Machine learning component management in federated learning

1. An apparatus for wireless communication at a client device, comprising:

a memory; and

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

receive, using at least one lower layer of a wireless communication protocol stack, a machine learning component from a server device;

transmit, to the server device and using the at least one lower layer, an update associated with the machine learning component, wherein transmitting the update comprises transmitting a plurality of transport blocks; and

receive, using one or more lower layers of the wireless communication protocol stack, a plurality of data packets associated with a globally updated machine learning component from the server device, the globally updated machine learning component being segmented into an additional plurality of transport blocks using one or more upper layers of a wireless communication protocol stack of the server device.

2. The apparatus of claim 1 , wherein the machine learning component comprises at least one neural network model.

3. The apparatus of claim 1 , wherein the at least one lower layer comprises at least one of a physical layer or a medium access control layer.

4. The apparatus of claim 1 , wherein the one or more processors are further configured to prepare the update for transmission using an upper layer of the wireless communication protocol stack.

5. The apparatus of claim 4 , wherein the wireless communication protocol stack comprises a user plane protocol stack.

6. The apparatus of claim 5 , wherein the upper layer comprises at least one of a radio link control layer, a packet data convergence protocol layer, or a service data adaptation protocol layer.

7. The apparatus of claim 4 , wherein the wireless communication protocol stack comprises a control plane protocol stack.

8. The apparatus of claim 7 , wherein the upper layer comprises at least one of a radio link control layer, a packet data convergence protocol layer, a radio resource control layer, or a non-access stratum layer.

9. The apparatus of claim 1 , wherein the one or more processors are further configured to prepare the update for transmission using the at least one lower layer of the wireless communication protocol stack.

10. The apparatus of claim 1 , wherein the one or more processors are further configured to:

collect a set of training data; and

determine the update associated with the machine learning component by training the machine learning component using the set of training data.

11. The apparatus of claim 10 , wherein the update comprises an updated machine learning component.

12. The apparatus of claim 1 , wherein the one or more processors are further configured to segment the update into the plurality of transport blocks using one or more upper layers of the wireless communication protocol stack.

13. The apparatus of claim 12 , wherein the one or more processors are further configured to:

generate, based at least in part on the plurality of transport blocks and using the one or more upper layers of the wireless communication protocol stack, an additional plurality of data packets; and

prepare the additional plurality of data packets for transmission based at least in part on a specified quality of service.

14. The apparatus of claim 1 , wherein the one or more processors are further configured to:

obtain the globally updated machine learning component from the plurality of data packets using one or more upper layers of the wireless communication protocol stack; and

configure the globally updated machine learning component to a physical layer of the wireless communication protocol stack.

15. The apparatus of claim 1 , wherein the client device comprises a user equipment, and wherein the server device comprises a base station.

16. An apparatus for wireless communication at a server device, comprising:

a memory; and

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

transmit, using at least one lower layer of a wireless communication protocol stack, a machine learning component to a client device;

receive, from the client device and using the at least one lower layer, an update associated with the machine learning component, wherein receiving the update comprises receiving a plurality of transport blocks; and

transmit, using one or more lower layers of the wireless communication protocol stack, a plurality of data packets associated with a globally updated machine learning component to the client device, the globally updated machine learning component being segmented into an additional plurality of transport blocks using one or more upper layers of the wireless communication protocol stack.

17. The apparatus of claim 16 , wherein the machine learning component comprises at least one neural network.

18. The apparatus of claim 16 , wherein the at least one lower layer comprises at least one of a physical layer or a medium access control layer.

19. The apparatus of claim 16 , wherein to receive the update, the one or more processors is configured to receive an additional plurality of data packets, and wherein the one or more processors is further configured to:

obtain the update from the additional plurality of data packets using the one or more upper layers of the wireless communication protocol stack; and

configure the update to a physical layer of the wireless communication protocol stack.

20. The apparatus of claim 19 , wherein the wireless communication protocol stack comprises a user plane protocol stack, and wherein the one or more upper layers comprise at least one of a radio link control layer, a packet data convergence protocol layer, or a service data adaptation protocol layer.

21. The apparatus of claim 19 , wherein the wireless communication protocol stack comprises a control plane protocol stack, and wherein the one or more upper layers comprise at least one of a radio link control layer, a packet data convergence protocol layer, a radio resource control layer, or a non-access stratum layer.

22. The apparatus of claim 16 , wherein the one or more processors are further configured to:

receive, from at least one additional client device, at least one additional update corresponding to the machine learning component; and

aggregate the update and the at least one additional update to generate the globally updated machine learning component.

23. The apparatus of claim 22 , wherein the one or more processors, to aggregate the update and the at least one additional update, are configured to average the update and the at least one additional update.

24. The apparatus of claim 16 , wherein the one or more processors are further configured to:

generate, based at least in part on the additional plurality of transport blocks and using the one or more upper layers of the wireless communication protocol stack, the plurality of data packets; and

prepare the plurality of data packets for transmission based at least in part on a specified quality of service.

25. The apparatus of claim 16 , wherein the one or more processors are further configured to prepare the globally updated machine learning component for transmission using the at least one lower layer of the wireless communication protocol stack.

26. A method of wireless communication performed by a client device, comprising:

receiving, using at least one lower layer of a wireless communication protocol stack, a machine learning component from a server device;

transmitting, to the server device and using the at least one lower layer, an update associated with the machine learning component, wherein transmitting the update comprises transmitting a plurality of transport blocks; and

receiving, using one or more lower layers of the wireless communication protocol stack, a plurality of data packets associated with a globally updated machine learning component from the server device, the globally updated machine learning component being segmented into an additional plurality of transport blocks using one or more upper layers of a wireless communication protocol stack of the server device.

27. A method of wireless communication performed by a server device, comprising:

transmitting, using at least one lower layer of a wireless communication protocol stack, a machine learning component to a client device;

receiving, from the client device and using the at least one lower layer, an update associated with the machine learning component, wherein receiving the update comprises receiving a plurality of transport blocks; and

transmitting, using one or more lower layers of the wireless communication protocol stack, a plurality of data packets associated with a globally updated machine learning component to the client device, the globally updated machine learning component being segmented into an additional plurality of transport blocks using one or more upper layers of the wireless communication protocol stack.

28. The method of claim 26 , further comprising:

collecting a set of training data; and

determining the update associated with the machine learning component by training the machine learning component using the set of training data.

29. The method of claim 28 , wherein the update comprises an updated machine learning component.

30. The method of claim 27 , further comprising:

receiving, from at least one additional client device, at least one additional update corresponding to the machine learning component; and

aggregating the update and the at least one additional update to generate the globally updated machine learning component.