§ 瀏覽學位論文書目資料
  
系統識別號 U0002-1809202308094700
DOI 10.6846/tku202300664
論文名稱(中文) 「基於類級的最大均值差異之無監督域適應深度網路」
論文名稱(英文) Unsupervised Domain Adaptation Deep Network Based on Class-Wise MMD
第三語言論文名稱
校院名稱 淡江大學
系所名稱(中文) 資訊工程學系碩士班
系所名稱(英文) Department of Computer Science and Information Engineering
外國學位學校名稱
外國學位學院名稱
外國學位研究所名稱
學年度 111
學期 2
出版年 112
研究生(中文) 劉孟鑫
研究生(英文) Meng-Hsing Liu
學號 609410450
學位類別 碩士
語言別 繁體中文
第二語言別
口試日期 2023-07-18
論文頁數 82頁
口試委員 指導教授 - 林慧珍(086204@gms.tku.edu.tw)
口試委員 - 顏淑惠(105390@mail.tku.edu.tw)
口試委員 - 凃瀞珽(cttu@nchu.edu.tw)
關鍵字(中) 遷移學習
域適應
深度網路
特徵學習
最大均值差異
類級最大均值差異
關鍵字(英) Transfer learning
Domain adaptation
Deep network
Feature learning
Maximum Mean Discrepanc
Class-wise Maximum Mean Discrepanc
第三語言關鍵字
學科別分類
中文摘要
本篇論文旨在非監督域適應(unsupervised domain adaptation, UDA),即對不具標籤的目標域資料,從具有標籤的源域資料中學習域不變性特徵(domain-invariant feature)。透過最小化兩域樣本的最大均值差異(maximum mean discrepancy, MMD)已被證明可以有效地拉近兩域樣本在特徵空間的分布,進而學習到域不變性特徵;然而在最小化兩域資料的MMD的過程中,不保證能對齊各類別資料。Long等人提出類級MMD(class-wise MMD)來解決這個問題,不過利用最小化類級MMD來拉近同一類的兩域資料會同時最大化類內距離(intra-class distance)造成特徵可辨性降低。Wang等人提出調整類級MMD裡面隱含的類內距離之權重來減輕此問題,不過該方法是在線性轉換的特徵空間中計算兩域資料均值之歐式距離來定義的MMD,這樣的定義並不符合Gretton等人在雙樣本檢定所定義之MMD性質。本篇論文將改進Wang等人所提的方法,將兩域樣本透過卷積網路(CNN)的非線性轉換映射到特徵空間,提出了在一個再現核希爾伯特空間計算的具基於類級的可辨性MMD,此MMD可以用核計巧(kernel trick)簡單計算投影空間中的向量內積。使得相同類別的兩域樣本在特徵空間的分布能夠有效對齊,同時減低特徵可辨性降低的風險,因而增強了目標域資料分類明確性,進而達到域適應的目的。
英文摘要
This paper aims to address unsupervised domain adaptation (UDA), specifically learning domain invariant features from labeled source domain data for unlabeled target domain data. By minimizing the Maximum Mean Discrepancy (MMD) between the two domains, it has been demonstrated that the distributions of samples in the feature space can effectively be brought closer together, thus facilitating the learning of domain invariant features. However, when minimizing the MMD between the two domains, there is no guar antee that data from different classes will be aligned properly. Long et al. proposed class wise MMD to tackle this issue, but minimizing class wise MMD to bring together samples from the same class in both domains simultaneously maximizes the intra class distance, which leads to a reduction in feature discriminability. Wang et al. proposed adjusting the weights of the implicit intra class distances within class wise MMD to mitigate this problem. However, their method calculates the MMD in a linearly transformed feature space, defining MMD as the Euclidean distance between the mean of two domains, which does not adhere to the properties of MMD as defined in the two sample test by Gretton et al. This paper improves upon Wang et al.'s method by non linearly transforming samples from both domains into a feature space using Convolutional Neural Networks (CNNs). It introduces a class wise discriminative MMD computed in a Reproducing Kernel Hilbert Space (RKHS). This MMD can b e easily calculated using the kernel trick, simplifying the computation of vector inner products in the projection space. This approach effectively aligns the distributions of samples from the same class in the feature space while reducing the risk of feat ure discriminability reduction. As a result, it enhances the clarity of target domain data classification, thus achieving the goal of domain adaptation.
