Improving Semi-supervised Federated Learning by Reducing the Gradient Diversity of Models
Open accessThe DOI matches its Crossref record; 32 of the 32 references it cites were found in a registry or an index.
Abstract
Federated learning (FL) is a promising way to use the computing power of mobile devices while maintaining the privacy of users. Current work in FL, however, makes the unrealistic assumption that the users have ground-truth labels on their devices, while also assuming that the server has neither data nor labels. In this work, we consider the more realistic scenario where the users have only unlabeled data, while the server has some labeled data, and where the amount of labeled data is smaller than the amount of unlabeled data. We call this learning problem semi-supervised federated learning (SSFL). For SSFL, we demonstrate that a critical issue that affects the test accuracy is the large gradient diversity of the models from different users. Based on this, we investigate several design choices. First, we find that the so-called consistency regularization loss (CRL), which is widely used in semi-supervised learning, performs reasonably well but has large gradient diversity. Second, we find that Batch Normalization (BN) increases gradient diversity. Replacing BN with the recently-proposed Group Normalization (GN) can reduce gradient diversity and improve test accuracy. Third, we show that CRL combined with GN still has a large gradient diversity when the number of users is large. Based on these results, we propose a novel grouping-based model averaging method to replace the FedAvg averaging method. Overall, our grouping-based averaging, combined with GN and CRL, achieves better test accuracy than not just a contemporary paper on SSFL in the same settings (>10%), but also four supervised FL algorithms.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474
ImageNet: A large-scale hierarchical image database
Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., & Fei-Fei, L. (2009) · 2009 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2009.5206848 · Cited by 65,333
EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks
Tan, M., & Le, Q. V. (2019) · arXiv · doi:10.48550/arxiv.1905.11946 · Cited by 17,072 · cited 1×
“which improves the state-of-the-art accuracy on ImageNet [57], even compared to supervised learning [29], [58].”
BackgroundAdvances and Open Problems in Federated Learning
Kairouz, P., McMahan, H. B., Avent, B., Bellet, A., Bennis, M., Bhagoji, A. N., Bonawitz, K., Charles, Z., Cormode, G., Cummings, R., et al. (2019) · arXiv · doi:10.48550/arxiv.1912.04977 · Cited by 7,857
Advances and Open Problems in Federated Learning
Kairouz, P., & McMahan, H. B. (2021) · Foundations and Trends in Machine Learning · doi:10.1561/2200000083 · Cited by 7,772
mixup: Beyond Empirical Risk Minimization
Zhang, H., Cisse, M., Dauphin, Y. N., & Lopez-Paz, D. (2017) · arXiv · doi:10.48550/arxiv.1710.09412 · Cited by 5,831
SGDR: Stochastic Gradient Descent with Warm Restarts
Loshchilov, I., & Hutter, F. (2016) · arXiv · doi:10.48550/arxiv.1608.03983 · Cited by 5,309
Tarvainen, A., & Valpola, H. (2017) · arXiv · doi:10.48550/arxiv.1703.01780 · Cited by 4,275
Federated Optimization in Heterogeneous Networks
Li, T., Sahu, A. K., Zaheer, M., Sanjabi, M., Talwalkar, A., & Smith, V. (2018) · arXiv · doi:10.48550/arxiv.1812.06127 · Cited by 4,023
Practical Secure Aggregation for Privacy-Preserving Machine Learning
Bonawitz, K., Ivanov, V., Kreuter, B., Marcedone, A., McMahan, H. B., Patel, S., Ramage, D., Segal, A., & Seth, K. (2017) · Proceedings of the 2017 ACM SIGSAC Conference on Computer and Communications Security · doi:10.1145/3133956.3133982 · Cited by 3,956 · cited 1×
“In addition to the challenge of the non-iidness of the data distribution and the need for local ground truth labels, communication efficiency is another critical problem in FL [37], [42]–[45].”
BackgroundFederated Learning: Strategies for Improving Communication Efficiency
Konečný, J., McMahan, H. B., Yu, F. X., Richtárik, P., Suresh, A. T., & Bacon, D. (2016) · arXiv · doi:10.48550/arxiv.1610.05492 · Cited by 3,766
Wu, Y., & He, K. (2020) · International Journal of Computer Vision · doi:10.1007/s11263-019-01198-w · Cited by 3,601 · cited 1×
“Then, we show that replacing the batch normalization (BN) [12] in the model with group normalization (GN) [13] can reduce gradient diversity and enhance test accuracy in the SSFL setting.”
