self training with noisy student improves imagenet classification

To achieve this result, we first train an EfficientNet model on labeled ImageNet images and use it as a teacher to generate pseudo labels on 300M unlabeled images. Noisy Student Training is based on the self-training framework and trained with 4 simple steps: Train a classifier on labeled data (teacher). Train a larger classifier on the combined set, adding noise (noisy student). Then, that teacher is used to label the unlabeled data. This result is also a new state-of-the-art and 1% better than the previous best method that used an order of magnitude more weakly labeled data [ 44, 71]. Noisy Student (B7, L2) means to use EfficientNet-B7 as the student and use our best model with 87.4% accuracy as the teacher model. Hence, a question that naturally arises is why the student can outperform the teacher with soft pseudo labels. When data augmentation noise is used, the student must ensure that a translated image, for example, should have the same category with a non-translated image. IEEE Trans. Soft pseudo labels lead to better performance for low confidence data. Selected images from robustness benchmarks ImageNet-A, C and P. Test images from ImageNet-C underwent artificial transformations (also known as common corruptions) that cannot be found on the ImageNet training set. If nothing happens, download GitHub Desktop and try again. For this purpose, we use a much larger corpus of unlabeled images, where some images may not belong to any category in ImageNet. As shown in Table3,4 and5, when compared with the previous state-of-the-art model ResNeXt-101 WSL[44, 48] trained on 3.5B weakly labeled images, Noisy Student yields substantial gains on robustness datasets. The most interesting image is shown on the right of the first row. Our work is based on self-training (e.g.,[59, 79, 56]). mFR (mean flip rate) is the weighted average of flip probability on different perturbations, with AlexNets flip probability as a baseline. On ImageNet-C, it reduces mean corruption error (mCE) from 45.7 to 31.2. Noisy Student Training extends the idea of self-training and distillation with the use of equal-or-larger student models and noise added to the student during learning. EfficientNet with Noisy Student produces correct top-1 predictions (shown in. To noise the student, we use dropout[63], data augmentation[14] and stochastic depth[29] during its training. Unlike previous studies in semi-supervised learning that use in-domain unlabeled data (e.g, ., CIFAR-10 images as unlabeled data for a small CIFAR-10 training set), to improve ImageNet, we must use out-of-domain unlabeled data. Since a teacher models confidence on an image can be a good indicator of whether it is an out-of-domain image, we consider the high-confidence images as in-domain images and the low-confidence images as out-of-domain images. Please refer to [24] for details about mFR and AlexNets flip probability. We will then show our results on ImageNet and compare them with state-of-the-art models. Learn more. In particular, we set the survival probability in stochastic depth to 0.8 for the final layer and follow the linear decay rule for other layers. Lastly, we will show the results of benchmarking our model on robustness datasets such as ImageNet-A, C and P and adversarial robustness. Iterative training is not used here for simplicity. The top-1 accuracy of prior methods are computed from their reported corruption error on each corruption. A self-training method that better adapt to the popular two stage training pattern for multi-label text classification under a semi-supervised scenario by continuously finetuning the semantic space toward increasing high-confidence predictions, intending to further promote the performance on target tasks. We verify that this is not the case when we use 130M unlabeled images since the model does not overfit the unlabeled set from the training loss. We thank the Google Brain team, Zihang Dai, Jeff Dean, Hieu Pham, Colin Raffel, Ilya Sutskever and Mingxing Tan for insightful discussions, Cihang Xie for robustness evaluation, Guokun Lai, Jiquan Ngiam, Jiateng Xie and Adams Wei Yu for feedbacks on the draft, Yanping Huang and Sameer Kumar for improving TPU implementation, Ekin Dogus Cubuk and Barret Zoph for help with RandAugment, Yanan