Electronic devices and methods used in associative learning

Abstract

An electronic device and method for use in associative learning are provided. [Solution] A federated learning process (300) in a central processing unit determines (S305) a set of distributed nodes from a plurality of distributed nodes for generating global model parameters. Correlations between the local training data of the set of distributed nodes satisfy specific correlation requirements, and channel resources (e.g., time resources, frequency resources, and / or space resources) are obtained for transmitting the local model parameters to the central processing unit. Global model parameters are generated (S310) based on the local model parameters of the set of distributed nodes. The local model parameters are generated by the set of distributed nodes based on their respective local training data. The central processing unit then distributes the generated global model parameters to the plurality of distributed nodes (S315).

Claims

1. An electronic device used for federative learning in a central processing unit, comprising a processing circuit, wherein the processing circuit is Generating global model parameters based on local model parameters of a set of distributed nodes, wherein the local model parameters are generated by the set of distributed nodes based on their respective local training data. Of the multiple distributed nodes, the set of distributed nodes for generating the global model parameters is determined based on the correlation between the set of distributed nodes. The aforementioned correlation is an electronic device that satisfies a specific correlation requirement.

2. The aforementioned specific correlation requirement is, The electronic device according to claim 1, further comprising the condition that the distance between the node locations of the set of distributed nodes is greater than a predetermined distance threshold.

3. Of the plurality of distributed nodes, the set of distributed nodes for generating global model parameters is determined based on the fact that one or more of the parameters of the distributed nodes satisfy a specific requirement. The electronic device according to claim 1, wherein the one or more parameters include the location of the set of distributed nodes, the network performance parameters of the set of distributed nodes, the time taken to collect the local training data, a generative model, or a data source.

4. The electronic device according to claim 3, wherein the specific requirement for the location of the set of distributed nodes includes that the distributed nodes are within the coverage area of ​​the central processing unit.

5. The electronic device according to claim 3, wherein the network performance parameter of the set of distributed nodes is QoS.

6. An electronic device used for federative learning in a central processing unit, comprising a processing circuit, the processing circuit is: Generating global model parameters based on local model parameters of a set of distributed nodes, wherein the local model parameters are generated by the set of distributed nodes based on their respective local training data. An electronic device in which, among a plurality of distributed nodes, the set of distributed nodes for generating the global model parameters is determined based on the fact that the distance between the node locations of the set of distributed nodes is greater than a predetermined distance threshold.

7. A method used for associative learning in a central processing unit, The process involves determining a set of distributed nodes from among multiple distributed nodes to be used to generate global model parameters, wherein the correlation between the set of distributed nodes satisfies a specific correlation requirement. A method comprising generating global model parameters based on local model parameters of a set of distributed nodes, wherein the local model parameters are generated by the set of distributed nodes based on their respective local training data.

8. An electronic device used for federative learning in a central processing unit, comprising a processing circuit, the processing circuit is: The process involves determining a set of distributed nodes from among multiple distributed nodes to generate global model parameters, wherein the local model parameters are generated by the set of distributed nodes based on their respective local training data. The correlation between the aforementioned set of distributed nodes satisfies a specific correlation requirement in an electronic device.

9. The specific correlation requirement is, The electronic device according to claim 8, further comprising the condition that the distance between the node locations of the set of distributed nodes is greater than a predetermined distance threshold.

10. Of the plurality of distributed nodes, the set of distributed nodes for generating global model parameters is determined on the basis that one or more of the parameters of the distributed nodes satisfy a specific requirement. The electronic device according to claim 8, wherein the one or more parameters include the location of the set of distributed nodes, the network performance parameters of the set of distributed nodes, the time taken to collect the local training data, a generative model, or a data source.

11. The electronic device according to claim 10, wherein the specific requirement for the location of the set of distributed nodes includes that the distributed nodes are located within the coverage area of ​​the central processing unit.

12. The electronic device according to claim 10, wherein the network performance parameter of the set of distributed nodes is QoS.

13. An electronic device used for federative learning in a central processing unit, comprising a processing circuit, the processing circuit is: The process involves determining a set of distributed nodes from among multiple distributed nodes to generate global model parameters, wherein the local model parameters are generated by the set of distributed nodes based on their respective local training data. An electronic device in which, among a plurality of distributed nodes, the set of distributed nodes for generating the global model parameters is determined based on the fact that the distance between the node locations of the set of distributed nodes is greater than a predetermined distance threshold.

14. A method used for associative learning in a central processing unit, The process involves determining a set of distributed nodes from among multiple distributed nodes to generate global model parameters, wherein the local model parameters are generated by the set of distributed nodes based on their respective local training data. A method by which the correlation between the aforementioned pair of distributed nodes satisfies a specific correlation requirement.

15. An electronic device used for associative learning, comprising a processing circuit, the processing circuit is The determination that a particular distributed node will be used to generate global model parameters, wherein the correlation between the particular distributed node and other distributed nodes for generating the global model parameters satisfies a specific correlation requirement. Electronic device that transmits local model parameters of a particular distributed node to a central processing unit, wherein the local model parameters are arranged to be generated by the particular distributed node based on its local training data.

16. The specific correlation requirement is, The electronic device according to claim 15, wherein the distance between the node locations of a set of distributed nodes is greater than a predetermined distance threshold.

17. The determination that the particular distributed node is used to generate global model parameters means that Transmitting one or more parameters of the aforementioned specific distributed node to the central processing unit, wherein the one or more parameters may be used to determine the correlation between the aforementioned specific distributed node and other distributed nodes that have been determined to be used for generating the global model parameters, The electronic device according to claim 15, wherein the one or more parameters include the location of a set of distributed nodes, the network performance parameters of the set of distributed nodes, the time taken to collect the local training data, a generative model, or a data source.

18. The electronic device according to claim 17, wherein a specific requirement for the location of the set of distributed nodes includes that the distributed nodes are located within the coverage area of ​​the central processing unit.

19. The electronic device according to claim 17, wherein the network performance parameter of the set of distributed nodes is QoS.

20. The electronic device according to claim 15, wherein the particular distributed node is determined to be used to generate the global model parameters based on an instruction from the central processing unit to upload local model parameters.

21. Obtaining a channel resource for transmitting local model parameters to the central processing unit, The electronic device according to claim 15, further comprising determining, after successfully acquiring the channel resources, that the particular distributed node is to be used to generate global model parameters.

22. An electronic device used for associative learning, comprising a processing circuit, the processing circuit is The determination that a particular distributed node will be used to generate global model parameters, wherein the distance between the particular distributed node and other distributed nodes that have been determined to be used to generate the global model parameters is greater than a predetermined distance threshold. Electronic device that transmits local model parameters of a particular distributed node to a central processing unit, wherein the local model parameters are arranged to be generated by the particular distributed node based on its local training data.

23. A method used for federative learning on a specific distributed node, The determination that a particular distributed node is used to generate global model parameters, wherein the correlation between the particular distributed node and other distributed nodes determined to be used to generate the global model parameters satisfies a specific correlation requirement. A method comprising transmitting local model parameters of a particular distributed node to a central processing unit, wherein the local model parameters are generated by the particular distributed node based on its local training data.

Images