MethodRandAugment: Practical automated data augmentation with a reduced search space
Cubuk, E. D., Zoph, B., Shlens, J., & Le, Q. V. (2019) · arXiv · doi:10.48550/arxiv.1909.13719 · Cited by 3,223 · cited 4×
“Here, the latter RandAugment uses two different augmentation methods (i.e., shift and crop) out of twelve possible augmentation methods (e.g., rotate, shift, solarize, etc.) for one image.”
BackgroundMethodInfluentialRandaugment: Practical automated data augmentation with a reduced search space
Cubuk, E. D., Zoph, B., Shlens, J., & Le, Q. V. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw50498.2020.00359 · Cited by 3,223
FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence
Sohn, K., Berthelot, D., Li, C.-L., Zhang, Z., Carlini, N., Cubuk, E. D., Kurakin, A., Zhang, H., & Raffel, C. (2020) · arXiv · doi:10.48550/arxiv.2001.07685 · Cited by 2,892 · cited 10×
“This is shown in [11] to boost the testing performance.”
BackgroundMethodResultInfluentialSelf-citationSelf-Training With Noisy Student Improves ImageNet Classification
Xie, Q., Luong, M.-T., Hovy, E., & Le, Q. V. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.01070 · Cited by 2,393
Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour
Goyal, P., Dollár, P., Girshick, R., Noordhuis, P., Wesolowski, L., Kyrola, A., Tulloch, A., Jia, Y., & He, K. (2017) · arXiv · doi:10.48550/arxiv.1706.02677 · Cited by 2,294
Federated Learning with Non-IID Data
Zhao, Y., Li, M., Lai, L., Suda, N., Civin, D., & Chandra, V. (2018) · arXiv · doi:10.48550/arxiv.1806.00582 · Cited by 2,077 · cited 11×
“Federated learning (FL) [1], [2], [4], [16], [18], [34]–[38] is a decentralized computing framework that enables multiple users to learn a shared model while potentially protecting the privacy of users (although recent”
BackgroundMethodInfluentialCheap and fast---but is it good?: evaluating non-expert annotations for natural language tasks
Snow, R., O'Connor, B., Jurafsky, D., & Ng, A. Y. (2008) · Proceedings of the Conference on Empirical Methods in Natural Language Processing - EMNLP '08 · doi:10.3115/1613715.1613751 · Cited by 1,887
Robust and Communication-Efficient Federated Learning From Non-i.i.d. Data
Sattler, F., Wiedemann, S., Muller, K.-R., & Samek, W. (2020) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2019.2944481 · Cited by 1,725
AutoAugment: Learning Augmentation Policies from Data
Cubuk, E. D., Zoph, B., Mane, D., Vasudevan, V., & Le, Q. V. (2018) · arXiv · doi:10.48550/arxiv.1805.09501 · Cited by 1,106
Adaptive Federated Optimization
Reddi, S., Charles, Z., Zaheer, M., Garrett, Z., Rush, K., Konečný, J., Kumar, S., & McMahan, H. B. (2020) · arXiv · doi:10.48550/arxiv.2003.00295 · Cited by 840 · cited 2×
“We use ResNet as the training model on both Cifar10 and SVHN datasets; and we use the same CNN model as [27] on EMNIST.”
BackgroundMethodSimilarity of Neural Network Representations Revisited
Kornblith, S., Norouzi, M., Lee, H., & Hinton, G. (2019) · arXiv · doi:10.48550/arxiv.1905.00414 · Cited by 724 · cited 1×
“It is worth noting that metrics based on the similarity between weights, feature representations, and gradients have been widely explored to analyze the trainability and the generalization performance of learning models [10], [64], [65].”
BackgroundSemi-supervised Knowledge Transfer for Deep Learning from Private Training Data
Papernot, N., Abadi, M., Erlingsson, Ú., Goodfellow, I., & Talwar, K. (2016) · arXiv · doi:10.48550/arxiv.1610.05755 · Cited by 492 · cited 1×
“Another paper [63] studies semi-supervised private aggregation of an ensemble of teacher models trained on separate subsets of the whole dataset, which is not in the FL setting but is closely related.”
