Knowledge graph embedding compression method based on knowledge graph distillation

A technology of knowledge graph and compression method, which is applied in knowledge expression, special data processing application, unstructured text data retrieval, etc. It can solve problems such as poor prediction accuracy, difficulty in capturing important information, and lack of practical value, so as to improve performance , reduce computing overhead, and improve the effect of practical value

Pending Publication Date: 2022-01-28
ZHEJIANG UNIV
View PDF0 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, it is usually difficult to directly train a small-size Embedding model to capture important information in the knowledge graph, with poor prediction accuracy and lack of practical value.

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
  • Knowledge graph embedding compression method based on knowledge graph distillation
  • Knowledge graph embedding compression method based on knowledge graph distillation
  • Knowledge graph embedding compression method based on knowledge graph distillation

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0019] In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention, and do not limit the protection scope of the present invention.

[0020] figure 1 It is a flowchart of a knowledge graph embedding compression method based on knowledge graph distillation provided by an embodiment of the present invention. Such as figure 1 As shown, the knowledge map embedding compression method based on knowledge map distillation provided by the embodiment includes the following steps:

[0021] Step 1, prepare the knowledge graph, obtain the pre-trained high-dimensional knowledge graph embedding model as the teacher model, and randomly initialize a low-dimensional knowledge graph embedding model as the student model...

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 knowledge graph embedding compression method based on knowledge graph distillation. According to triple information and embedded structure information fully captured in a high-dimensional knowledge graph embedding model (Teacher model) are distilled into a knowledge graph embedding model (Student model), the expression ability of the Student model is improved under the condition that the storage and reasoning efficiency of the Student model is ensured, in the distillation process, the dual influence between the Teacher model and the Student model is considered, a soft label evaluation mechanism is provided to distinguish the quality of soft labels of different triads, a training mode of fixing the Teacher model first and then releasing the fixed Teacher model is provided, the adaptability of the Student model to the Teacher model is improved, and finally the performance of the Student model is improved.

Description

technical field [0001] The invention belongs to the technical field of knowledge graph representation, and in particular relates to a knowledge graph embedding compression method based on knowledge graph distillation. Background technique [0002] Knowledge graphs (KG Knowledge Graph), such as FreeBase, YAGO, and WordNet, have been gradually constructed, which provide an effective basis for many important AI tasks, such as semantic search, recommendation, and question answering. A knowledge graph is usually a multi-relational graph, which mainly includes entities, relationships, and triples. Each triple uses entities as nodes and relationships as edges to represent a piece of knowledge. Triplets are represented in the form of (head entity, relation, tail entity) (abbreviated as (h, r, t)). [0003] However, it is well known that most current knowledge graphs are far from complete, which in turn has prompted many studies on the completeness of knowledge graphs, a task aimed ...

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 & AuthorityApplications(China)
IPC IPC(8): G06F16/36G06N5/02
CPCG06F16/367G06N5/027Y02T10/40
Inventor张文朱渝珊赖亦璇徐雅静陈华钧
OwnerZHEJIANG UNIV