Identification of a machine learning model in a cellular communication network

Abstract

During an radio resource control (RRC) connected mode during a first RRC connection period, an apparatus determines 1110 an identifier of at least one configuration related to a machine learning (ML) model used or to be used by the apparatus in the connected mode. After transitioning 1120 to a an RRC idle or RRC inactive mode, and whilst transitioning or attempting to transition 1130 from that mode to the RRC connected mode for a second RRC connection period, information is transmitted to a wireless network node that relates to the identifier of the at least one configuration related to the ML model via RRC signalling. The identifier of the configuration may be stored as a transient model identifier. In some embodiments, information is received at the apparatus indicating whether it is to use the configuration, and the identifier is either determined to be valid or discarded.

Description

[002] Machine Learning, ML, models may be used in various cellular communication networks, such as, in cellular communication networks operating according to 5G radio access technology. 5G radio access technology may also be referred to as New Radio, NR, access technology. 3rd Generation Partnership Project, 3GPP, develops standards for 5G / NR and for future radio access technologies. In case of ML models, identification of the ML model is an important issue. Thus, there is a need to provide enhancements for identification of the ML model at least for 5G / NR and future radio access technologies, such as 6G. SUMMARY

[003] According to some aspects, there is provided the subject-matter of the independent claims. Some example embodiments are defined in the dependent claims.

[004] The scope of protection sought for various example embodiments of the disclosure is set out by the independent claims. The example embodiments and features, if any, described in this specification that do not fall under the scope of the independent claims are to be interpreted as examples useful for understanding various example embodiments of the disclosure.

[005] According to an aspect of the present disclosure, there is provided an apparatus comprising at least one processor and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to determine, when the apparatus is in a connected mode during a first connection period, an identifier of at least one configuration related to a machine learning model used or to be used by the apparatus in the connected mode, transition from the connected mode to a non-connected mode and transmit to a wireless network node, when the apparatus is transitioning or attempting to transition from the non-connected mode to the connected mode for a second connection period, information related to the identifier of the at least one configuration related to the machine learning model. The apparatus may be a user equipment or a control device configured to control the functioning thereof, when installed therein. Example embodiments of the aspect may comprise at least one feature from the following bulleted list or any combination of the following features: • wherein the machine learning model is associated with at least one of a measurement configuration or a data collection configuration; • wherein the at least one processor and the at least one memory further cause the apparatus at least to transmit an indication about whether the apparatus has the identifier of the at least one configuration related to the machine learning model; • wherein the at least one processor and the at least one memory further cause the apparatus at least to store the identifier of the at least one configuration related to the machine learning model when the apparatus is in the connected mode during the first connection period; • wherein the at least one processor and the at least one memory further cause the apparatus at least to store the identifier of the at least one configuration related to the machine learning model as a transient model identifier; • wherein the at least one processor and the at least one memory further cause the apparatus at least to determine that the apparatus is configured to store the identifier of the at least one configuration related to the machine learning model; • wherein the apparatus is in the connected mode first with a first wireless network node during the first connection period and transitioning or attempting to transition to the connected mode with a second wireless network node for the second connection period, wherein the first wireless network node is different than the second wireless network node; • wherein the at least one processor and the at least one memory further cause the apparatus at least to transmit an indication to the wireless network node, said indication indicating whether the apparatus prefers reusing the identifier of the at least one configuration related to the machine learning model or initializing identification of the machine learning model; • wherein the at least one processor and the at least one memory further cause the apparatus at least to transmit to the wireless network node information about at least one condition which has changed between the first and second connection periods; • wherein the at least one processor and the at least one memory further cause the apparatus at least to receive, in response to transmitting said information related to the identifier of the at least one configuration related to the machine learning model, information indicating whether the apparatus is to use the at least one configuration related to the machine learning model corresponding to the identifier during the second connection period; • wherein the at least one processor and the at least one memory further cause the apparatus at least to determine, when said information indicating whether the apparatus is to use the at least one configuration indicates that the apparatus is to use the at least one configuration related to the machine learning model corresponding to the identifier during the second connection period, whether the identifier of the at least one configuration related to the machine learning model is valid or discard, when said information indicating whether the apparatus is to use the at least one configuration indicates that the apparatus is not to use the at least one configuration related to the machine learning model corresponding to the identifier during the second connection period, the identifier of the at least one configuration related to the machine learning model; • wherein said information indicating whether the apparatus is to use the at least one configuration related to the machine learning model corresponding to the identifier during the second connection period is received in a connection resume message or a connection setup message; • wherein the at least one processor and the at least one memory further cause the apparatus at least to receive, in response to transmitting said information related to the identifier of the at least one configuration related to the machine learning model, a connection release message and discard the at least one configuration associated with the machine learning model corresponding to the identifier; • wherein the at least one processor and the at least one memory further cause the apparatus at least to transmit, to the wireless network node, an indication about whether the at least one configuration related to the machine learning model has been used by the apparatus during the first connection period.

[006] According to an aspect of the present disclosure, there is provided an apparatus comprising at least one processor and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to receive from a user equipment, when the user equipment is transitioning or attempting to transition from a non-connected mode to a connected mode for a second connection period, information related to an identifier of at least one configuration related to a machine learning model that was used or to be used by the user equipment during a first connection period before the second connection period and evaluate whether the at least one configuration related to the machine learning model is suitable to be used by at least one of the user equipment in the connected mode during the second connection period or by the apparatus while the user equipment is in the connected mode during the second connection period. The apparatus may be a wireless network node or a control device configured to control the functioning thereof, when installed therein. Example embodiments of the aspect may comprise at least one feature from the following bulleted list or any combination of the following features: • wherein the machine learning model is associated with at least one of a measurement or a data collection configuration; • wherein the at least one processor and the at least one memory further cause the apparatus at least to receive an indication about whether the user equipment has the identifier of the at least one configuration related to the machine learning model; • wherein the at least one processor and the at least one memory further cause the apparatus at least to store the identifier of the at least one configuration related to the machine learning model during the first connection period; • wherein the at least one processor and the at least one memory further cause the apparatus at least to receive an indication from the user equipment, said indication indicating whether the user equipment prefers re-using the at least one configuration related to the machine learning model or initializing identification of the machine learning model and evaluate, based at least partly on said indication, whether the at least one configuration related to the machine learning model is to be used by the user equipment during the second connection period; • wherein the at least one processor and the at least one memory further cause the apparatus at least to receive, from the user equipment, information about at least one condition which has changed between the first and second connection periods and evaluate, based at least partly on said information, whether the at least one configuration related to the machine learning model associated with the identifier is to be used by the user equipment during the second connection period; • wherein the at least one processor and the at least one memory further cause the apparatus at least to transmit to the user equipment, based on the evaluation, information indicating whether the user equipment is to use the at least one configuration related to the machine learning model corresponding to the identifier during the second connection period; • wherein said information indicating whether the user equipment is to use the at least one configuration related to the machine learning model corresponding to the identifier during the second connection period is transmitted in a connection resume message or a connection setup message; • wherein the at least one processor and the at least one memory further cause the apparatus at least to transmit to the user equipment, based on the evaluation, a connection release message; • wherein the at least one processor and the at least one memory further cause the apparatus at least to receive, from the user equipment, an indication about whether the at least one configuration related to the machine learning model has been used by the user equipment during the first connection period; • wherein the at least one processor and the at least one memory further cause the apparatus at least to discard the at least one configuration related to the machine learning model when the evaluation indicates that the at least one configuration related to the machine learning model is not suitable to be used by at least one of the user equipment or the apparatus during the second connection period.

[007] According to an aspect of the present disclosure, there is provided an apparatus comprising at least one processor and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to determine, when the apparatus is in a Radio Resource Control, RRC, connected mode during a first RRC connection period, an identifier of at least one configuration related to a machine learning model used or to be used by the apparatus in the RRC connected mode, transition from the RRC connected mode to an RRC inactive mode or an RRC idle mode and transmit to a wireless network node, when the apparatus is transitioning or attempting to transition from the RRC inactive mode or the RRC idle mode to the RRC connected mode for a second RRC connection period, the identifier of the at least one configuration related to the machine learning model via RRC signalling. The apparatus may be a user equipment or a control device configured to control the functioning thereof, when installed therein. Example embodiments of the aspect may comprise at least one feature from the following bulleted list or any combination of the following features: • wherein the machine learning model is associated with at least one of a measurement or a data collection configuration; • wherein the at least one processor and the at least one memory further cause the apparatus at least to store the identifier of the at least one configuration related to the machine learning model when the apparatus is in the RRC connected mode during the first RRC connection period; • wherein the at least one processor and the at least one memory further cause the apparatus at least to store the identifier of the at least one configuration related to the machine learning model as a transient model identifier; • wherein the at least one processor and the at least one memory further cause the apparatus at least to determine that the apparatus is configured to store the identifier of the at least one configuration related to the machine learning model; • wherein the apparatus is in the RRC connected mode first with a first wireless network node during the first RRC connection period and transitioning or attempting to transition to the RRC connected mode with a second wireless network node for the RRC second connection period, wherein the first wireless network node is different than the second wireless network node; • wherein the at least one processor and the at least one memory further cause the apparatus at least to transmit an indication to the wireless network node via RRC signalling, said indication indicating whether the apparatus prefers re-using the identifier of the at least one configuration related to the machine learning model or initializing identification of the machine learning model; • wherein the at least one processor and the at least one memory further cause the apparatus at least to transmit to the wireless network node via RRC signalling information about at least one condition which has changed between the first and second RRC connection periods; • wherein the at least one processor and the at least one memory further cause the apparatus at least to receive via RRC signalling, in response to transmitting the identifier of the at least one configuration related to the machine learning model, information indicating whether the apparatus is to use the at least one configuration related to the machine learning model corresponding to the identifier during the second RRC connection period; • wherein the at least one processor and the at least one memory further cause the apparatus at least to determine, when said information indicates that the apparatus is to use the at least one configuration related to the machine learning model corresponding to the identifier during the second RRC connection period, whether the identifier of the at least one configuration related to the machine learning model is valid or discard, when said information indicates that the apparatus is not to use the at least one configuration related to the machine learning model corresponding to the identifier during the second RRC connection period, the identifier of the at least one configuration related to the machine learning model; • wherein said information indicating whether the apparatus is to use the at least one configuration related to the machine learning model corresponding to the identifier during the second RRC connection period is received in an RRC resume message or an RRC setup message; • wherein the at least one processor and the at least one memory further cause the apparatus at least to receive, in response to transmitting the identifier of the at least one configuration related to the machine learning model, an RRC release message and discard the at least one configuration associated with the machine learning model corresponding to the identifier; • wherein the at least one processor and the at least one memory further cause the apparatus at least to transmit via RRC signalling, to the wireless network node, an indication about whether the at least one configuration related to the machine learning model has been used by the apparatus during the first RRC connection period.