BackgroundJeong, E., Oh, S., Kim, H., Park, J., Bennis, M., & Kim, S.-L. (2018) · arXiv · doi:10.48550/arxiv.1811.11479 · Cited by 465 · cited 1×
“In fact, dealing with non-iid distributions is one of the most critical challenges in FL [16], [35], [40].”
BackgroundCommunication-efficient federated learning
Chen, M., Shlezinger, N., Poor, H. V., Eldar, Y. C., & Cui, S. (2021) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2024789118 · Cited by 397
Improving Federated Learning Personalization via Model Agnostic Meta Learning
Jiang, Y., Konečný, J., Rush, K., & Kannan, S. (2019) · arXiv · doi:10.48550/arxiv.1909.12488 · Cited by 381 · cited 1×
“In addition, personalization [67], [68] is important”
Itahara, S., Nishio, T., Koda, Y., Morikura, M., & Yamamoto, K. (2020) · arXiv · doi:10.48550/arxiv.2008.06180 · Cited by 338
Itahara, S., Nishio, T., Koda, Y., Morikura, M., & Yamamoto, K. (2023) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2021.3070013 · Cited by 338 · cited 1×
“For example, the paper [52] considers using shared unlabeled data for distillation-based message exchanging.”
BackgroundExpanding the Reach of Federated Learning by Reducing Client Resource Requirements
Caldas, S., Konečny, J., McMahan, H. B., & Talwalkar, A. (2018) · arXiv · doi:10.48550/arxiv.1812.07210 · Cited by 312 · cited 1×
“Federated learning (FL) [1, 2, 4, 27, 16, 28, 29, 30, 18, 31] is a decentralized computing framework that enables multiple users to learn a shared model while potentially protecting the privacy of users (although recent work [32] shows this may not be the case).”
BackgroundYang, D., Xu, Z., Li, W., Myronenko, A., Roth, H. R., Harmon, S., Xu, S., Turkbey, B., Turkbey, E., Wang, X., et al. (2021) · Medical Image Analysis · doi:10.1016/j.media.2021.101992 · Cited by 279
Exploiting Unlabeled Data in Smart Cities using Federated Edge Learning
Albaseer, A., Ciftler, B. S., Abdallah, M., & Al-Fuqaha, A. (2020) · 2020 International Wireless Communications and Mobile Computing (IWCMC) · doi:10.1109/iwcmc48107.2020.9148475 · Cited by 84
Cited by
Federated learning on non-IID data: A survey
Zhu, H., Xu, J., Liu, S., & Jin, Y. (2021) · Neurocomputing · doi:10.1016/j.neucom.2021.07.098 · Cited by 1,062
Semi-Supervised Federated Learning for Travel Mode Identification From GPS Trajectories
Zhu, Y., Liu, Y., Yu, J. J. Q., & Yuan, X. (2022) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2021.3092015 · Cited by 79
“Typically, FedAvg [29] can be used in the SSFL system to aggregate model updates : However, if we direct use the vanilla FedAvg given in the Algorithm 1, the model diversity between different clients brings great challenges to training and aggregation [47].”
MethodModel-Heterogeneous Semi-Supervised Federated Learning for Medical Image Segmentation
Ma, Y., Wang, J., Yang, J., & Wang, L. (2024) · IEEE Transactions on Medical Imaging · doi:10.1109/tmi.2023.3348982 · Cited by 55
A Label Efficient Semi Self-Supervised Learning Framework for IoT Devices in Industrial Process
Bharti, V., Kumar, A., Purohit, V., Singh, R., Singh, A. K., & Singh, S. K. (2024) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2023.3289184 · Cited by 15
Metadata
- Authors
- Zhang, Z., Yang, Y., Yao, Z., Yan, Y., Gonzalez, J. E., Ramchandran, K., & Mahoney, M. W.
- Venue
- 2021 IEEE International Conference on Big Data (Big Data)
- Publisher
- IEEE
- Published
- 2021-12-15
- Pages
- 1214-1225
- OpenAlex
- W4225482284
- Cited by
- 83
- References
- 32
- FWCI
- 6.54
- Type
- conference paper
- Language
- en
- Open access
- green
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.