Quantitative index construction method, device and equipment based on knowledge distillation
The invention provides a quantitative index construction method, device and equipment based on knowledge distillation, which can be applied to the technical fields of artificial intelligence, computers and data mining. The method comprises the steps of inputting a target query vector into a target l...
Gespeichert in:
Hauptverfasser: | , , |
---|---|
Format: | Patent |
Sprache: | chi ; eng |
Schlagworte: | |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
container_end_page | |
---|---|
container_issue | |
container_start_page | |
container_title | |
container_volume | |
creator | CHEN ENHONG LIAN DEFU LU ZEPU |
description | The invention provides a quantitative index construction method, device and equipment based on knowledge distillation, which can be applied to the technical fields of artificial intelligence, computers and data mining. The method comprises the steps of inputting a target query vector into a target low-dimensional index model so as to execute the following operations: based on an initial quantization codebook, respectively quantizing the target query vector and outputting a quantized query vector, the initial quantization codebook is pre-stored in a query database, and the quantized query vector is stored in the target low-dimensional index model; the target low-dimensional index model is obtained based on training of a high-dimensional index model and a target loss function, the target loss function comprises a knowledge distillation loss function, and the quantitative query vector is used for representing compression representation of a target query vector; updating the initial quantized codebook according t |
format | Patent |
fullrecord | <record><control><sourceid>epo_EVB</sourceid><recordid>TN_cdi_epo_espacenet_CN117312325A</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>CN117312325A</sourcerecordid><originalsourceid>FETCH-epo_espacenet_CN117312325A3</originalsourceid><addsrcrecordid>eNqNyzEKwkAQRuE0FqLeYey1SIJYS1CsBMFSCOvOrw5uZqM7iR5fBQ9g9ZrvDbPjvnNqYs6kB4kyXuSjJnt03iQqNbBr5BkxevEgp0y4d9I2UKOTS2D6qJvGZwBfQCzJJAT3ncfZ4OxCwuTXUTbdrA_Vdo421kit81BYXe3yfFnmRVksVuU_5g0e4jyL</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>patent</recordtype></control><display><type>patent</type><title>Quantitative index construction method, device and equipment based on knowledge distillation</title><source>esp@cenet</source><creator>CHEN ENHONG ; LIAN DEFU ; LU ZEPU</creator><creatorcontrib>CHEN ENHONG ; LIAN DEFU ; LU ZEPU</creatorcontrib><description>The invention provides a quantitative index construction method, device and equipment based on knowledge distillation, which can be applied to the technical fields of artificial intelligence, computers and data mining. The method comprises the steps of inputting a target query vector into a target low-dimensional index model so as to execute the following operations: based on an initial quantization codebook, respectively quantizing the target query vector and outputting a quantized query vector, the initial quantization codebook is pre-stored in a query database, and the quantized query vector is stored in the target low-dimensional index model; the target low-dimensional index model is obtained based on training of a high-dimensional index model and a target loss function, the target loss function comprises a knowledge distillation loss function, and the quantitative query vector is used for representing compression representation of a target query vector; updating the initial quantized codebook according t</description><language>chi ; eng</language><subject>CALCULATING ; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS ; COMPUTING ; COUNTING ; ELECTRIC DIGITAL DATA PROCESSING ; PHYSICS</subject><creationdate>2023</creationdate><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://worldwide.espacenet.com/publicationDetails/biblio?FT=D&date=20231229&DB=EPODOC&CC=CN&NR=117312325A$$EHTML$$P50$$Gepo$$Hfree_for_read</linktohtml><link.rule.ids>230,308,780,885,25564,76547</link.rule.ids><linktorsrc>$$Uhttps://worldwide.espacenet.com/publicationDetails/biblio?FT=D&date=20231229&DB=EPODOC&CC=CN&NR=117312325A$$EView_record_in_European_Patent_Office$$FView_record_in_$$GEuropean_Patent_Office$$Hfree_for_read</linktorsrc></links><search><creatorcontrib>CHEN ENHONG</creatorcontrib><creatorcontrib>LIAN DEFU</creatorcontrib><creatorcontrib>LU ZEPU</creatorcontrib><title>Quantitative index construction method, device and equipment based on knowledge distillation</title><description>The invention provides a quantitative index construction method, device