[008] According to an aspect of the present disclosure, there is provided an apparatus comprising at least one processor and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to receive via RRC signalling from a user equipment, when the user equipment is transitioning or attempting to transition from an RRC inactive mode or an RRC idle mode to a RRC connected mode for a second RRC connection period, an identifier of at least one configuration related to a machine learning model that was used or to be used by the user equipment during a first RRC connection period before the second RRC connection period and evaluate whether the at least one configuration related to the machine learning model is suitable to be used by at least one of the user equipment in the RRC connected mode during the second RRC connection period or by the apparatus while the user equipment is in the RRC connected mode during the second RRC connection period. The apparatus may be a wireless network node or a control device configured to control the functioning thereof, when installed therein. Example embodiments of the aspect may comprise at least one feature from the following bulleted list or any combination of the following features: • wherein the machine learning model is associated with at least one of a measurement or a data collection configuration; • wherein the at least one processor and the at least one memory further cause the apparatus at least to store the identifier of the at least one configuration related to the machine learning model during the first RRC connection period; • wherein the at least one processor and the at least one memory further cause the apparatus at least to receive via RRC signalling an indication from the user equipment, said indication indicating whether the user equipment prefers re-using the at least one configuration related to the machine learning model or initializing identification of the machine learning model and evaluate, based at least partly on said indication, whether the at least one configuration related to the machine learning model is to be used by the user equipment during the second RRC connection period; • wherein the at least one processor and the at least one memory further cause the apparatus at least to receive via RRC signalling, from the user equipment, information about at least one condition which has changed between the first and second RRC connection periods and evaluate, based at least partly on said information, whether the at least one configuration related to the machine learning model associated with the identifier is to be used by the user equipment during the second RRC connection period; • wherein the at least one processor and the at least one memory further cause the apparatus at least to transmit to the user equipment via RRC signalling, based on the evaluation, information indicating whether the user equipment is to use the at least one configuration related to the machine learning model corresponding to the identifier during the second RRC connection period; • wherein said information indicating whether the user equipment is to use the at least one configuration related to the machine learning model corresponding to the identifier during the second RRC connection period is transmitted in an RRC resume message or an RRC setup message; • wherein the at least one processor and the at least one memory further cause the apparatus at least to transmit to the user equipment, based on the evaluation, an RRC release message; • wherein the at least one processor and the at least one memory further cause the apparatus at least to receive via RRC signalling, from the user equipment, an indication about whether the at least one configuration related to the machine learning model has been used by the user equipment during the first RRC connection period; • wherein the at least one processor and the at least one memory further cause the apparatus at least to discard the at least one configuration related to the machine learning model when the evaluation indicates that the at least one configuration related to the machine learning model is not suitable to be used by at least one of the user equipment or the apparatus during the second RRC connection period.

[009] According to an aspect of the present disclosure, there is provided an apparatus comprising at least one processor and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to receive from a wireless network node, when the apparatus is transitioning or attempting to transition from a non-connected mode to a connected mode for a second connection period, an identifier of at least one configuration related to a machine learning model that was used or to be used by the apparatus during a first connection period before the second connection period and determine whether to use the at least one configuration related to the machine learning model in the connected mode during the second connection period. The apparatus may be a user equipment or a control device configured to control the functioning thereof, when installed therein. Example embodiments of the aspect may comprise at least one feature from the following bulleted list or any combination of the following features: • wherein the machine learning model is associated with at least one of a measurement configuration or a data collection configuration; • wherein the connected mode is a radio resource control connected mode and the non-connected mode is a radio resource control inactive mode or a radio resource control idle mode; • wherein the at least one processor and the at least one memory further cause the apparatus at least to store the identifier of the at least one configuration related to the machine learning model when the apparatus is in the connected mode during the first connection period; • wherein the at least one processor and the at least one memory further cause the apparatus at least to store the identifier of the at least one configuration related to the machine learning model as a transient model identifier; • wherein the at least one processor and the at least one memory further cause the apparatus at least to determine that the apparatus is configured to store the identifier of the at least one configuration related to the machine learning model; • wherein the apparatus is in the connected mode first with a first wireless network node during the first connection period and transitioning or attempting to transition to the connected mode with a second wireless network node for the second connection period, wherein the first wireless network node is different than the second wireless network node; • wherein the at least one processor and the at least one memory further cause the apparatus at least to transmit an indication, said indication indicating whether the at least one configuration related to the machine learning model is valid to be used in the connected mode during the second connection period; • wherein the at least one processor and the at least one memory further cause the apparatus at least to transmit said indication indicating whether the at least one configuration related to the machine learning model is valid to be used in the connected mode during the second connection period in a connection resume complete message or a connection setup complete message; • wherein the at least one processor and the at least one memory further cause the apparatus at least to determine that the at least one configuration related to the machine learning model is valid to be used in the connected mode during the second connection period and use the at least one configuration related to the machine learning model in the connected mode for the second connection period or determine that the machine learning model is not valid to be used in the connected mode during the second connection period and discard the at least one configuration related to the machine learning model; • wherein the identifier of the at least one configuration related to the machine learning model is received in a connection resume message or a connection setup message; • wherein the at least one processor and the at least one memory further cause the apparatus at least to receive the identifier of the at least one configuration related to the machine learning model in a connection release message; and • discard the at least one configuration associated with the identifier of the machine learning model; • wherein the at least one processor and the at least one memory further cause the apparatus at least to transmit an indication about whether the at least one configuration related to the machine learning model has been used by the apparatus during the first connection period.

[0010] According to an aspect of the present disclosure, there is provided an apparatus comprising at least one processor and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to determine, when a user equipment is transitioning or attempting to transition from a nonconnected mode to a connected mode for a second connection period, an identifier of at least one configuration related to a machine learning model that was used or to be used by the user equipment during a first connection period before the second connection period and transmit the identifier of the at least one configuration related to the machine learning model to the user equipment. The apparatus may be a wireless network node or a control device configured to control the functioning thereof, when installed therein. Example embodiments of the aspect may comprise at least one feature from the following bulleted list or any combination of the following features: • wherein the machine learning model is associated with at least one of a measurement or a data collection configuration; • wherein the connected mode is a radio resource control connected mode and the non-connected mode is a radio resource control inactive mode or a radio resource control idle mode; • wherein the at least one processor and the at least one memory further cause the apparatus at least to store the identifier of the at least one configuration related to the machine learning model when the user equipment is in the connected mode during the first connection period; • wherein the at least one processor and the at least one memory further cause the apparatus at least to store the identifier of the at least one configuration related to the machine learning model as a transient model identifier; • wherein the at least one processor and the at least one memory further cause the apparatus at least to receive an indication from the user equipment, said indication indicating whether the at least one configuration related to the machine learning model is valid to be used in the connected mode during the second connection period; • wherein the at least one processor and the at least one memory further cause the apparatus at least to receive said indication indicating whether the at least one configuration related to the machine learning model is valid to be used in the connected mode during the second connection period in a connection resume complete message or a connection setup complete message; • wherein the identifier of the machine learning model is transmitted in a connection resume message or a connection setup message; • wherein the at least one processor and the at least one memory further cause the apparatus at least to receive the identifier of the at least one configuration related to the machine learning model in a connection release message when a connection between the apparatus and the user equipment is suspended; • wherein the at least one processor and the at least one memory further cause the apparatus at least to receive, from the user equipment, an indication about whether the at least one configuration related to the machine learning model has been used by the user equipment during the first connection period.

[0011] According to an aspect, there is provided a method comprising, determining, when the apparatus is in a connected mode during a first connection period, an identifier of at least one configuration related to a machine learning model used or to be used by the apparatus in the connected mode, transitioning from the connected mode to a non-connected mode and transmitting to a wireless network node, when the apparatus is transitioning or attempting to transition from the non-connected mode to the connected mode for a second connection period, information related to the identifier of the at least one configuration related to the machine learning model. The method may be performed by a user equipment or a control device configured to control the functioning thereof, when installed therein.

[0012] According to an aspect, there is provided a method comprising, receiving from a user equipment, when the user equipment is transitioning or attempting to transition from a non-connected mode to a connected mode for a second connection period, information related to an identifier of at least one configuration related to a machine learning model that was used or to be used by the user equipment during a first connection period before the second connection period and evaluating whether the at least one configuration related to the machine learning model is suitable to be used by at least one of the user equipment in the connected mode during the second connection period or by the apparatus while the user equipment is in the connected mode during the second connection period. The method may be performed by a wireless network node or a control device configured to control the functioning thereof, when installed therein.

[0013] According to an aspect, there is provided a method comprising, determining, when the apparatus is in a Radio Resource Control, RRC, connected mode during a first RRC connection period, an identifier of at least one configuration related to a machine learning model used or to be used by the apparatus in the RRC connected mode, transitioning from the RRC connected mode to an RRC inactive mode or an RRC idle mode and transmitting to a wireless network node, when the apparatus is transitioning or attempting to transition from the RRC inactive mode or the RRC idle mode to the RRC connected mode for a second RRC connection period, the identifier of the at least one configuration related to the machine learning model via RRC signalling. The method may be performed by a user equipment or a control device configured to control the functioning thereof, when installed therein.

[0014] According to an aspect, there is provided a method comprising, receiving via RRC signalling from a user equipment, when the user equipment is transitioning or attempting to transition from an RRC inactive mode or an RRC idle mode to a RRC connected mode for a second RRC connection period, an identifier of at least one configuration related to a machine learning model that was used or to be used by the user equipment during a first RRC connection period before the second RRC connection period and evaluating whether the at least one configuration related to the machine learning model is suitable to be used by at least one of the user equipment in the RRC connected mode during the second RRC connection period or by the apparatus while the user equipment is in the RRC connected mode during the second RRC connection period. The method may be performed by a wireless network node or a control device configured to control the functioning thereof, when installed therein.

[0015] According to an aspect, there is provided a method comprising, receiving from a wireless network node, when the apparatus is transitioning or attempting to transition from a non-connected mode to a connected mode for a second connection period, an identifier of at least one configuration related to a machine learning model that was used or to be used by the apparatus during a first connection period before the second connection period and determining whether to use the at least one configuration related to the machine learning model in the connected mode during the second connection period.The method may be performed by a user equipment or a control device configured to control the functioning thereof, when installed therein.

[0016] According to an aspect, there is provided a method comprising, determining, when a user equipment is transitioning or attempting to transition from a non-connected mode to a connected mode for a second connection period, an identifier of at least one configuration related to a machine learning model that was used or to be used by the user equipment during a first connection period before the second connection period and transmitting the identifier of the at least one configuration related to the machine learning model to the user equipment. The method may be performed by a wireless network node or a control device configured to control the functioning thereof, when installed therein.