第三語言摘要
論文目次
目錄
目錄		IV
圖目錄		V
表目錄		VI
第一章 緒論	1
第二章 相關研究	4
2.1 偽標籤	4
2.2 最大均值差異	5
2.2.1 最大均值差異之計算	6
2.2.2 類級最大均值差異	7
2.2.3 可辨性的類級最大均值差異	9
第三章 研究方法	11
3.1. 符號定義	11
3.2. 類間距離、類內距離與變異量之間的關係	12
3.4. 可辨性的類級損失函數	15
3.5. 可辨性類級域適應訓練	17
第四章 實驗結果	21
4.1 實驗設定	21
4.2 正確率比較	23
4.3 消融實驗	26
第五章 結論與未來展望	27
參考文獻		28
附錄:英文論文	35
圖目錄
圖一、MMD的工作原理	9
圖二、DCWDA網路架構圖	19
圖三、數字資料集	22
圖四、Office-31 資料集	22
表目錄
表一、數字資料集用於域適應正確率比較	24
表二、基於Resnet-50網路用於Office-31資料集域適應正確率比較 	25
表三、在數字資料集測試之消融實驗 	26
表四、在Office-31資料集測試之消融實驗	 26

參考文獻
[1] K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in Proceedings of Conference on Computer Vision and Pattern Recognition (CVPR), pp. 770–778, 2016.
[2] S. Ren, K. He, R. Girshick, and J. Sun, “Faster r-cnn: towards real-time object detection with region proposal networks,” Workshop on Neural Information Processing Systems (NIPS), pp. 91–99, 2015.
[3] K. He, G. Gkioxari, P. Dollár, and R. Girshick, “Mask r-cnn,” in Proceedings of Conference on Computer Vision and Pattern Recognition (CVPR), 2018.
[4] S. J. Pan and Q. Yang, “A survey on transfer learning,” Transactions on Knowledge and Data Engineering, Vol. 22, No. 10, pp. 1345–1359, 2010.
[5] Bernhard Schölkopf; John Platt; and Thomas Hofmann, “Correcting sample selection bias by unlabeled data,” in Proceedings of the 2006 Conference on Advances in Neural Information Processing Systems, 19, pp. 601–608, 2007. 
[6] S. Li, S. Song, and G. Huang, “Prediction reweighting for domain adaptation,” IEEE Transactions on Neural Networks and Learning Systems, Vol. 28, No. 7, pp. 1682–1695, Jul. 2017.
[7] M. Baktashmotlagh, M. T. Harandi, B. C. Lovell, and M. Salzmann, “Domain adaptation on the statistical manifold,” in Proceedings of IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 2481–2488, Jun. 2014.
[8] M. Long, J. Wang, G. Ding, J. Sun, and P. S. Yu, “Transfer feature learning with joint distribution adaptation,” in Proceedings of IEEE International Conference on Computer Vision (ICCV), pp. 2200–2207, Dec. 2013.
[9] M. Long, J. Wang, G. Ding, J. Sun, and P. S. Yu, “Transfer joint matching for unsupervised domain adaptation,” in Proceedings of IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 1410–1417, Jun. 2014.
[10] M. Baktashmotlagh, M. T. Harandi, B. C. Lovell, and M. Salzmann, “Unsupervised domain adaptation by domain invariant projection,” in Proceedings of IEEE International Conference of Computer Vision, Dec. 2013, pp. 769–776.
[11] S. J. Pan, J. T. Kwok, and Q. Yang, “Transfer learning via dimensionality reduction,” in Proceedings of Association for the Advancement of Artificial Intelligence (AAAI), Vol. 8, 2008, pp. 677–682.
[12] M. Long, J. Wang, G. Ding, S. J. Pan, and P. S. Yu, “Adaptation regularization: A general framework for transfer learning,” IEEE Transactions on Knowledge and Data Engineering, Vol. 26, No. 5, pp. 1076–1089, May 2014.
[13] L. Bruzzone and M. Marconcini, “Domain adaptation problems: A DASVM classification technique and a circular validation strategy,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 32, No. 5, pp. 770 – 787, 2010.
[14] Weichen Zhang, Wanli Ouyang, Wen Li, and Dong Xu, “Collaborative and adversarial network for unsupervised domain adaptation,” in Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2018.
[15] K. Bousmalis, N. Silberman, D. Dohan, D. Erhan, and D. Krishnan, “Unsupervised pixel-level domain adaptation with generative adversarial networks,” in Proceedings of Conference on Computer Vision and Pattern Recognition (CVPR), pp. 3722–3731, 2017.
[16] Y. Ganin, E. Ustinova, H. Ajakan, P. Germain, H. Larochelle, F. Laviolette, M. Marchand, and V. Lempitsky, “Domain adversarial training of neural networks,” Journal of Machine Learning Research (JMLR), 17(59): 1–35, 2016.
