 A, 11 B and 12-14.

[0286] The cause value of S1601 may comprise properties of a service for which a connection is being requested by the user equipment.

[0287] Stated differently the user equipment may request to establish or resume a connection with the apparatus. The request to establish or resume the connection may comprise a cause value, wherein the cause value indicates that the request for establishing or resuming a connection with the apparatus is for the purpose of exchanging an ML associated task and / or an ML model.

[0288] The cause value may be a binary indication. Stated differently, the cause value may indicate whether the connection request is related to an ML associated task and / or ML model using a binary indication. This may be as described above in connection with Figure 11A and 13-14.

[0289] The cause value may be an index value that maps a machine learning associated task and / or a machine learning model to a priority level. Stated differently, the UE may map the ML associated task and / or ML model to a priority level. For example, the UE may map a ML associated task ID and / or ML model ID to a priority level. The UE may subsequently map the priority level to a cause value. For example the UE may map a priority level (associated with a ML associated task and / or ML model) to a cause value. This may be as described above in connection with Figures 11 B, 12-14.

[0290] The ML associated task may comprise at least one of the ML associated tasks as described in relation to Table 1 . For example, the ML associated task may comprise at least one of: transfer of a machine learning model; transfer of one or more conditionsof a machine learning model; transfer of machine learning model validation data; transfer of training data related to a machine learning model; or transfer of a machine learning data set.

[0291] The apparatus may transmit to the UE a connection release message, prior to receiving the connection request of S1601 .

[0292] The connection release message may be as described above in connection with Figures 5, 7 and 13-14. For example, the connection release message may indicate to the UE to release a connection with the second network node or to switch from a connected state to another state. In this way, based on the connection release message, the UE may release a connection with the apparatus or may switch from a connected state to another state.

[0293] In some examples, the another state is an idle state or an inactive state.

[0294] Stated differently, the connection release message received by the UE from the apparatus, may indicate to the UE to switch from a connected to state to either an idle state or an inactive state. For example, the connection release message may indicate to the UE to release its connection with the apparatus and enter into either an RRC idle state or an RRC inactive state.

[0295] In some examples, when the another state is the idle state, the connection release message is a radio resource control release message, the connection request message is a radio resource control setup request message, a message indicating that the connection request has been accepted by the apparatus is a radio resource control setup message and the cause value is an establishment cause value.

[0296] Stated differently, when the connection release message indicates to the UE to switch to an idle state, the connection release message may be an RRC release message. The UE may subsequently transmit a connection request message (as described in S1601 ), to the apparatus, wherein the connection request is a RRC setup request message. The connection setup request may comprise an establishment cause value. For example, the establishment cause value may indicate to the apparatus that the UE is requesting to establish a connection with the apparatus for the purpose of exchanging an ML associated task and / or an ML model. When apparatus accepts the connection request, the apparatus may send a RRC setup message to the UE to indicate to the UE that the connection setup request has been accepted.

[0297] In some examples, when the another state is the inactive state, the connection release message is a radio resource control release with a suspend configuration message, the connection request message is a radio resource control resume request message, a message indicating that the connection request has been accepted by the apparatus is a radio resource control resume message and the cause value is a resume cause value.

[0298] Stated differently, when the connection release message indicates to the UE to switch to an inactive state, the connection release message may be an RRC release message with a suspend configuration message. The UE may subsequently transmit a connection request message (as described in S1601 ), to the apparatus, wherein the connection request is a RRC resume request message. The connection resume request may comprise a resume cause value. For example, the resume cause value may indicate to the apparatus that the UE is requesting to establish or resume a connection with the apparatus for the purpose of exchanging an ML associated task and / or an ML model. When the apparatus accepts the connection request, the apparatus may send a RRC resume message to the UE to indicate to the UE that the connection resume request has been accepted.

[0299] Prior to S1601 , the apparatus may determine that signaling for a ML associated task and / or a ML model will be performed between the apparatus and the UE, wherein transmitting the connection release message is performed based on said determination that signaling for a machine learning associated task and / or a machine learning model will be performed.

[0300] In some examples, the apparatus determines that signaling for a machine learning associated task and / or a machine learning model will be performed based on receiving, from the user equipment, information related to an upcoming machine learning associated task and / or machine learning model.

[0301] Stated differently, the apparatus may receive information related to signaling to be performed for a ML associated task and / or ML model from the user equipment.

[0302] In some examples, the apparatus determines that signaling for a machine learning associated task and / or a machine learning model will be performed autonomously and transmits, to the user equipment, information related to an upcoming machine learning associated task and / or machine learning model.

[0303] Stated differently, the apparatus may autonomously determine that signaling for an ML associated task and / or an ML model is to be performed. Autonomously determining may comprise the apparatus determining that signaling for an ML associated task and / or an ML model is to be performed without receiving triggering signaling from the UE. The apparatus may subsequently transmit information related to signaling to be performed for a ML associated task and / or ML model to the UE. This may be as described above in connection with Figure 13 and 14.