[0017] According to an aspect of the present disclosure, there is provided an apparatus comprising means for determining, when the apparatus is in a connected mode during a first connection period, an identifier of at least one configuration related to a machine learning model used or to be used by the apparatus in the connected mode, means for transitioning from the connected mode to a non-connected mode and means for transmitting to a wireless network node, when the apparatus is transitioning or attempting to transition from the non-connected mode to the connected mode for a second connection period, information related to the identifier of the at least one configuration related to the machine learning model.The apparatus of the aspect may be a user equipment or a control device configured to control the functioning thereof, when installed therein.

[0018] According to an aspect of the present disclosure, there is provided an apparatus comprising means for receiving from a user equipment, when the user equipment is transitioning or attempting to transition from a non-connected mode to a connected mode for a second connection period, information related to an identifier of at least one configuration related to a machine learning model that was used or to be used by the user equipment during a first connection period before the second connection period and means for evaluating whether the at least one configuration related to the machine learning model is suitable to be used by at least one of the user equipment in the connected mode during the second connection period or by the apparatus while the user equipment is in the connected mode during the second connection period. The apparatus of the aspect may be a wireless network node or a control device configured to control the functioning thereof, when installed therein.

[0019] According to an aspect of the present disclosure, there is provided an apparatus comprising means for determining, when the apparatus is in a Radio Resource Control, RRC, connected mode during a first RRC connection period, an identifier of at least one configuration related to a machine learning model used or to be used by the apparatus in the RRC connected mode, means for transitioning from the RRC connected mode to an RRC inactive mode or an RRC idle mode and means for transmitting to a wireless network node, when the apparatus is transitioning or attempting to transition from the RRC inactive mode or the RRC idle mode to the RRC connected mode for a second RRC connection period, the identifier of the at least one configuration related to the machine learning model via RRC signalling. The apparatus of the aspect may be a user equipment or a control device configured to control the functioning thereof, when installed therein.

[0020] According to an aspect of the present disclosure, there is provided an apparatus comprising means for receiving via RRC signalling from a user equipment, when the user equipment is transitioning or attempting to transition from an RRC inactive mode or an RRC idle mode to a RRC connected mode for a second RRC connection period, an identifier of at least one configuration related to a machine learning model that was used or to be used by the user equipment during a first RRC connection period before the second RRC connection period and means for evaluating whether the at least one configuration related to the machine learning model is suitable to be used by at least one of the user equipment in the RRC connected mode during the second RRC connection period or by the apparatus while the user equipment is in the RRC connected mode during the second RRC connection period.

[0021] According to an aspect of the present disclosure, there is provided an apparatus comprising means for receiving from a wireless network node, when the apparatus is transitioning or attempting to transition from a non-connected mode to a connected mode for a second connection period, an identifier of at least one configuration related to a machine learning model that was used or to be used by the apparatus during a first connection period before the second connection period and means for determining whether to use the at least one configuration related to the machine learning model in the connected mode during the second connection period. The apparatus of the aspect may be a user equipment or a control device configured to control the functioning thereof, when installed therein.

[0022] According to an aspect of the present disclosure, there is provided an apparatus comprising means for determining, when a user equipment is transitioning or attempting to transition from a non-connected mode to a connected mode for a second connection period, an identifier of at least one configuration related to a machine learning model that was used or to be used by the user equipment during a first connection period before the second connection period and means for transmitting the identifier of the at least one configuration related to the machine learning model to the user equipment.

[0023] According to an aspect of the present disclosure, there is provided a computer program comprising instructions which, when the program is executed by an apparatus, cause the apparatus to carry out determining, when the apparatus is in a connected mode during a first connection period, an identifier of at least one configuration related to a machine learning model used or to be used by the apparatus in the connected mode, transitioning from the connected mode to a non-connected mode and transmitting to a wireless network node, when the apparatus is transitioning or attempting to transition from the non-connected mode to the connected mode for a second connection period, information related to the identifier of the at least one configuration related to the machine learning model.

[0024] According to an aspect of the present disclosure, there is provided a computer program comprising instructions which, when the program is executed by an apparatus, cause the apparatus to carry out receiving from a user equipment, when the user equipment is transitioning or attempting to transition from a non-connected mode to a connected mode for a second connection period, information related to an identifier of at least one configuration related to a machine learning model that was used or to be used by the user equipment during a first connection period before the second connection period and evaluating whether the at least one configuration related to the machine learning model is suitable to be used by at least one of the user equipment in the connected mode during the second connection period or by the apparatus while the user equipment is in the connected mode during the second connection period.

[0025] According to an aspect of the present disclosure, there is provided a computer program comprising instructions which, when the program is executed by an apparatus, cause the apparatus to carry out determining, when the apparatus is in a Radio Resource Control, RRC, connected mode during a first RRC connection period, an identifier of at least one configuration related to a machine learning model used or to be used by the apparatus in the RRC connected mode, transitioning from the RRC connected mode to an RRC inactive mode or an RRC idle mode and transmitting to a wireless network node, when the apparatus is transitioning or attempting to transition from the RRC inactive mode or the RRC idle mode to the RRC connected mode for a second RRC connection period, the identifier of the at least one configuration related to the machine learning model via RRC signalling.

[0026] According to an aspect of the present disclosure, there is provided a computer program comprising instructions which, when the program is executed by an apparatus, cause the apparatus to carry out receiving via RRC signalling from a user equipment, when the user equipment is transitioning or attempting to transition from an RRC inactive mode or an RRC idle mode to a RRC connected mode for a second RRC connection period, an identifier of at least one configuration related to a machine learning model that was used or to be used by the user equipment during a first RRC connection period before the second RRC connection period and evaluating whether the at least one configuration related to the machine learning model is suitable to be used by at least one of the user equipment in the RRC connected mode during the second RRC connection period or by the apparatus while the user equipment is in the RRC connected mode during the second RRC connection period.

[0027] According to an aspect of the present disclosure, there is provided a computer program comprising instructions which, when the program is executed by an apparatus, cause the apparatus to carry out receiving from a wireless network node, when the apparatus is transitioning or attempting to transition from a non-connected mode to a connected mode for a second connection period, an identifier of at least one configuration related to a machine learning model that was used or to be used by the apparatus during a first connection period before the second connection period and determining whether to use the at least one configuration related to the machine learning model in the connected mode during the second connection period.

[0028] According to an aspect of the present disclosure, there is provided a computer program comprising instructions which, when the program is executed by an apparatus, cause the apparatus to carry out determining, when a user equipment is transitioning or attempting to transition from a non-connected mode to a connected mode for a second connection period, an identifier of at least one configuration related to a machine learning model that was used or to be used by the user equipment during a first connection period before the second connection period and transmitting the identifier of the at least one configuration related to the machine learning model to the user equipment.

[0029] According to an aspect of the present disclosure, there is provided a non-transitory computer readable medium having stored thereon a set of computer readable instructions that, when executed by at least one processor, cause an apparatus to at least perform determining, when the apparatus is in a connected mode during a first connection period, an identifier of at least one configuration related to a machine learning model used or to be used by the apparatus in the connected mode, transitioning from the connected mode to a non-connected mode and transmitting to a wireless network node, when the apparatus is transitioning or attempting to transition from the non-connected mode to the connected mode for a second connection period, information related to the identifier of the at least one configuration related to the machine learning model.

[0030] According to an aspect of the present disclosure, there is provided a non-transitory computer readable medium having stored thereon a set of computer readable instructions that, when executed by at least one processor, cause an apparatus to at least perform receiving from a user equipment, when the user equipment is transitioning or attempting to transition from a non-connected mode to a connected mode for a second connection period, information related to an identifier of at least one configuration related to a machine learning model that was used or to be used by the user equipment during a first connection period before the second connection period and evaluating whether the at least one configuration related to the machine learning model is suitable to be used by at least one of the user equipment in the connected mode during the second connection period or by the apparatus while the user equipment is in the connected mode during the second connection period.

[0031] According to an aspect of the present disclosure, there is provided a non-transitory computer readable medium having stored thereon a set of computer readable instructions that, when executed by at least one processor, cause an apparatus to at least perform determining, when the apparatus is in a Radio Resource Control, RRC, connected mode during a first RRC connection period, an identifier of at least one configuration related to a machine learning model used or to be used by the apparatus in the RRC connected mode, transitioning from the RRC connected mode to an RRC inactive mode or an RRC idle mode and transmitting to a wireless network node, when the apparatus is transitioning or attempting to transition from the RRC inactive mode or the RRC idle mode to the RRC connected mode for a second RRC connection period, the identifier of the at least one configuration related to the machine learning model via RRC signalling.

[0032] According to an aspect of the present disclosure, there is provided a non-transitory computer readable medium having stored thereon a set of computer readable instructions that, when executed by at least one processor, cause an apparatus to at least perform receiving via RRC signalling from a user equipment, when the user equipment is transitioning or attempting to transition from an RRC inactive mode or an RRC idle mode to a RRC connected mode for a second RRC connection period, an identifier of at least one configuration related to a machine learning model that was used or to be used by the user equipment during a first RRC connection period before the second RRC connection period and evaluating whether the at least one configuration related to the machine learning model is suitable to be used by at least one of the user equipment in the RRC connected mode during the second RRC connection period or by the apparatus while the user equipment is in the RRC connected mode during the second RRC connection period.

[0033] According to an aspect of the present disclosure, there is provided a non-transitory computer readable medium having stored thereon a set of computer readable instructions that, when executed by at least one processor, cause an apparatus to at least perform receiving from a wireless network node, when the apparatus is transitioning or attempting to transition from a non-connected mode to a connected mode for a second connection period, an identifier of at least one configuration related to a machine learning model that was used or to be used by the apparatus during a first connection period before the second connection period and determining whether to use the at least one configuration related to the machine learning model in the connected mode during the second connection period.

