MAPLE: Semi-Supervised Learning with Multi-Alignment and Pseudo-Learning
Authors: 
Juncheng Yang, Chao Li, Zuchao Li, Wei Yu, Bo Du, Shijun LiAuthors Info & Claims
Juncheng Yang, Chao Li, Zuchao Li, Wei Yu, Bo Du, Shijun LiAuthors Info & ClaimsPages 2941 - 2952
Abstract
Data augmentation has undoubtedly enabled a significant leap forward in training a high-accuracy deep network. Besides the commonly used augmentation to target data, e.g., random cropping, flipping, and rotation, recent works have been dedicated to mining generalized knowledge by using multiple sources. However, along with plentiful data comes the huge data distribution gap between the target and different sources (hybrid shift). To mitigate this problem, existing methods tend to manually annotate more data. Unlike previous methods, this paper focuses on the study of learning deep models by gathering knowledge from multiple sources in a labor-free fashion and further proposes the "Multi-Alignment and Pseudo-Learning'' method, dubbed MAPLE. MAPLE constructs the multi-alignment module, which consists of multiple discriminators to align different data distributions via an adversarial process. In addition, a novel semi-supervised learning (SSL) manner is introduced to further facilitate the utility of our MAPLE. Extensive evaluations conducted on four benchmarks show the effectiveness of the proposed MAPLE, which achieves state-of-the-art performance outperforming existing methods by an obvious margin.
Supplementary Material
Current SSL methods focus on the study of an ideal setting where both the feature distribution and class distribution are identical between training and test sets. Unlike these methods, a more rigorous setting should be where data of different distributions are collected in a wild and dynamic environment, and without any additional annotation can be used in the training phase. We come up with a confirmed answer by proposing ''Multi-Alignment and Pseudo-Learning (MAPLE),'' which aligns different data distributions by constructing the multi-alignment module consisting of multiple discriminators, and is optimized in a generative adversarial learning fashion. In the meantime, for the mentioned multi-source hybrid shift problem, a novel semi-supervised learning manner is introduced to exploit pseudo-labels of unlabeled data, which proved to be valid in further facilitating the learning of our MAPLE.
- Download
- 2.74 MB
Current SSL methods focus on the study of an ideal setting where both the feature distribution and class distribution are identical between training and test sets. Unlike these methods, a more rigorous setting should be where data of different distributions are collected in a wild and dynamic environment, and without any additional annotation can be used in the training phase. We come up with a confirmed answer by proposing ''Multi-Alignment and Pseudo-Learning (MAPLE),'' which aligns different data distributions by constructing the multi-alignment module consisting of multiple discriminators, and is optimized in a generative adversarial learning fashion. In the meantime, for the mentioned multi-source hybrid shift problem, a novel semi-supervised learning manner is introduced to exploit pseudo-labels of unlabeled data, which proved to be valid in further facilitating the learning of our MAPLE.
- Download
- 15.39 MB
References
[1]
David Berthelot, Nicholas Carlini, Ekin D Cubuk, Alex Kurakin, Kihyuk Sohn, Han Zhang, and Colin Raffel. 2019a. Remixmatch: Semi-supervised learning with distribution alignment and augmentation anchoring. arXiv preprint arXiv:1911.09785 (2019).
[2]
David Berthelot, Nicholas Carlini, Ian Goodfellow, Nicolas Papernot, Avital Oliver, and Colin A Raffel. 2019b. Mixmatch: A holistic approach to semi-supervised learning. Advances in Neural Information Processing Systems, Vol. 32 (2019).
[3]
Avrim Blum and Tom Mitchell. 1998. Combining labeled and unlabeled data with co-training. In Proceedings of the eleventh annual conference on Computational learning theory. 92--100.
[4]
Ekin D Cubuk, Barret Zoph, Dandelion Mane, Vijay Vasudevan, and Quoc V Le. 2019. Autoaugment: Learning augmentation strategies from data. In CVPR. 113--123.
[5]
Ekin D Cubuk, Barret Zoph, Jonathon Shlens, and Quoc V Le. 2020. Randaugment: Practical automated data augmentation with a reduced search space. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops. 702--703.
[6]
Yaroslav Ganin and Victor Lempitsky. 2015. Unsupervised domain adaptation by backpropagation. In ICML. PMLR, 1180--1189.
[7]
Yaroslav Ganin, Evgeniya Ustinova, Hana Ajakan, Pascal Germain, Hugo Larochelle, Francc ois Laviolette, Mario Marchand, and Victor Lempitsky. 2016. Domain-adversarial training of neural networks. The journal of machine learning research, Vol. 17, 1 (2016), 2096--2030.
[8]
Behnam Gholami, Pritish Sahu, Ognjen Rudovic, Konstantinos Bousmalis, and Vladimir Pavlovic. 2020. Unsupervised multi-target domain adaptation: An information theoretic approach. IEEE Transactions on Image Processing, Vol. 29 (2020), 3993--4002.
[9]
Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. 2014. Generative adversarial nets. NIPS, Vol. 27 (2014).