Bao, Zheyun Feng and Daiyi Peng for help with the JFT dataset, Olga Wichrowska and Ola Spyra for help with infrastructure. During the learning of the student, we inject noise such as dropout, stochastic depth, and data augmentation via RandAugment to the student so that the student generalizes better than the teacher. Noisy Student improves adversarial robustness against an FGSM attack though the model is not optimized for adversarial robustness. We sample 1.3M images in confidence intervals. Noisy Student self-training is an effective way to leverage unlabelled datasets and improving accuracy by adding noise to the student model while training so it learns beyond the teacher's knowledge. Figure 1(b) shows images from ImageNet-C and the corresponding predictions. On robustness test sets, it improves ImageNet-A top . on ImageNet ReaL ImageNet-A top-1 accuracy from 16.6 . In other words, the student is forced to mimic a more powerful ensemble model. ; 2006)[book reviews], Semi-supervised deep learning with memory, Proceedings of the European Conference on Computer Vision (ECCV), Xception: deep learning with depthwise separable convolutions, K. Clark, M. Luong, C. D. Manning, and Q. V. Le, Semi-supervised sequence modeling with cross-view training, E. D. Cubuk, B. Zoph, D. Mane, V. Vasudevan, and Q. V. Le, AutoAugment: learning augmentation strategies from data, Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, E. D. Cubuk, B. Zoph, J. Shlens, and Q. V. Le, RandAugment: practical data augmentation with no separate search, Z. Dai, Z. Yang, F. Yang, W. W. Cohen, and R. R. Salakhutdinov, Good semi-supervised learning that requires a bad gan, T. Furlanello, Z. C. Lipton, M. Tschannen, L. Itti, and A. Anandkumar, A. Galloway, A. Golubeva, T. Tanay, M. Moussa, and G. W. Taylor, R. Geirhos, P. Rubisch, C. Michaelis, M. Bethge, F. A. Wichmann, and W. Brendel, ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness, J. Gilmer, L. Metz, F. Faghri, S. S. Schoenholz, M. Raghu, M. Wattenberg, and I. Goodfellow, I. J. Goodfellow, J. Shlens, and C. Szegedy, Explaining and harnessing adversarial examples, Semi-supervised learning by entropy minimization, Advances in neural information processing systems, K. Gu, B. Yang, J. Ngiam, Q. However an important requirement for Noisy Student to work well is that the student model needs to be sufficiently large to fit more data (labeled and pseudo labeled). Our experiments showed that self-training with Noisy Student and EfficientNet can achieve an accuracy of 87.4% which is 1.9% higher than without Noisy Student. In the following, we will first describe experiment details to achieve our results. To achieve this result, we first train an EfficientNet model on labeled ImageNet images and use it as a teacher to generate pseudo labels on 300M unlabeled images. We use our best model Noisy Student with EfficientNet-L2 to teach student models with sizes ranging from EfficientNet-B0 to EfficientNet-B7. Infer labels on a much larger unlabeled dataset. Specifically, we train the student model for 350 epochs for models larger than EfficientNet-B4, including EfficientNet-L0, L1 and L2 and train the student model for 700 epochs for smaller models. The ADS is operated by the Smithsonian Astrophysical Observatory under NASA Cooperative On robustness test sets, it improves ImageNet-A top-1 accuracy from 61.0% to 83.7%, reduces ImageNet-C mean corruption error from 45.7 to 28.3, and reduces ImageNet-P mean flip rate from 27.8 to 12.2. Our largest model, EfficientNet-L2, needs to be trained for 3.5 days on a Cloud TPU v3 Pod, which has 2048 cores. Conclusion, Abstract , ImageNet , web-scale extra labeled images weakly labeled Instagram images weakly-supervised learning . task. Self-training with Noisy Student. This work introduces two challenging datasets that reliably cause machine learning model performance to substantially degrade and curates an adversarial out-of-distribution detection dataset called IMAGENET-O, which is the first out- of-dist distribution detection dataset created for ImageNet models. Chowdhury et al. This model investigates a new method for incorporating unlabeled data into a supervised learning pipeline. Self-Training With Noisy Student Improves ImageNet Classification @article{Xie2019SelfTrainingWN, title={Self-Training With Noisy Student