[0034] According to an aspect of the present disclosure, there is provided a non-transitory computer readable medium having stored thereon a set of computer readable instructions that, when executed by at least one processor, cause an apparatus to at least perform determining, when a user equipment is transitioning or attempting to transition from a non-connected mode to a connected mode for a second connection period, an identifier of at least one configuration related to a machine learning model that was used or to be used by the user equipment during a first connection period before the second connection period and transmitting the identifier of the at least one configuration related to the machine learning model to the user equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] FIG. 1 illustrates an example of a network scenario in accordance with at least some example embodiments;

[0036] FIG. 2 illustrates a first flow chart in accordance with at least some example embodiments;

[0037] FIG. 3 illustrates a second flow chart in accordance with at least some example embodiments;

[0038] FIG. 4 illustrates a first signalling graph in accordance with at least some example embodiments;

[0039] FIG. 5 illustrates a second signalling graph in accordance with at least some example embodiments;

[0040] FIG. 6 illustrates a third signalling graph in accordance with at least some example embodiments;

[0041] FIG. 7 illustrates a fourth signalling graph in accordance with at least some example embodiments;

[0042] FIG. 8 illustrates a fifth signalling graph in accordance with at least some example embodiments;

[0043] FIG. 9 illustrates an example apparatus capable of supporting at least some example embodiments;

[0044] FIG. 10 illustrates a first flow graph of a method in accordance with at least some example embodiments;

[0045] FIG. 11 illustrates a second flow graph of a method in accordance with at least some example embodiments; and

[0046] FIG. 12 illustrates a third flow graph of a method in accordance with at least some example embodiments. EXAMPLE EMBODIMENTS

[0047] Embodiments of the present disclosure provide enhancements for Machine Learning, ML, model identification in cellular communication networks. More specifically, if a complete ML model identification would be required every time a User Equipment, UE, transitions from a non-connected mode to a connected mode, the operation would be inefficient because it would require transmission of large amount of data and waste of time due to excessive signalling. Embodiments of the present disclosure therefore provide efficient ML model identification when the UE transitions from the non-connected mode to the connected mode. In some example embodiments, the connected mode may be a Radio Resource Control, RRC, connected mode and the nonconnected mode may an RRC idle mode or an RRC inactive mode. In such a case, RRC signalling may be used for transmitting.

[0048] FIG. 1 illustrates an example of a network scenario in accordance with at least some example embodiments. According to the example scenario of FIG. I, there may be a beam-based wireless communication system, which comprises UE 110, wireless network node 120 and core network element 130. UE 110 may be connected to wireless network node 120 via air interface using beams 115, either simultaneously or one at a time.

[0049] UE 110 may comprise, for example, a smartphone, a cellular phone, a Machine-to-Machine, M2M, node, Machine-Type Communications, MTC, node, an Internet of Things, loT, node, a car telemetry unit, a laptop computer, a tablet computer or, indeed, any kind of suitable wireless terminal. In the example system of FIG. 1, UE 110 may communicate wirelessly with wireless network node 120 via at least one beam 115. Wireless network node 120 may be considered as a serving node for UE 110 and one cell of wireless network node 120 may be a serving cell for UE 110.

[0050] Air interface between UE 110 and wireless network node 120 may be configured in accordance with a Radio Access Technology, RAT, which both UE 110 and wireless network node 120 are configured to support. Examples of cellular RATs include Long Term Evolution, LTE, New Radio, NR, which may also be known as fifth generation, 5G, radio access technology, 6G radio access technology, and MulteFire.

[0051] For example, in the context of LTE wireless network node 120 may be referred to as eNB while wireless network node 120 may be referred to as gNB in the context of NR. In some example embodiments, wireless network node 120 may be referred to as a Transmission and Reception Point, TRP, or control multiple TRPs that may be co-located or non-co-located. In any case, example embodiments of the present disclosure are not restricted to any particular wireless technology. Instead, example embodiments may be exploited in any wireless communication system, wherein ML models are used.

[0052] Wireless network node 120 may be connected, directly or via at least one intermediate node, with core network 130 via interface 125. Core network 130 may be, in turn, coupled via interface 135 with another network (not shown in FIG. 1), via which connectivity to further networks may be obtained, for example via a worldwide interconnection network. Wireless network node 120 may be connected, directly or via at least one intermediate node, with core network 130 or with another core network.

[0053] In some example embodiments, the network scenario may comprise a relay node instead of, or in addition to, UE 110 and / or wireless network node 120. Relaying may be used for example when operating on millimeter-wave frequencies. One example of the relay node may be an Integrated Access and Backhaul, IAB, node. The IAB node may be referred to as a self-backhauling relay as well. Another example of a relay may be an out-band relay. In general, the relay node may comprise two parts: 1) Distributed Unit, DU, part which may facilitate functionalities of wireless network node 120, such as a gNB. Thus, in some example embodiments, the DU part of a relay may be referred to as wireless network node 120 and the DU may perform tasks of wireless network node 120; 2) Mobile Termination, MT, part which may facilitate functionalities of UE 110, i.e., a backhaul link which may be the communication link between a parent node (DU), such as a DU part of wireless network node 120, and the relay, such as an IAB node. In some example embodiments, the MT part may be referred to as UE 110 and perform tasks of UE 110.

[0054] Procedures related to ML model identification may comprise, for example, model transfer, data collection and data transfer. The ML configuration flexibility may be dependent on connection modes, such RRC modes. UE 110 may be in different connection modes at different times and if a very complex ML model identification procedure is adopted, UE 110 would be required to perform ML model identification each time UE 110 transitions from the non-connected mode, such as an RRC idle or an RRC inactive mode, to the connected mode, such as an RRC connected mode. Even though RRC modes are used as examples in various embodiments of the present disclosure, the embodiments may be applied similarly for any similar connection modes.

[0055] Such a procedure would lead to impractically complex solution(s) and huge signaling overhead in a cell. The reason is that data collection configuration, data collection and checking would be required every time for each ML model of each UE 110 and for each cell of wireless network node 120. Thus, large amount of data would be required and a lot of time spent in due to redundant signalling. Another challenge may be that the ML model Identifiers, IDs, may be very large. In such a case, it would be good if the network could transact fewer bits when identifying and / or referring to ML Model ID(s) of UE 110 for Life Cycle Management (LCM) purposes.

[0056] Embodiments of the present disclosure therefore enable avoiding a complete online model identification procedure every time when UE 110 transitions from the non-connected mode, such as the RRC idle or the RRC inactive mode, to the connected mode, such as the RRC connected mode. In addition, or alternatively, model identification may be avoided in handover cases when UE 110 changes cells. For example, embodiments of the present disclosure may be exploited to avoid complexity and non-scalability issues related to online ML model identification options presented in the 3GPP standard document TR 38.843 V. 18.0.0.

[0057] UE 1 10 and wireless network node 120 may be allowed to efficiently re-use the ML models at UE 110 with an existing ML model ID(s). The existing ML model ID(s) may have been allocated by wireless network node 120 to UE 110 during a previous ML model identification procedure in the connected state. UE 110 may thus determine, when UE 110 is in the connected mode during a first connection period (also called RRC connected period), an ID of at least configuration related to a ML model used or to be used in the connected mode. For example, the at least one configuration may be related to CSI-ReportConfig which allows Channel State Information, CSI, - Reference Signal, RS, to be configured to UE 110. UE 110 may measure the resources for input to the ML model and predicted output may be predicted Ll-RSRP or predicted CRI / SSBRI to be reported to the NW.

[0058] UE 1 10 may store the ID of the at least one configuration related to the ML model as a transient model ID. The transient model ID may be the model ID which is already available at UE 110 and wireless network node 120, and has been identified in the previous, first connected state. The transient or model ID may be at least temporarily used in the new, second connected state. Transient may indicate that either UE 110 or wireless network node 120 may decide to change and trigger model identification, however until then the previously identified model ID is being used.

[0059] Alternatively, the ID of the at least one configuration related to the ML model, allocated for UE 110 during its last connected mode, may be referred to as an old Model ID, a temporary ID, transitional ID, transitory ID or short-term ID. The ID of the at least one configuration related to the ML model may refer to, or may be represented by an index, tag, label, hash, secure bitstring or fingerprint.

[0060] UE 110 may then transmit, when UE 110 is transitioning or attempting to transition from the non-connected mode to the connected mode for a second connection period, information related to the ID of at least one configuration related to the ML Model. In some example embodiments, UE 110 may transmit the ID of at least one configuration related to the ML Model that was determined for the first connection period, before the second connection period.

[0061] Even though transmission of the ID by UE 110 is used as an example in various example embodiments, the procedures may be applied similarly when said information related to the ID is transmitted by UE 110. When said information related to the ID is transmitted by UE 110, wireless network node 120 may perform subsequent actions similarly as when the ID is transmitted by UE 110. Said information related to the ID may comprise information about the radio context in which the model ID is expected to operate optimally, e.g., by indicating a set of functionality / feature configurations (or their indices) related to the input measurements and expected outputs, etc.

[0062] In some example embodiments, to reduce the bit size requirements of a ML Model ID, wireless network node 120 may decide to map the ML model ID to a shorter version. For example, wireless network node 120 may map a 128-bit ML Model ID to a 16 bit “temporary” ID. This may be the transient ID for the ML model used in some of the embodiments. This may require the context information from above to be available. Else Model IDs might get mixed up due to the 16-bit representation, of the original 128-bit representation.

[0063] The ID of the at least one configuration related to the ML model, which has been exchanged, or aligned, between UE 110 and wireless network node 120 during the previous full online model identification procedure in the connected mode during a first connection period, may be configured by wireless network node 120 to be the transient ID for the successive, second connection period after one or more periods in the nonconnected mode(s). UE 110 may transmit to wireless network node 120, when UE 110 is transitioning or attempting to transition from the non-connected mode to the connected mode for a second connection period, the ID of the at least one configuration related to the ML model.

[0064] In some example embodiments, UE 110 may be in the connected mode during the first and second connection periods with the same or different wireless network nodes. For example, UE 110 may be in the connected mode first with a first wireless network node during the first connection period and transitioning or attempting to transition to the connected mode with a second wireless network node for the second connection period, wherein the first wireless network node is different than the second wireless network node. Thus, handover of UE 110 may be supported. In some example embodiments, the first and second wireless network nodes may be the same node.

[0065] In some example embodiments, the ID of the at least one configuration related to the ML model may be allocated as a local ID, within a Radio Access Network, RAN, the ID may be exploited to provide a flexible mechanism and allow sufficient control of the online model identification within the RAN domain. The local ID may refer to a cell-specific, i.e., per cell. Thus, it is possible to re-use for example existing measurement identifiers as local unique model IDs and also to use RRC messages with some enhancement to convey Information Elements, IEs, related to the mapping between a model ID and the ID of the at least one configuration related to the ML model. Alternatively, the ID of the at least one configuration related to the ML model may be a globally unique Model ID, e.g., as mentioned in the 3GPP standard document TR 38.843 vl 8.0.0, in section 4.2.2. The ID of the at least one configuration related to the ML model may be designated as a transient model ID and correspond to an indicated and / or configured feature, feature group or functionality. The ID of the at least one configuration related to the ML model may be associated with additional conditions, such as configuration, parameters, etc., and context information which may be exchanged during the online model identification procedure.

[0066] FIG. 2 illustrates a first flow chart in accordance with at least some example embodiments. The first flow chart is a flow chart for transient ID mechanisms when a local ID is used as the ID of the at least one configuration related to the ML model in different transitions from the RRC inactive mode to the RRC connected mode.

