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

Small sample fine-grained entity classification method based on relational graph convolutional network

A technology of convolutional network and classification method, applied in the field of small-sample fine-grained entity classification based on relational graph convolutional network, can solve problems such as poor effect, and achieve the effect of improving robustness and high classification accuracy

Active Publication Date: 2021-06-29
ZHEJIANG UNIV
View PDF3 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

They tend to be less effective when labeled data is limited

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
  • Small sample fine-grained entity classification method based on relational graph convolutional network
  • Small sample fine-grained entity classification method based on relational graph convolutional network
  • Small sample fine-grained entity classification method based on relational graph convolutional network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0123] Below in conjunction with the method of the present invention describe in detail the concrete steps that this embodiment implements, as follows:

[0124] In this embodiment, the method of the present invention is applied to FIGER, a commonly used data set for fine-grained entity classification, and 10 categories are randomly selected as small sample categories, and each category has K (K=5 or 10) labels exemplars to classify other target entities in these few-shot categories.

[0125] 1) Divide the dataset for each episode. The FIGER dataset contains a total of 128 categories. After excluding 10 small sample categories, 118 categories are actually used for training. In each episode, imitating the setting of small sample learning, randomly select 10 categories from 118 categories as small sample categories, and randomly select K (K=5 or 10) samples for each category to form a support set . The remaining 108 categories are used as frequent sample categories to form the...

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 small sample fine-grained entity classification method based on a relational graph convolutional network. Firstly, to-be-classified entities in sentences are encoded by adopting an entity-context encoder, model robustness is enhanced by using adversarial learning, and new data are automatically generated by using data enhancement. Secondly, a category co-occurrence graph capable of effectively sensing small samples is constructed, and categories are coded by using a relational graph convolutional network; the entities are then classified by matching the codes of the entities and categories. The whole model is trained in a meta-learning mode. Finally, the category data of the small samples into the model are input to finely adjust the parameters of the model, then other target entities of the small sample categories can be classified by using the model.

Description

technical field [0001] The invention belongs to the technical field of entity classification, and in particular relates to a small-sample fine-grained entity classification method based on a relational graph convolution network. Background technique [0002] Fine-grained entity classification is a fundamental task in natural language processing, which is to assign an appropriate category to a specific entity with surrounding context. The category concept is the most basic unit for constructing a thinking system when human beings understand the world. With categories, people can extract the common essential characteristics of entities under the category without being obsessed with the nuances. In addition, correct entity category information will also help people better understand and recognize new entities, and can also serve more downstream tasks. [0003] Nowadays, a large number of deep learning models composed of convolutional neural networks have been proposed to solv...

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): G06F16/35G06F40/126G06F40/295G06K9/62G06N3/04G06N3/08
CPCG06F16/35G06F40/126G06F40/295G06N3/084G06N3/044G06N3/045G06F18/241G06F18/214
Inventor 鲁伟明陈晨庄越挺
Owner ZHEJIANG UNIV