Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

Multi-mechanism collaborative learning system and method based on hierarchical parameter server

A learning system and learning method technology, applied in the field of multi-institution collaborative learning system based on layered parameter server, can solve the problems of big data islands, high maintenance costs, low resource utilization, etc.

Active Publication Date: 2020-04-10
UNIV OF ELECTRONICS SCI & TECH OF CHINA
View PDF5 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] Aiming at the above-mentioned deficiencies in the prior art, a multi-institution collaborative learning system and method based on a layered parameter server provided by the present invention solves the data island problem of big data, solves the data privacy security problem during multi-party collaboration, and solves the problem of High communication costs, high maintenance costs, high security risks, and low resource utilization of existing systems

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
  • Multi-mechanism collaborative learning system and method based on hierarchical parameter server
  • Multi-mechanism collaborative learning system and method based on hierarchical parameter server
  • Multi-mechanism collaborative learning system and method based on hierarchical parameter server

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0081] In the cross-domain multi-center scenario, due to the characteristics of high bandwidth and low latency, homogeneous computing and communication resources, security and reliability within the domain, and the characteristics of low bandwidth and high latency, heterogeneous computing and communication resources, and insecurity and reliability between domains , Isolating intra-domain and inter-domain can maximize intra-domain resource utilization, minimize inter-domain communication pressure, and provide flexibility for organizations to choose a suitable communication topology according to their own computing cluster environment. The present invention proposes a multi-party collaborative learning framework HiPS based on a layered parameter server, and isolates intra-domain and inter-domain data interactions through the layered parameter server. The intra-domain fused model update is sent to the central organization by the intra-domain parameter server, and the global model ...

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 multi-mechanism collaborative learning system based on a hierarchical parameter server. The multi-mechanism collaborative learning system comprises a central mechanism and aplurality of participation mechanisms connected with the central mechanism through a WAN network. Based on the system, the invention further discloses a multi-mechanism collaborative learning method based on the hierarchical parameter server. According to the method, the data island problem of big data is solved, the data privacy safety problem during multi-party cooperation is solved, and the problems of high communication cost, high maintenance cost, high safety risk and low resource utilization rate of an existing system are solved. On the premise of ensuring data privacy safety, efficientcommunication and efficient calculation multi-party collaborative learning is realized, and the method is suitable for cross-domain interconnection of multiple independent mechanisms and multiple datacenters. The system provided by the invention supports a platform mode and a participation mode, can be used as a platform to provide multi-party knowledge fusion service, and can also be used as a tool to support sharing cooperation among a plurality of independent mechanisms.

Description

technical field [0001] The invention belongs to the field of electronic technology, and in particular relates to a multi-institution collaborative learning system and method based on a layered parameter server. Background technique [0002] In the 5G era of high-speed interconnection of everything, the speed of data collection and the amount of data accumulation have exploded, marking that human society has truly entered the era of big data. Big data puts forward higher requirements for data mining capabilities, and the rapid development of artificial intelligence provides powerful data mining and analysis capabilities for many frontier scientific fields, enabling intelligent applications to extract core knowledge from huge amounts of data, and Organically combine these knowledge to perform complex tasks such as detection, recognition, prediction, decision-making, generation, etc., such as Alipay's face recognition, China Customs' face detection, Douyin short video human bod...

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): H04L12/24H04L29/06H04L29/08
CPCH04L41/145H04L41/0826H04L41/0893H04L63/20H04L67/51
Inventor 虞红芳李宗航李晴孙罡周华漫
Owner UNIV OF ELECTRONICS SCI & TECH OF CHINA
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Eureka Blog
Learn More
PatSnap group products