We present Noisy Student Training, a semi-supervised learning approach that works well even when labeled data is abundant. Noisy Student Training achieves 88.4% top-1 accuracy on ImageNet, which is 2.0% better than the state-of-the-art model that requires 3.5B weakly labeled Instagram images. 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.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. On Im-ageNet, we first train an EfficientNet model on labeled images and use it as a teacher to generate pseudo labels for 300M unlabeled images. We then train a larger Efficient-Net as a student model on the combination of labeled and pseudo labeled images. We iterate this process by putting back the student as the teacher. 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. 1 * This work was conducted at Google.

Semi-supervised learning (SSL) provides an effective means of leveraging unlabeled data to improve a model's performance. This domain has seen fast progress recently, at the cost of requiring more complex methods. In this paper we propose FixMatch, an algorithm that is a significant simplification of existing SSL methods. FixMatch first generates pseudo-labels using the model's predictions on weaklyaugmented unlabeled images. For a given image, the pseudo-label is only retained if the model produces a high-confidence prediction. The model is then trained to predict the pseudo-label when fed a strongly-augmented version of the same image. Despite its simplicity, we show that FixMatch achieves state-of-the-art performance across a variety of standard semi-supervised learning benchmarks, including 94.93% accuracy on CIFAR-10 with 250 labels and 88.61% accuracy with 40 -just 4 labels per class. We carry out an extensive ablation study to tease apart the experimental factors that are most important to FixMatch's success. The code is available at https://github.com/google-research/fixmatch.

Semi-supervised learning lately has shown much promise in improving deep learning models when labeled data is scarce. Common among recent approaches is the use of consistency training on a large amount of unlabeled data to constrain model predictions to be invariant to input noise. In this work, we present a new perspective on how to effectively noise unlabeled examples and argue that the quality of noising, specifically those produced by advanced data augmentation methods, plays a crucial role in semi-supervised learning. By substituting simple noising operations with advanced data augmentation methods such as RandAugment and back-translation, our method brings substantial improvements across six language and three vision tasks under the same consistency training framework. On the IMDb text classification dataset, with only 20 labeled examples, our method achieves an error rate of 4.20, outperforming the state-of-the-art model trained on 25,000 labeled examples. On a standard semi-supervised learning benchmark, CIFAR-10, our method outperforms all previous approaches and achieves an error rate of 5.43 with only 250 examples. Our method also combines well with transfer learning, e.g., when finetuning from BERT, and yields improvements in high-data regime, such as ImageNet, whether when there is only 10% labeled data or when a full labeled set with 1.3M extra unlabeled examples is used. 1

We present Self Meta Pseudo Labels, a novel semi-supervised learning method similar to Meta Pseudo Labels but without the teacher model. We introduce a novel way to use a single model for both generating pseudo labels and classification, allowing us to store only one model in memory instead of two. Our method attains similar performance to the Meta Pseudo Labels method while drastically reducing memory usage.

在实现最先进的性能和在实际应用中负担得起的大型模型之间，计算机视觉的差异越来越大。在本文中，我们解决了这个问题，并显着弥合了这两种模型之间的差距。在我们的实证研究中，我们不一定要提出一种新方法，而是要努力确定一个可靠的有效食谱，以使最先进的大型模型在实践中负担得起。我们证明，当正确执行时，知识蒸馏可以成为减少大型尺寸而不损害其性能的强大工具。特别是，我们发现存在某些隐式设计选择，这可能会严重影响蒸馏的有效性。我们的关键贡献是对这些设计选择的明确识别，这些选择以前在文献中尚未阐明。我们通过一项全面的实证研究备份了我们的发现，在广泛的视觉数据集上展示了令人信服的结果，尤其是获得了最先进的Imagenet Resnet-50模型，该模型可实现82.8％的Top-1准确性。 。

培训深层神经网络以识别图像识别通常需要大规模的人类注释数据。为了减少深神经溶液对标记数据的依赖，文献中已经提出了最先进的半监督方法。尽管如此，在面部表达识别领域（FER）领域，使用这种半监督方法非常罕见。在本文中，我们介绍了一项关于最近提出的在FER背景下的最先进的半监督学习方法的全面研究。我们对八种半监督学习方法进行了比较研究当使用各种标记的样品时。我们还将这些方法的性能与完全监督的培训进行了比较。我们的研究表明，当培训现有的半监督方法时，每类标记的样本只有250个标记的样品可以产生可比的性能，而在完整标记的数据集中训练的完全监督的方法。为了促进该领域的进一步研究，我们在：https：//github.com/shuvenduroy/ssl_fer上公开提供代码

The recently proposed Temporal Ensembling has achieved state-of-the-art results in several semi-supervised learning benchmarks. It maintains an exponential moving average of label predictions on each training example, and penalizes predictions that are inconsistent with this target. However, because the targets change only once per epoch, Temporal Ensembling becomes unwieldy when learning large datasets. To overcome this problem, we propose Mean Teacher, a method that averages model weights instead of label predictions. As an additional benefit, Mean Teacher improves test accuracy and enables training with fewer labels than Temporal Ensembling. Without changing the network architecture, Mean Teacher achieves an error rate of 4.35% on SVHN with 250 labels, outperforming Temporal Ensembling trained with 1000 labels. We also show that a good network architecture is crucial to performance. Combining Mean Teacher and Residual Networks, we improve the state of the art on CIFAR-10 with 4000 labels from 10.55% to 6.28%, and on ImageNet 2012 with 10% of the labels from 35.24% to 9.11%.