[0067] At step 202, UE 110 and wireless network node 120 of FIG. 1 may perform initial ML model identification. UE 110 may be in the RRC connected mode (e.g. during a first connection period) and wireless network node 120 may assign for UE 110 an ID of at least one configuration related to a ML model used or to be used by the apparatus in the RRC connected mode. Wireless network node 120 assign the model ID based on the available and supported radio configurations, e.g. for radio configurations is the configuration of the CSI -RS resources (downlink transmission) to be used by UE 110, to estimate the channel and / or select / predict radio beams, etc. During assignment, wireless network node 120 may store the configuration for subsequent LCM operations (activation, deactivation, switching) associated with the configuration.

[0068] Wireless network node 120 may transmit the ID of the at least one configuration related to the ML model to UE 110. UE 110 may thus determine the ID of the at least one configuration related to the ML model based on the received ID. The ID may be e.g. the transient ID, globally unique ID, for example.

[0069] At step 204, UE 110 may use the at least one configuration related to the ML model identified by the determined ID during a first connection period in the RRC connected mode. The ML model may be associated with at least one of a measurement configuration or a data collection configuration. UE 110 may hence perform measurements or collect data according to the at least one configuration related to the ML model. That is, UE 110 may apply the configuration associated with the ID related to the ML model, e.g., start using the measurement configuration received from wireless network node 120 which is associated with the ID, and the corresponding radio measurements are the used directly / indirectly (w / wo pre-processing) to the ML algorithm. Alternatively, or in addition, UE 110 may use the ID for reporting performance monitoring results. Wireless network node 120 may then identify the performance related to the configured reports / resources. After using the at least one configuration, UE 110 may transition from the RRC connected mode to the RRC idle mode.

[0070] At step 206, UE 110 may decide how to proceed depending on whether it is in the RRC idle mode, or not. If UE 110 is in the RRC idle mode, UE 110 may release resources, at step 208. If UE 110 is not in the RRC idle mode, UE 110 may determine that it is in the RRC inactive mode. In such a case, UE 110 may transition from the RRC inactive mode to the RRC connected mode, at step 210.

[0071] At step 212, UE 110 may determine whether it has stored the ID of the at least one configuration related to the ML model to its memory. If UE 110 has not stored the ID of the at least one configuration related to the ML model, UE 110 may perform initial model identification again, at step 202. If UE 110 has stored the ID of the at least one configuration related to the ML model, UE 110 may determine whether the ID of the at least one configuration related to the ML model is valid, at step 214.

[0072] If the ID of the at least one configuration related to the ML model is not valid, UE 110 may perform initial model identification again, at step 202. If the ID of the at least one configuration related to the ML model is valid, UE 110 may use, during the second connection period, the at least one configuration related to the ML model identified by the determined ID for the first connection period in the RRC connected mode. UE 110 may determine that the ID of the at least one configuration related to the ML model is valid when current information or configuration associated with the ML enabled features and models are available and matches with the previous configurati on / informati on.

[0073] After that, UE 110 may transition from the RRC connected mode to the RRC idle state.

[0074] FIG. 3 illustrates a second flow chart in accordance with at least some example embodiments. The second flow chart is a flow chart for transient ID mechanisms when a local ID is used as the ID of the at least one configuration related to the ML model in different transitions from the RRC idle mode to the RRC connected mode

[0075] At step 302, UE 110 and wireless network node 120 of FIG. 1 may perform initial ML model identification. At step 304, UE 110 may use the at least one configuration related to the ML model identified by the determined ID during a first connection period in the RRC connected mode. At step 306, UE 110 may transition from the RRC connected mode to the RRC idle mode. After a while, UE 110 may transition from the RRC idle mode back to the RRC connected mode.

[0076] At step 308, UE 110 may determine whether it has stored the ID of the at least one configuration related to the ML model to its memory. If UE 110 has not stored the ID of the at least one configuration related to the ML model, UE 110 may release resources, at step 310. In such a case, UE 110 may perform initial model identification again, at step 302.

[0077] If UE 110 has stored the ID of the at least one configuration related to the ML model, UE 110 determine whether the ID of the at least one configuration related to the ML model is valid, at step 312. If the ID of the at least one configuration related to the ML model is not valid, UE 110 may perform initial model identification again, at step 302. If the ID of the at least one configuration related to the ML model is valid, UE 110 may use, during a second connection period, the at least one configuration related to the ML model identified by the determined ID during the first connection period in the RRC connected mode.

[0078] FIG. 4 illustrates a first signalling graph in accordance with at least some example embodiments. On the vertical axes are disposed, from the left to the right, UE 110 and wireless network node 120 of FIG. 1. FIG. 4 illustrates the usage of the ID of the at least one configuration related to the ML model for the RRC inactive mode.

[0079] At steps 402 and 404, UE 110 may be in the RRC connected mode during a first connection period. At step 402, UE 110 and wireless network node 120 may perform initial ML model identification. UE 110 may determine the ID of the at least one configuration related to the ML model based on the performed identification. At step 404, UE 110 may perform inference using the ID of the at least one configuration related to the ML model. Performing inference may comprise, e.g., beam prediction or CSI compression, according to the configurations received from wireless network node 120. Alternatively, or in addition, at least in AI / ML enabled BM-Case 1, UE 110 may measure the RSs configured by wireless network node 120, for example, to predict channel related measurements (Ll-RSRP, beam ID). After performing inference, UE 110 may transition from the RRC connected mode to the RRC inactive mode.

[0080] At step 406, UE 110 and wireless network node 120 may determine that the ID of the at least one configuration related to the ML model has been assigned. At step 408, UE 110 may store the ID of the at least one configuration related to the ML model to its memory. UE 110 may also stop using at least one configuration related to the ML model corresponding to the ID. The determination of whether to store the ID for subsequent purposes may be decided during mode transition to tell UE 110 to add / store to its local storage. For example, the determination may happen just before UE 110 enters idle mode. For example, UE 110 may store information from the RRC configurations, and the ID may be stored as well with or without signalling to or from wireless network node 120. Wireless network node 120 may do the same when it detects that UE 110 entered idle mode.

[0081] In some example embodiments, UE 110 may store the ID of the at least one configuration related to the ML model to its memory in the RRC connected mode during the first connection period already, possibly as a transient mode ID. UE 110 may also determine that it is configured to store the ID of the at least one configuration related to the ML model to its memory. UE 110 may be configured for said storing by wireless network node 120.

[0082] In some example embodiments, when UE 110 transitions to the RRC inactive mode, UE 110 may stop its LCM operations associated with the ID of the at least one configuration related to the ML model. UE 110 may release any dedicated resources associated with the at least one configuration related to the ML model. UE 110 may store the ID of the at least one configuration related to the ML model as a transient model ID if UE 110 has been configured to perform said storing during the initial online model dentification, at step 402.

[0083] At step 410, UE 110 may transmit a resume request message to wireless network node 120. For example, UE 110 may transmit an RRC resume request. The resume request message may be transmitted with the ID of the at least one configuration related to the ML model. In some example embodiments, new conditions may apply. UE 110 may thus transmit to wireless network node 120, when UE 110 is transitioning or attempting to transition from the non-connected mode to the connected mode for a second connection period, the ID of the at least one configuration related to the ML model.

[0084] In some example embodiments, UE 110 may transmit to wireless network node 120 information about at least one condition which has changed between the first and second connection periods. The at least one condition may comprise, eg., applicability of the at least one configuration. That is, the least one condition may comprise information about whether the at least one configuration is available or whether UE 110 still has to execute inference monitoring according to the at least one configuration / functionality. In some example embodiments, the at least one condition may comprise radio channel conditions detectable by UE 110, e.g. based on Doppler shift estimation. In such a case, UE 110 may decide that its movements, or the movement of wireless network node 120, is not within the conditions where the ML model would operate optimally to process the CSI-RS.

[0085] In some example embodiments, UE 110 may transmit an indication to wireless network node 120, said indication indicating whether UE 110 prefers re-using the ID of the at least one configuration related to the ML model or initializing identification of the ML model.

[0086] In some example embodiments, UE 110 may transmit an indication about whether the apparatus has the ID of the at least one configuration related to the ML model. The indication may be transmitted, e.g., when UE 110 is handed over from the first wireless network node to the second network node, so that the second network node knows whether UE 110 has the ID. Thus, the use of ML models may be enhanced in case of handovers.

[0087] In some example embodiments, UE 110 may transmit, to wireless network node 120, an indication about whether the at least one configuration related to the ML model has been used by UE 110 during the first connection period. Thus, wireless network node 120 then knows that UE 110 has the required configuration already, so wireless network node 120 does not need to signal it to UE 110 again. Instead, wireless network node 120 may activate it via RRC / MAC signalling. For example, wireless network node 120 knows by receiving RRC complete message along with the ID used that this configuration is applicable.

[0088] In some example embodiments, when UE 110 transitions from the RRC inactive mode to the RRC connected mode, one or more IE(s) may be included in the RRCResumeRequest (short Inactive-Radio Network Temporary Identifier, I-RNTI) and RRCResumeRequest 1 (full I-RNTI) message from UE 110 to wireless network node 120. The IE(s) may indicate a preferred initiation of online model identification or a preferred reuse of the previously stored ID of the at least one configuration related to the ML model, when UE 110 transitioned to the RRC inactive mode.

[0089] One IE may be 1 bit, e.g., 0 to indicate online identification preferred and 1 to indicate that the usage of the ID of the at least one configuration related to the ML model is preferred by UE 110. The ID of the at least one configuration related to the ML model itself might not need to be transmitted because both, UE 110 and wireless network node 120, may already know about the existence of the ML model ID and associated conditions, possibly as stored in a context of UE 110 for the RRC inactive mode.

[0090] In some example embodiments, one IE may occupy more than one bit and / or byte. In such a case, UE 110 may be able to indicate to wireless network node 120 information about conditions which UE 110 has determined as changed since step 406. To reduce the size of the transacted information, UE 110 may indicate by a bit position an index of the impacted ML model(s). For example, if UE 110 has three ML model ID(s) that have been assigned transient ID(s), a bit map of “010” may indicate that the second ML model ID needs to be refreshed. If UE 110 adds more ML model(s) and requires a fresh allocation, a separate bit map may be used to indicate it by mentioning a count of such ML model(s).

[0091] At step 412, wireless network node 120 may evaluate relevant conditions of the UE 110 and the usage of the ID of the at least one configuration related to the ML model. Wireless network node 120 may evaluate whether the at least one configuration related to the ML model is suitable to be used by at least one of UE 110 in the RRC connected mode during the second connection period or by wireless network node 120 while UE 110 is in the RRC connected mode during the second connection period. The at least one configuration related to the ML model may be suitable to be used when both evaluations are true, but not to be used if only one or none.

[0092] Wireless network node 120 may perform said evaluation based at least partly on at least one of the following: • the received indication indicating whether UE 110 prefers re-using the ID of the at least one configuration related to the ML model or initializing identification of the ML model; or • the received information about at least one condition which has changed between the first and second connection periods.