[0304] In some examples, the apparatus determines that signaling for a machine learning associated task and / or a machine learning model will be performed based on receiving, from a core network, information related to an upcoming machine learning associated task and / or machine learning model. This may be as described above in connection with Figure 13 and 14.

[0305] Stated differently, when the apparatus is a network access node, the apparatus may receive information related to signaling to be performed for a ML associated task and / or ML model from a core network node.

[0306] In some examples, the apparatus may receive a release command from the core network, the release command triggering the apparatus to provide the connection release message to the user equipment. This may be as described above in connection with Figure 9.

[0307] In some examples, the connection release message that is transmitted by the apparatus to the UE may comprise the information related to the upcoming machine learning associated task and / or machine learning model. For example, the connection release message may comprise information related to signaling to be performed for a ML associated task and / or ML model from a core network node.

[0308] The method may further comprise, as referenced at S1602, determining, based on the cause value, whether to accept the connection request by configuring a connection to the user equipment.

[0309] Stated differently, the apparatus determines whether to establish or resume a connection with the UE, based on the received cause value.

[0310] For example, when the apparatus has a load that is below a configured threshold value, the apparatus may accept the connection request received from the UE. For example, when an overload condition is not triggered, the apparatus mayaccept the connection request received from the UE. The overload condition may be as described above in connection with Figure 12.

[0311] In other examples, when the apparatus has a load that is at or above a configured threshold value, the apparatus may reject the connection request received from the UE. For example, when an overload condition is triggered, the apparatus may reject the connection request received from the UE. The overload condition may be as described above in connection with Figure 12.

[0312] In some examples, the apparatus transmits, to the user equipment, a message indicating whether the connection request has been accepted or rejected by the apparatus, based on the determination of the apparatus.

[0313] For example, when the apparatus determines to accept the connection request, the apparatus may transmit an indication to the UE, indicating that the apparatus has determined to accept the connection request.

[0314] For example, when the apparatus determines to reject the connection request, the apparatus may transmit an indication to the UE, indicating that the apparatus has determined to reject the connection request.

[0315] In some examples, when the apparatus determines to accept the connection request and after a connection is established with the user equipment, the apparatus exchanges information to be used when executing a machine learning associated task and / or a machine learning model with the user equipment over the established connection.

[0316] Stated differently, when the apparatus determines to establish a connection with the UE and indicates to the UE of said determination, the UE may establish a connection with the apparatus. This established connection may be used for exchanging information to be used when executing a machine learning associated task and / or a machine learning model.

[0317] It is to be understood that establishing a connection between the apparatus and UE may comprise establishing a new connection or resuming an existing connection.

[0318] In some examples, the apparatus may determine that a load is above or at a first threshold value and the apparatus may prioritize an RRM related connection over an established connection corresponding to the connection request related to a machine learning task and / or machine learning model.

[0319] In some examples, the apparatus determines that a load at the apparatus is above or at a first threshold value, and transmits to the user equipment, a command to release the established connection with the apparatus.

[0320] In other words, when the apparatus accepts the connection request of S1601 , the UE may establish a connection with the apparatus. When a load at the apparatus becomes above or equal to a configured threshold, and the apparatus receives a connection request for an RRM related task, the apparatus may indicate to the UE to release the established connection for signalling information related to a ML associated task and / or ML model. The apparatus may subsequently accept the connection request for the RRM related task.

[0321] In some examples, when the apparatus determines to reject the connection request, the apparatus transmits a backoff timer to the user equipment.

[0322] For example, the backoff timer may indicate how long the apparatus will have a load at or above the configured threshold value. In some examples, after the time indicated by the backoff timer has expired, the apparatus may receive a connection request message from the UE, wherein the connection request message comprises a cause value as indicated in S1601.

[0323] In some examples, the apparatus pre-empts machine learning related connections when it is determined that a load at the apparatus is above or at a first threshold value.

[0324] The apparatus may be one of: a network access node; a core network node; or an operations, administration, and management function. Reference is made to Figure17 which shows a method of some examples.

[0325] This method of Figure 17 may be performed by an apparatus. The apparatus may comprise or be a user equipment, such as described above in connection with Figure 3.

[0326] Alternatively, or additionally, the apparatus may be such as discussed in relation to Figure 19.

[0327] The method may comprise, as referenced at S1701 receiving, from a first network node, information related to signaling to be performed for a machine learning associated task and / or a machine learning model.

[0328] In some examples, the information related to signaling to be performed for a machine learning associated task and / or a machine learning model comprisesinformation of an upcoming machine learning associated task and / or machine learning model.

[0329] The method may further comprise, at reference at S1702, transmitting, to a second network node, a connection request, the connection request comprising a cause value, the cause value indicating that the connection request is related to a machine learning associated task and / or a machine learning model.

[0330] It may be understood that the connection request of S1701 relates to the connection request described for Figures 4, 10, 11 A, 11 B and 12-14.

[0331] The cause value of S1701 may comprise properties of a service for which a connection is being requested by the apparatus.

