Approximate nearest neighbor search method based on hierarchical navigable small world map

A technology that approximates nearest neighbors and search methods, applied in database indexing, structured data retrieval, digital data information retrieval, etc., can solve problems such as unbalanced index structure, achieve the goal of solving unbalanced index structure, improving speed and stability Effect

Inactive Publication Date: 2019-07-12
HANGZHOU DIANZI UNIV
View PDF0 Cites 12 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

B-HNSW solves the problem of unbalanced HNSW index structure by modifying the way of inserting new nod

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
  • Approximate nearest neighbor search method based on hierarchical navigable small world map
  • Approximate nearest neighbor search method based on hierarchical navigable small world map
  • Approximate nearest neighbor search method based on hierarchical navigable small world map

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0026] The present invention proposes an approximate nearest neighbor search method based on a layered navigable small world graph, which is called B-HNSW. B-HNSW modifies the way of inserting new nodes and the way of layering, and then improves the HNSW algorithm to achieve a more balanced index structure, making the query of approximate nearest neighbor nodes faster and more stable. In order to make the object of the present invention, technical scheme and effect clearer and more definite, the present invention is combined below figure 1 and figure 2 Further details are as follows:

[0027] 1. B-HNSW index generation

[0028] part 1 as figure 1 shown in the following steps:

[0029] 1.1 Build a node set

[0030] Construct a vector set P, and take each vector as a node in B-HNSW, and the vector set P can be regarded as a node set P.

[0031] 1.2 Insert the initial node

[0032] Traversing step 1.1 Traversing the node set P, first insert the first node in the node 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 invention discloses an approximate nearest neighbor search method based on a hierarchical navigable small world map. The method comprises the following steps: step 1, aB-HNSW index generation method comprises the steps that a node set P is constructed and traversed, each node is inserted into the corresponding layer of the B-HNSW index, and the B-HNSW index is stored as a binary file; Step 2,B-HNSW index approximate nearest neighbor search method comprises the steps that a target node is input, one nearest neighbor node with the target node is sequentially searched from the maximum layerto the first layer of the B-HNSW index, the nearest neighbor node obtained from the first layer is taken as an entrance node, and k nearest neighbor nodes of the target node are searched and obtainedat the 0-th layer by using a greedy algorithm. According to the method, the mode of inserting new nodes and the layered layout are modified, a more balanced index structure is achieved, and the time for querying the approximate nearest neighbor nodes is faster and more stable.

Description

technical field [0001] The invention relates to the field of approximate nearest neighbor search, in particular to an approximate nearest neighbor search method based on a layered navigable small world graph. Background technique [0002] Nearest neighbor search is the most widely used data search technology, widely used in computer machine learning, pattern recognition, text and multimedia search and other fields. However, in practical applications, because the search set is usually very large, it cannot meet the time performance requirements for large-scale data retrieval. Therefore, approximate nearest neighbor search has received more and more research and attention. [0003] Approximate nearest neighbor search is currently implemented in many ways, such as tree method, hash method, vector quantization method, nearest neighbor graph method, etc. Among them, the nearest neighbor graph method is a recently proposed method with the best performance in query speed and reca...

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
IPC IPC(8): G06F16/2458G06F16/22
Inventor 徐小良李畅王宇翔何宏夏一行
Owner HANGZHOU DIANZI UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap