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

A Sequence Model Building Method Based on Segmented Recurrent Neural Network

A cyclic neural network and model building technology, applied in biological neural network models, neural learning methods, neural architectures, etc., can solve problems such as slow RNN speed, and achieve the effect of improving accuracy, improving accuracy, and improving training speed.

Active Publication Date: 2022-01-28
上海直画科技有限公司
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The present invention can solve the problem of slow speed of traditional RNN

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
  • A Sequence Model Building Method Based on Segmented Recurrent Neural Network
  • A Sequence Model Building Method Based on Segmented Recurrent Neural Network
  • A Sequence Model Building Method Based on Segmented Recurrent Neural Network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0061] The present invention will be described in detail below in conjunction with specific embodiments. The following examples will facilitate further understanding of the present invention in any form of techniques, but will not limit the invention in any form. It should be noted that several variations and improvements can also be made without departing from the concept of the present invention. These are all of the scope of protection of the present invention.

[0062] The present invention defines the following technical term: RNN (Recurrent Neural Network): Circulancy Neural Network: Convolutional Neural Network: Convolution Neural Network; LSTM (Long-Short Term Memory): Long Short-term Memory Network; GRU (Gated Recurrent Unit : Gateway cycle unit; SRNN (SLICED Recurrentneural Network): Split Circulating Neural Network.

[0063] Figure 4 Structure for SRNN, this is like Figure 4 As shown, the original sequence is divided into many minimum subsequences, and the RNN of the s...

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 provides a method for establishing a sequence model based on a segmented cyclic neural network. The SRNN of the invention improves the overall structure of the RNN so that it can be trained in parallel, and the speed of the SRNN is greatly improved compared with the traditional RNN. The SRNN of the present invention can obtain high-level information of the sequence. For example, when the number of layers is 3, the RNN at the bottom layer can obtain information at the vocabulary level, the RNN at the middle layer can obtain information at the sentence level, and the RNN at the top layer can obtain information at the paragraph level. information, and SRNN limits each RNN to the length of the minimum subsequence, which effectively improves the ability to retain important information in the sequence.

Description

Technical field [0001] The present invention relates to the field of artificial intelligence, and Background technique [0002] With the development of artificial intelligence and computer hardware, the circular neural network can be widely applied to sequence models such as natural language processing, speech recognition, etc., due to the sequence model of natural language processing, speech recognition, etc., due to the sequence model of natural language processing, speech recognition, etc. Amplitude increase. Circulating neural network structure figure 1 As shown, in which the sequence length is 8, where the lower X represents the input of each time, if the natural language processing task, X represents the word or word, if the voice recognition task, then X represents the phoneme. A represents the cycle unit, which can be Simplernn, GRU, LSTM, etc. I represents the initial state, usually set to 0. h represents the hidden state of each time. If you output a sequence, the H is ...

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
Patent Type & Authority Patents(China)
IPC IPC(8): G06N3/04G06N3/08
CPCG06N3/08G06N3/048G06N3/045
Inventor 于泽平刘功申
Owner 上海直画科技有限公司
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