[0332] Stated differently the apparatus may request to establish or resume a connection with a second network node. The request to establish or resume the connection may comprise a cause value, wherein the cause value indicates that the request for establishing or resuming a connection with the second network node is for the purpose of exchanging an ML associated task and / or an ML model.

[0333] The cause value may be a binary indication. Stated differently, the cause value may indicate whether the connection request is related to an ML associated task and / or ML model using a binary indication. This may be as described above in connection with Figure 11A and 13-14.

[0334] The cause value may be an index value that maps a machine learning associated task and / or a machine learning model to a priority level. Stated differently, the apparatus may map the ML associated task and / or ML model to a priority level. For example, the apparatus may map a ML associated task ID and / or ML model ID to a priority level. The apparatus may subsequently map the priority level to a cause value. For example the apparatus may map a priority level (associated with a ML associated task and / or ML model) to a cause value. This may be as described above in connection with Figures 11 B, 12-14.

[0335] The ML associated task may comprise at least one of the ML associated tasks as described in relation to Table 1 . For example, the ML associated task may comprise at least one of: transfer of a machine learning model; transfer of one or more conditions of a machine learning model; transfer of machine learning model validation data; transfer of training data related to a machine learning model; or transfer of a machine learning data set.

[0336] The apparatus may receive, from the first network node, a connection release message, prior to the method step at S1702.

[0337] The connection release message may comprise the information related to the upcoming machine learning associated task and / or machine learning model.

[0338] The connection release message may be as described above in connection with Figures 5, 7 and 13-14. For example, the connection release message may indicate to the apparatus to release a connection with the first network node or to switch from a connected state to another state. In this way, based on the connection release message, the apparatus may release a connection with the first network node or may switch from a connected state to another state.

[0339] In some examples, the another state is an idle state or an inactive state.

[0340] Stated differently, the connection release message received by the apparatus from the first network node, may indicate to the apparatus to switch from a connected to state to either an idle state or an inactive state. For example, the connection release message may indicate to the apparatus to release its connection with the first network node and enter into either an RRC idle state or an RRC inactive state.

[0341] In some examples, when the another state is the idle state, the connection release message is a radio resource control release message, the connection request message is a radio resource control setup request message, a message indicating that the connection request has been accepted by the second network node is a radio resource control setup message and the cause value is an establishment cause value.

[0342] Stated differently, when the connection release message indicates to the apparatus to switch to an idle state, the connection release message may be an RRC release message. The apparatus may subsequently transmit a connection request message (as described in S1702), to a second network node, wherein the connection request is a RRC setup request message. The connection setup request may comprise an establishment cause value. For example, the establishment cause value may indicate to the second network node that the apparatus is requesting to establish a connection with the second network node for the purpose of exchanging an ML associated task and / or an ML model. When the second network node accepts the connection request, the second network node may send a RRC setup message to the apparatus to indicate to the apparatus that the connection setup request has been accepted.

[0343] In some examples, when the another state is the inactive state, the connection release message is a radio resource control release with a suspend configuration message, the connection request message is a radio resource control resume request message, a message indicating that the connection request has been accepted by the second network node is a radio resource control resume message and the cause value is a resume cause value.

[0344] Stated differently, when the connection release message indicates to the apparatus to switch to an inactive state, the connection release message may be an RRC release message with a suspend configuration message. The apparatus may subsequently transmit a connection request message (as described in S1702), to a second network node, wherein the connection request is a RRC resume request message. The connection resume request may comprise a resume cause value. For example, the resume cause value may indicate to the second network node that the apparatus is requesting to establish or resume a connection with the second network node for the purpose of exchanging an ML associated task and / or an ML model. When the second network node accepts the connection request, the second network node may send a RRC resume message to the apparatus to indicate to the apparatus that the connection resume request has been accepted.

[0345] In some examples, after the apparatus transmits the connection request of S1702 to the second network node, the apparatus may receive, from the second network node, a message indicating whether the connection request has been accepted or rejected by the second network node. This may be as described above in connection with Figure 13 and 14.

[0346] Stated differently, when the second network node has a load that is below a configured threshold value, the second network node may accept the connection request received from the apparatus. For example, when an overload condition is not triggered, the second network node may accept the connection request received from the apparatus. The overload condition may be as described above in connection with Figure 12.

[0347] In other examples, when the second network node has a load that is at or above a configured threshold value, the second network node may reject the connection request received from the apparatus. For example, when an overload condition is triggered, the second network node may reject the connection request received fromthe apparatus. The overload condition may be as described above in connection with Figure 12.

[0348] In some examples, when the message received from the second network node indicates that the connection request has been accepted, the apparatus may establish a connection with the second network node and may exchange information to be used when executing a machine learning associated task and / or a machine learning model with the second network node over the established connection.

[0349] It is to be understood that establishing a connection with the second network node may comprise establishing a new connection or resuming an existing connection.

