Monitoring and updating machine learning models

Patent metadata

Field	Value
Publication number	US-12563425-B2
Application number	US-202217956200-A
Country	US
Kind code	B2
Filing date	Sep 29, 2022
Priority date	Sep 29, 2022
Publication date	Feb 24, 2026
Grant date	Feb 24, 2026

Abstract

Methods, systems, and devices for wireless communications are described. The method may include a user equipment (UE) may receive a control signal indicating an event trigger for reporting a performance parameter associated with a machine learning model. Further, the UE may receive one or more signals indicating input data for monitoring a performance of the machine learning model by the UE. Upon detecting the event trigger, the UE may transmit a report comprising the performance parameter.

First claim

What is claimed is:

1 . A method for wireless communications at a user equipment (UE), comprising:
receiving a control signal indicating a configuration for reporting a performance parameter associated with a machine learning model that is implemented at the UE; receiving one or more signals indicating input data for monitoring, by the UE, a performance of the machine learning model that is implemented at the UE; and transmitting a report comprising the performance parameter, wherein the performance parameter is based at least in part on monitoring the performance of the machine learning model and on a comparison between the input data and output data of the machine learning model.

2 . The method of claim 1 , wherein transmitting the report comprises:
transmitting the report based at least in part on detecting an event trigger, wherein the event trigger comprises a change in one or more communication parameters associated with the UE, wherein the one or more communication parameters comprise a number of antennas used for communication between the UE and a node, active component carriers used for communication between the UE and the node, a location of the UE with respect to the node, an orientation of the UE, a velocity of the UE, network slicing, a quality of service flow, a session, or a combination thereof.

3 . The method of claim 1 , wherein transmitting the report comprises:
transmitting the report based at least in part on detecting an event trigger, wherein the event trigger comprises detecting that the performance parameter satisfies a threshold.

4 . The method of claim 1 , further comprising:
receiving a second control signal indicating one or more parameters associated with the machine learning model based at least in part on transmitting the report; and updating the machine learning model based at least in part on the one or more parameters associated with the machine learning model.

5 . The method of claim 1 , further comprising:
receiving a second control signal configuring the UE with a second machine learning model or activating the second machine learning model based at least in part on transmitting the report; and implementing the second machine learning model to perform one or more wireless communication procedures based at least in part the second control signal.

6 . The method of claim 1 , further comprising:
transmitting a second control signal indicating that the UE successfully received and implemented the control signal, wherein receiving the one or more signals is based at least in part on transmitting the second control signal.

7 . The method of claim 1 , wherein receiving the one or more signals comprises:
receiving a plurality of signals according to a periodicity.

8 . The method of claim 1 , wherein the input data comprises metadata corresponding to evaluating the performance of the machine learning model, ground truth for the machine learning model, one or more thresholds associated with the performance parameter, or a combination thereof.

9 . The method of claim 1 , wherein the performance parameter comprises a system key performance indicator or an interference key performance indicator.

10 . The method of claim 1 , wherein the report further comprises the input data and the output data of the machine learning model.

11 . A method for wireless communications at a node, comprising:
transmitting a control signal indicating a configuration for reporting a performance parameter associated with a machine learning model that is for implementation at a UE; transmitting one or more signals indicating input data for monitoring, by the UE, of a performance of the machine learning model that is for implementation at the UE; and receiving a report comprising the performance parameter, wherein the performance parameter is based at least in part on a comparison between the input data and output data of the machine learning model.

12 . The method of claim 11 , wherein transmitting the control signal indicating the configuration is based at least in part on an event trigger that comprises a change in one or more in one or more communication parameters associated with the node, the one or more communication parameters comprise a number of antennas used for communication between the UE and the node, active component carriers used for communication between the UE and the node, a location of the UE with respect to the node, an orientation of the UE, a velocity of the UE, network slicing, a quality of service flow, a session, or a combination thereof.

13 . The method of claim 11 , wherein transmitting the control signal indicating the configuration is based at least in part on an event trigger that comprises the performance parameter satisfying a threshold.

14 . The method of claim 11 , further comprising:
transmitting, based at least in part on receiving the report, the report to a second node.

15 . The method of claim 14 , further comprising:
generating a second performance parameter based at least in part on detecting an event trigger, wherein the report further comprises the second performance parameter.

16 . The method of claim 11 , further comprising:
receiving the control signal from a second node, wherein transmitting the control signal is based at least in part on receiving the control signal from the second node.

17 . The method of claim 11 , further comprising:
transmitting a second control signal indicating one or more parameters associated with the machine learning model based at least in part on receiving the report.

18 . The method of claim 11 , further comprising:
transmitting a second control signal configuring the UE with a second machine learning model or activating the second machine learning model based at least in part on receiving the report.

19 . The method of claim 11 , wherein transmitting the one or more signals comprises:
transmitting a plurality of signals according to a periodicity.

20 . The method of claim 1 , wherein implementing the machine learning model at the UE comprises using the input data and the machine learning model to generate the output data of the machine learning model at the UE.

21 . The method of claim 11 , wherein the node comprises a network entity, a central unit included in the network entity, a distributed unit associated with the network entity, an operations, administration, and maintenance server, a third party server, a machine learning model repository, a network data analytics function server, or a random access network intelligent controller.

22 . A method for wireless communication at a node, comprising:
receiving a first control signal indicating input data for monitoring, by the node, a performance of a machine learning model that is for implementation at a user equipment (UE); generating a performance parameter associated with the machine learning model based at least in part on detecting an event trigger, wherein the performance parameter is based at least in part on a comparison between the input data and output data of the machine learning model; and transmitting a second control signal configuring the UE with a second machine learning model that is for implementation at the UE or activating the second machine learning model that is for implementation at the UE based at least in part on the performance parameter.

23 . The method of claim 22 , wherein detecting the event trigger comprises:
detecting a

Assignees

• Qualcomm Inc

Inventors

• Kumar Rajeev
• Horn Gavin Bernard
• Gholmieh Aziz

Classifications

• H04W24/02

  Arrangements for optimising operational condition · CPC title

• H04L41/16

  using machine learning or artificial intelligence · CPC title

• H04W24/08 Primary

  Testing, {supervising or monitoring} using real traffic · CPC title

• G06N20/00

  Machine learning · CPC title

• H04W24/10 Primary

  Scheduling measurement reports {; Arrangements for measurement reports} · CPC title