18486482. GRADIENT TRANSMISSION METHOD AND RELATED APPARATUS simplified abstract (Huawei Technologies Co., Ltd.)

From WikiPatents
Jump to navigation Jump to search

GRADIENT TRANSMISSION METHOD AND RELATED APPARATUS

Organization Name

Huawei Technologies Co., Ltd.

Inventor(s)

Chen Xu of Hangzhou (CN)

Gongzheng Zhang of Hangzhou (CN)

Jian Wang of Hangzhou (CN)

Rong Li of Boulogne Billancourt (FR)

Jun Wang of Hangzhou (CN)

GRADIENT TRANSMISSION METHOD AND RELATED APPARATUS - A simplified explanation of the abstract

This abstract first appeared for US patent application 18486482 titled 'GRADIENT TRANSMISSION METHOD AND RELATED APPARATUS

Simplified Explanation

The abstract describes a patent application for a communication apparatus that uses training data to update a parameter of a neural network in another communication apparatus. The first communication apparatus maps the intermediate gradient to an air interface resource to generate a gradient signal, which is then sent to the second communication apparatus. The gradient signal includes gradient symbols corresponding to gradient values.

  • The patent application involves a communication apparatus that updates a neural network parameter based on training data.
  • The first communication apparatus maps the intermediate gradient to an air interface resource to generate a gradient signal.
  • The gradient signal, containing gradient symbols, is sent to the second communication apparatus.
  • Each gradient symbol corresponds to one or more gradient values.

Potential Applications:

  • Wireless communication systems
  • Neural network training and optimization

Problems Solved:

  • Efficient parameter updating in neural networks
  • Transmission of gradient information in communication systems

Benefits:

  • Improved performance of neural networks
  • Enhanced communication efficiency in wireless systems


Original Abstract Submitted

A first communication apparatus receives training data, and determines a first intermediate gradient based on the training data. The first intermediate gradient is used to update a parameter of a first neural network located in a second communication apparatus. The first communication apparatus maps the first intermediate gradient to an air interface resource to generate a first gradient signal, and sends the first gradient signal to the second communication apparatus. The first gradient signal includes one or more first gradient symbols, and each of the first gradient symbols is corresponding to one or more gradient values.