[0350] In some examples, when a load at the second network node is above or at a first threshold value, the apparatus may prioritize an RRM related connection over an established connection corresponding to the connection request related to a machine learning task and / or machine learning model.

[0351] In other words, when the second network node accepts the connection request of S1702, the apparatus may establish a connection with the second network node. When a load at the second network node becomes above or equal to a configured threshold, and the second network node receives a connection request for an RRM related task, the apparatus may receive an indication from the second network node to release the established connection for signalling information related to a ML associated task and / or ML model. The second network node may subsequently accept the connection request for the RRM related task.

[0352] In some examples, when the message received from the second network node indicates that the connection request has been rejected, the apparatus may initiate a backoff timer, and may transmit, to the second network node, a second connection request comprising the cause value when the backoff timer expires.

[0353] Stated differently, the apparatus may be configured with a backoff timer. For example, the apparatus may receive a backoff timer from the second network node. For example, the backoff timer may indicate how long the second network node will have a load at or above the configured threshold value. In some examples, after the time indicated by the backoff timer has expired, the apparatus may transmit a connection request message to the second network node, wherein the connection request message comprises a cause value as indicated in S1702.

[0354] In some examples, when the message received from the second network node indicates that the connection request has been rejected, the apparatus may transmit a second connection request comprising the cause value to another network node or cell.

[0355] Stated differently, when the second network node rejects the connection request received from the apparatus, the apparatus may transmit a connection request to a different network node or cell, wherein the connection request comprises the cause value as described in relation to S1702.

[0356] The first network node may be the second network node. For example, the first network node of S1701 may be the same entity as the second network node of S1702. In other examples, the first network node may be different to the second network node. For example, the first network node of S1701 is not the same entity as the second network node of S1702.

[0357] The first network node and / or second network node may be one of: a network access node; a core network node; or an operations, administration, and management function.

[0358] Reference is made to Figure 18 which shows a method of some examples.

[0359] This method of Figure 18 may be performed by an apparatus. The apparatus may comprise or be a network node, such as described above in connection with Figure 3.

[0360] Alternatively, or additionally, the apparatus may be such as discussed in relation to Figure 19.

[0361]

[0362] The method may comprise, as referenced at S1801 , transmitting, to a user equipment, information related to signaling to be performed for a machine learning task and / or a machine learning model.

[0363] In some examples, the information related to signaling to be performed for a machine learning associated task and / or a machine learning model comprises information of an upcoming machine learning associated task and / or machine learning model.

[0364] The apparatus may determine the information related to the signaling to be performed for the machine learning associated task and / or machine learning model autonomously.

[0365] Stated differently, the apparatus may autonomously determine that signaling for an ML associated task and / or an ML model is to be performed. Autonomously determining may comprise the apparatus determining that signaling for an ML associated task and / or an ML model is to be performed without receiving triggering signaling from the UE. The apparatus may subsequently transmit information related to signaling to be performed for a ML associated task and / or ML model to the UE. This may be as described above in connection with Figure 13 and 14.

[0366] In some examples, the apparatus determines that signaling for a machine learning associated task and / or a machine learning model will be performed based on receiving, from a core network, information related to an upcoming machine learning associated task and / or machine learning model. This may be as described above in connection with Figure 13 and 14.

[0367] Stated differently, when the apparatus is a network access node, the apparatus may receive information related to signaling to be performed for a ML associated task and / or ML model from a core network node. The method may further comprise, as referenced at S1802, receiving, from the user equipment, a connection request, the connection request comprising a cause value, the cause value indicating that the connection request is related to a machine learning associated task and / or a machine learning model.

[0368] It may be understood that the connection request of S1802 relates to the connection request described for Figures 4, 10, 11 A, 11 B and 12-14.

[0369] The cause value of S1802 may comprise properties of a service for which a connection is being requested by the user equipment.

[0370] Stated differently the user equipment may request to establish or resume a connection with the apparatus. The request to establish or resume the connection may comprise a cause value, wherein the cause value indicates that the request for establishing or resuming a connection with the apparatus is for the purpose of exchanging an ML associated task and / or an ML model.

[0371] The cause value may be a binary indication. Stated differently, the cause value may indicate whether the connection request is related to an ML associated task and / or ML model using a binary indication. This may be as described above in connection with Figure 11A and 13-14.

[0372] The cause value may be an index value that maps a machine learning associated task and / or a machine learning model to a priority level. Stated differently, the UE may map the ML associated task and / or ML model to a priority level. For example, the UE may map a ML associated task ID and / or ML model ID to a priority level. The UE may subsequently map the priority level to a cause value. For example the UE may map a priority level (associated with a ML associated task and / or ML model) to a cause value. This may be as described above in connection with Figures 11 B, 12-14.

[0373] The ML associated task may comprise at least one of the ML associated tasks as described in relation to Table 1 . For example, the ML associated task may comprise at least one of: transfer of a machine learning model; transfer of one or more conditions of a machine learning model; transfer of machine learning model validation data; transfer of training data related to a machine learning model; or transfer of a machine learning data set.