[0093] At step 414, wireless network node 120 may transmit a connection resume message to UE 110. For example, wireless network node 120 may transmit an RRC resume message. Wireless network node 120 may transmit the connection resume message together with an acknowledgement about the ID of the at least one configuration related to the ML model. Wireless network node 120 may transmit the RRC resume message to UE 110, comprising an acknowledgement or a negative acknowledgement about the use of the ID of the at least one configuration related to the ML model.

[0094] In some example embodiments, a condition which triggers a new model identification may comprise at least one of: UE context message has been updated, UEInformation requested by UEcapabilityEnquiry has been changed, UE moves to a new cell, UE has a new model or the old model is not more valid (performance issue, configuration changed, etc.) If any of the conditions is valid, a new transient ID is needed. Otherwise, the ID of the at least one configuration related to the ML model may be reused.

[0095] In some example embodiments, wireless network node 120 may transmit information indicating whether UE 110 is to use the at least one configuration related to the ML model corresponding to the ID during the second connection period. UE 110 may thus receive said information in response to transmitting information related to the ID of the at least one configuration related to the ML model.

[0096] At step 416, UE 10 may evaluate conditions related to the ID of the at least one configuration related to the ML model. After that, UE 110 may determine whether the ID of the at least one configuration related to the ML model is valid. For example, if the RRCResume comprises an acknowledgement about the use of the ID of the at least one configuration related to the ML model, then UE 110 may check if the new conditions are aligned with the configurations for the ID of the at least one configuration related to the ML model. UE 110 may determine the validity of the ID of the at least one configuration using the same approach as wireless network node 120, at step 412.

[0097] In some example embodiments, the evaluation of the validity may be done before step 410, to ensure that UE 110 does not transmit an invalid ID. For example, the evaluation of the validity may be done before step 410 if UE 110 can perform the required radio measurements. If these can be done based on the broadcast channel, then the evaluation may take place before. Else, a connection resume, such as a the RRC resume, may need to configure UE 110 to be able to perform radio measurements and after that UE 110 may estimate if the conditions are acceptable or not.

[0098] In some example embodiments, UE 110 may determine, when said information indicating whether UE 110 is to use the at least one configuration indicates that UE 110 is to use the at least one configuration related to the ML model corresponding to the ID during the second connection period, whether the ID of the at least one configuration related to the ML model is valid. Valid may mean that the current information or configuration associated with the ML enabled features and models are available, and match with the previous configuration / information.

[0099] If the ID of the at least one configuration related to the ML model is valid, UE 110 may restore all configurations related to the ML model. After that, UE 110 may transmit a connection resume completed message to wireless network node 120, at step 418. The connection resume completed message may be an RRC resume complete message. The connection resume message may indicate that the ID of the at least one configuration related to the ML model is valid. After that, UE 110 may transition to the RRC connected mode for the second connection period and perform inference with the ID of the at least one configuration related to the ML model.

[00100] Alternatively, UE 110 may discard, when said information indicating whether UE 110 is to use the at least one configuration indicates that UE 110 is not to use the at least one configuration related to the ML model corresponding to the ID during the second connection period, the ID of the at least one configuration related to the ML model. If the ID of the at least one configuration related to the ML model is invalid, UE 110 may discard all configurations related to the ML model.

[00101] After that, UE 110 may transmit a connection resume completed message to wireless network node 120, at step 420. The connection resume completed message may be an RRC resume complete message. The connection resume message may indicate that the ID of the at least one configuration related to the ML model is invalid. After that, UE 110 may transition to the RRC connected mode for the second connection period and perform another online model identification to determine another configuration related to the ML model.

[00102] In some example embodiments, the connection resume message may comprise a negative acknowledgement about the use of the ID of the at least one configuration related to the ML model. In such a case, UE 110 may discard the ID of the at least one configuration related to the ML model and release the associated configurations. UE 110 may be expected to enter the RRC connected mode and wireless network node 120 may then initiate the online model identification procedure.

[00103] For a given functionality or per use case, the ID of the at least one configuration related to the ML model may be applied only to an active model (logical model), but not for physical models. If UE 110 has multiple physical models, the physical models may be transparent to wireless network node 120. A logical model may be comprised of a collection of physical models. The physical models may be hardware dependent and transparent to wireless network node 120.

[00104] In some example embodiments, wireless network node 120 may decide to respond with the connection resume message to UE 110, regardless of the use of AI / ML-enabled feature in UE 110. For handling the AI / ML model ID, prior to sending the connection resume message to UE 110, wireless network node 120 may evaluate the relevant ML-related conditions of UE 110 to determine if a new model identification procedure is needed to be triggered. That is, if identification of the ML model needs to be initialized or if the ID of the at least one configuration related to the ML model may be used.

[00105] FIG. 5 illustrates a second signalling graph in accordance with at least some example embodiments. On the vertical axes are disposed, from the left to the right, UE 110 and wireless network node 120 of FIG. 1. FIG. 5 illustrates a scenario, wherein wireless network node 120 decides to response with a connection setup message, such as an RRC setup message, to UE 110.

[00106] Steps 502 - 512 in FIG. 5 may be the same as steps 402 - 412 in FIG. 4. As an outcome of the evaluation of wireless network node 120, at step 512, wireless network node 120 may transmit a connection setup message, such as an RRCSetup message, to UE. The connection setup message may comprise an acknowledgement about the use of the ID of the at least configuration of the ML model (ACK temp-model-ID) or a negative acknowledgement about the use of the ID of the at least configuration of the ML model (NACK temp-model-ID).

[00107] At step 514, wireless network node 120 may transmit the connection setup message with the acknowledgement. In such a case, UE 110 may restore all the configurations related to the ML model. UE 110 may further transmit, at step 516, a setup complete message, such as an RRC setup complete message. After that, UE 110 may transmit a connection setup completed message to wireless network node 120, at step 516. The connection setup completed message may be an RRC setup complete message. After that, UE 110 may transition to the RRC connected mode for the second connection period and perform inference with the ID of the at least one configuration related to the ML model.

[00108] Alternatively, at step 518, wireless network node 120 may transmit the connection setup message with the negative acknowledgement. After that, UE 110 may discard the ID of the at least one configuration related to the ML model and release all the configurations related to the ML model and transmit a connection setup completed message to wireless network node 120, at step 520. The connection setup completed message may be an RRC setup complete message. After that, UE 110 may transition to the RRC connected mode for the second connection period and perform another online model identification to determine another ML model. UE 110 may perform inference using said another ML model.

[00109] In some example embodiments, a similar procedure may be used by UE 110 as at step 416 of FIG. 4. That is, UE 110 may evaluate the conditions related to the use of the ID of the at least configuration of the ML model and reply to wireless network node 120 whether the ID of the at least configuration of the ML model is valid or invalid. The procedure may then proceed according to step 418 or step 420.

[00110] FIG. 6 illustrates a third signalling graph in accordance with at least some example embodiments. On the vertical axes are disposed, from the left to the right, UE 110 and wireless network node 120 of FIG. 1. FIG. 6 illustrates a scenario, wherein wireless network node 120 responds to UE 110 with a connection release message, such as an RRC release message.

[00111] Steps 602 - 612 in FIG. 6 may be the same as steps 402 - 412 in FIG. 4. As outcome of the evaluation of wireless network node 120, at step 612, wireless network node 120 may transmit a connection release message to UE 110, at step 614. The connection release message may be an RRC release message. UE 110 may receive, in response to transmitting said information related to the ID of the at least one configuration related to the ML model, the connection release message and discard the at least one configuration associated with the ML model corresponding to the ID.

[00112] Wireless network node 120 may decide to respond with the connection release message regardless of the use of AI / ML-enabled feature in UE 110. At step 616, UE 110 may discard the configurations identified by the ID of the at least one configuration related to the ML model. The configurations may be released in accordance with the 3GPP standard document TS 38.331, clauses 5.3.8 and 5.3.12. Any subsequent UE indication of a Physical Uplink Control Channel, PUCCH, and / or Sounding Reference Signal, SRS, request may trigger a new online model identification procedure in the RRC connected mode.

[00113] FIG. 7 illustrates a fourth signalling graph in accordance with at least some example embodiments. On the vertical axes are disposed, from the left to the right, UE 110 and wireless network node 120 of FIG. 1. FIG. 7 illustrates a scenario, wherein wireless network node 120 responds to UE 110 with a connection release message, such as an RRC release message, with a connection suspend command.

[00114] Steps 702 - 712 in FIG. 7 may be the same as steps 402 - 412 in FIG. 4. As outcome of the evaluation of wireless network node 120, at step 712, wireless network node 120 may transmit a connection release message with a connection suspend command to UE 110, at step 714.

[00115] The connection release message may be an RRC release message with the connection suspend command. At step 716, UE 110 may store the configurations identified by the ID of the at least one configuration related to the ML model. The configurations may be stored in accordance with the 3GPP standard document TS 38.331, clauses 5.3.8 and 5.3.12. Any subsequent UE indication of a PUCCH and / or SRS release request may trigger a new online model identification procedure in the RRC connected mode.

[00116] FIG. 8 illustrates a fifth signalling graph in accordance with at least some example embodiments. On the vertical axes are disposed, from the left to the right, UE 110 and wireless network node 120 of FIG. 1. FIG. 8 illustrates a scenario, wherein wireless network node 120 responds to UE 110 with a connection reject message, such as an RRC reject message.

[00117] Steps 802 - 812 in FIG. 8 may be the same as steps 402 - 412 in FIG. 4. As outcome of the evaluation of wireless network node 120, at step 812, wireless network node 120 may transmit a connection reject message to UE 110, at step 814. The connection reject message may be an RRC reject message. At step 816, UE 110 may discard the configurations identified by the ID of the at least one configuration related to the ML model. Any configuration and / or resources associated with the ID of the at least one configuration related to the ML model may be released. Thus, UE 110 may need to re-establish connection with the network and trigger new online model identification procedures.

[00118] In some example embodiments, the spare bit in the RRCResumeRequest message may be used to indicate a new message to wireless network node 120. Wireless network node 120 may wait until it has received the new message. The new message may be used to indicate the usage of the ID of the at least one configuration related to the ML model, instead of embedding the indication in the RRC messages. For example, RRCResum eRequest 1-IEs ::= SEQUENCE {resumeidentity LRNTI-Value, resumeMAC-I BIT STRING (SIZE (16)), resumeCause ResumeCause, spare BIT STRING (SIZE (1))}.

[00119] In some example embodiments, a UE-triggered or network-triggered indication may be used after the connection resume request message. For example, in the case of the UE-triggered approach, where UE 110 transmits an indication to wireless network node 120, indicating an existing / previous assignment of the ID of the at least one configuration related to the ML model. The indication may be transmitted after the connection resume request message via another RRC message, an uplink control information, or an uplink Medium Access Control Control Element, MAC CE, message from UE 110 to wireless network node 120.