and equipment based on knowledge distillation, which can be applied to the technical fields of artificial intelligence, computers and data mining. The method comprises the steps of inputting a target query vector into a target low-dimensional index model so as to execute the following operations: based on an initial quantization codebook, respectively quantizing the target query vector and outputting a quantized query vector, the initial quantization codebook is pre-stored in a query database, and the quantized query vector is stored in the target low-dimensional index model; the target low-dimensional index model is obtained based on training of a high-dimensional index model and a target loss function, the target loss function comprises a knowledge distillation loss function, and the quantitative query vector is used for representing compression representation of a target query vector; updating the initial quantized codebook according t</description><subject>CALCULATING</subject><subject>COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS</subject><subject>COMPUTING</subject><subject>COUNTING</subject><subject>ELECTRIC DIGITAL DATA PROCESSING</subject><subject>PHYSICS</subject><fulltext>true</fulltext><rsrctype>patent</rsrctype><creationdate>2023</creationdate><recordtype>patent</recordtype><sourceid>EVB</sourceid><recordid>eNqNyzEKwkAQRuE0FqLeYey1SIJYS1CsBMFSCOvOrw5uZqM7iR5fBQ9g9ZrvDbPjvnNqYs6kB4kyXuSjJnt03iQqNbBr5BkxevEgp0y4d9I2UKOTS2D6qJvGZwBfQCzJJAT3ncfZ4OxCwuTXUTbdrA_Vdo421kit81BYXe3yfFnmRVksVuU_5g0e4jyL</recordid><startdate>20231229</startdate><enddate>20231229</enddate><creator>CHEN ENHONG</creator><creator>LIAN DEFU</creator><creator>LU ZEPU</creator><scope>EVB</scope></search><sort><creationdate>20231229</creationdate><title>Quantitative index construction method, device and equipment based on knowledge distillation</title><author>CHEN ENHONG ; LIAN DEFU ; LU ZEPU</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-epo_espacenet_CN117312325A3</frbrgroupid><rsrctype>patents</rsrctype><prefilter>patents</prefilter><language>chi ; eng</language><creationdate>2023</creationdate><topic>CALCULATING</topic><topic>COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS</topic><topic>COMPUTING</topic><topic>COUNTING</topic><topic>ELECTRIC DIGITAL DATA PROCESSING</topic><topic>PHYSICS</topic><toplevel>online_resources</toplevel><creatorcontrib>CHEN ENHONG</creatorcontrib><creatorcontrib>LIAN DEFU</creatorcontrib><creatorcontrib>LU ZEPU</creatorcontrib><collection>esp@cenet</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>CHEN ENHONG</au><au>LIAN DEFU</au><au>LU ZEPU</au><format>patent</format><genre>patent</genre><ristype>GEN</ristype><title>Quantitative index construction method, device and equipment based on knowledge distillation</title><date>2023-12-29</date><risdate>2023</risdate><abstract>The invention provides a quantitative index construction method, device and equipment based on knowledge distillation, which can be applied to the technical fields of artificial intelligence, computers and data mining. The method comprises the steps of inputting a target query vector into a target low-dimensional index model so as to execute the following operations: based on an initial quantization codebook, respectively quantizing the target query vector and outputting a quantized query vector, the initial quantization codebook is pre-stored in a query database, and the quantized query vector is stored in the target low-dimensional index model; the target low-dimensional index model is obtained based on training of a high-dimensional index model and a target loss function, the target loss function comprises a knowledge distillation loss function, and the quantitative query vector is used for representing compression representation of a target query vector; updating the initial quantized codebook according t</abstract><oa>free_for_read</oa></addata></record> |
fulltext | fulltext_linktorsrc |
identifier | |
ispartof | |
issn | |
language | chi ; eng |
recordid | cdi_epo_espacenet_CN117312325A |
source | esp@cenet |
subjects | CALCULATING COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS COMPUTING COUNTING ELECTRIC DIGITAL DATA PROCESSING PHYSICS |
title | Quantitative index construction method, device and equipment based on knowledge distillation |
url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-05T01%3A27%3A01IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-epo_EVB&rft_val_fmt=info:ofi/fmt:kev:mtx:patent&rft.genre=patent&rft.au=CHEN%20ENHONG&rft.date=2023-12-29&rft_id=info:doi/&rft_dat=%3Cepo_EVB%3ECN117312325A%3C/epo_EVB%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true |