[0374] The apparatus may transmit to the UE a connection release message, prior to receiving the connection request of S1802. The connection release message may comprise the information related to signalling to be performed for the ML associated task and / or ML model (as indicated in S1801 ). In some examples, the apparatus may receive a release command from the core network, the release command triggering the apparatus to provide the connection release message to the user equipment. This may be as described above in connection with Figure 9.

[0375] The connection release message may be as described above in connection with Figures 5, 7 and 13-14. For example, the connection release message may indicate to the UE to release a connection with the second network node or to switch from a connected state to another state. In this way, based on the connection release message, the UE may release a connection with the apparatus or may switch from a connected state to another state.

[0376] In some examples, the another state is an idle state or an inactive state.

[0377] Stated differently, the connection release message received by the UE from the apparatus, may indicate to the UE to switch from a connected to state to either an idle state or an inactive state. For example, the connection release message may indicate to the UE to release its connection with the apparatus and enter into either an RRC idle state or an RRC inactive state.

[0378] In some examples, when the another state is the idle state, the connection release message is a radio resource control release message, the connection request message is a radio resource control setup request message, a message indicating that the connection request has been accepted by the apparatus is a radio resource control setup message and the cause value is an establishment cause value.

[0379] Stated differently, when the connection release message indicates to the UE to switch to an idle state, the connection release message may be an RRC release message. The UE may subsequently transmit a connection request message (as described in S1802), to the apparatus, wherein the connection request is a RRC setup request message. The connection setup request may comprise an establishment cause value. For example, the establishment cause value may indicate to the apparatus that the UE is requesting to establish a connection with the apparatus for the purpose of exchanging an ML associated task and / or an ML model. When apparatus accepts the connection request, the apparatus may send a RRC setup message to the UE to indicate to the UE that the connection setup request has been accepted.

[0380] In some examples, when the another state is the inactive state, the connection release message is a radio resource control release with a suspend configuration message, the connection request message is a radio resource control resume request message, a message indicating that the connection request has been accepted by the apparatus is a radio resource control resume message and the cause value is a resume cause value.

[0381] Stated differently, when the connection release message indicates to the UE to switch to an inactive state, the connection release message may be an RRC release message with a suspend configuration message. The UE may subsequently transmit a connection request message (as described in S1802), to the apparatus, wherein the connection request is a RRC resume request message. The connection resume request may comprise a resume cause value. For example, the resume cause value may indicate to the apparatus that the UE is requesting to establish or resume a connection with the apparatus for the purpose of exchanging an ML associated task and / or an ML model. When the apparatus accepts the connection request, the apparatus may send a RRC resume message to the UE to indicate to the UE that the connection resume request has been accepted.

[0382] The method may further comprise, as referenced at S1803, determining, based on the machine learning related cause value, whether to accept the connection request by configuring a connection to the user equipment.

[0383] Stated differently, the apparatus determines whether to establish or resume a connection with the UE, based on the received cause value.

[0384] For example, when the apparatus has a load that is below a configured threshold value, the apparatus may accept the connection request received from the UE. For example, when an overload condition is not triggered, the apparatus may accept the connection request received from the UE. The overload condition may be as described above in connection with Figure 12.

[0385] In other examples, when the apparatus has a load that is at or above a configured threshold value, the apparatus may reject the connection request received from the UE. For example, when an overload condition is triggered, the apparatus may reject the connection request received from the UE. The overload condition may be as described above in connection with Figure 12.

[0386] In some examples, the apparatus transmits, to the user equipment, a message indicating whether the connection request has been accepted or rejected by the apparatus, based on the determination of the apparatus.

[0387] For example, when the apparatus determines to accept the connection request, the apparatus may transmit an indication to the UE, indicating that the apparatus has determined to accept the connection request.

[0388] For example, when the apparatus determines to reject the connection request, the apparatus may transmit an indication to the UE, indicating that the apparatus has determined to reject the connection request.

[0389] In some examples, when the apparatus determines to accept the connection request and after a connection is established with the user equipment, the apparatus exchanges information to be used when executing a machine learning associated task and / or a machine learning model with the user equipment over the established connection.

[0390] Stated differently, when the apparatus determines to establish a connection with the UE and indicates to the UE of said determination, the UE may establish a connection with the apparatus. This established connection may be used forexchanging information to be used when executing a machine learning associated task and / or a machine learning model.

[0391] It is to be understood that establishing a connection between the apparatus and UE may comprise establishing a new connection or resuming an existing connection.

[0392] In some examples, the apparatus may determine that a load is above or at a first threshold value and the apparatus may prioritize an RRM related connection over an established connection corresponding to the connection request related to a machine learning task and / or machine learning model.

[0393] In some examples, the apparatus determines that a load at the apparatus is above or at a first threshold value, and transmits to the user equipment, a command to release the established connection with the apparatus.