Improves ImageNet Classification}, author={Qizhe Xie and Eduard H. Hovy and Minh-Thang Luong and Quoc V. Le}, journal={2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)}, year={2019 . Self-training with Noisy Student improves ImageNet classificationCVPR2020, Codehttps://github.com/google-research/noisystudent, Self-training, 1, 2Self-training, Self-trainingGoogleNoisy Student, Noisy Studentstudent modeldropout, stochastic depth andaugmentationteacher modelNoisy Noisy Student, Noisy Student, 1, JFT3ImageNetEfficientNet-B00.3130K130K, EfficientNetbaseline modelsEfficientNetresnet, EfficientNet-B7EfficientNet-L0L1L2, batchsize = 2048 51210242048EfficientNet-B4EfficientNet-L0l1L2350epoch700epoch, 2EfficientNet-B7EfficientNet-L0, 3EfficientNet-L0EfficientNet-L1L0, 4EfficientNet-L1EfficientNet-L2, student modelNoisy, noisystudent modelteacher modelNoisy, Noisy, Self-trainingaugmentationdropoutstochastic depth, Our largest model, EfficientNet-L2, needs to be trained for 3.5 days on a Cloud TPU v3 Pod, which has 2048 cores., 12/self-training-with-noisy-student-f33640edbab2, EfficientNet-L0EfficientNet-B7B7, EfficientNet-L1EfficientNet-L0, EfficientNetsEfficientNet-L1EfficientNet-L2EfficientNet-L2EfficientNet-B75. w Summary of key results compared to previous state-of-the-art models. As noise injection methods are not used in the student model, and the student model was also small, it is more difficult to make the student better than teacher. For simplicity, we experiment with using 1128,164,132,116,14 of the whole data by uniformly sampling images from the the unlabeled set though taking the images with highest confidence leads to better results. Chum, Label propagation for deep semi-supervised learning, D. P. Kingma, S. Mohamed, D. J. Rezende, and M. Welling, Semi-supervised learning with deep generative models, Semi-supervised classification with graph convolutional networks. The top-1 accuracy is simply the average top-1 accuracy for all corruptions and all severity degrees. (or is it just me), Smithsonian Privacy This is a recurring payment that will happen monthly, If you exceed more than 500 images, they will be charged at a rate of $5 per 500 images. Then we finetune the model with a larger resolution for 1.5 epochs on unaugmented labeled images. The abundance of data on the internet is vast. Our main results are shown in Table1. Work fast with our official CLI. Use, Smithsonian Noisy Student Training is based on the self-training framework and trained with 4-simple steps: This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository. Their noise model is video specific and not relevant for image classification. The paradigm of pre-training on large supervised datasets and fine-tuning the weights on the target task is revisited, and a simple recipe that is called Big Transfer (BiT) is created, which achieves strong performance on over 20 datasets. Noisy Student Training is a semi-supervised learning method which achieves 88.4% top-1 accuracy on ImageNet (SOTA) and surprising gains on robustness and adversarial benchmarks. For classes where we have too many images, we take the images with the highest confidence. Due to duplications, there are only 81M unique images among these 130M images. A. Krizhevsky, I. Sutskever, and G. E. Hinton, Temporal ensembling for semi-supervised learning, Pseudo-label: the simple and efficient semi-supervised learning method for deep neural networks, Workshop on Challenges in Representation Learning, ICML, Certainty-driven consistency loss for semi-supervised learning, C. Liu, B. Zoph, M. Neumann, J. Shlens, W. Hua, L. Li, L. Fei-Fei, A. Yuille, J. Huang, and K. Murphy, R. G. Lopes, D. Yin, B. Poole, J. Gilmer, and E. D. Cubuk, Improving robustness without sacrificing accuracy with patch gaussian augmentation, Y. Luo, J. Zhu, M. Li, Y. Ren, and B. Zhang, Smooth neighbors on teacher graphs for semi-supervised learning, L. Maale, C. K. Snderby, S. K. Snderby, and O. Winther, A. Madry, A. Makelov, L. Schmidt, D. Tsipras, and A. Vladu, Towards deep learning models resistant to adversarial attacks, D. Mahajan, R. Girshick, V. Ramanathan, K. He, M. Paluri, Y. Li, A. Bharambe, and L. van der Maaten, Exploring the limits of weakly supervised pretraining, T. Miyato, S. Maeda, S. Ishii, and M. Koyama, Virtual adversarial training: a regularization