Online training and augmentation of neural networks for channel state feedback

1. A method for wireless communications at a user equipment (UE), comprising:

transmitting, to one or more network entities, a request for a plurality of sets of neural network weights corresponding to a plurality of neural networks, wherein transmitting the request for the plurality of sets of neural network weights is based at least in part on a change in one or more channel conditions at the UE;

receiving the plurality of sets of neural network weights in response to the request;

selecting, using a first neural network at the UE, a set of neural network weights from the plurality of sets of neural network weights, the selected set of neural network weights corresponding to a second neural network of the plurality of neural networks;

updating the selected set of neural network weights corresponding to the second neural network according to one or more measurements identified at the UE; and

transmitting, to the one or more network entities, an identifier corresponding to the second neural network and the updated selected set of neural network weights corresponding to the second neural network.

2. The method of claim 1 , wherein the one or more measurements comprise a plurality of measurements corresponding to a plurality of base stations, a plurality of sensors at the UE, a plurality of radio access technologies, or a combination thereof.

3. The method of claim 1 , further comprising:

identifying a network entity of the one or more network entities is a target network entity in a handover procedure based at least in part on selecting the set of neural network weights;

receiving, from the network entity and during a time period, one or more reference signals; and

determining the one or more measurements based at least in part on the one or more reference signals.

4. The method of claim 1 , the transmitting the request for the plurality of sets of neural network weights comprises:

identifying that the one or more channel conditions fail to satisfy a threshold value.

5. The method of claim 4 , further comprising:

receiving, from the one or more network entities, an indication that the one or more channel conditions fail to satisfy the threshold value.

6. The method of claim 4 , further comprising:

identifying a timer associated with transmitting the request for the plurality of sets of neural network weights has expired.

7. The method of claim 1 , further comprising:

receiving the plurality of sets of neural network weights from the one or more network entities, the one or more network entities comprising a serving network entity, a neighboring network entity, or a combination thereof.

8. The method of claim 1 , the transmitting the request for the plurality of sets of neural network weights comprises:

identifying the plurality of sets of neural network weights based at least in part on a location of the UE, a channel type, a UE antenna configuration, a line of sight condition, or a combination thereof.

9. The method of claim 1 , further comprising:

receiving, from the one or more network entities, scheduling information corresponding to the identifier; and

transmitting the identifier during a set of resources according to the scheduling information.

10. A method for wireless communications at a network entity, comprising:

receiving, from a user equipment (UE), a request for one or more sets of neural network weights corresponding to one or more neural networks, wherein receiving the request for the one or more sets of neural network weights is based at least in part on a change in one or more channel conditions at the UE;

transmitting, to the UE, the one or more sets of neural network weights in response to the request; and

receiving, from the UE, an identifier corresponding to a neural network of the one or more neural networks for training data and an updated set of neural network weights corresponding to the neural network, the updated set of neural network weights being in accordance with one or more measurements.

11. The method of claim 10 , further comprising:

identifying the network entity is a target network entity in a handover procedure; and

transmitting, to the UE and during a time period, one or more reference signals based at least in part on transmitting the one or more sets of neural network weights.

12. The method of claim 10 , further comprising:

transmitting, to the UE, scheduling information corresponding to the identifier; and

receiving the identifier during a set of resources according to the scheduling information.

13. The method of claim 10 , further comprising:

storing the updated set of neural network weights corresponding to the neural network.

14. The method of claim 10 , the receiving the request for the one or more sets of neural network weights comprises:

transmitting, to the UE, an indication that the one or more channel conditions fail to satisfy a threshold value.

15. An apparatus for wireless communications at a user equipment (UE), comprising:

one or more processors; and

memory coupled with the one or more processors, the one or more processors are further configured to:

transmit, to one or more network entities, a request for a plurality of sets of neural network weights corresponding to a plurality of neural networks, wherein transmitting the request for the plurality of sets of neural network weights is based at least in part on a change in one or more channel conditions at the UE;

receive the plurality of sets of neural network weights in response to the request;

select, using a first neural network, a set of neural network weights from the plurality of sets of neural network weights, the selected set of neural network weights corresponding to a second neural network of the plurality of neural networks;

update the selected set of neural network weights corresponding to the second neural network according to one or more measurements identified at the UE; and

transmit, to the one or more network entities, an identifier corresponding to the second neural network and the updated selected set of neural network weights corresponding to the second neural network.

16. The apparatus of claim 15 , wherein the one or more measurements comprise a plurality of measurements corresponding to a plurality of base stations, a plurality of sensors at the UE, a plurality of radio access technologies, or a combination thereof.

17. The apparatus of claim 15 , wherein the one or more processors are configured to:

identify a network entity of the one or more network entities is a target network entity in a handover procedure based at least in part on selecting the set of neural network weights;

receive, from the network entity and during a time period, one or more reference signals; and

determine the one or more measurements based at least in part on the one or more reference signals.

18. The apparatus of claim 15 , wherein, to transmit the request for the plurality of sets of neural network weights, the one or more processors are configured to:

identify that the one or more channel conditions fail to satisfy a threshold value.

19. The apparatus of claim 18 , wherein the one or more processors are configured to:

receive, from the one or more network entities, an indication that the one or more channel conditions fail to satisfy the threshold value.

20. The apparatus of claim 18 , wherein the one or more processors are configured to:

identify a timer associated with transmitting the request for the plurality of sets of neural network weights has expired.

21. The apparatus of claim 15 , wherein the one or more processors are configured to:

receive the plurality of sets of neural network weights from the one or more network entities, the one or more network entities comprising a serving network entity, a neighboring network entity, or a combination thereof.

22. The apparatus of claim 15 , wherein, to transmit the request for the plurality of sets of neural network weights, the one or more processors are configured to:

identify the plurality of sets of neural network weights based at least in part on a location of the UE, a channel type, a UE antenna configuration, a line of sight condition, or a combination thereof.

23. The apparatus of claim 15 , wherein the one or more processors are configured to:

receive, from the one or more network entities, scheduling information corresponding to the identifier; and

transmit the identifier during a set of resources according to the scheduling information.

24. An apparatus for wireless communications at a network entity, comprising:

one or more processors; and

memory coupled with the one or more processors, the one or more processors configured to:

receive, from a user equipment (UE), a request for one or more sets of neural network weights corresponding to one or more neural networks, wherein receiving the request for the one or more sets of neural network weights is based at least in part on a change in one or more channel conditions at the UE;

transmit, to the UE, the one or more sets of neural network weights in response to the request; and

receive, from the UE, an identifier corresponding to a neural network of the one or more neural networks for training data and an updated set of neural network weights corresponding to the neural network, the updated set of neural network weights being in accordance with one or more measurements.

25. The apparatus of claim 24 , wherein the one or more processors are configured to:

identify the network entity is a target network entity in a handover procedure; and

transmit, to the UE and during a time period, one or more reference signals based at least in part on transmitting the one or more sets of neural network weights.

26. The apparatus of claim 24 , wherein the one or more processors are configured to:

transmit, to the UE, scheduling information corresponding to the identifier; and

receive the identifier during a set of resources according to the scheduling information.