[0394] In other words, when the apparatus accepts the connection request of S1802, the UE may establish a connection with the apparatus. When a load at the apparatus becomes above or equal to a configured threshold, and the apparatus receives a connection request for an RRM related task, the apparatus may indicate to the UE to release the established connection for signalling information related to a ML associated task and / or ML model. The apparatus may subsequently accept the connection request for the RRM related task.

[0395] In some examples, when the apparatus determines to reject the connection request, the apparatus transmits a backoff timer to the user equipment.

[0396] For example, the backoff timer may indicate how long the apparatus will have a load at or above the configured threshold value. In some examples, after the time indicated by the backoff timer has expired, the apparatus may receive a connection request message from the UE, wherein the connection request message comprises a cause value as indicated in S1802.

[0397] In some examples, the apparatus pre-empts machine learning related connections when it is determined that a load at the apparatus is above or at a first threshold value.

[0398] The apparatus may be one of: a network access node; a core network node; or an operations, administration, and management function.

[0399] Figure 19 shows, by way of example, a block diagram of an apparatus 10. The apparatus 10 comprises, for example, at least one processor 12 and at least one memory 14 storing instructions 15 that, when executed by the at least one processor,cause the apparatus 10 at least to perform the method or methods as disclosed herein, and any of the examples thereof. In an example, the at least one memory and the instructions (e.g. a computer program code, software), are configured, with the at least one processor, to cause the apparatus 10 to perform the method or methods as disclosed herein, and any of the examples thereof.

[0400] A processor 12 may comprise circuitry, or be constituted as circuitry or circuitries, the circuitry or circuitries being configured to perform phases of methods in accordance with examples described herein. As used in this application, the term “circuitry” may refer to one or more or all of the following: (a) hardware-only circuit implementations, such as implementations in only analog and / or digital circuitry, and (b) combinations of hardware circuits and software, such as, as applicable: (i) a combination of analog and / or digital hardware circuit(s) with software / firmware and (ii) any portions of hardware processor(s) with software (including digital signal processor(s)), software, and memory(ies) that work together to cause an apparatus, such as a user equipment, to perform various functions) and (c) hardware circuit(s) and or processor(s), such as a microprocessor(s) or a portion of a microprocessor(s), that requires software (e.g., firmware) for operation, but the software may not be present when it is not needed for operation. This definition of circuitry applies to all uses of this term in this application, including in any claims. As a further example, as used in this application, the term circuitry also covers an implementation of merely a hardware circuit or processor (or multiple processors) or portion of a hardware circuit or processor and its (or their) accompanying software and / or firmware. The term circuitry also covers, for example and if applicable to the particular claim element, a baseband integrated circuit or processor integrated circuit for a mobile device or a similar integrated circuit in server, a cellular network device, or other computing or network device.

[0401] The memory 14 may be implemented using any suitable data storage technology. The memory may comprise a database for storing data. The memory 14 may be at least in part external to apparatus 10 but accessible to apparatus 10.

[0402] The instructions 15 may be comprised in a computer readable medium or a non-transitory computer readable medium. A term non-transitory, as used herein, is a limitation of the medium itself (i.e. tangible, not a signal) as opposed to a limitation on data storage persistency (e.g. random access memory, RAM, vs. read onlymemory, ROM).

[0403] For example, the apparatus 10 is a terminal device, such as the UE of any of the Figures. As another example, the apparatus is comprised in such a terminal device, e.g. as a chipset configured to control the terminal device. The apparatus 10 may be caused or configured to perform at least the method of Figures 15 and / or 17 and / or any one or more of the examples described.

[0404] As another example, the apparatus 10 is a network node, e.g. the network node of any of the Figures. In another example, the apparatus is comprised in such a network node, e.g. as a chipset configured to control the network node. The apparatus 10 may be caused or configured to perform at least the method of Figures 16 and / or 18 and / or any one or more of the examples described.

[0405] The apparatus may comprise one or more entities of any of protocol layers, such as a MAC entity, an RRC entity, an RLC entity, a PDCP entity or a PHY entity. In some examples, the entity is configured to perform at least the method of Figures 15-18, and / or any one or more of the examples described.

[0406] The apparatus 10 comprises a radio interface 16. The radio interface 16 may provide the apparatus 10 with communication capabilities. The radio interface 16 may comprise a receiver configured to receive information in accordance with at least one cellular or non-cellular standard. The radio interface 16 may comprise a transmitter configured to transmit information in accordance with at least one cellular or non-cellular standard. The receiver may comprise more than one receiver. The transmitter may comprise more than one transmitter. The radio interface 16 may comprise a transceiver configured to receive and transmit information in accordance with at least one cellular or non-cellular standard. The transceiver may comprise more than one transceiver.

[0407] The apparatus 10 may comprise a user interface 18 comprising, for example, at least one of a keypad, a microphone, a touch display, a display, a speaker, etc. The user interface 18 may be used to control the apparatus by the user. The user interface 18 may be external to the apparatus 10. For example, the apparatus 10 may be connected to another device, such as a computer, either via wireless or wired connection, and the apparatus 10 is controlled by the user via the computer.

