Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Approximate query processing algorithm based on conditional generative model

A technology for conditional generation and processing algorithms, applied in the field of information retrieval, can solve difficult problems such as internal network balance and model collapse, and achieve the effects of performance improvement, collapse elimination, and error reduction

Active Publication Date: 2022-06-17
HARBIN INST OF TECH AT WEIHAI
View PDF9 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, it is difficult for GAN to ensure the balance of the internal network during the training process, and it is prone to model collapse

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 query processing algorithm based on conditional generative model
  • Approximate query processing algorithm based on conditional generative model
  • Approximate query processing algorithm based on conditional generative model

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0036] The approximate query processing algorithm based on the conditional generation model provided by the present invention, the specific implementation process and steps are described in detail as follows:

[0037] 1. Construction of the Conditional Variational Generative Adversarial Network Model Based on Wasserstein The Conditional Variational Wasserstein Generative Adversarial Network (CVWGAN) provided by the present invention is based on the network structure of CGAN and is integrated into the coding in CVAE network to ensure the stability of the overall model. The specific structure of the model is as figure 1 shown.

[0038] The model consists of an encoder network (Encoder, E), a generation network (Generator, G) and a discriminator network (Discriminator, D). Among them, the encoding network maps the unknown distribution of the real data to the common distribution in the latent space (LS). There are three layers in total. The real data X and the corresponding cond...

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 belongs to the technical field of information retrieval, and in particular relates to an approximate query processing algorithm. An efficient approximate query processing algorithm based on a conditional generation model, including: using aggregated precomputation to obtain the pre-aggregated value of the user query; processing the user query to obtain a new query newQ that estimates the difference between the user query and the pre-aggregated range and the selected pre-aggregated value. Aggregate values; build a Wasserstein-based conditional variational generative adversarial network model, use the trained model to generate data samples for the new query newQ; filter the generated data samples, and compare the filtered data samples with the selected pre-aggregated values Combined, the calculation gets the final query estimate. The method of the present invention constructs an efficient depth generation model, and introduces Wasserstein distance as an error measure to eliminate model collapse; the model is applied to approximate query, combined with aggregate pre-calculation, and a voting algorithm is adopted to reduce approximate query error.

Description

technical field [0001] The invention belongs to the technical field of information retrieval, and in particular relates to an approximate query processing algorithm. Background technique [0002] With the rapid development of information technology, the amount of data continues to grow at an explosive rate, making it difficult for traditional database system software to answer users' aggregated queries within the interactive response time. In specific decision analysis tasks, users usually only need to obtain general trends from the data, and do not require precise results. Moreover, in practical situations, the data distribution is not uniform, and there is a serious skew problem. Therefore, it is of great significance to obtain query results with higher precision with faster response speed in massive skewed data. [0003] Approximate Query Processing (AQP) algorithm (CHAUDHURI S, DINGB, KANDULA S.Approximate query processing:no silver bullet[C] / / Proceedings of the 2017AC...

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): G06F16/2458G06K9/62G06N3/04G06N3/08
CPCG06F16/2462G06N3/08G06N3/045G06F18/23213G06F18/2411G06F18/24323
Inventor 白文超韩希先何京璇
Owner HARBIN INST OF TECH AT WEIHAI
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com