Managing ML processing model
A UE may indicate, to a wireless network, a capability to support one or more versions of an ML processing model, and a wireless network entity of the wireless network may configure the UE to apply a version of the ML processing model that is supported by the UE. The indication from the UE may indicate the capability to support one or more of a first version of the ML processing model that is common across multiple wireless networks and at least one alternate version of the ML processing model, such as a version that is specific to a network. The wireless network entity may configure at least one of a network specific algorithm structure or one or more weights for an network specific machine learning model that is supported by the UE.
1 . A method of wireless communication at a user equipment (UE), comprising:
transmitting, in a first wireless signal to a wireless network, an indication of a capability of the UE to support one or more versions of a machine learning (ML) processing model, wherein the indication indicates whether the UE supports one or more of a first version of the ML processing model that is common across multiple wireless networks and whether the UE supports at least one alternative version of the ML processing model that is optimized for the wireless network; and
receiving, in a second wireless signal, a configuration from the wireless network to apply a version of the ML processing model supported by the UE, wherein the configuration indicates the version of the ML processing model determined based on the transmitted indication of the capability of the UE to support the one or more versions of the ML processing model,
wherein the configuration configures the UE to use the first version of the ML processing model when the at least one alternative version of the ML processing model is not supported by the UE and configures the UE to use the at least one alternative version of the ML processing model when the at least one alternative version is supported by the UE.
2 . The method of claim 1 , wherein each alternative ML processing model includes at least one of a network specific algorithm structure or one or more network specific weights.
3 . The method of claim 1 , wherein the UE indicates a version identifier of each alternative ML processing model supported by the UE.
4 . The method of claim 1 , wherein the UE indicates a Public Land Mobile Network (PLMN) identifier (ID) associated with each alternative ML processing model supported by the UE.
5 . The method of claim 1 , wherein the configuration that the UE receives from the wireless network indicates a version identifier for the ML processing model configured for the UE.
6 . The method of claim 1 , wherein the configuration that the UE receives from the wireless network indicates a Public Land Mobile Network (PLMN) identifier (ID) associated with an alternative ML processing model that is configured for the UE.
7 . The method of claim 1 , wherein the configuration that the UE receives from the wireless network indicates at least one of a network specific algorithm structure or one or more weights for an alternative ML processing model that is supported by the UE.
8 . The method of claim 1 , wherein the ML processing model is for processing at least one of sensor data or wireless channel measurements.
9 . The method of claim 1 , wherein the UE indicates, to the wireless network, a scenario for which the UE supports the one or more versions of the ML processing model.
10 . An apparatus for wireless communication at a user equipment (UE), comprising:
a memory; and
one or more processors coupled to the memory and configured to cause the UE to:
transmit, in a first wireless signal to a wireless network, an indication of a capability of the UE to support one or more versions of a machine learning (ML) processing model, wherein the indication indicates whether the UE supports one or more of a first version of the ML processing model that is common across multiple wireless networks and whether the UE supports at least one alternative version of the ML processing model that is optimized for the wireless network; and
receive, in a second wireless signal, a configuration from the wireless network to apply a version of the ML processing model supported by the UE, wherein the configuration indicates the version of the ML processing model determined based on the transmitted indication of the capability of the UE to support the one or more versions of the ML processing model,
wherein the configuration configures the UE to use the first version of the ML processing model when the at least one alternative version of the ML processing model is not supported by the UE and configures the UE to use the at least one alternative version of the ML processing model when the at least one alternative version is supported by the UE.
11 . The apparatus of claim 10 , wherein each alternative ML processing model includes at least one of a network specific algorithm structure or one or more network specific weights.
12 . The apparatus of claim 10 , wherein the UE indicates a version identifier of each alternative ML processing model supported by the UE.
13 . The apparatus of claim 10 , wherein the UE indicates a Public Land Mobile Network (PLMN) identifier (ID) associated with each alternative ML processing model supported by the UE.
14 . The apparatus of claim 10 , wherein the configuration that the UE receives from the wireless network indicates a version identifier for the ML processing model configured for the UE.
15 . The apparatus of claim 10 , wherein the configuration that the UE receives from the wireless network indicates a Public Land Mobile Network (PLMN) identifier (ID) associated with an alternative ML processing model that is configured for the UE.
16 . The apparatus of claim 10 , wherein the configuration that the UE receives from the wireless network indicates at least one of a network specific algorithm structure or one or more weights for an alternative ML processing model that is supported by the UE.
17 . The apparatus of claim 10 , wherein the ML processing model is for processing at least one of sensor data or wireless channel measurements.
18 . The apparatus of claim 10 , wherein the one or more processors are further configured to cause the UE to indicate, to the wireless network, a scenario for which the UE supports the one or more versions of the ML processing model.
19 . A method of wireless communication at a wireless network entity, comprising:
receiving an indication of a capability for a user equipment (UE) to support one or more versions of a machine learning (ML) processing model, wherein the indication indicates whether the UE supports one or more of a first version of the ML processing model that is common across multiple wireless networks and whether the UE supports at least one alternative version of the ML processing model that is optimized for a wireless network;
determining a version of the ML processing model to configure the UE to apply based on the indication of the capability received from the UE, wherein a first determination determines to configure the UE to use the first version of the ML processing model when the at least one alternative version of the ML processing model is not supported by the UE and a second determination determines to configure the UE to use the at least one alternative version of the ML processing model when the at least one alternative version is supported by the UE; and
configuring the UE to apply the version of the ML processing model that is supported by the UE, wherein the configuration configures the UE to use the first version of the ML processing model when the at least one alternative version of the ML processing model is not supported by the UE and configures the UE to use the at least one alternative version of the ML processing model when the at least one alternative version is supported by the UE.
20 . The method of claim 19 , wherein each alternative ML processing model includes at least one of a network specific algorithm structure or one or more network specific weights.
21 . The method of claim 19 , wherein the indication indicates at least one of a version identifier of each alternative ML processing model supported by the UE or a Public Land Mobile Network (PLMN) identifier (ID) associated with each alternative ML processing model supported by the UE.
22 . The method of claim 19 , further comprising indicating at least one of:
a version identifier for the ML processing model configured for the UE, or
a Public Land Mobile Network (PLMN) identifier (ID) associated with an alternative ML processing model that is configured for the UE.
23 . The method of claim 19 , further comprising configuring at least one of a network specific algorithm structure or one or more weights for an alternative ML processing model that is supported by the UE.
24 . An apparatus for wireless communication at a wireless network entity, comprising:
memory; and
one or more processors coupled to the memory and configured to cause the wireless network entity to:
receive an indication of a capability for a user equipment (UE) to support one or more versions of a machine learning (ML) processing model, wherein the indication indicates whether the UE supports one or more of a first version of the ML processing model that is common across multiple wireless networks and whether the UE supports at least one alternative version of the ML processing model that is optimized for a wireless network;
determine a version of the ML processing model to configure the UE to apply based on the indication of the capability received from the UE, wherein a first determination determines to configure the UE to use the first version of the ML processing model when the at least one alternative version of the ML processing model is not supported by the UE and a second determination determines to configure the UE to use the at least one alternative version of the ML processing model when the at least one alternative version is supported by the UE; and
configure the UE to apply the version of the ML processing model that is supported by the UE, wherein the configuration configures the UE to use the first version of the ML processing model when the at least one alternative version of the ML processing model is not supported by the UE and configures the UE to use the at least one alternative version of the ML processing model when the at least one alternative version is supported by the UE.
25 . The apparatus of claim 24 , wherein each alternative ML processing model includes at least one of a network specific algorithm structure or one or more network specific weights.
26 . The apparatus of claim 24 , wherein the indication indicates a version identifier of each alternative ML processing model supported by the UE.
27 . The apparatus of claim 24 , wherein the indication indicates a Public Land Mobile Network (PLMN) identifier (ID) associated with each alternative ML processing model supported by the UE.
28 . The apparatus of claim 24 , wherein the one or more processors are configured to cause the wireless network entity to indicate a version identifier for the ML processing model configured for the UE.
29 . The apparatus of claim 24 , wherein the one or more processors are configured to cause the wireless network entity to indicate a Public Land Mobile Network (PLMN) identifier (ID) associated with an alternative ML processing model that is configured for the UE.
30 . The apparatus of claim 24 , wherein the one or more processors are configured to cause the wireless network entity to configure at least one of a network specific algorithm structure or one or more weights for an alternative ML processing model that is supported by the UE.
31 . The apparatus of claim 24 , wherein the ML processing model is for processing at least one of sensor data or wireless channel measurements.
32 . The apparatus of claim 24 , wherein the indication indicates to the wireless network entity a scenario for which the UE supports the one or more versions of the ML processing model.