[0408] In an example, at least some of the processes described herein may be carried out by an apparatus comprising means for carrying out at least some of thedescribed processes. Means for performing method steps as disclosed herein may comprise software and / or hardware components of the apparatus 10. For example, the at least one processor 12, the memory 14, and the computer program code form means for carrying out the method or methods as disclosed herein, and any of the examples thereof. As used herein the term “means” is to be construed in singular form, i.e. referring to a single element, or in plural form, i.e. referring to a combination of single elements. Therefore, terminology “means for [performing A, B, C]”, is to be interpreted to cover an apparatus in which there is only one means for performing A, B and C, or where there are separate means for performing A, B and C, or partially or fully overlapping means for performing A, B, C. Further, terminology “means for performing A, means for performing B, means for performing C” is to be interpreted to cover an apparatus in which there is only one means for performing A, B and C, or where there are separate means for performing A, B and C, or partially or fully overlapping means for performing A, B, C.

Claims

1. CLAIMS1 . An apparatus comprising means for: transmitting, to a first network node, a connection request, the connection request comprising a cause value, the cause value indicating that the connection request is related to a machine learning associated task and / or a machine learning model.

2. The apparatus of claim 1 , wherein the cause value is a binary indication or an index value that maps a machine learning associated task and / or a machine learning model to a priority level.

3. The apparatus of claim 1 or claim 2, wherein the machine learning associated task comprises at least one of: transfer of a machine learning model; transfer of one or more conditions of a machine learning model; transfer of machine learning model validation data; transfer of training data related to a machine learning model; or transfer of a machine learning data set.

4. The apparatus of any of claims 1 to 3, further comprising means for receiving, from the first network node, a message indicating whether the connection request has been accepted or rejected by the first network node.

5. The apparatus of claim 4, wherein the apparatus further comprises means for, when the message received from the first network node indicates that the connection request has been accepted, establishing a connection with the first network node and exchanging information to be used when executing a machine learning associated task and / or a machine learning model with the first network node over the established connection.

6. The apparatus of claim 5, further comprising means for, when a load at the first network node is above or at a first threshold value, prioritizing a radio resource management related connection over the established connection corresponding to64the connection request related to a machine learning task and / or machine learning model.

7. The apparatus of claim 4, wherein the apparatus further comprises means for, when the message received from the first network node indicates that the connection request has been rejected, initiating a backoff timer, and, transmitting, to the first network node, a second connection request comprising the cause value when the backoff timer expires.

8. The apparatus of claim 4, wherein the apparatus further comprises means for, when the message received from the first network node indicates that the connection request has been rejected, transmitting a second connection request comprising the cause value to another network node or cell.

9. The apparatus of any of claims 1 to 8, further comprising means for receiving, from a second network node, a connection release message prior to transmitting the connection request.

10. The apparatus of claim 9, based on the connection release message, further comprising means for releasing a connection with the second network node or switching from a connected state to another state.11 . The apparatus of claim 10, wherein the another state is an idle state or an inactive state.

12. The apparatus of claim 11 , wherein the another state is the idle state, the connection release message is a radio resource control release message, the connection request message is a radio resource control setup request message, a message indicating that the connection request has been accepted by the first network node is a radio resource control setup message and the cause value is an establishment cause value.

13. The apparatus of claim 11 , wherein the another state is the inactive state, the connection release message is a radio resource control release with a suspend configuration message, the connection request message is a radio resource control resume request message, a message indicating that the connection request has been accepted by the first network node is a radio resource control resume message and the cause value is a resume cause value.

14. The apparatus of any of claims 9 to 13, further comprising means for determining that signaling for a machine learning associated task and / or a machine learning model will be performed between the apparatus and a network node, wherein transmitting the connection request is performed based on said determination that signaling for a machine learning associated task and / or a machine learning model will be performed.

15. The apparatus of claim 14, wherein the means for determining, that signaling for a machine learning associated task and / or a machine learning model will be performed, is performed autonomously by the apparatus, and wherein the apparatus comprises means for transmitting, to the second network node, information related to an upcoming machine learning associated task and / or machine learning model.

16. The apparatus of claim 14, wherein the means for determining that signaling for a machine learning associated task and / or a machine learning model will be performed is performed based on receiving, from the second network node, information related to an upcoming machine learning associated task and / or machine learning model.

17. The apparatus of claim 16, wherein the connection release message received from the second network node comprises the information related to the upcoming machine learning associated task and / or machine learning model.

18. The apparatus of any of claims 1 to 17, wherein the first network node is the second network node, or wherein the first network node is different to the second network node.

19. The apparatus of any of claims 1 to 18, wherein the first network node and / or second network node is one of: a network access node; a core network node; or an operations, administration and management function.

20. An apparatus comprising means for: receiving, from a user equipment, a connection request, the connection request comprising a cause value, the cause value indicating that the connection request is related to a machine learning associated task and / or a machine learning model; and determining, based on the cause value, whether to accept the connection request by configuring a connection to the user equipment.21 . The apparatus of claim 20, wherein the cause value is a binary indication or an index value that maps a machine learning associated task and / or a machine learning model to a priority level.

