Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

General machine learning algorithm model training method and system and computing nodes

A computing node, machine learning technology, applied in computing, computer parts, instruments, etc., can solve the problem of low computing efficiency

Active Publication Date: 2016-11-23
ALIBABA GRP HLDG LTD
View PDF7 Cites 66 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to provide a general machine learning algorithm model training method, system and computing node, to solve the technical problem that the training method and system in the prior art can only train specific machine learning algorithms, and the calculation efficiency is not high

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
  • General machine learning algorithm model training method and system and computing nodes
  • General machine learning algorithm model training method and system and computing nodes
  • General machine learning algorithm model training method and system and computing nodes

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0042] The technical solution of the present invention will be described in further detail below in conjunction with the accompanying drawings and embodiments, and the following embodiments do not constitute a limitation of the present invention.

[0043] The general machine learning algorithm model training method and system of this embodiment are applicable to the training of all machine learning algorithms, especially for the training of machine learning algorithms with massive input data, for example, for training the acoustic model of speech recognition with millions of input data. The present invention divides the machine algorithm model into a plurality of model partitions, and obtains the model parameters through the distributed parallel computing of the computing cluster, wherein the computing cluster can be one or more physical devices, such as including one or more central processing units (CPUs) In the form of a server computer, desktop computer, etc., or a server c...

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 general machine learning algorithm model training method and system, and computing nodes. The general machine learning algorithm model training method comprises the following steps: before initialization, a machine learning algorithm model is divided into m model partitions, each model partition copy is configured to one corresponding computing node, and the computing nodes obtain input data according to the configured model partition copies to compute the model partition copies; the computing nodes upload model parameter information obtained by computation to corresponding parameter servers, the parameter servers compute model parameters according to the obtained model parameter information, and the computing nodes obtain the model parameters from the parameter servers to update local model parameters. The system comprises configuration equipment, the computing nodes and the parameter servers, wherein each computing node comprises an input and output unit, a computing unit and a synchronization unit. The general machine learning algorithm model training method and system disclosed by the invention are unrelated to a machine learning algorithm and are high in computing efficiency.

Description

technical field [0001] The invention belongs to the technical field of artificial intelligence, and in particular relates to a general machine learning algorithm model training method, a system and a computing node for running the method. Background technique [0002] Machine learning is a branch of artificial intelligence, and in many cases, it has almost become synonymous with artificial intelligence. Simply put, machine learning is to use machine learning algorithm models to enable machines to learn laws from a large amount of historical data, so as to intelligently identify new samples or make predictions for the future. [0003] The general process of machine learning is to calculate the machine learning algorithm model parameters from the input data (input data), form the machine algorithm model according to the calculated model parameters, and intelligently identify new samples or make predictions for the future. When training model parameters, if the input data is l...

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): G06K9/66
Inventor 李锴
Owner ALIBABA GRP HLDG LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com