Unlock instant, AI-driven research and patent intelligence for your innovation.

Federated learning client intelligent selection method and system based on deep reinforcement learning

A reinforcement learning and client-side technology, applied in neural learning methods, neural architectures, biological neural network models, etc., can solve problems such as not fully considering the impact of client-side data quality on federated learning performance

Active Publication Date: 2021-07-30
TSINGHUA UNIV +1
View PDF6 Cites 12 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The present invention provides a federated learning client intelligent selection method (hereinafter referred to as AUCTION) and system based on deep reinforcement learning, which is used to solve the problem that the existing client selection scheme does not fully consider the data quantity, data quality, and computing resources of the client. Technical issues such as the impact of factors on federated learning performance

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Federated learning client intelligent selection method and system based on deep reinforcement learning
  • Federated learning client intelligent selection method and system based on deep reinforcement learning
  • Federated learning client intelligent selection method and system based on deep reinforcement learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0044] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings, but the present invention can be implemented in many different ways defined and covered by the claims.

[0045] figure 1 It is a schematic diagram of a typical federated service market framework referred to in this embodiment, which includes a federated platform and some candidate clients who are willing to participate in federated learning. The federated platform recruits clients with a certain budget to complete tasks and participate in federated learning. Learning clients can submit federated learning tasks to the federated platform. For a given federated learning task, there exists a set of N clients Willing to {b 1 ,b 2 ,...b n} the price involved, each client C i Maintain a set of private local data samples related to the federated learning task However, some client training samples may be mislabeled, which is common in reality but will s...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a federated learning client intelligent selection method and system based on deep reinforcement learning, and the method comprises the steps: enabling a federated platform to collect the state of a client from a federated service market environment as an input, inputting the state into a client selection agent based on a strategy network, and outputting a client selection scheme; enabling the federation platform to select a group of optimal clients from a plurality of candidate clients according to the current environment condition and the client selection scheme to cooperatively train a federation learning model, feeding back federation learning performance to the client selection agent as an award, and using the award for optimizing and updating a strategy network; obtaining the strategy network through offline training of a reinforcement learning method. According to the method, high-quality equipment can be selected from the candidate mobile edge equipment to participate in federated learning, so that the problem of low-quality data of a distributed client is solved, and the federated learning quality is remarkably improved.

Description

technical field [0001] The invention relates to the technical field of performance optimization of a large-scale distributed edge intelligent learning system, in particular to a method and system for intelligently selecting federated learning clients based on deep reinforcement learning. Background technique [0002] The popularity of mobile edge devices has led to the rapid growth of data generated at the edge, and it has also promoted the prosperity and development of modern artificial intelligence applications. However, the traditional mechanism of collecting large amounts of data in the cloud for centralized model training has become less desirable due to privacy concerns and high data transmission costs. In order to make full use of data resources without revealing privacy, a new learning paradigm emerges at the historic moment, namely federated learning (Federated Learning, FL), which enables mobile edge devices to train collaboratively without sharing their raw data. ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06N3/04G06N3/08
CPCG06N3/08G06N3/045
Inventor 张尧学邓永恒吕丰任炬
Owner TSINGHUA UNIV