Abstract
Active Domain Adaptation (ADA) attempts to improve the adaptation performance on a target domain by annotating informative target data with a limited budget. Previous ADA methods have significantly advanced by incorporating domain representativeness and predictive uncertainty. However, they only focus on domain-level alignment and ignore the category discriminability of two domains, which may cause classwise mismatching. These mismatched data are overlooked by the above query strategy. To solve this, a Learning Category Discriminability approach is proposed for active domain adaptation. Specifically, it achieves semantic-level alignment and selects the informative target data consistent with the domain adaptation based on task-specific classifiers. To overcome the class imbalance from the small queried data, progressive augmentation of the queried set with confident pseudo labels is designed in our work. In addition, discriminability and diversity learning for unlabeled target samples are performed to improve the reliability of pseudo labels, which makes the classification boundaries more applicable to the target domain. Extensive experiments on two benchmarks, Office-31 and Office-Home, demonstrate the superiority of the proposed method.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Bachman, P., Sordoni, A., Trischler, A.: Learning algorithms for active learning. In: International Conference on Machine Learning, pp. 301–310 (2017)
Cui, S., Wang, S., Zhuo, J., Li, L., Huang, Q., Tian, Q.: Towards discriminability and diversity: batch nuclear-norm maximization under label insufficient situations. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 3941–3950 (2020)
Dagan, I., Engelson, S.P.: Committee-based sampling for training probabilistic classifiers. In: Machine Learning Proceedings 1995, pp. 150–157. Elsevier (1995)
Dasgupta, S.: Two faces of active learning. Theoret. Comput. Sci. 412(19), 1767–1781 (2011)
Fu, B., Cao, Z., Wang, J., Long, M.: Transferable query selection for active domain adaptation. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 7272–7281 (2021)
Grandvalet, Y., Bengio, Y.: Semi-supervised learning by entropy minimization. In: Proceedings of the 17th International Conference on Neural Information Processing Systems, pp. 529–536 (2004)
Gretton, A., Borgwardt, K., Rasch, M.J., Scholkopf, B., Smola, A.J.: A kernel method for the two-sample problem. arXiv preprint arXiv:0805.2368 (2008)
Joshi, A.J., Porikli, F., Papanikolopoulos, N.: Multi-class active learning for image classification. In: 2009 IEEE Conference on Computer Vision and Pattern Recognition, pp. 2372–2379. IEEE (2009)
Lee, C.Y., Batra, T., Baig, M.H., Ulbricht, D.: Sliced Wasserstein discrepancy for unsupervised domain adaptation. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 10285–10295 (2019)
Li, S., Gong, K., Xie, B., Liu, C.H., Cao, W., Tian, S.: Critical classes and samples discovering for partial domain adaptation. IEEE Trans. Cybern. 1–14 (2022)
Li, S., et al.: Transferable semantic augmentation for domain adaptation. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 11516–11525 (2021)
Long, M., Cao, Y., Cao, Z., Wang, J., Jordan, M.I.: Transferable representation learning with deep adaptation networks. IEEE Trans. Pattern Anal. Mach. Intell. 41(12), 3071–3085 (2018)
Long, M., Cao, Y., Wang, J., Jordan, M.: Learning transferable features with deep adaptation networks. In: International Conference on Machine Learning, pp. 97–105 (2015)
Long, M., Cao, Z., Wang, J., Jordan, M.I.: Conditional adversarial domain adaptation. arXiv e-prints, arXiv-1705 (2017)
Long, M., Zhu, H., Wang, J., Jordan, M.I.: Unsupervised domain adaptation with residual transfer networks. arXiv e-prints, arXiv-1602 (2016)
Luo, Z., Zou, Y., Hoffman, J., Fei-Fei, L.: Label efficient learning of transferable representations across domains and tasks. arXiv e-prints, arXiv-1712 (2017)
de Mathelin, A., Deheeger, F., Mougeot, M., Vayatis, N.: Discrepancy-based active learning for domain adaptation. arXiv preprint arXiv:2103.03757 (2021)
Motiian, S., Jones, Q., Iranmanesh, S.M., Doretto, G.: Few-shot adversarial domain adaptation. arXiv e-prints, arXiv-1711 (2017)
Nie, Z., Lin, Y., Yan, M., Cao, Y., Ning, S.: An adversarial training method for improving model robustness in unsupervised domain adaptation. In: Knowledge Science, Engineering and Management (KSEM), pp. 3–13 (2021)
Prabhu, V., Chandrasekaran, A., Saenko, K., Hoffman, J.: Active domain adaptation via clustering uncertainty-weighted embeddings. In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pp. 8505–8514 (October 2021)
Rangwani, H., Jain, A., Aithal, S.K., Babu, R.V.: S3Vaada: submodular subset selection for virtual adversarial active domain adaptation. In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pp. 7516–7525, October 2021
Ren, P., et al.: A survey of deep active learning. ACM Comput. Surv. 54(9), 1–40 (2021)
Saito, K., Watanabe, K., Ushiku, Y., Harada, T.: Maximum classifier discrepancy for unsupervised domain adaptation. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 3723–3732 (2018)
Sener, O., Savarese, S.: Active learning for convolutional neural networks: a core-set approach. arXiv preprint arXiv:1708.00489 (2017)
Seung, H.S., Opper, M., Sompolinsky, H.: Query by committee. In: Proceedings of the Fifth Annual Workshop on Computational Learning Theory, pp. 287–294 (1992)
Shu, R., Bui, H.H., Narui, H., Ermon, S.: A dirt-t approach to unsupervised domain adaptation. arXiv preprint arXiv:1802.08735 (2018)
Su, J.C., Tsai, Y.H., Sohn, K., Liu, B., Maji, S., Chandraker, M.: Active adversarial domain adaptation. In: Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pp. 739–748 (2020)
Tzeng, E., Hoffman, J., Saenko, K., Darrell, T.: Adversarial discriminative domain adaptation. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 7167–7176 (2017)
Wang, D., Shang, Y.: A new active labeling method for deep learning. In: 2014 International Joint Conference on Neural Networks (IJCNN), pp. 112–119. IEEE (2014)
Wang, J., Chen, Y., Hao, S., Feng, W., Shen, Z.: Balanced distribution adaptation for transfer learning. In: 2017 IEEE International Conference on Data Mining (ICDM), pp. 1129–1134. IEEE (2017)
Xie, B., Li, S., Lv, F., Liu, C.H., Wang, G., Wu, D.: A collaborative alignment framework of transferable knowledge extraction for unsupervised domain adaptation. IEEE Trans. Knowl. Data Eng. (2022)
Xu, R., Li, G., Yang, J., Lin, L.: Larger norm more transferable: an adaptive feature norm approach for unsupervised domain adaptation. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 1426–1435 (2019)
Yoo, D., Kweon, I.S.: Learning loss for active learning. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 93–102 (2019)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Zhang, J., Li, M., Zhang, W., Gong, L., Zhang, Z. (2023). Learning Category Discriminability for Active Domain Adaptation. In: Jin, Z., Jiang, Y., Buchmann, R.A., Bi, Y., Ghiran, AM., Ma, W. (eds) Knowledge Science, Engineering and Management. KSEM 2023. Lecture Notes in Computer Science(), vol 14120. Springer, Cham. https://doi.org/10.1007/978-3-031-40292-0_25
Download citation
DOI: https://doi.org/10.1007/978-3-031-40292-0_25
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-40291-3
Online ISBN: 978-3-031-40292-0
eBook Packages: Computer ScienceComputer Science (R0)