Federated learning of clients
A method, a computer program product, and a computer system determine when to perform a federated learning process. The method includes identifying currently available contributors among contributors of a federated learning task for which the federated learning process is to be performed. The method includes determining a usefulness metric of the currently available contributors for respective datasets from each of the currently available contributors used in performing the federated learning process. The method includes, as a result of the usefulness metric of the currently available contributors being at least a usefulness threshold, generating a recommendation to perform the federated learning process with the datasets of the currently available contributors. The method includes transmitting the recommendation to a processing component configured to perform the federated learning process.
1. A computer-implemented method for determining when to perform a federated learning process, the method comprising:
identifying currently available contributors among contributors of a federated learning task for which the federated learning process is to be performed;
determining a usefulness metric of the currently available contributors for respective datasets from each of the currently available contributors used in performing the federated learning process;
as a result of the usefulness metric of the currently available contributors being at least a usefulness threshold, generating a recommendation to perform the federated learning process with the datasets of the currently available contributors; and
transmitting the recommendation to a processing component configured to perform the federated learning process.
2. The computer-implemented method of claim 1 , further comprising:
as a result of the usefulness metric of the currently available contributors being less than the usefulness threshold, generating a further recommendation to perform the federated learning process after a time duration.
3. The computer-implemented method of claim 2 , further comprising:
determining a further usefulness metric of currently unavailable contributors for respective further datasets from each of the currently unavailable contributors;
determining a predicted availability of select ones of the currently unavailable contributors who have the further usefulness metric being at least the usefulness threshold.
4. The computer-implemented method of claim 3 , wherein the time duration is based on the predicted availability of the currently unavailable contributors among the contributors.
5. The computer-implemented method of claim 3 , wherein the predicted availability is based on a temporal characteristic, a spatial characteristic, or a combination thereof associated with the currently unavailable contributors, the currently available contributors, or a combination thereof.
6. The computer-implemented method of claim 1 , wherein the usefulness metric is based on a contribution impact to the federated learning process, a complementary nature of the datasets to the federated learning process, a space associated with a global model generated by the federated learning process, a diversity of the datasets, characteristics of the federated learning task, or a combination thereof.
7. The computer-implemented method of claim 1 , wherein the usefulness metric includes a local usefulness metric, a global usefulness metric, or a combination thereof, and the usefulness threshold includes a local usefulness threshold, a global usefulness threshold, or a combination thereof, respectively.
8. A computer program product for determining when to perform a federated learning process, the computer program product comprising:
one or more non-transitory computer-readable storage media and program instructions collectively stored on the one or more non-transitory computer-readable storage media capable of performing a method, the method comprising:
identifying currently available contributors among contributors of a federated learning task for which the federated learning process is to be performed;
determining a usefulness metric of the currently available contributors for respective datasets from each of the currently available contributors used in performing the federated learning process;
as a result of the usefulness metric of the currently available contributors being at least a usefulness threshold, generating a recommendation to perform the federated learning process with the datasets of the currently available contributors; and
transmitting the recommendation to a processing component configured to perform the federated learning process.
9. The computer program product of claim 8 , wherein the method further comprises:
as a result of the usefulness metric of the currently available contributors being less than the usefulness threshold, generating a further recommendation to perform the federated learning process after a time duration.
10. The computer program product of claim 9 , wherein the method further comprises:
determining a further usefulness metric of currently unavailable contributors for respective further datasets from each of the currently unavailable contributors;
determining a predicted availability of select ones of the currently unavailable contributors who have the further usefulness metric being at least the usefulness threshold.
11. The computer program product of claim 10 , wherein the time duration is based on the predicted availability of the currently unavailable contributors among the contributors.
12. The computer program product of claim 10 , wherein the predicted availability is based on a temporal characteristic, a spatial characteristic, or a combination thereof associated with the currently unavailable contributors, the currently available contributors, or a combination thereof.
13. The computer program product of claim 8 , wherein the usefulness metric is based on a contribution impact to the federated learning process, a complementary nature of the datasets to the federated learning process, a space associated with a global model generated by the federated learning process, a diversity of the datasets, characteristics of the federated learning task, or a combination thereof.
14. The computer program product of claim 8 , wherein the usefulness metric includes a local usefulness metric, a global usefulness metric, or a combination thereof, and the usefulness threshold includes a local usefulness threshold, a global usefulness threshold, or a combination thereof, respectively.
15. A computer system for determining when to perform a federated learning process, the computer system comprising:
one or more computer processors, one or more computer-readable storage media, and program instructions collectively stored on the one or more of the computer-readable storage media for execution by at least one of the one or more processors capable of performing a method, the method comprising:
identifying currently available contributors among contributors of a federated learning task for which the federated learning process is to be performed;
determining a usefulness metric of the currently available contributors for respective datasets from each of the currently available contributors used in performing the federated learning process;
as a result of the usefulness metric of the currently available contributors being at least a usefulness threshold, generating a recommendation to perform the federated learning process with the datasets of the currently available contributors; and
transmitting the recommendation to a processing component configured to perform the federated learning process.
16. The computer system of claim 15 , wherein the method further comprises:
as a result of the usefulness metric of the currently available contributors being less than the usefulness threshold, generating a further recommendation to perform the federated learning process after a time duration.
17. The computer system of claim 16 , wherein the method further comprises:
determining a further usefulness metric of currently unavailable contributors for respective further datasets from each of the currently unavailable contributors;
determining a predicted availability of select ones of the currently unavailable contributors who have the further usefulness metric being at least the usefulness threshold.
18. The computer system of claim 17 , wherein the time duration is based on the predicted availability of the currently unavailable contributors among the contributors.
19. The computer system of claim 17 , wherein the predicted availability is based on a temporal characteristic, a spatial characteristic, or a combination thereof associated with the currently unavailable contributors, the currently available contributors, or a combination thereof.
20. The computer system of claim 15 , wherein the usefulness metric is based on a contribution impact to the federated learning process, a complementary nature of the datasets to the federated learning process, a space associated with a global model generated by the federated learning process, a diversity of the datasets, characteristics of the federated learning task, or a combination thereof.