method for supervised and semi-supervised learning, IEEE transactions on pattern analysis and machine intelligence, A. Najafi, S. Maeda, M. Koyama, and T. Miyato, Robustness to adversarial perturbations in learning from incomplete data, J. Ngiam, D. Peng, V. Vasudevan, S. Kornblith, Q. V. Le, and R. Pang, Robustness properties of facebooks resnext wsl models, Adversarial dropout for supervised and semi-supervised learning, Lessons from building acoustic models with a million hours of speech, IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), S. Qiao, W. Shen, Z. Zhang, B. Wang, and A. Yuille, Deep co-training for semi-supervised image recognition, I. Radosavovic, P. Dollr, R. Girshick, G. Gkioxari, and K. He, Data distillation: towards omni-supervised learning, A. Rasmus, M. Berglund, M. Honkala, H. Valpola, and T. Raiko, Semi-supervised learning with ladder networks, E. Real, A. Aggarwal, Y. Huang, and Q. V. Le, Proceedings of the AAAI Conference on Artificial Intelligence, B. Recht, R. Roelofs, L. Schmidt, and V. Shankar. But during the learning of the student, we inject noise such as data Our experiments showed that self-training with Noisy Student and EfficientNet can achieve an accuracy of 87.4% which is 1.9% higher than without Noisy Student. On ImageNet-P, it leads to an mean flip rate (mFR) of 17.8 if we use a resolution of 224x224 (direct comparison) and 16.1 if we use a resolution of 299x299.111For EfficientNet-L2, we use the model without finetuning with a larger test time resolution, since a larger resolution results in a discrepancy with the resolution of data and leads to degraded performance on ImageNet-C and ImageNet-P. This attack performs one gradient descent step on the input image[20] with the update on each pixel set to . We present Noisy Student Training, a semi-supervised learning approach that works well even when labeled data is abundant. On, International journal of molecular sciences. Their framework is highly optimized for videos, e.g., prediction on which frame to use in a video, which is not as general as our work. We iterate this process by putting back the student as the teacher. C. Szegedy, S. Ioffe, V. Vanhoucke, and A. Next, with the EfficientNet-L0 as the teacher, we trained a student model EfficientNet-L1, a wider model than L0. Self-training was previously used to improve ResNet-50 from 76.4% to 81.2% top-1 accuracy[76] which is still far from the state-of-the-art accuracy. ImageNet-A test set[25] consists of difficult images that cause significant drops in accuracy to state-of-the-art models. The algorithm is basically self-training, a method in semi-supervised learning (. Although they have produced promising results, in our preliminary experiments, consistency regularization works less well on ImageNet because consistency regularization in the early phase of ImageNet training regularizes the model towards high entropy predictions, and prevents it from achieving good accuracy. You can also use the colab script noisystudent_svhn.ipynb to try the method on free Colab GPUs. Train a classifier on labeled data (teacher). During the generation of the pseudo labels, the teacher is not noised so that the pseudo labels are as accurate as possible. A. Alemi, Thirty-First AAAI Conference on Artificial Intelligence, C. Szegedy, W. Liu, Y. Jia, P. Sermanet, S. Reed, D. Anguelov, D. Erhan, V. Vanhoucke, and A. Rabinovich, C. Szegedy, V. Vanhoucke, S. Ioffe, J. Shlens, and Z. Wojna, Rethinking the inception architecture for computer vision, C. Szegedy, W. Zaremba, I. Sutskever, J. Bruna, D. Erhan, I. Goodfellow, and R. Fergus, EfficientNet: rethinking model scaling for convolutional neural networks, Mean teachers are better role models: weight-averaged consistency targets improve semi-supervised deep learning results, H. Touvron, A. Vedaldi, M. Douze, and H. Jgou, Fixing the train-test resolution discrepancy, V. Verma, A. Lamb, J. Kannala, Y. Bengio, and D. Lopez-Paz, Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence (IJCAI-19), J. Weston, F. Ratle, H. Mobahi, and R. Collobert, Deep learning via semi-supervised embedding, Q. Xie, Z. Dai, E. Hovy, M. Luong, and Q. V. Le, Unsupervised data augmentation for consistency training, S. Xie, R. Girshick, P. Dollr, Z. Tu, and K. He, Aggregated residual transformations for deep neural networks, I.