22. The apparatus of claim 20 or claim 21 , wherein the machine learning associated task comprises at least one of: transfer of a machine learning model; transfer of one or more conditions of a machine learning model; transfer of machine learning model validation data; transfer of training data related to a machine learning model; or transfer of a machine learning data set.

23. The apparatus of any of claims 20 to 22, further comprising means for transmitting, to the user equipment, a message indicating whether the connection request has been accepted or rejected by the apparatus, based on the determination of the apparatus.

24. The apparatus of any of claims 20 to 23, wherein the apparatus further comprises means for, when determining to accept the connection request and after a connection is established with the user equipment, exchanging information to be used when executing a machine learning associated task and / or a machine learning model with the user equipment over the established connection.

25. The apparatus of claim 24, further comprising means for determining that a load at the apparatus is above or at a first threshold value, and prioritizing a radio resource management related connection over the established connection corresponding to the connection request related to a machine learning associated task and / or a machine learning model.

26. The apparatus of claim 24 or claim 25, further comprising means for determining that a load at the apparatus is above or at a first threshold value, and transmitting, to the user equipment, a command to release the established connection with the apparatus.

27. The apparatus of any of claims 20 to 23, wherein the apparatus further comprises means for, when said determining determines to reject the connection request, transmitting a backoff timer to the user equipment.

28. The apparatus of any of claims 20 to 27, further comprising means for transmitting a connection release message to the user equipment, prior to receiving the connection request.

29. The apparatus of claim 28, wherein the connection release message comprises an indication to release the connection with the apparatus, or to switch from a connected state to another state.

30. The apparatus of claim 29, wherein the another state is an idle state or an inactive state.31 . The apparatus of claim 30, wherein the another state is the idle state, the connection release message is a radio resource control release message, the connection request message is a radio resource control setup request message, a message indicating that the connection request has been accepted by the apparatus is a radio resource control setup message and the cause value is an establishment cause value.

32. The apparatus of claim 31 , wherein the another state is the inactive state, the connection release message is a radio resource control release with a suspend configuration message, the connection request message is a radio resource control resume request message, a message indicating that the connection request has been accepted by the apparatus is a radio resource control resume message and the cause value is a resume cause value.

33. The apparatus of any of claims 28 to 32, further comprising means for determining that signaling for a machine learning task and / or a machine learning model will be performed between the apparatus and the user equipment, wherein transmitting the connection release message is performed based on said determination that signaling for a machine learning associated task and / or a machine learning model will be performed.

34. The apparatus of claim 33, wherein the means for determining that signaling for a machine learning associated task and / or a machine learning model will be performed, is performed based on receiving, from the user equipment, information related to an upcoming machine learning associated task and / or machine learning model.

35. The apparatus of claim 33, wherein the means for determining that signaling for a machine learning associated task and / or a machine learning model will be performed, is performed autonomously by the apparatus, and wherein the apparatus comprises means for transmitting, to the user equipment, information related to an upcoming machine learning associated task and / or machine learning model.

36. The apparatus of claim 33, wherein the means for determining that signaling for a machine learning associated task and / or a machine learning model will be performed, is performed based on receiving, from a core network, information related to an upcoming machine learning associated task and / or machine learning model.

37. The apparatus of claim 36, further comprising means for receiving a release command from the core network, the release command triggering the apparatus to provide the connection release message to the user equipment.

38. The apparatus of any of claims 35 to 37, wherein the connection release message comprises the information related to the upcoming machine learning associated task and / or machine learning model.

39. The apparatus of any of claims 20 to 38, further comprising means for preempting machine learning related connections when it is determined that a load at the apparatus is above or at a first threshold value.

40. The apparatus of any of claims 20 to 39, wherein the apparatus is one of: a network access node; a core network node; or an operations, administration and management function.41 . A method comprising: transmitting, to a first network node, a connection request, the connection request comprising a cause value, the cause value indicating that the connection request is related to a machine learning associated task and / or a machine learning model.

42. A method comprising: receiving, from a user equipment, a connection request, the connection request comprising a cause value, the cause value indicating that the connection request is related to a machine learning associated task and / or a machine learning model; and determining, based on the cause value, whether to accept the connection request by configuring a connection to the user equipment.

43. A computer program comprising instructions which, when the program is executed by a computer, cause an apparatus to carry out:transmitting, to a first network node, a connection request, the connection request comprising a cause value, the cause value indicating that the connection request is related to a machine learning associated task and / or a machine learning model.

44. A computer program comprising instructions which, when the program is executed by a computer, cause an apparatus to carry out: receiving, from a user equipment, a connection request, the connection request comprising a cause value, the cause value indicating that the connection request is related to a machine learning associated task and / or a machine learning model; and determining, based on the cause value, whether to accept the connection request by configuring a connection to the user equipment.

Images