2025 Vol.8 Apr.N02

• Title: A adversarial domain adaptation extreme learning machine for image classification
• Name: Shunmin Wang1,2, Qingyu Peng3*
• Company: 1School of Information & Intelligence Engineering, University of Sanya, Sanya, 572022, China 2Academician Guoliang Chen Team Innovation Center, University of Sanya, Sanya, 572022, China 3School of computer science and technology, Hainan Tropical Ocean Un
• Abstract:

  In this paper, adversarial domain adaptation extreme learning machine (ADAELM) is proposed for unsupervised domain adaptation. The ADAELM consists of three parts: feature extractor, adversarial domain adaptation network and the classifier. We design two single hidden-layer neural networks as feature extractors to extract the preliminary domain features from source domain and target domain respectively. The key problems in domain adaptation method are to learn the domain invariant feature and solve the domain gaps between source domain and target domain. In order to get the effective domain invariant feature, we introduce adversarial learning in ADAELM. We construct adversarial domain adaptation network and make the two domain features from source domain and target domain play against each other in the adversarial game. The adversarial domain adaptation network consists of the feature extractors from two domains and the discriminator. The joint loss function of the adversarial domain adaptation network is designed to minimize the difference between source domain and target domain. When the adversarial domain adaptation network is trained, the best feature matrix of the target domain can be acquired. Finally, the classifier: conjugate gradient kernel extreme learning machine (CG-KELM) is embedded in to classify the target domain data. Comprehensive image classification on several commonly used domain adaptation datasets is presented to evaluate the effectiveness of ADAELM. The results show that ADAELM significantly outperforms the transfer or non-transfer ELM networks.

  
  

• Keyword: unsupervised domain adaptation, extreme learning machine, adversarial learning
• DOI: 10.12250/jpciams2025090411
• Citation form: Shunmin Wang,Qingyu Peng.题目[J]. Computer Informatization and Mechanical System,2025,Vol.8,pp.

[1]Wu, P., Dietterich, T. G, Improving SVM accuracy by training on auxiliary data sources, C. Proceedings of the Twenty-first International Conference on Machine Learning. (2004) 871-878. https://doi.org/10. 1145/1015330.1015436.

[2]J. Blitzer, R. McDonald, F. Pereira, Domain adaptation with structural correspondence learning, C. Proceedings of the 2006 Conference on Empirical Methods in Natural Language Processing. (2016) 120-128. https://doi.org/10.3115/1610075.1610094.

[3]Li, S. Song, G. Huang and C. Wu, Cross-Domain Extreme Learning Machines for Domain Adaptation, J.  IEEE Transactions on Systems, Man, and Cybernetics: Systems. 49 (2019) 1194-1207. https://doi.org/10. 1109/TSMC.2017.2735997.

[4]I. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, A. Courville, and Y. Bengio, Generative adversarial nets, J. Advances in Neural Information Processing Systems 27. 2014. https://arxiv.org/pdf/1406.2661v1.pdf.

[5]G.B. Huang, Q.Y. Zhu, C.K. Siew, Extreme learning machine: Theory and applications, J. Neurocomputing. 70 (2006) 489-501. https://arxiv.org/10.1016/j.neucom.2005.12.126.

[6]C. He, Y. Liu, T. Yao, F. Xu, Y. Hu and J. Zheng, A fast learning algorithm based on extreme learning machine for regular fuzzy neural network, J. Journal of Intelligent & Fuzzy Systems. 36(2019) 3263-3269. https://doi.org/10.3233/JIFS-18046.

[7]C. Marconi, C. A. Di Buduo, S. Barozzi et al, SLFN14-related thrombocytopenia: identification within a large series of patients with inherited thrombocytopenia, J. Thrombosis and Haemostasis. 115 (2016) 1076-1079. https://doi.org/10.1160/TH15-11-0884.

[8]L. Zhang and D. Zhang, Domain Adaptation Extreme Learning Machines for Drift Compensation in E-Nose Systems, J. IEEE Transactions on Instrumentation and Measurement. 64 (2015) 1790–1801. https://doi.org/10.1109/TIM.2014.2367775.

[9]X. Li, W. Mao, and W. Jiang, Extreme learning machine based transfer learning for data classification, J. Neurocomputing. 174 (2016) 203-210. https://doi.org/10.1016/j.neucom.2015.01.096.

[10]C. H. Q. Ding, D. Zhou, X. He, and H. Zha, R1-pca: rotational invariant L1-norm principal component analysis for robust subspace for robust subspace factorization, C. International Conference on Machine Learning 2006 - Proceedings of the 23rd International Conference on Machine Learning, (2006) 281-288. https://doi.org/10.1145/1143844.1143880.

[11]C. Chen, B. Jiang, X. Jin, Parameter Transfer Extreme Learning Machine based on Projective Model, C. International Joint Conference on Neural Networks(IJCNN). 2018. https://doi.org/10.1109/IJCNN. 2018.8489244.

[12]Y. Chen, S. Song, S. Li, L. Yang and C. Wu, Domain Space Transfer Extreme Learning Machine for Domain Adaptation, J. IEEE Transactions on Cybernetics. 49 (2019) 1909-1922. https://doi.org/10.1109/ TCYB.2018.2816981.

[13]L. Zhang, Transfer Adaptation Learning: A Decade Survey, J. Computer Vision and Pattern Recognition, (2019) 1903-4687. https://arxiv.org/abs/1903.04687.  

[14]Y. Ganin and V. Lempitsky, Unsupervised domain adaptation by backpropagation, J. arXiv. 2014. https://arxiv.org/abs/1409.7495v2.

[15]J. Zhang, Z. Ding, W. Li and P. Ogunbona, Importance Weighted Adversarial Nets for Partial Domain Adaptation, C. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. (2018) 8156-8164. https://doi.org/10.1109/CVPR.2018.00851.

[16] Z. Pei, Z. Cao, M. Long, and J. Wang, Multi-adversarial domain adaptation, J. the Association for the Advance of Artificial Intelligence (AAAI). 2018. https://arxiv.org/pdf/1809.02176.pdf.

[17]E. Tzeng, J. Hoffman, T. Darrell, and K. Saenko, Simultaneous Deep Transfer Across Domains and Tasks, C. 2015 IEEE International Conference on Computer Vision (ICCV). 30 (2015) 4068-4076.

[18]H. Ajakan, P. Germain, H. Larochelle, F. Laviolette, and M. Marc hand. Domain-adversarial neural network. arXiv. (2015). https://arxiv.org/pdf/1412.4446.pdf

[19]E. Tzeng, J. Hoffman, K. Saenko and T. Darrell, Adversarial Discriminative Domain Adaptation, C. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). (2017) 2962-2971. https: //doi.ieeecomputersociety.org/10.1109/CVPR.2017.316.

[20]S. Motiian, Q. Jones, S. Iranmanesh, and G. Doretto, Few-shot adversarial domain adaptation, C. Conference and Workshop on Neural Information Processing Systems (NIPS). (2017). https://arxiv.org/ abs/1711.02536.