The overall structure of SDMAE.

<p>In sentence-aspect pairs encoder, mainly a BERT-based encoding process. DualGCN modules contain SemGCN and SynGCN. <i>Loss</i><sub><i>ar</i></sub> and <i>Loss</i><sub><i>scl</i></sub> refer to the multi-strategy auxilia...

وصف كامل

محفوظ في:
التفاصيل البيبلوغرافية
المؤلف الرئيسي: Lu Liu (171341) (author)
مؤلفون آخرون: Da Li (194457) (author), Chuanxu Yue (22041810) (author), Xiaojin Gao (11264337) (author), Yunhai Zhu (19467015) (author)
منشور في: 2025
الموضوعات:
الوسوم: إضافة وسم
لا توجد وسوم, كن أول من يضع وسما على هذه التسجيلة!
_version_ 1852017669901910016
author Lu Liu (171341)
author2 Da Li (194457)
Chuanxu Yue (22041810)
Xiaojin Gao (11264337)
Yunhai Zhu (19467015)
author2_role author
author
author
author
author_facet Lu Liu (171341)
Da Li (194457)
Chuanxu Yue (22041810)
Xiaojin Gao (11264337)
Yunhai Zhu (19467015)
author_role author
dc.creator.none.fl_str_mv Lu Liu (171341)
Da Li (194457)
Chuanxu Yue (22041810)
Xiaojin Gao (11264337)
Yunhai Zhu (19467015)
dc.date.none.fl_str_mv 2025-08-12T17:38:34Z
dc.identifier.none.fl_str_mv 10.1371/journal.pone.0329018.g002
dc.relation.none.fl_str_mv https://figshare.com/articles/figure/The_overall_structure_of_SDMAE_/29895024
dc.rights.none.fl_str_mv CC BY 4.0
info:eu-repo/semantics/openAccess
dc.subject.none.fl_str_mv Neuroscience
Science Policy
Mental Health
Environmental Sciences not elsewhere classified
Biological Sciences not elsewhere classified
Information Systems not elsewhere classified
supervised contrastive learning
strategy auxiliary enhancement
model &# 8217
head attention mechanism
feature aggregation system
comprehensive experiments conducted
art methods across
sentiment polarity associated
incorporates sentiment lexicons
based sentiment analysis
accurately detect sentiment
original dependency tree
xlink "> aspect
aspect sentiment recognition
extract syntactic features
speech features
dependency trees
dependency parsing
aspect terms
syntactic structures
syntactic denoising
specific part
significantly impairs
several state
semantic information
public datasets
primarily focused
online reviews
often insufficient
often derived
noise introduced
maam ),
integrate semantic
existing studies
context words
constructing graphs
adversely impacted
dc.title.none.fl_str_mv The overall structure of SDMAE.
dc.type.none.fl_str_mv Image
Figure
info:eu-repo/semantics/publishedVersion
image
description <p>In sentence-aspect pairs encoder, mainly a BERT-based encoding process. DualGCN modules contain SemGCN and SynGCN. <i>Loss</i><sub><i>ar</i></sub> and <i>Loss</i><sub><i>scl</i></sub> refer to the multi-strategy auxiliary enhancement loss.</p>
eu_rights_str_mv openAccess
id Manara_2f4f80b2fbe609376fd773bef42e59f7
identifier_str_mv 10.1371/journal.pone.0329018.g002
network_acronym_str Manara
network_name_str ManaraRepo
oai_identifier_str oai:figshare.com:article/29895024
publishDate 2025
repository.mail.fl_str_mv
repository.name.fl_str_mv
repository_id_str
rights_invalid_str_mv CC BY 4.0
spelling The overall structure of SDMAE.Lu Liu (171341)Da Li (194457)Chuanxu Yue (22041810)Xiaojin Gao (11264337)Yunhai Zhu (19467015)NeuroscienceScience PolicyMental HealthEnvironmental Sciences not elsewhere classifiedBiological Sciences not elsewhere classifiedInformation Systems not elsewhere classifiedsupervised contrastive learningstrategy auxiliary enhancementmodel &# 8217head attention mechanismfeature aggregation systemcomprehensive experiments conductedart methods acrosssentiment polarity associatedincorporates sentiment lexiconsbased sentiment analysisaccurately detect sentimentoriginal dependency treexlink "> aspectaspect sentiment recognitionextract syntactic featuresspeech featuresdependency treesdependency parsingaspect termssyntactic structuressyntactic denoisingspecific partsignificantly impairsseveral statesemantic informationpublic datasetsprimarily focusedonline reviewsoften insufficientoften derivednoise introducedmaam ),integrate semanticexisting studiescontext wordsconstructing graphsadversely impacted<p>In sentence-aspect pairs encoder, mainly a BERT-based encoding process. DualGCN modules contain SemGCN and SynGCN. <i>Loss</i><sub><i>ar</i></sub> and <i>Loss</i><sub><i>scl</i></sub> refer to the multi-strategy auxiliary enhancement loss.</p>2025-08-12T17:38:34ZImageFigureinfo:eu-repo/semantics/publishedVersionimage10.1371/journal.pone.0329018.g002https://figshare.com/articles/figure/The_overall_structure_of_SDMAE_/29895024CC BY 4.0info:eu-repo/semantics/openAccessoai:figshare.com:article/298950242025-08-12T17:38:34Z
spellingShingle The overall structure of SDMAE.
Lu Liu (171341)
Neuroscience
Science Policy
Mental Health
Environmental Sciences not elsewhere classified
Biological Sciences not elsewhere classified
Information Systems not elsewhere classified
supervised contrastive learning
strategy auxiliary enhancement
model &# 8217
head attention mechanism
feature aggregation system
comprehensive experiments conducted
art methods across
sentiment polarity associated
incorporates sentiment lexicons
based sentiment analysis
accurately detect sentiment
original dependency tree
xlink "> aspect
aspect sentiment recognition
extract syntactic features
speech features
dependency trees
dependency parsing
aspect terms
syntactic structures
syntactic denoising
specific part
significantly impairs
several state
semantic information
public datasets
primarily focused
online reviews
often insufficient
often derived
noise introduced
maam ),
integrate semantic
existing studies
context words
constructing graphs
adversely impacted
status_str publishedVersion
title The overall structure of SDMAE.
title_full The overall structure of SDMAE.
title_fullStr The overall structure of SDMAE.
title_full_unstemmed The overall structure of SDMAE.
title_short The overall structure of SDMAE.
title_sort The overall structure of SDMAE.
topic Neuroscience
Science Policy
Mental Health
Environmental Sciences not elsewhere classified
Biological Sciences not elsewhere classified
Information Systems not elsewhere classified
supervised contrastive learning
strategy auxiliary enhancement
model &# 8217
head attention mechanism
feature aggregation system
comprehensive experiments conducted
art methods across
sentiment polarity associated
incorporates sentiment lexicons
based sentiment analysis
accurately detect sentiment
original dependency tree
xlink "> aspect
aspect sentiment recognition
extract syntactic features
speech features
dependency trees
dependency parsing
aspect terms
syntactic structures
syntactic denoising
specific part
significantly impairs
several state
semantic information
public datasets
primarily focused
online reviews
often insufficient
often derived
noise introduced
maam ),
integrate semantic
existing studies
context words
constructing graphs
adversely impacted