Language model training method and device, language model prediction method and device and electronic equipment

The invention relates to a language model training method and device, a language model prediction method and device and electronic equipment. The language model training method comprises the steps of obtaining a training positive sample and a training negative sample; inputting the training positive...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: ZENG PEIYANG, LIU ZHENGUO
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 ZENG PEIYANG
LIU ZHENGUO
description The invention relates to a language model training method and device, a language model prediction method and device and electronic equipment. The language model training method comprises the steps of obtaining a training positive sample and a training negative sample; inputting the training positive sample and the training negative sample into a to-be-trained language model, and training the to-be-trained language model; under the condition that a loss function value obtained based on an output result of the trained language model is smaller than a preset threshold value, determining the trained language model as a target language model; the target language model is used for predicting the semantic integrity of the to-be-recognized search information. The training negative sample is obtained by deleting and/or replacing at least one search word in the preset search information, that is, the negative label in the training negative sample is obtained by deleting or replacing the search word, so that the negativ
format Patent
fullrecord <record><control><sourceid>epo_EVB</sourceid><recordid>TN_cdi_epo_espacenet_CN114386396A</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>CN114386396A</sourcerecordid><originalsourceid>FETCH-epo_espacenet_CN114386396A3</originalsourceid><addsrcrecordid>eNqNyrEKwjAURuEsDqK-w3XXoUSKjlIUB3FyLyH5jReSm5imPr8gLuLidM7wTVU6G_Gj8aCYHALVYlhYPEXUe3JkxJHDky1WFL5pLnBsKyf5xe9FgK0lCVvCY-QcIXWuJjcTBiw-nanl8XDtTmvk1GPIxkJQ--7SNBu9bfWu3et_zAvn2ULm</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>patent</recordtype></control><display><type>patent</type><title>Language model training method and device, language model prediction method and device and electronic equipment</title><source>esp@cenet</source><creator>ZENG PEIYANG ; LIU ZHENGUO</creator><creatorcontrib>ZENG PEIYANG ; LIU ZHENGUO</creatorcontrib><description>The invention relates to a language model training method and device, a language model prediction method and device and electronic equipment. The language model training method comprises the steps of obtaining a training positive sample and a training negative sample; inputting the training positive sample and the training negative sample into a to-be-trained language model, and training the to-be-trained language model; under the condition that a loss function value obtained based on an output result of the trained language model is smaller than a preset threshold value, determining the trained language model as a target language model; the target language model is used for predicting the semantic integrity of the to-be-recognized search information. The training negative sample is obtained by deleting and/or replacing at least one search word in the preset search information, that is, the negative label in the training negative sample is obtained by deleting or replacing the search word, so that the negativ</description><language>chi ; eng</language><subject>CALCULATING ; COMPUTING ; COUNTING ; ELECTRIC DIGITAL DATA PROCESSING ; PHYSICS</subject><creationdate>2022</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&amp;date=20220422&amp;DB=EPODOC&amp;CC=CN&amp;NR=114386396A$$EHTML$$P50$$Gepo$$Hfree_for_read</linktohtml><link.rule.ids>230,308,780,885,25564,76547</link.rule.ids><linktorsrc>$$Uhttps://worldwide.espacenet.com/publicationDetails/biblio?FT=D&amp;date=20220422&amp;DB=EPODOC&amp;CC=CN&amp;NR=114386396A$$EView_record_in_European_Patent_Office$$FView_record_in_$$GEuropean_Patent_Office$$Hfree_for_read</linktorsrc></links><search><creatorcontrib>ZENG PEIYANG</creatorcontrib><creatorcontrib>LIU ZHENGUO</creatorcontrib><title>Language model training method and device, language model prediction method and device and electronic equipment</title><description>The invention relates to a language model training method and device, a language model prediction method and device and electronic equipment. The language model training method comprises the steps of obtaining a training positive sample and a training negative sample; inputting the training positive sample and the training negative sample into a to-be-trained language model, and training the to-be-trained language model; under the condition that a loss function value obtained based on an output result of the trained language model is smaller than a preset threshold value, determining the trained language model as a target language model; the target language model is used for predicting the semantic integrity of the to-be-recognized search information. The training negative sample is obtained by deleting and/or replacing at least one search word in the preset search information, that is, the negative label in the training negative sample is obtained by deleting or replacing the search word, so that the negativ</description><subject>CALCULATING</subject><subject>COMPUTING</subject><subject>COUNTING</subject><subject>ELECTRIC DIGITAL DATA PROCESSING</subject><subject>PHYSICS</subject><fulltext>true</fulltext><rsrctype>patent</rsrctype><creationdate>2022</creationdate><recordtype>patent</recordtype><sourceid>EVB</sourceid><recordid>eNqNyrEKwjAURuEsDqK-w3XXoUSKjlIUB3FyLyH5jReSm5imPr8gLuLidM7wTVU6G_Gj8aCYHALVYlhYPEXUe3JkxJHDky1WFL5pLnBsKyf5xe9FgK0lCVvCY-QcIXWuJjcTBiw-nanl8XDtTmvk1GPIxkJQ--7SNBu9bfWu3et_zAvn2ULm</recordid><startdate>20220422</startdate><enddate>20220422</enddate><creator>ZENG PEIYANG</creator><creator>LIU ZHENGUO</creator><scope>EVB</scope></search><sort><creationdate>20220422</creationdate><title>Language model training method and device, language model prediction method and device and electronic equipment</title><author>ZENG PEIYANG ; LIU ZHENGUO</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-epo_espacenet_CN114386396A3</frbrgroupid><rsrctype>patents</rsrctype><prefilter>patents</prefilter><language>chi ; eng</language><creationdate>2022</creationdate><topic>CALCULATING</topic><topic>COMPUTING</topic><topic>COUNTING</topic><topic>ELECTRIC DIGITAL DATA PROCESSING</topic><topic>PHYSICS</topic><toplevel>online_resources</toplevel><creatorcontrib>ZENG PEIYANG</creatorcontrib><creatorcontrib>LIU ZHENGUO</creatorcontrib><collection>esp@cenet</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>ZENG PEIYANG</au><au>LIU ZHENGUO</au><format>patent</format><genre>patent</genre><ristype>GEN</ristype><title>Language model training method and device, language model prediction method and device and electronic equipment</title><date>2022-04-22</date><risdate>2022</risdate><abstract>The invention relates to a language model training method and device, a language model prediction method and device and electronic equipment. The language model training method comprises the steps of obtaining a training positive sample and a training negative sample; inputting the training positive sample and the training negative sample into a to-be-trained language model, and training the to-be-trained language model; under the condition that a loss function value obtained based on an output result of the trained language model is smaller than a preset threshold value, determining the trained language model as a target language model; the target language model is used for predicting the semantic integrity of the to-be-recognized search information. The training negative sample is obtained by deleting and/or replacing at least one search word in the preset search information, that is, the negative label in the training negative sample is obtained by deleting or replacing the search word, so that the negativ</abstract><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier
ispartof
issn
language chi ; eng
recordid cdi_epo_espacenet_CN114386396A
source esp@cenet
subjects CALCULATING
COMPUTING
COUNTING
ELECTRIC DIGITAL DATA PROCESSING
PHYSICS
title Language model training method and device, language model prediction method and device and electronic equipment
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-26T14%3A10%3A34IST&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=ZENG%20PEIYANG&rft.date=2022-04-22&rft_id=info:doi/&rft_dat=%3Cepo_EVB%3ECN114386396A%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