[00120] In the case of the network-triggered approach, UE 110 may receive from wireless network node 120, when UE 110 is transitioning or attempting to transition from the non-connected mode to the connected mode for the second connection period, the ID of the at least one configuration related to the ML model that was used or to be used by UE 110 during the first connection period before the second connection period. UE 110 may then determine whether to use the at least one configuration related to the ML model in the connected mode during the second connection period.

[00121] In case of the network-triggered approach, wireless network node 120 may transmit an indication to UE 110, indicating an existing / previous assignment of the ID of the at least one configuration related to the ML model. The indication may be transmitted after the connection request resume message. For example, the indication may be transmitted after the RRCResumeRequest message via an RRC message (e.g., RRCResume, RRCSetup or another RRC message), a downlink control information, or a downlink MAC CE message from wireless network node 120 to UE 110. The network-triggered approach may otherwise work similarly as the UE-triggered approach.

[00122] If UE 110 connects to a different wireless network node than wireless network node 120 when transitioning from the non-connected mode to the connected mode, the ID of the at least one configuration related to the ME model may be exchanged between wireless network node 120 and said another wireless network node. That is, the current wireless network node being (attempted to be) connected may be different than the previous wireless network node with which the initial assignment of the ID of the at least one configuration related to the ML model was done. In some example embodiments, wireless network node 120 may be referred to as a first wireless network node, such as a source BS, and said another wireless network node may be referred to as a second wireless network node, such as a target BS. The information may be exchanged via UE access stratum context message.

[00123] In some example embodiments, the triggering of the assignment of the ID of the at least one configuration related to the ML model may be done explicitly by the network or UE 110 may request it. Successive (e.g., new) assignment(s) of the ID of the at least one configuration related to the ML model may be triggered by a previous triggered entity (i.e., the entity that triggered the previous (or first) assignment) explicitly or by any of wireless network node 120 or UE 110 entities regardless of the entity triggered the previous (or first) assignment.

[00124] Example embodiments of the present disclosure may be applied similarly for the RRC idle mode as for the RRC inactive mode. In the RRC idle mode triggered by a RRCRelease message with release suspend configuration, any UE inactive access stratum context may be discarded regardless of the use of AEML-enabled feature in UE 110. However, if any indicative message is triggered, such as acknowledgement or negative acknowledgement of the ID of the at least one configuration related to the ML model, then the message may trigger restoring of the ID of the at least one configuration related to the ML model, evaluated by wireless network node 120 or UE 110.

[00125] FIG. 9 illustrates an example apparatus capable of supporting at least some example embodiments. Illustrated is device 900, which may comprise, for example, UE 110 or wireless network node 120, or a control device configured to control the functioning thereof, possibly when installed therein. Comprised in device 900 is processor 910, which may comprise, for example, a single- or multi-core processor wherein a single-core processor comprises one processor and a multi-core processor comprises more than one processor. Processor 910 may comprise, in general, a control device. Processor 910 may comprise more than one processor. Processor 910 may be a control device. Processor 910 may comprise at least one application-specific integrated circuit, ASIC. Processor 910 may comprise at least one field-programmable gate array, FPGA. Processor 910 may be means for performing method steps in device 900. Processor 910 may be configured, at least in part by computer instructions, to perform actions.

[00126] A processor may comprise circuitry, or be constituted as circuitry or circuitries, the circuitry or circuitries being configured to perform phases of methods in accordance with example embodiments 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 mobile phone or server, to perform various functions) and (c) hardware circuit(s) and or processor(s), such as a microprocessor(s) or a portion of a microprocessors), that requires software (e.g., firmware) for operation, but the software may not be present when it is not needed for operation.

[00127] 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.

[00128] Device 900 may comprise memory 920. Memory 920 may comprise random-access memory and / or permanent memory. Memory 920 may comprise at least one RAM chip. Memory 920 may comprise solid-state, magnetic, optical and / or holographic memory, for example. Memory 920 may be at least in part accessible to processor 910. Memory 920 may be at least in part comprised in processor 910. Memory 920 may be means for storing information. Memory 920 may comprise computer instructions that processor 910 is configured to execute. When computer instructions configured to cause processor 910 to perform certain actions are stored in memory 920, and device 900 overall is configured to run under the direction of processor 910 using computer instructions from memory 920, processor 910 and / or its at least one processor may be considered to be configured to perform said certain actions. Memory 920 may be at least in part comprised in processor 910. Memory 920 may be at least in part external to device 900 but accessible to device 900.

[00129] Device 900 may comprise a transmitter 930. Device 900 may comprise a receiver 940. Transmitter 930 and receiver 940 may be configured to transmit and receive, respectively, information in accordance with at least one cellular or non-cellular standard. Transmitter 930 may comprise more than one transmitter. Receiver 940 may comprise more than one receiver. Transmitter 930 and / or receiver 940 may be configured to operate in accordance with Global System for Mobile communication, GSM, Wideband Code Division Multiple Access, WCDMA, Long Term Evolution, LTE, and / or 5G / NR standards, for example.

[00130] Device 900 may comprise aNear-Field Communication, NFC, transceiver 950. NFC transceiver 950 may support at least one NFC technology, such as Bluetooth, Wibree or similar technologies.

[00131] Device 900 may comprise User Interface, UI, 960. UI 960 may comprise at least one of a display, a keyboard, a touchscreen, a vibrator arranged to signal to a user by causing device 900 to vibrate, a speaker and a microphone. A user may be able to operate device 900 via UI 960, for example to accept incoming telephone calls, to originate telephone calls or video calls, to browse the Internet, to manage digital files stored in memory 920 or on a cloud accessible via transmitter 930 and receiver 940, or via NFC transceiver 950, and / or to play games.

[00132] Device 900 may comprise or be arranged to accept a user identity module 970. User identity module 970 may comprise, for example, a Subscriber Identity Module, SIM, card installable in device 900. A user identity module 970 may comprise information identifying a subscription of a user of device 900. A user identity module 970 may comprise cryptographic information usable to verify the identity of a user of device 900 and / or to facilitate encryption of communicated information and billing of the user of device 900 for communication effected via device 900.

[00133] Processor 910 may be furnished with a transmitter arranged to output information from processor 910, via electrical leads internal to device 900, to other devices comprised in device 900. Such a transmitter may comprise a serial bus transmitter arranged to, for example, output information via at least one electrical lead to memory 920 for storage therein. Alternatively to a serial bus, the transmitter may comprise a parallel bus transmitter. Likewise processor 910 may comprise a receiver arranged to receive information in processor 910, via electrical leads internal to device 900, from other devices comprised in device 900. Such a receiver may comprise a serial bus receiver arranged to, for example, receive information via at least one electrical lead from receiver 940 for processing in processor 910. Alternatively to a serial bus, the receiver may comprise a parallel bus receiver.

[00134] Device 900 may comprise further devices not illustrated in FIG. 3. For example, where device 900 comprises a smartphone, it may comprise at least one digital camera. Some devices 900 may comprise a back-facing camera and a front-facing camera, wherein the back-facing camera may be intended for digital photography and the frontfacing camera for video telephony. Device 900 may comprise a fingerprint sensor arranged to authenticate, at least in part, a user of device 900. In some example embodiments, device 900 lacks at least one device described above. For example, some devices 900 may lack a NFC transceiver 950 and / or user identity module 970.

[00135] Processor 910, memory 920, transmitter 930, receiver 940, NFC transceiver 950, UI 960 and / or user identity module 970 may be interconnected by electrical leads internal to device 900 in a multitude of different ways. For example, each of the aforementioned devices may be separately connected to a master bus internal to device 900, to allow for the devices to exchange information. However, as the skilled person will appreciate, this is only one example and depending on the example embodiment various ways of interconnecting at least two of the aforementioned devices may be selected without departing from the scope of the example embodiments.

[00136] FIG. 10 is a first flow graph of a method in accordance with at least some example embodiments. The apparatus of the first method may be UE 110 or by a control device configured to control the functioning thereof, when installed therein. The phases of the first method may be performed UE 110 or by a control device configured to control the functioning thereof, when installed therein.

[00137] The method may comprise, at step 1010, determining, when the apparatus is in a connected mode during a first connection period, an identifier of at least one configuration related to a machine learning model used or to be used by the apparatus in the connected mode. The method may also comprise, at step 1020, transitioning from the connected mode to a non-connected mode. Finally, the method may comprise, at step 1030, transmitting to a wireless network node, when the apparatus is transitioning or attempting to transition from the non-connected mode to the connected mode for a second connection period, information related to the identifier of the at least one configuration related to the machine learning model.

[00138] FIG. 11 is a second flow graph of a method in accordance with at least some example embodiments. The apparatus of the second method may be UE 110 or by a control device configured to control the functioning thereof, when installed therein. The phases of the second method may be performed UE 110 or by a control device configured to control the functioning thereof, when installed therein.

[00139] The second method may comprise, at step 1110, determining, when the apparatus is in a Radio Resource Control, RRC, connected mode during a first RRC connection period, an identifier of at least one configuration related to a machine learning model used or to be used by the apparatus in the RRC connected mode. The second method may also comprise, at step 1120, transitioning from the RRC connected mode to an RRC inactive mode or an RRC idle mode. Finally, the second method may comprise, at step 1130, transmitting to a wireless network node, when the apparatus is transitioning or attempting to transition from the RRC inactive mode or the RRC idle mode to the RRC connected mode for a second RRC connection period, the identifier of the at least one configuration related to the machine learning model via RRC signalling.

[00140] FIG. 12 is a third flow graph of a method in accordance with at least some example embodiments. The apparatus of the third method may be UE 110 or by a control device configured to control the functioning thereof, when installed therein. The phases of the third method may be performed UE 110 or by a control device configured to control the functioning thereof, when installed therein.

[00141] The method may comprise, at step 1210, receiving from a wireless network node, when the user equipment is transitioning or attempting to transition from a nonconnected mode to a connected mode for a second connection period, an identifier of at least one configuration related to a machine learning model that was used or to be used by the user equipment during a first connection period before the second connection period. The method may also comprise, at step 1220, determining whether to use the at least one configuration related to the machine learning model in the connected mode during the second connection period.

[00142] It is to be understood that the example embodiments disclosed are not limited to the particular structures, process steps, or materials disclosed herein, but are extended to equivalents thereof as would be recognized by those ordinarily skilled in the relevant arts. It should also be understood that terminology employed herein is used for the purpose of describing particular example embodiments only and is not intended to be limiting.

[00143] Reference throughout this specification to one example embodiment or an example embodiment means that a particular feature, structure, or characteristic described in connection with the example embodiment is included in at least one example embodiment. Thus, appearances of the phrases “in one example embodiment” or “in an example embodiment” in various places throughout this specification are not necessarily all referring to the same example embodiment. Where reference is made to a numerical value using a term such as, for example, about or substantially, the exact numerical value is also disclosed.