[10]
Yves Grandvalet, Yoshua Bengio, et al. 2005. Semi-supervised learning by entropy minimization. CAP, Vol. 367 (2005), 281--296.
[11]
Sivan Harary, Eli Schwartz, Assaf Arbelle, Peter Staar, Shady Abu-Hussein, Elad Amrani, Roei Herzig, Amit Alfassy, Raja Giryes, Hilde Kuehne, et al. 2022. Unsupervised Domain Generalization by Learning a Bridge Across Domains. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 5280--5290.
[12]
Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. 2016. Deep residual learning for image recognition. In CVPR. 770--778.
[13]
Jiaxing Huang, Dayan Guan, Aoran Xiao, and Shijian Lu. 2021. Fsdr: Frequency space domain randomization for domain generalization. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 6891--6902.
[14]
Hiroshi Inoue. 2018. Data augmentation by pairing samples for images classification. arXiv preprint arXiv:1801.02929 (2018).
[15]
Philip TG Jackson, Amir Atapour Abarghouei, Stephen Bonner, Toby P Breckon, and Boguslaw Obara. 2019. Style augmentation: data augmentation via style randomization. In CVPR Workshops. 83--92.
[16]
Taekyung Kim and Changick Kim. 2020. Attract, perturb, and explore: Learning a feature alignment network for semi-supervised domain adaptation. In Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23-28, 2020, Proceedings, Part XIV 16. Springer, 591--607.
[17]
Samuli Laine and Timo Aila. 2016.Temporal ensembling for semi-supervised learning. arXiv preprint arXiv:1610.02242 (2016).
[18]
Yann LeCun, Léon Bottou, Yoshua Bengio, and Patrick Haffner. 1998. Gradient-based learning applied to document recognition. Proc. IEEE, Vol. 86, 11 (1998), 2278--2324.
[19]
Joseph Lemley, Shabab Bazrafkan, and Peter Corcoran. 2017. Smart augmentation learning an optimal data augmentation strategy. Ieee Access, Vol. 5 (2017), 5858--5869.
[20]
Daojun Liang, Feng Yang, Tian Zhang, and Peter Yang. 2018. Understanding mixup training methods. IEEE Access, Vol. 6 (2018), 58774--58783.
[21]
Tsung-Yi Lin, Michael Maire, Serge Belongie, James Hays, Pietro Perona, Deva Ramanan, Piotr Dollár, and C Lawrence Zitnick. 2014. Microsoft coco: Common objects in context. In European conference on computer vision. Springer, 740--755.
[22]
Mingsheng Long, Yue Cao, Jianmin Wang, and Michael Jordan. 2015. Learning transferable features with deep adaptation networks. In International conference on machine learning. PMLR, 97--105.
[23]
Yawei Luo, Ping Liu, Tao Guan, Junqing Yu, and Yi Yang. 2020. Adversarial style mining for one-shot unsupervised domain adaptation. Advances in Neural Information Processing Systems, Vol. 33 (2020), 20612--20623.
[24]
Christoph Mayer, Matthieu Paul, and Radu Timofte. 2021. Adversarial feature distribution alignment for semi-supervised learning. Computer Vision and Image Understanding, Vol. 202 (2021), 103109.
[25]
Zhongyi Pei, Zhangjie Cao, Mingsheng Long, and Jianmin Wang. 2018. Multi-adversarial domain adaptation. In AAAI.
[26]
Xingchao Peng, Qinxun Bai, Xide Xia, Zijun Huang, Kate Saenko, and Bo Wang. 2019. Moment matching for multi-source domain adaptation. In ICCV. 1406--1415.
[27]
Luis Perez and Jason Wang. 2017. The effectiveness of data augmentation in image classification using deep learning. arXiv preprint arXiv:1712.04621 (2017).
[28]
Antti Rasmus, Mathias Berglund, Mikko Honkala, Harri Valpola, and Tapani Raiko. 2015. Semi-supervised learning with ladder networks. Advances in neural information processing systems, Vol. 28 (2015).
[29]
Olga Russakovsky, Jia Deng, Hao Su, Jonathan Krause, Sanjeev Satheesh, Sean Ma, Zhiheng Huang, Andrej Karpathy, Aditya Khosla, Michael Bernstein, et al. 2015. Imagenet large scale visual recognition challenge. International journal of computer vision, Vol. 115, 3 (2015), 211--252.
[30]
Kuniaki Saito, Donghyun Kim, Stan Sclaroff, Trevor Darrell, and Kate Saenko. 2019. Semi-supervised domain adaptation via minimax entropy. In Proceedings of the IEEE/CVF International Conference on Computer Vision. 8050--8058.
[31]
Kuniaki Saito, Kohei Watanabe, Yoshitaka Ushiku, and Tatsuya Harada. 2018. Maximum classifier discrepancy for unsupervised domain adaptation. In CVPR. 3723--3732.
[32]
Kihyuk Sohn, David Berthelot, Nicholas Carlini, Zizhao Zhang, Han Zhang, Colin A Raffel, Ekin Dogus Cubuk, Alexey Kurakin, and Chun-Liang Li. 2020. Fixmatch: Simplifying semi-supervised learning with consistency and confidence. Advances in Neural Information Processing Systems, Vol. 33 (2020), 596--608.
