View source for 18461410. ZONE GRADIENT DIFFUSION (ZGD) FOR ZONE-BASED FEDERATED LEARNING simplified abstract (QUALCOMM Incorporated)

=ZONE GRADIENT DIFFUSION (ZGD) FOR ZONE-BASED FEDERATED LEARNING=


==Organization Name==

[[:Category:QUALCOMM Incorporated|QUALCOMM Incorporated]]


[[Category:QUALCOMM Incorporated]]

==Inventor(s)==

[[:Category:Vijaya Datta Mayyuri of San Diego CA (US)|Vijaya Datta Mayyuri of San Diego CA (US)]][[Category:Vijaya Datta Mayyuri of San Diego CA (US)]]

[[:Category:An Chen of San Diego CA (US)|An Chen of San Diego CA (US)]][[Category:An Chen of San Diego CA (US)]]

==ZONE GRADIENT DIFFUSION (ZGD) FOR ZONE-BASED FEDERATED LEARNING - A simplified explanation of the abstract==


This abstract first appeared for US patent application 18461410 titled 'ZONE GRADIENT DIFFUSION (ZGD) FOR ZONE-BASED FEDERATED LEARNING

==Simplified Explanation==

The abstract describes a processor-implemented method for updating a central machine learning model in a federated learning system based on updates from clients in different zones.

* The method involves determining fixed local and neighbor zones for clients, each with their own boundaries.
* Model weights of the central machine learning model are updated based on local machine learning updates from clients in the local zone.
* Model weights are also updated based on neighbor machine learning updates from clients in the neighbor zone, with a different weight assigned to these updates.
* The different weight corresponds to a similarity parameter, which influences the impact of neighbor updates on the central model.

== Potential Applications ==
This technology could be applied in various fields such as healthcare, finance, and marketing where privacy concerns are paramount, and data needs to be processed locally before being aggregated for model updates.

== Problems Solved ==
This method addresses privacy concerns by allowing clients to update a central model without sharing their raw data. It also enables efficient model updates by considering both local and neighbor zones for clients.

== Benefits ==
- Enhanced privacy protection for client data
- Improved model accuracy through collaborative learning from local and neighbor updates
- Efficient model updates without the need for centralized data storage

== Potential Commercial Applications ==
"Privacy-Preserving Federated Learning Method for Enhanced Model Updates"

== Possible Prior Art ==
One possible prior art could be the use of federated learning methods in machine learning systems to update models based on client contributions while preserving data privacy.

=== What are the specific machine learning algorithms used in this method? ===
The abstract does not specify the exact machine learning algorithms used in this method.

=== How does the method handle clients with varying levels of data quality or quantity? ===
The abstract does not provide information on how the method handles clients with varying levels of data quality or quantity.


==Original Abstract Submitted==

A processor-implemented method includes receiving machine learning model updates from clients in a federated learning system. The method also includes determining a fixed local zone associated with each of the clients, the fixed local zone having a first fixed boundary. The method includes updating model weights of a central machine learning model based on local machine learning updates for a local subset of the clients corresponding to the fixed local zone. The method includes updating the model weights of the central machine learning model based on neighbor machine learning updates for a neighbor subset of the clients. The neighbor subset corresponds to a fixed neighbor zone that neighbors the fixed local zone and has a second fixed boundary. The neighbor machine learning updates have a different weight than the local machine learning updates when updating model weights. A value of the different weight corresponds to a similarity parameter.



[[Category:G06N3/098]]