Co-Regularized Adversarial Learning for Multi-Domain Text Classification

被引:0
作者
Wu, Yuan [1 ]
Inkpen, Diana [2 ]
El-Roby, Ahmed [1 ]
机构
[1] Carleton Univ, Ottawa, ON, Canada
[2] Univ Ottawa, Ottawa, ON, Canada
来源
INTERNATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE AND STATISTICS, VOL 151 | 2022年 / 151卷
关键词
D O I
暂无
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Multi-domain text classification (MDTC) aims to leverage all available resources from multiple domains to learn a predictive model that can generalize well on these domains. Recently, many MDTC methods adopt adversarial learning, shared-private paradigm, and entropy minimization to yield state-of-the-art results. However, these approaches face three issues: (1) Minimizing domain divergence can not fully guarantee the success of domain alignment; (2) Aligning marginal feature distributions can not fully guarantee the discriminability of the learned features; (3) Standard entropy minimization may make the predictions on unlabeled data over-confident, deteriorating the discriminability of the learned features. In order to address the above issues, we propose a co-regularized adversarial learning (CRAL) mechanism for MDTC. This approach constructs two diverse shared latent spaces, performs domain alignment in each of them, and punishes the disagreements of these two alignments with respect to the predictions on unlabeled data. Moreover, virtual adversarial training (VAT) with entropy minimization is incorporated to impose consistency regularization to the CRAL method. Experiments show that our model outperforms state-of-the-art methods on two MDTC benchmarks.
引用
收藏
页数:12
相关论文
共 50 条
[41]   Multi-Task Multi-Domain Learning for Digital Staining and Classification of Leukocytes [J].
Tomczak, Agnieszka ;
Ilic, Slobodan ;
Marquardt, Gaby ;
Engel, Thomas ;
Forster, Frank ;
Navab, Nassir ;
Albarqouni, Shadi .
IEEE TRANSACTIONS ON MEDICAL IMAGING, 2021, 40 (10) :2897-2910
[42]   Co-regularized multi-view sparse reconstruction embedding for dimension reduction [J].
Wang, Huibing ;
Peng, Jinjia ;
Fu, Xianping .
NEUROCOMPUTING, 2019, 347 :191-199
[43]   Low-resource text classification using domain-adversarial learning [J].
Griesshaber, Daniel ;
Ngoc Thang Vu ;
Maucher, Johannes .
COMPUTER SPEECH AND LANGUAGE, 2020, 62
[44]   Low-Resource Text Classification Using Domain-Adversarial Learning [J].
Griesshaber, Daniel ;
Ngoc Thang Vu ;
Maucher, Johannes .
STATISTICAL LANGUAGE AND SPEECH PROCESSING, SLSP 2018, 2018, 11171 :129-139
[45]   An algorithm for multi-domain website classification [J].
Ullah M.A. ;
Tahrin A. ;
Marjan S. .
International Journal of Web-Based Learning and Teaching Technologies, 2020, 15 (04) :57-65
[46]   Collaborative Multi-Domain Sentiment Classification [J].
Wu, Fangzhao ;
Huang, Yongfeng .
2015 IEEE INTERNATIONAL CONFERENCE ON DATA MINING (ICDM), 2015, :459-468
[47]   Semi-supervised Multi-domain Learning for Medical Image Classification [J].
Chavhan, Ruchika ;
Banerjee, Biplab ;
Das, Nibaran .
RECENT TRENDS IN IMAGE PROCESSING AND PATTERN RECOGNITION, RTIP2R 2022, 2023, 1704 :22-33
[48]   Fuzzy Semantic Classification of Multi-Domain E-Learning Concept [J].
Ahmed, Rafeeq ;
Ahmad, Tanvir ;
Almutairi, Fadiyah M. ;
Qahtani, Abdulrahman M. ;
Alsufyani, Abdulmajeed ;
Almutiry, Omar .
MOBILE NETWORKS & APPLICATIONS, 2021, 26 (05) :2206-2215
[49]   Co-regularized Kernel k-means for Multi-view Clustering [J].
Ye, Yongkai ;
Liu, Xinwang ;
Yin, Jianping ;
Zhu, En .
2016 23RD INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION (ICPR), 2016, :1583-1588
[50]   Transfer Learning for the Multilingual and Multi-Domain Classification of Messages Relating to Crises [J].
Sanchez, Cinthia .
SIGIR '21 - PROCEEDINGS OF THE 44TH INTERNATIONAL ACM SIGIR CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION RETRIEVAL, 2021, :2708-2708