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

A method for optimizing a word vector based on a non-retention optimal individual genetic algorithm

A genetic algorithm and word vector technology, applied in the field of text generators, can solve problems such as the local optimal solution of the gradient descent algorithm, and achieve the effect of improving the robustness of the model

Active Publication Date: 2019-06-21
SOUTH CHINA UNIV OF TECH
View PDF14 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

But the gradient descent algorithm is easy to fall into the local optimal solution

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 method for optimizing a word vector based on a non-retention optimal individual genetic algorithm
  • A method for optimizing a word vector based on a non-retention optimal individual genetic algorithm
  • A method for optimizing a word vector based on a non-retention optimal individual genetic algorithm

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0029] The present invention will be further described in detail below in conjunction with the embodiments and the accompanying drawings, but the embodiments of the present invention are not limited thereto.

[0030] A method of optimizing the word vector based on not retaining the optimal individual genetic algorithm of the embodiment, such as figure 1 The flow shown:

[0031] Step 1: Construct the word vector recurrent neural network model, randomly generate word vector weights and recurrent neural network values ​​through the random function improved by c language, and create the weights of P word vector matrices as the population of the genetic algorithm, where P is The number of individuals in the population, where the set value of P is 50;

[0032] Step 2: Convert the word vector matrix to a one-dimensional chromosome: expand the weight of the word vector matrix W into a one-dimensional vector The number string, and use the number string as the chromosome in the genet...

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 method for optimizing a word vector based on a non-retention optimal individual genetic algorithm. The method comprises the following steps: constructing a word vector recurrent neural network model; Expanding the weight of the word vector matrix into a numeric string of a one-dimensional vector, and taking the numeric string as a chromosome in a genetic algorithm; Calculating the fitness of each chromosome; Randomly selecting individuals to reconstruct the population; Exchanging fragments on the chromosomes; Randomly selecting one or more variables on the chromosome;Generating a variable value through a random function, and replacing the variable value; And re-calculating the fitness of each individual. According to the method, the optimal individual retention strategy in the genetic algorithm is removed, the exploration capability is improved, and the improved genetic algorithm is applied to a text generator to search word vector matrix parameters.

Description

technical field [0001] The invention belongs to the application of genetic algorithms in the field of computer applications, and specifically relates to a method for optimizing word vectors based on genetic algorithms that do not retain the optimal individual, which is applied to a text generator. Background technique [0002] Genetic algorithm plays a big role in artificial intelligence, parameter optimization and other fields. However, in order to accelerate the convergence speed, most genetic algorithms use the strategy of retaining the optimal individual. Although this approach can speed up the convergence speed of the algorithm, it is easy to make the population fall into local optimum. Therefore, applications using this strategy are more difficult to obtain better results. At present, most text generators use neural networks to construct word vectors and use gradient descent algorithm to update the word vector matrix to obtain parameter vectors with contextual meanin...

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 Applications(China)
IPC IPC(8): G06F17/27G06N3/12
Inventor 胡劲松邓国健叶宏雄
Owner SOUTH CHINA UNIV OF TECH