[17] E. Tzeng, J. Hoffman, K. Saenko, and T. Darrell, “Adversarial discriminative domain adaptation,” in Proceedings of Conference on Computer Vision and Pattern Recognition (CVPR), pp. 7167–7176, 2017.
[18] M. Long, Y. Cao, J. Wang, and M. Jordan, “Learning transferable features with deep adaptation networks,” in Proceedings of the 40th International Conference on Machine Learning (ICML), pp. 97–105, 2015.
[19] M. Long, H. Zhu, J. Wang, and M. I. Jordan, “Unsupervised domain adaptation with residual transfer networks,” Workshop on Neural Information Processing Systems (NIPS), pp. 136–144, 2016.
[20] B. Sun and K. Saenko, “Deep coral: correlation alignment for deep domain adaptation,” in Proceedings of European Conference on Computer Vision (ECCV), pp. 443–450, 2016.
[21] Muhammad Ghifary, W. Bastiaan Kleijn, and Mengjie Zhang, “Deep reconstruction-classification networks for unsupervised domain adaptation,” in Proceedings of European Conference on Computer Vision (ECCV), 2016.
[22] L. Zhang, W. Zuo, and D. Zhang, “LSDT: Latent sparse domain transfer learning for visual adaptation,” IEEE Transactions on Image Processing, Vol. 25, No. 3, pp. 1177–1191, Mar. 2016.
[23] Y. Chen, W. Li, C. Sakaridis, D. Dai, and L. Van Gool, “Domain adaptive faster R-CNN for object detection in the wild,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 3339–3348, 2018.
[24] Konstantinos Bousmalis, Nathan Silberman, and David Dohan, “Unsupervised pixel–level domain adaptation with generative adversarial networks,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017.
[25] H. Xu, J. Zheng, A. Alavi, and R. Chellappa, “Cross-domain visual recognition via domain adaptive dictionary learning,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018.
[26] A. Gretton, K. Borgwardt, M. Rasch, B. Sch¨olkopf, and A. Smola, “A kernel two-sample test,” Journal of Machine Learning Research (JMLR), 13, pp. 723–773, March 2012.
[27] S. J. Pan, I. W. Tsang, J. T. Kwok, and Q. Yang, “Domain adaptation via transfer component analysis,” IEEE Transactions on Neural Networks, Vol. 22, No. 2, pp. 199–210, 2011.
[28] Karsten M. Borgwardt, Arthur Gretton, Malte J. Rasch, Hans-Peter Kriegel, Bernhard Sch¨olkopf, and Alexander J. Smola, “Integrating structured biological data by kernel maximum mean discrepancy,” Bioinformatics, Vol. 22, Issue 14, pp. e49–e57, Jul. 2006.
[29] S. Si, D. Tao, and B. Geng, “Bregman divergence-based regularization for transfer subspace learning,” IEEE Transactions on Knowledge and Data Engineering, Vol. 22, No. 7, pp. 929–942, 2010.
[30] J. Blitzer, K. Crammer, A. Kulesza, F. Pereira, and J. Wortman, “Learning bounds for domain adaptation,” in Proceedings of Annual Conference on Neural Information Processing Systems (NIPS), Vancouver, British Columbia, Canada, pp. 129–136, 2007.
[31] Wei Wang, Haojie Li, Zhengming Ding, and Zhihui Wang, “Rethink maximum mean discrepancy for domain adaptation,” in Proceedings of Conference on Computer Vision and Pattern Recognition (CVPR), 2020.
[32] J. Liang, D. Hu, and J. Feng, “Do we really need to access the source data? source hypothesis transfer for unsupervised domain adaptation,” in Proceedings of International Conference on Machine Learning (ICML), 2020.
[33] L. Song, A. Gretton, D. Bickson, Y. Low, and C. Guestrin, “Kernel belief propagation,” in Proceedings of the Fourteenth International Conference on Artificial Intelligence and Statistics (AISTATS), Fort Lauderdale, USA, pp. 707–715, 2011.
[34] M. Park, W. Jitkrittum, and D. Sejdinovic, “K2-ABC: approximate bayesian computation with kernel embeddings,” Journal of Machine Learning Research (JMLR), pp. 398–407, 2016.
[35] W. Jitkrittum, W. Xu, Z. Szab´o, K. Fukumizu, and A. Gretton, “A linear-time kernel goodness-of-fit test,” in Proceedings of the 30th Annual Conference on Neural Information Processing Systems (NIPS), pp. 262– 271, 2017.
[36] Y. Li, K. Swersky, and R. S. Zemel, “Generative moment matching networks,” Journal of Machine Learning Research (JMLR), pp. 1718–1727, 2015.
