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

A Transfer Learning Ranking Method Based on Recurrent Generative Adversarial Networks

A technology of transfer learning and sorting method, applied in the field of transfer learning sorting based on recurrent generative confrontation network

Active Publication Date: 2021-09-28
SUN YAT SEN UNIV +1
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

This method does not directly use the information of the target domain during the training process, and it will not perform well if the common information of the two domains is relatively small.

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 Transfer Learning Ranking Method Based on Recurrent Generative Adversarial Networks
  • A Transfer Learning Ranking Method Based on Recurrent Generative Adversarial Networks
  • A Transfer Learning Ranking Method Based on Recurrent Generative Adversarial Networks

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0036] Such as figure 1 As shown, the present invention is a transfer learning sorting method based on cyclic generative confrontation network. Since it is transfer learning, it is necessary to understand the data set first. We use the LETOR3.0 dataset containing TREC 2003 and 2004, which is a standard dataset for information retrieval research learning to rank published by Microsoft Research Asia. We decompose this data into several feature domains by ranking task, namely home page lookup (HP), named page lookup (NP) and topic extraction (TD) each as a separate domain. The number of queries in each individual domain is shown in Table 1, where the similarity between HP and NP tasks is greater, while TD is quite different from the other two tasks. 64-dimensional features are used to describe query document instances. Each task has five folders, which are used for five times of cross-validation. Each folder has three sub-datasets: training set, verification set, and test set. ...

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 present invention provides a method for migration learning and sorting based on the cyclic confrontation generation network. The present invention performs migration sorting learning on the learning and sorting data set LETOR3.0, and uses the cyclic confrontation generation network in the field of computer vision as the migration learning framework. That is to use the features of the A domain to generate the features of B, and use the features of the B domain to generate the features of the A domain, so that the generated features contain the feature information of another domain; then use the RankNet learning sorting algorithm to learn the transferred data. , using the learned ranking model to test in the target domain; the label information of the target domain is not involved in the learning process.

Description

technical field [0001] The present invention relates to the related fields of learning sorting and computer vision, and more specifically, relates to a transfer learning sorting method based on recurrent generative confrontation network. Background technique [0002] In recent years, with the rapid development of big data and information technology, the data information generated every day cannot be estimated. How to search for the information you want in this vast data information is particularly important. At the same time, information retrieval technology has to be greatly developed and applied. One of the more important technologies in the field of information retrieval is learning to sort. The purpose of learning to rank is to retrieve documents that are relevant to the query. The goal of learning to rank is to optimize a ranking function that incorporates a wide range of relevant properties and avoids extensive tuning of parameters empirically. Like other supervised ...

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): G06F16/332G06N3/04G06N3/08
CPCG06N3/08G06N3/045
Inventor 赵伟强赖韩江印鉴高静
Owner SUN YAT SEN UNIV