Adversarial perturbation and bidirectional attention mechanism for few-shot English text classification

Xianghua  Wu

doi:10.6180/jase.202608_31.010

Adversarial perturbation and bidirectional attention mechanism for few-shot English text classification

Research Categories

Xianghua WuThis email address is being protected from spambots. You need JavaScript enabled to view it.

School of Foreign Languages, Zhengzhou University of Science and Technology, Zhengzhou 450064 China

Received: November 15, 2025
Accepted: January 3, 2026
Publication Date: February 3, 2026

Copyright The Author(s). This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are cited.

Download Citation: ||https://doi.org/10.6180/jase.202608_31.010

In few-shot text classification, how well the query and support sets are encoded largely decides the final accuracy. Yet, most prior methods overlook the pairwise correspondences between them and treat all features as equally important, neglecting the varying informativeness within each set. Therefore, we propose a novel adversarial perturbation and bidirectional attention mechanism for few-shot English text classification. The model combines the global information extraction ability of GRU and the local detail learning ability of attention mechanism to model text features. The adversarial training strategy is introduced to improve the stability and accuracy of the model when facing fuzzy expression or perturbation input. To capture nuanced interactions, we employ a bidirectional attention module that aligns support and query instances, highlighting within-class distinctions and forging category prototypes with enhanced discriminative power. The proposed model is tested on the public few-shot classification data sets, and results show that the classification accuracy rates with proposed method are better than other advanced models.

Keywords: Adversarial perturbation, bidirectional attention mechanism, few-shot English text classification, GRU