[33]
Antti Tarvainen and Harri Valpola. 2017. Meanteachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results. In NeurIPS. 1195--1204.
[34]
Eric Tzeng, Judy Hoffman, Kate Saenko, and Trevor Darrell. 2017. Adversarial discriminative domain adaptation. In CVPR. 7167--7176.
[35]
Eric Tzeng, Judy Hoffman, Ning Zhang, Kate Saenko, and Trevor Darrell. 2014. Deep domain confusion: Maximizing for domain invariance. arXiv preprint arXiv:1412.3474 (2014).
[36]
Laurens Van der Maaten and Geoffrey Hinton. 2008. Visualizing data using t-SNE. Journal of machine learning research, Vol. 9, 11 (2008).
[37]
Ruiqi Wang, Lei Qi, Yinghuan Shi, and Yang Gao. 2023. Better pseudo-label: Joint domain-aware label and dual-classifier for semi-supervised domain generalization. Pattern Recognition, Vol. 133 (2023), 108987.
[38]
Qizhe Xie, Zihang Dai, Eduard Hovy, Thang Luong, and Quoc Le. 2020. Unsupervised data augmentation for consistency training. Advances in neural information processing systems, Vol. 33 (2020), 6256--6268.
[39]
Sergey Zagoruyko and Nikos Komodakis. 2016. Wide residual networks. arXiv preprint arXiv:1605.07146 (2016).
[40]
Michał Zajac, Konrad Zołna, Negar Rostamzadeh, and Pedro O Pinheiro. 2019. Adversarial framing for image and video classification. In Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 33. 10077--10078.
[41]
Bowen Zhang, Yidong Wang, Wenxin Hou, Hao Wu, Jindong Wang, Manabu Okumura, and Takahiro Shinozaki. 2021. Flexmatch: Boosting semi-supervised learning with curriculum pseudo labeling. NIPS, Vol. 34 (2021).
[42]
Hongyi Zhang, Moustapha Cisse, Yann N Dauphin, and David Lopez-Paz. 2017. mixup: Beyond empirical risk minimization. arXiv preprint arXiv:1710.09412 (2017).
[43]
Zhun Zhong, Liang Zheng, Guoliang Kang, Shaozi Li, and Yi Yang. 2020. Random erasing data augmentation. In AAAI, Vol. 34. 13001--13008.
[44]
Kaiyang Zhou, Yongxin Yang, Yu Qiao, and Tao Xiang. 2021a. Domain adaptive ensemble learning. IEEE Transactions on Image Processing, Vol. 30 (2021), 8008--8018.
[45]
Kaiyang Zhou, Yongxin Yang, Yu Qiao, and Tao Xiang. 2021b. Mixstyle neural networks for domain generalization and adaptation. arXiv preprint arXiv:2107.02053 (2021).
Recommendations
Inductive Semi-supervised Multi-Label Learning with Co-Training
KDD '17: Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data MiningIn multi-label learning, each training example is associated with multiple class labels and the task is to learn a mapping from the feature space to the power set of label space. It is generally demanding and time-consuming to obtain labels for training ...
Semi-Supervised Deep Learning Using Pseudo Labels for Hyperspectral Image Classification
Deep learning has gained popularity in a variety of computer vision tasks. Recently, it has also been successfully applied for hyperspectral image classification tasks. Training deep neural networks, such as a convolutional neural network for ...
On the relation between multi-instance learning and semi-supervised learning
ICML '07: Proceedings of the 24th international conference on Machine learningMulti-instance learning and semi-supervised learning are different branches of machine learning. The former attempts to learn from a training set consists of labeled bags each containing many unlabeled instances; the latter tries to exploit abundant ...
Comments
Information & Contributors
Information
Published In
August 2023
5996 pages
ISBN:9798400701030
DOI:10.1145/3580305
- General Chairs:
- Ambuj Singh,
- Yizhou Sun,
- Program Chairs:
- Leman Akoglu,
- Dimitrios Gunopulos,
- Xifeng Yan,
- Ravi Kumar,
- Fatma Ozcan,
- Jieping Ye
Copyright © 2023 ACM.
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than the author(s) must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Request permissions from [email protected].
Sponsors
Publisher
Association for Computing Machinery
New York, NY, United States
Publication History
Published: 04 August 2023
Check for updates
Author Tags
Qualifiers
- Research-article
Funding Sources
- the Fundamental Research Funds for the Central Universities
- the National Science Fund for Distinguished Young Scholars
Conference
KDD '23: The 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining
August 6 - 10, 2023
CA, Long Beach, USA
Acceptance Rates
Overall Acceptance Rate 1,133 of 8,635 submissions, 13%
Upcoming Conference
Contributors
Other Metrics
Bibliometrics & Citations
Bibliometrics
Article Metrics
- 0Total Citations
- 332Total Downloads
- Downloads (Last 12 months)175
- Downloads (Last 6 weeks)11
Reflects downloads up to 25 Jan 2025
Other Metrics
Citations
View Options
Login options
Check if you have access through your login credentials or your institution to get full access on this article.
Sign in