[00144] As used herein, a plurality of items, structural elements, compositional elements, and / or materials may be presented in a common list for convenience. However, these lists should be construed as though each member of the list is individually identified as a separate and unique member. Thus, no individual member of such list should be construed as a de facto equivalent of any other member of the same list solely based on their presentation in a common group without indications to the contrary. In addition, various example embodiments and examples may be referred to herein along with alternatives for the various components thereof. It is understood that such example embodiments, examples, and alternatives are not to be construed as de facto equivalents of one another, but are to be considered as separate and autonomous representations.

[00145] In an example embodiment, an apparatus, such as, for example, UE 110 or wireless network node 120, may comprise means for carrying out the example embodiments described above and any combination thereof.

[00146] In an example embodiment, a computer program may be configured to cause a method in accordance with the example embodiments described above and any combination thereof. In an example embodiment, a computer program product, embodied on a non-transitory computer readable medium, may be configured to control a processor to perform a process comprising the example embodiments described above and any combination thereof.

[00147] In an example embodiment, an apparatus, such as, for example, UE 110 or wireless network node 120, may comprise at least one processor, and at least one memory including computer program code, wherein the at least one memory and the computer program code are configured to, with the at least one processor, cause the apparatus at least to perform the example embodiments described above and any combination thereof.

[00148] Furthermore, the described features, structures, or characteristics may be combined in any suitable manner in one or more example embodiments. In the preceding description, numerous specific details are provided, such as examples of lengths, widths, shapes, etc., to provide a thorough understanding of example embodiments of the disclosure. One skilled in the relevant art will recognize, however, that the disclosure can be practiced without one or more of the specific details, or with other methods, components, materials, etc. In other instances, well-known structures, materials, or operations are not shown or described in detail to avoid obscuring aspects of the disclosure.

[00149] While the forgoing examples are illustrative of the principles of the example embodiments in one or more particular applications, it will be apparent to those of ordinary skill in the art that numerous modifications in form, usage and details of implementation can be made without the exercise of inventive faculty, and without departing from the principles and concepts of the disclosure. Accordingly, it is not intended that the disclosure be limited, except as by the claims set forth below.

[00150] The verbs “to comprise” and “to include” are used in this document as open 5 limitations that neither exclude nor require the existence of also un-recited features. The features recited in depending claims are mutually freely combinable unless otherwise explicitly stated. Furthermore, it is to be understood that the use of "a" or "an", that is, a singular form, throughout this document does not exclude a plurality. 10 INDUSTRIAL APPLICABILITY

[00151] At least some example embodiments find industrial application in cellular communication networks, for example in 3GPP networks. ACRONYMS LIST 3 GPP 3rd Generation Partnership Project rx DO Base Station 5 CSI Channel State Information DU Distributed Unit GSM Global System for Mobile communication LRNTI Inactive-Radio Network Temporary Identifier IAB Integrated Access and Backhaul 10 ID Identifier IE Information Element loT Internet of Things LTE Long-Term Evolution M2M Machine-to-Machine 15 MAC CE Medium Access Control Control Element ML Machine Learning MT Mobile Termination MTC Machine-Type Communications NFC Near-Field Communication 20 NR New Radio RAN Radio Access Network RRC Radio Resource Control RS Reference Signal PUCCH Physical Uplink Control Channel 25 SRS Sounding Reference Signal TRP Transmission and Reception Point UE User Equipment UI User Interface WCDMA Wideband Code Division Multiple Access 30 WiMAX Worldwide Interoperability for Microwave Access WLAN Wireless Local Area Network REFERENCE SIGNS LIST 110 UE 115 Beams 120 Wireless network node 125, 135 Wired interfaces 130 Core Network 202-214 Steps in FIG. 2 302312 Steps in FIG. 3 402 - 420 Steps in FIG. 4 502 - 520 Steps in FIG. 5 602-616 Steps in FIG. 6 702-716 Steps in FIG. 7 802-816 Steps in FIG. 8 900 - 970 Structure of the apparatus of FIG. 3 1010-1030 Phases of the method in FIG. 10 1110-1130 Phases of the method in FIG. 11 1210-1220 Phases of the method in FIG. 12

Claims

1. An apparatus comprising at least one processor and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to:- determine, when the apparatus is in a Radio Resource Control, RRC, connected mode during a first RRC connection period, an identifier of at least one configuration related to a machine learning model used or to be used by the apparatus in the RRC connected mode;- transition from the RRC connected mode to an RRC inactive mode or an RRC idle mode; and- transmit to a wireless network node, when the apparatus is transitioning or attempting to transition from the RRC inactive mode or the RRC idle mode to the RRC connected mode for a second RRC connection period, the identifier of the at least one configuration related to the machine learning model via RRC signalling.

2. The apparatus according to claim 1, wherein the machine learning model is associated with at least one of a measurement or a data collection configuration.

3. The apparatus according to claim 1 or claim 2, wherein the at least one processor and the at least one memory further cause the apparatus at least to:- store the identifier of the at least one configuration related to the machine learning model when the apparatus is in the RRC connected mode during the first RRC connection period.

4. The apparatus according to claim 3, wherein the at least one processor and the at least one memory further cause the apparatus at least to:- store the identifier of the at least one configuration related to the machine learning model as a transient model identifier.

5. The apparatus according to claim 3 or claim 4, wherein the at least one processor and the at least one memory further cause the apparatus at least to:- determine that the apparatus is configured to store the identifier of the at least one configuration related to the machine learning model.

6. The apparatus according to any of the preceding claims, wherein the apparatus is in the RRC connected mode first with a first wireless network node during the first RRC connection period and transitioning or attempting to transition to the RRC connected mode with a second wireless network node for the RRC second connection period, wherein the first wireless network node is different than the second wireless network node.

7. The apparatus according to any of the preceding claims, wherein the at least one processor and the at least one memory further cause the apparatus at least to:- transmit an indication to the wireless network node via RRC signalling, said indication indicating whether the apparatus prefers re-using the identifier of the at least one configuration related to the machine learning model or initializing identification of the machine learning model.

8. The apparatus according to any of the preceding claims, wherein the at least one processor and the at least one memory further cause the apparatus at least to:- transmit to the wireless network node via RRC signalling information about at least one condition which has changed between the first and second RRC connection periods.

9. The apparatus according to any of the preceding claims, wherein the at least one processor and the at least one memory further cause the apparatus at least to:- receive via RRC signalling, in response to transmitting the identifier of the at least one configuration related to the machine learning model, information indicating whether the apparatus is to use the at least one configuration related to the machine learning model corresponding to the identifier during the second RRC connection period.

10. The apparatus according to claim 9, wherein the at least one processor and the at least one memory further cause the apparatus at least to:- determine, when said information indicates that the apparatus is to use the at least one configuration related to the machine learning model corresponding to theidentifier during the second RRC connection period, whether the identifier of the at least one configuration related to the machine learning model is valid; or- discard, when said information indicates that the apparatus is not to use the at least one configuration related to the machine learning model corresponding to the identifier during the second RRC connection period, the identifier of the at least one configuration related to the machine learning model.

11. The apparatus according to claim 9 or claim 10, wherein said information indicating whether the apparatus is to use the at least one configuration related to the machine learning model corresponding to the identifier during the second RRC connection period is received in an RRC resume message or an RRC setup message.

12. The apparatus according to any of claims 1 to 8, wherein the at least one processor and the at least one memory further cause the apparatus at least to:- receive, in response to transmitting the identifier of the at least one configuration related to the machine learning model, an RRC release message; and- discard the at least one configuration associated with the machine learning model corresponding to the identifier.

13. The apparatus according to any of the preceding claims, wherein the at least one processor and the at least one memory further cause the apparatus at least to:- transmit via RRC signalling, to the wireless network node, an indication about whether the at least one configuration related to the machine learning model has been used by the apparatus during the first RRC connection period.

14. An apparatus comprising at least one processor and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to:- receive via RRC signalling from a user equipment, when the user equipment is transitioning or attempting to transition from an RRC inactive mode or an RRC idle mode to a RRC connected mode for a second RRC connection period, an identifier of at least one configuration related to a machine learning model that was used or to be used by the user equipment during a first RRC connection period before the second RRC connection period; and- evaluate whether the at least one configuration related to the machine learning model is suitable to be used by at least one of the user equipment in the RRC connected mode during the second RRC connection period or by the apparatus while the user equipment is in the RRC connected mode during the second RRC connection period.

15. The apparatus according to claim 14, wherein the machine learning model is associated with at least one of a measurement or a data collection configuration.

16. The apparatus according to claim 14 or claim 15, wherein the at least one processor and the at least one memory further cause the apparatus at least to:- store the identifier of the at least one configuration related to the machine learning model during the first RRC connection period.

17. The apparatus according to any of claims 14 to 16, wherein the at least one processor and the at least one memory further cause the apparatus at least to:- receive via RRC signalling an indication from the user equipment, said indication indicating whether the user equipment prefers re-using the at least one configuration related to the machine learning model or initializing identification of the machine learning model; and- evaluate, based at least partly on said indication, whether the at least one configuration related to the machine learning model is to be used by the user equipment during the second RRC connection period.

18. The apparatus according to any of claims 14 to 17, wherein the at least one processor and the at least one memory further cause the apparatus at least to:- receive via RRC signalling, from the user equipment, information about at least one condition which has changed between the first and second RRC connection periods; and- evaluate, based at least partly on said information, whether the at least one configuration related to the machine learning model associated with the identifier is to be used by the user equipment during the second RRC connection period.

19. The apparatus according to any of claims 14 to 18, wherein the at least one processor and the at least one memory further cause the apparatus at least to:- transmit to the user equipment via RRC signalling, based on the evaluation, information indicating whether the user equipment is to use the at least one configuration related to the machine learning model corresponding to the identifier during the second RRC connection period.

20. The apparatus according to claim 19, wherein said information indicating whether the user equipment is to use the at least one configuration related to the machine learning model corresponding to the identifier during the second RRC connection period is transmitted in an RRC resume message or an RRC setup message.

21. The apparatus according to any of claims 14 to 18, wherein the at least one processor and the at least one memory further cause the apparatus at least to:- transmit to the user equipment, based on the evaluation, an RRC release message.

22. The apparatus according to any of claims 14 to 21, wherein the at least one processor and the at least one memory further cause the apparatus at least to:- receive via RRC signalling, from the user equipment, an indication about whether the at least one configuration related to the machine learning model has been used by the user equipment during the first RRC connection period.

23. The apparatus according to any of claims 14 to 21, wherein the at least one processor and the at least one memory further cause the apparatus at least to:- discard the at least one configuration related to the machine learning model when the evaluation indicates that the at least one configuration related to the machine learning model is not suitable to be used by at least one of the user equipment or the apparatus during the second RRC connection period.

Images