[1] J. Just, (2024) “Natural language processing for innovation search–Reviewing an emerging non-human innovation intermediary" Technovation 129: 102883. DOI: 10.1016/j.technovation.2023.102883.
[2] J. Yu, L. Zhao, S. Yin, and M. Ivanovi´c, (2024) “News recommendation model based on encoder graph neural net work and bat optimization in online social multimedia art education" Computer Science and Information Systems 21(3): 989–1012. DOI: 10.2298/CSIS231225025Y.
[3] H. N. Chau, T. D. Bui, H. B. Nguyen, T. T. H. Duong, and Q. C. Nguyen, (2024) “A novel approach to multi channel speech enhancement based on graph neural net works" IEEE/ACM Transactions on Audio, Speech, and Language Processing 32: 1133–1144. DOI: 10.1109/TASLP.2024.3352259.
[4] S. Chopra, P. Agarwal, J. Ahmed, S. S. Biswas, and A. J. Obaid, (2024) “Roberta and BERT: Revolutionizing Mental Healthcare Through Natural Language" SN Computer Science 5(7): 889. DOI: 10.1007/s42979-024-03202-8.
[5] Y. Yuan, Q. Zhang, X. Zhou, and M. Gao, (2025) “A Chinese named entity recognition model: integrating la bel knowledge and lexicon information" International Journal of Machine Learning and Cybernetics 16(1): 253–266. DOI: 10.1007/s13042-024-02207-2.
[6] M. A. Metu, N. Akhter, S. Nasrin, T. Anzum, A. Khatun, and R. Mazumder, (2024) “Hybrid SVM Bidirectional Long Short-Term Memory Model for Fine Grained Software Requirement Classification" Journal of Advances in Information Technology 15(8): DOI: 10.12720/jait.15.8.914-922.
[7] S. Baik, M. Choi, J. Choi, H. Kim, and K. M. Lee, (2023) “Learning to learn task-adaptive hyperparameters for few-shot learning" IEEE Transactions on Pattern Analysis and Machine Intelligence 46(3): 1441–1454. DOI: 10.1109/TPAMI.2023.3261387.
[8] S. Song, C. Lan, J. Xing, W. Zeng, and J. Liu. “An end to-end spatio-temporal attention model for human action recognition from skeleton data”. In: Proceedings of the AAAI conference on artificial intelligence. 31. 1. 2017. DOI: 10.1609/aaai.v31i1.11212.
[9] Z. Ji, X. Chai, Y. Yu, Y. Pang, and Z. Zhang, (2020) “Improved prototypical networks for few-shot learning" Pattern Recognition Letters 140: 81–87. DOI: 10.1016/j.patrec.2020.07.015.
[10] Z.Chen,Y.Wang,J.Wu,C.Deng,andW.Jiang,(2022) “Wideresidual relation network-based intelligent fault diagnosis of rotating machines with small samples" Sensors 22(11): 4161. DOI: 10.3390/s22114161.
[11] Y. Jiang and S. Yin, (2023) “Heterogenous-view occluded expression data recognition based on cycle-consistent adversarial network and K-SVD dictionary learning under intelligent cooperative robot environment" Computer Science and Information Systems 20(4): 1869–1883. DOI: 10.2298/CSIS221228034J.
[12] W. Zhang and J. Wang, (2024) “English text sentiment analysis network based on CNN and U-Net" IFS/ACM Transactions on Machine Learning 1(1): 13–18. DOI: 10.70891/JSE.2024.100009.
[13] H. Mohammadian, A. A. Ghorbani, and A. H. Lashkari, (2023) “A gradient-based approach for adversarial attack on deep learning-based network intrusion detection systems" Applied Soft Computing 137: 110173. DOI: 10.1016/j.asoc.2023.110173.
[14] M. Park, T. Park, S. Park, S. J. Yoon, S. H. Koo, and Y.-L. Park, (2024) “Stretchable glove for accurate and robust hand pose reconstruction based on comprehensive motion data" Nature communications 15(1): 5821. DOI: 10.1038/s41467-024-50101-w.
[15] L. Zhang, M. Fu, and Y. Wang. “SProtoNet: self supervised ProtoNet for plant leaf disease few-shot classification”. In: Fourth International Conference on Computer Vision and Data Mining (ICCVDM 2023). 13063. SPIE. 2024, 138–141. DOI: 10.1117/12.3021300.
[16] D. Kang, H. Kwon, J. Min, and M. Cho. “Relational embedding for few-shot classification”. In: Proceedings of the IEEE/CVF international conference on computer vision. 2021, 8822–8833. DOI: 10.1109/ICCV48922.2021.00870.
[17] X. Li, Y. Li, J. Wang, C. Chen, L. Yang, and Z. Zheng, (2022) “Decentralized federated meta-learning framework for few-shot multitask learning" International Journal of Intelligent Systems 37(11): 8490–8522. DOI: 10.1002/int.22951.
[18] T. Wang, H. Xu, and S. Zhu. “Few-shot text classification with prototype network based on distribution propagation graph neural network”. In: International Conference on Cyber Security, Artificial Intelligence, and Digital Economy (CSAIDE 2023). 12718. SPIE. 2023, 125–131. DOI: 10.1117/12.2681670.
[19] Y. Shao, W. Wu, X. You, C. Gao, and N. Sang, (2022) “Improving the generalization of MAML in few-shot classification via bi-level constraint" IEEE Transactions on Circuits and Systems for Video Technology 33(7): 3284–3295. DOI: 10.1109/TCSVT.2022.3232717.
[20] D. Vetter, M. Ahsan, D. Delicado, T. A. Neubauer, T. Wilke, and G. Roig, (2024) “Classification of freshwater snails of the genus Radomaniola with multimodal triplet networks" arXiv preprint arXiv:2407.20013: DOI: 10. 48550/arXiv.2407.20013.
[21] Y. Zhang and Z. Kang. “Tpn: Transferable proto learning network towards few-shot document-level relation extraction”. In: 2024 International Joint Conference on Neural Networks (IJCNN). IEEE. 2024, 1–9. DOI: 10.1109/IJCNN60899.2024.10650913.
[22] M. Ren, E. Triantafillou, S. Ravi, J. Snell, K. Swersky, J. B. Tenenbaum, H. Larochelle, and R. S. Zemel, (2018) “Meta-learning for semi-supervised few-shot classification" arXiv preprint arXiv:1803.00676: DOI: 10.48550/arXiv.1803.00676.
[23] Q. Xu, (2025) “Application of an intelligent English text classification model with improved KNN algorithm in the context of big data in libraries" Systems and Soft Computing 7: 200186. DOI: 10.1016/j.sasc.2025. 200186.
[24] L. Zhang, Q. Lin, F. Meng, S. Liang, J. Lu, S. Liu, K. Chen, and Y. Zhan, (2025) “Leveraging Contrastive Semantics and Language Adaptation for Robust Financial Text Classification Across Languages" Computers 14(8): 338. DOI: 10.3390/computers14080338.