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

Adaptive sparseness quantification method of networked machine learning system

A machine learning and self-adaptive technology, applied in the computer field, can solve problems such as not universal, time-consuming and labor-intensive, and achieve the effects of reducing communication costs, ensuring communication quality, and fast convergence speed

Pending Publication Date: 2021-07-23
TONGJI UNIV
View PDF12 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] However, the above methods are all open-loop, and the parameters need to be adjusted according to specific problems, which is time-consuming and laborious, and not universal

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
  • Adaptive sparseness quantification method of networked machine learning system
  • Adaptive sparseness quantification method of networked machine learning system
  • Adaptive sparseness quantification method of networked machine learning system

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0022] Embodiment: The embodiment of the present invention includes an adaptive sparse quantization method for a networked machine learning system, which is applied to an agent of the networked machine learning system. Such as figure 1 As shown, the networked machine learning system includes multiple agents, figure 1 Each point represents an agent, each edge represents a communication link, and two agents connected by each edge can communicate with each other. Agents are entities with computing and communication capabilities such as computers, sensors, and drones. In a networked machine learning system, communication information is exchanged between adjacent (connected by communication links) agents at predetermined intervals (time steps). In some specific embodiments, the communication information includes information such as gradient vectors and gradient matrices during machine learning training.

[0023] In the process of exchanging communication information, the agent r...

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 an adaptive sparsity quantification method of a networked machine learning system, which is applied to an intelligent agent of the networked machine learning system and comprises the following steps of: sparsifying communication information by adopting a random sparsifier in the process of exchanging the communication information between a certain intelligent agent and a target intelligent agent at the current time step; and the intelligent agent calculates the difference between the communication information of the intelligent agent in the previous time step and the communication information of the target intelligent agent, and adopts the sparseness in negative correlation with the difference as the sparseness adopted by the random sparseness device in the current time step. And the sparseness adopted by the random sparsener is adaptively adjusted, so that the communication cost can be reduced as much as possible while the communication quality is ensured.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to an adaptive sparse quantization method of a networked machine learning system. Background technique [0002] In recent years, with the explosive growth of data in machine learning, the storage capacity of a single computer has been difficult to meet the requirements, and the method of improving the computing performance of a single computer is too costly, and machine learning in networked systems has been increasingly applied. The networked system includes multiple agents with perception, calculation and communication functions, and the network structure is as follows: figure 1 As shown in , the dots in the figure represent agents, and the straight lines represent communication links. There is no central node in this network. Each agent stores only part of the data, and they train machine learning models by performing local calculations and communicating the results to their n...

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): G06N20/00
CPCG06N20/00Y02D30/70
Inventor 衣鹏洪奕光雷金龙李莉陈杰梁舒李修贤马晓宇
Owner TONGJI UNIV