[37] S. Zhao, J. Song, and S. Ermon, “Infovae: Information maximizing variational autoencoders,” in Proceedings of Conference on Computer Vision and Pattern Recognition (CVPR), 2017.
[38] Rafael Müller, Simon Kornblith, and Geoffrey Hinton, “When does label smoothing help?,” in Proceedings of Conference on Computer Vision and Pattern Recognition (CVPR), 2020.
[39] Y. Lecun, L. Bottou, and Y. Bengio, “Gradient based learning applied to document recognition,” in Proceedings of the IEEE, Vol. 86, No. 11, pp. 2278-2324, Nov. 1998.
[40] J. J. Hull, “A database for handwritten text recognition research,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 16, No. 5, pp. 550-554, May 1994.
[41] Yuval Netzer, Tao Wang, and Adam Coates, “Reading digits in natural images with unsupervised feature learning,” Neural Information Processing Systems Workshop on Deep Learning and Unsupervised Feature Learning, 2011.
[42] K. Saenko, B. Kulis, M. Fritz, and T. Darrell, “Adapting visual category models to new domains,” in Proceedings of European Conference on Computer Vision (ECCV), pp. 213–226, 2010.
[43] K. Saito, Y. Ushiku, T. Harada, and K. Saenko, “Adversarial dropout regularization,” in Proceedings of Conference on Computer Vision and Pattern Recognition (CVPR), 2018.
[44] M. Long, Z. Cao, J. Wang, and M. I. Jordan, “Conditional adversarial domain adaptation,” in Proceedings of Conference on Neural Information Processing Systems (NeurIPS), pp. 1640–1650, 2018.
[45] J. Hoffman, E. Tzeng, T. Park, J.-Y. Zhu, P. Isola, K. Saenko, A. Efros, and T. Darrell, “Cycada: cycle-consistent adversarial domain adaptation,” in Proceedings of the 35th International Conference on Machine Learning (ICML), pp. 1989–1998, 2018.
[46] Z. Deng, Y. Luo, and J. Zhu, “Cluster alignment with a teacher for unsupervised domain adaptation,” in Proceedings of International Conference on Computer Vision (ICCV), pp. 9944–9953, 2019.
[47] C.-Y. Lee, T. Batra, M. H. Baig, and D. Ulbricht, “Sliced wasserstein discrepancy for unsupervised domain adaptation,” in Proceedings of Conference on Computer Vision and Pattern Recognition (CVPR), pp. 10 285–10 295,2019.
[48] Y. Ganin and V. Lempitsky, “Unsupervised domain adaptation by backpropagation,” International Conference on Machine Learning (ICML), 2015.
[49] Z. Pei, Z. Cao, M. Long, and J. Wang, “Multi-adversarial domain adaptation,” the Thirty-Second AAAI Conference on Artificial Intelligence (AAAI-18), pp. 3934–3941, 2018.
[50] Luca Bergamini, Yawei Ye, and Oliver Scheel, “SimNet: learning reactive self-driving simulations from real-world observations,” in Proceedings of International Conference on Robotics and Automation (ICRA), 2021.
[51] Xu Zhang, Dengbing Huang, and Hanyu Li, “Self‐training maximum classifier discrepancy for EEG emotion recognition”, CAAI Transactions on Intelligence Technology, Feb. 2023.
[52] Ruijia Xu, Guanbin Li, Jihan Yang, Liang Lin, “Larger norm more transferable: An adaptive feature norm approach for unsupervised domain adaptation,” International Conference on Computer Vision (ICCV), 2019.
[53] Yabin Zhang, Hui Tang, Kui Jia, and Mingkui Tan, “Domain-symmetric networks for adversarial domain adaptation,” in Proceedings of Conference on Computer Vision and Pattern Recognition (CVPR), 2019.
[54] Y. Zhang, T. Liu, M. Long, and M. Jordan, “Bridging theory and algorithm for domain adaptation,” in Proceedings of International Conference on Machine Learning (ICML), pp. 7404–7413, 2019.
[55] S. Cui, S. Wang, J. Zhuo, C. Su, Q. Huang, and Q. Tian, “Gradually vanishing bridge for adversarial domain adaptation,” in Proceedings of Conference on Computer Vision and Pattern Recognition (CVPR), pp. 12 455–12 464, 2020.
論文全文使用權限
國家圖書館
同意無償授權國家圖書館,書目與全文電子檔於繳交授權書後, 於網際網路立即公開
校內
校內紙本論文立即公開
同意電子論文全文授權於全球公開
校內電子論文立即公開
校外
同意授權予資料庫廠商
校外電子論文立即公開

如有問題,歡迎洽詢!
圖書館數位資訊組 (02)2621-5656 轉 2487 或 來信