Alignment-Based Pseudo-Label Generation With Collaborative Filtering Mechanism for Enhanced Cross-Domain Aspect-Based Sentiment Analysis

Aspect-based sentiment analysis (ABSA) in areas such as online shopping and restaurants can effectively facilitate specific service improvements. However, ABSA performance heavily relies on high-quality labeled data, posing a major challenge in data-scarce domains. To solve the data scarcity problem...

Full description

Saved in:
Bibliographic Details
Main Authors: Yadi Xu, Noor Farizah Ibrahim
Format: Article
Language:English
Published: IEEE 2024-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10697156/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850178299897053184
author Yadi Xu
Noor Farizah Ibrahim
author_facet Yadi Xu
Noor Farizah Ibrahim
author_sort Yadi Xu
collection DOAJ
description Aspect-based sentiment analysis (ABSA) in areas such as online shopping and restaurants can effectively facilitate specific service improvements. However, ABSA performance heavily relies on high-quality labeled data, posing a major challenge in data-scarce domains. To solve the data scarcity problem, existing cross-domain ABSA research employs semi-unsupervised learning using pseudo-labels. Unfortunately, low-quality pseudo-labels may mislead model learning, resulting in suboptimal outcomes. Two key questions are raised in this paper, how to improve the quality of pseudo-label generation and how to accurately filter invalid pseudo-labels. First, although traditional adversarial training aims to improve model generalization, its instability and convergence difficulties hinder its effectiveness in managing complex inter-domain differences. To address this challenge, we introduce a scene-style pre-alignment strategy. This method effectively reduces style feature differences between domains and solves adversarial training limitations in terms of convergence and stability, thus improving the quality of pseudo-tag generation. Secondly, we propose a pseudo-label collaborative filtering mechanism, which focuses on filtering labels and texts that would mislead the model by comparing the scene feature similarity between the pseudo-label-based generated texts and target domain texts. Our approach achieves significant performance gains on several public datasets, demonstrating that the combination of aligned pre-training and adversarial training, as well as collaborative filtering mechanisms, indeed improves the model’s domain adaptation.
format Article
id doaj-art-e1844e08e4c34497b1124711d9a9f076
institution OA Journals
issn 2169-3536
language English
publishDate 2024-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-e1844e08e4c34497b1124711d9a9f0762025-08-20T02:18:46ZengIEEEIEEE Access2169-35362024-01-011215561815563110.1109/ACCESS.2024.346987210697156Alignment-Based Pseudo-Label Generation With Collaborative Filtering Mechanism for Enhanced Cross-Domain Aspect-Based Sentiment AnalysisYadi Xu0https://orcid.org/0009-0009-7926-1935Noor Farizah Ibrahim1https://orcid.org/0000-0002-6319-5162School of Computer Science, Universiti Sains Malaysia, Penang, MalaysiaSchool of Computer Science, Universiti Sains Malaysia, Penang, MalaysiaAspect-based sentiment analysis (ABSA) in areas such as online shopping and restaurants can effectively facilitate specific service improvements. However, ABSA performance heavily relies on high-quality labeled data, posing a major challenge in data-scarce domains. To solve the data scarcity problem, existing cross-domain ABSA research employs semi-unsupervised learning using pseudo-labels. Unfortunately, low-quality pseudo-labels may mislead model learning, resulting in suboptimal outcomes. Two key questions are raised in this paper, how to improve the quality of pseudo-label generation and how to accurately filter invalid pseudo-labels. First, although traditional adversarial training aims to improve model generalization, its instability and convergence difficulties hinder its effectiveness in managing complex inter-domain differences. To address this challenge, we introduce a scene-style pre-alignment strategy. This method effectively reduces style feature differences between domains and solves adversarial training limitations in terms of convergence and stability, thus improving the quality of pseudo-tag generation. Secondly, we propose a pseudo-label collaborative filtering mechanism, which focuses on filtering labels and texts that would mislead the model by comparing the scene feature similarity between the pseudo-label-based generated texts and target domain texts. Our approach achieves significant performance gains on several public datasets, demonstrating that the combination of aligned pre-training and adversarial training, as well as collaborative filtering mechanisms, indeed improves the model’s domain adaptation.https://ieeexplore.ieee.org/document/10697156/Aspect-based sentiment analysisscene-style alignmentadversarial trainingpseudo-labelcollaborative filtering mechanisms
spellingShingle Yadi Xu
Noor Farizah Ibrahim
Alignment-Based Pseudo-Label Generation With Collaborative Filtering Mechanism for Enhanced Cross-Domain Aspect-Based Sentiment Analysis
IEEE Access
Aspect-based sentiment analysis
scene-style alignment
adversarial training
pseudo-label
collaborative filtering mechanisms
title Alignment-Based Pseudo-Label Generation With Collaborative Filtering Mechanism for Enhanced Cross-Domain Aspect-Based Sentiment Analysis
title_full Alignment-Based Pseudo-Label Generation With Collaborative Filtering Mechanism for Enhanced Cross-Domain Aspect-Based Sentiment Analysis
title_fullStr Alignment-Based Pseudo-Label Generation With Collaborative Filtering Mechanism for Enhanced Cross-Domain Aspect-Based Sentiment Analysis
title_full_unstemmed Alignment-Based Pseudo-Label Generation With Collaborative Filtering Mechanism for Enhanced Cross-Domain Aspect-Based Sentiment Analysis
title_short Alignment-Based Pseudo-Label Generation With Collaborative Filtering Mechanism for Enhanced Cross-Domain Aspect-Based Sentiment Analysis
title_sort alignment based pseudo label generation with collaborative filtering mechanism for enhanced cross domain aspect based sentiment analysis
topic Aspect-based sentiment analysis
scene-style alignment
adversarial training
pseudo-label
collaborative filtering mechanisms
url https://ieeexplore.ieee.org/document/10697156/
work_keys_str_mv AT yadixu alignmentbasedpseudolabelgenerationwithcollaborativefilteringmechanismforenhancedcrossdomainaspectbasedsentimentanalysis
AT noorfarizahibrahim alignmentbasedpseudolabelgenerationwithcollaborativefilteringmechanismforenhancedcrossdomainaspectbasedsentimentanalysis