# Construction method of deep neural network

## A deep neural network and construction method technology, which is applied in neural learning methods, biological neural network models, neural architectures, etc. The effect of avoiding convergence interference

Inactive Publication Date: 2017-01-04

SUZHOU INST FOR ADVANCED STUDY USTC +1

0 Cites 10 Cited by

## AI-Extracted Technical Summary

### Problems solved by technology

View more

### Method used

View more

## Abstract

The invention discloses a construction method of a deep neural network. The structure of the deep neural network is a tree structure of a shared parameter and comprises a plurality of shared parameter branches and a plurality of Softmax layers; when the convergence rate of a certain branch is lowered, a fissile node which owns a plurality of outputs is subjected to fission to obtain one fission node which has the same type with the fissile node, a new feature detector is created, different features are generated, a parent node and a child node of the fission node inherit the fissile node, and the parameter of the fission node is initialized. On a premise of multi-model combination, calculation cost can be reduced, and a plurality of high-quality models can be obtained by fission.

Application Domain

Neural architecturesNeural learning methods

Technology Topic

Deep neural networksNODAL +6

## Image

## Examples

- Experimental program(1)

### Example Embodiment

## PUM

## Description & Claims & Application Information

We can also present the details of the Description, Claims and Application information to help users get a comprehensive understanding of the technical details of the patent, such as background art, summary of invention, brief description of drawings, description of embodiments, and other original content. On the other hand, users can also determine the specific scope of protection of the technology through the list of claims; as well as understand the changes in the life cycle of the technology with the presentation of the patent timeline. Login to view more.

## Similar technology patents

## Road line detection method based on concatenated convolutional neural network

ActiveCN108830182Areduce computational cost

Owner:ZHEJIANG GONGSHANG UNIVERSITY

## Design method based on worst case genetic algorithm

Owner:DONGGUAN UNIV OF TECH +2

## Optical tomography reconstruction method based on iterative measurement

PendingCN109087372AImprove modeling accuracyreduce computational cost

Owner:SUZHOU UNIV

## Protein structure ab initio prediction method based on bacterial oraging optimization algorithm

InactiveCN107229840AIncreased Sampling Capabilityreduce computational cost

Owner:ZHEJIANG UNIV OF TECH

## Method for identifying plastic mechanical parameters of metal material by adopting spherical indentation morphology

ActiveCN113049425AAvoid Simulation Calculation Processreduce computational cost

Owner:XIDIAN UNIV

## Classification and recommendation of technical efficacy words

- reduce computational cost

## Support vector machines processing system

ActiveUS20050049990A1reduce computational costreduce data movement

Owner:ORACLE INT CORP

## Interactive computer simulation enhanced exercise machine

InactiveUS20070093360A1enhancement of exercise experiencereduce computational cost

Owner:CUBEX

## Generalized decision feedback equalizer precoder with receiver beamforming for matrix calculations in multi-user multiple-input multiple-output wireless transmission systems

InactiveUS20110058599A1reduce feedback overheadreduce computational cost

Owner:HITACHI LTD

## Mobile robot path planning method based on A * algorithm and RRT * algorithm

ActiveCN112393728Areduce computational costAccelerated Path Planning

Owner:ZHEJIANG UNIV OF TECH