Deep learning has emerged as an effective solution for solving the task of object detection in images but at the cost of requiring large labeled datasets. To mitigate this cost, semi-supervised object detection methods, which consist in leveraging abundant unlabeled data, have been proposed and have already shown impressive results. However, most of these methods require linking a pseudo-label to a ground-truth object by thresholding. In previous works, this threshold value is usually determined empirically, which is time consuming, and only done for a single data distribution. When the domain, and thus the data distribution, changes, a new and costly parameter search is necessary. In this work, we introduce our method Adaptive Self-Training for Object Detection (ASTOD), which is a simple yet effective teacher-student method. ASTOD determines without cost a threshold value based directly on the ground value of the score histogram. To improve the quality of the teacher predictions, we also propose a novel pseudo-labeling procedure. We use different views of the unlabeled images during the pseudo-labeling step to reduce the number of missed predictions and thus obtain better candidate labels. Our teacher and our student are trained separately, and our method can be used in an iterative fashion by replacing the teacher by the student. On the MS-COCO dataset, our method consistently performs favorably against state-of-the-art methods that do not require a threshold parameter, and shows competitive results with methods that require a parameter sweep search. Additional experiments with respect to a supervised baseline on the DIOR dataset containing satellite images lead to similar conclusions, and prove that it is possible to adapt the score threshold automatically in self-training, regardless of the data distribution.

One paradigm for learning from few labeled examples while making best use of a large amount of unlabeled data is unsupervised pretraining followed by supervised fine-tuning. Although this paradigm uses unlabeled data in a task-agnostic way, in contrast to common approaches to semi-supervised learning for computer vision, we show that it is surprisingly effective for semi-supervised learning on ImageNet. A key ingredient of our approach is the use of big (deep and wide) networks during pretraining and fine-tuning. We find that, the fewer the labels, the more this approach (task-agnostic use of unlabeled data) benefits from a bigger network. After fine-tuning, the big network can be further improved and distilled into a much smaller one with little loss in classification accuracy by using the unlabeled examples for a second time, but in a task-specific way. The proposed semi-supervised learning algorithm can be summarized in three steps: unsupervised pretraining of a big ResNet model using SimCLRv2, supervised fine-tuning on a few labeled examples, and distillation with unlabeled examples for refining and transferring the task-specific knowledge. This procedure achieves 73.9% ImageNet top-1 accuracy with just 1% of the labels (≤13 labeled images per class) using ResNet-50, a 10× improvement in label efficiency over the previous state-of-theart. With 10% of labels, ResNet-50 trained with our method achieves 77.5% top-1 accuracy, outperforming standard supervised training with all of the labels. 1

我们研究了视觉变压器的培训，用于半监督图像分类。变形金刚最近在众多监督的学习任务中表现出令人印象深刻的表现。令人惊讶的是，我们发现视觉变形金刚在半监督的想象中心设置上表现不佳。相比之下，卷积神经网络（CNNS）实现了小标记数据制度的卓越结果。进一步调查揭示了原因是CNN具有强大的空间归纳偏差。灵感来自这一观察，我们介绍了一个联合半监督学习框架，半统一，其中包含变压器分支，卷积分支和精心设计的融合模块，用于分支之间的知识共享。卷积分支在有限监督数据上培训，并生成伪标签，以监督变压器分支对未标记数据的培训。关于Imagenet的广泛实验表明，半统一达到75.5 \％的前1个精度，优于最先进的。此外，我们显示Semifirmer是一般框架，与大多数现代变压器和卷积神经结构兼容。

我们研究视觉变压器（VIT）的半监督学习（SSL），尽管VIT架构广泛采用了不同的任务，但视觉变形金刚（VIT）还是一个不足的主题。为了解决这个问题，我们提出了一条新的SSL管道，该管道由第一个联合国/自制的预训练组成，然后是监督的微调，最后是半监督的微调。在半监督的微调阶段，我们采用指数的移动平均线（EMA） - 教师框架，而不是流行的FixMatch，因为前者更稳定，并且为半手不见的视觉变压器提供了更高的准确性。此外，我们提出了一种概率的伪混合机制来插入未标记的样品及其伪标签以改善正则化，这对于训练电感偏差较弱的训练VIT很重要。我们所提出的方法被称为半vit，比半监督分类设置中的CNN对应物获得可比性或更好的性能。半vit还享受VIT的可伸缩性优势，可以很容易地扩展到具有越来越高的精度的大型模型。例如，半效率总数仅使用1％标签在Imagenet上获得令人印象深刻的80％TOP-1精度，使用100％ImageNet标签与Inception-V4相当。

Jitendra Malik once said, "Supervision is the opium of the AI researcher". Most deep learning techniques heavily rely on extreme amounts of human labels to work effectively. In today's world, the rate of data creation greatly surpasses the rate of data annotation. Full reliance on human annotations is just a temporary means to solve current closed problems in AI. In reality, only a tiny fraction of data is annotated. Annotation Efficient Learning (AEL) is a study of algorithms to train models effectively with fewer annotations. To thrive in AEL environments, we need deep learning techniques that rely less on manual annotations (e.g., image, bounding-box, and per-pixel labels), but learn useful information from unlabeled data. In this thesis, we explore five different techniques for handling AEL.

深度神经网络在大规模标记的数据集的帮助下，在各种任务上取得了出色的表现。然而，这些数据集既耗时又竭尽全力来获得现实的任务。为了减轻对标记数据的需求，通过迭代分配伪标签将伪标签分配给未标记的样本，自我训练被广泛用于半监督学习中。尽管它很受欢迎，但自我训练还是不可靠的，通常会导致训练不稳定。我们的实验研究进一步表明，半监督学习的偏见既来自问题本身，也来自不适当的训练，并具有可能不正确的伪标签，这会在迭代自我训练过程中累积错误。为了减少上述偏见，我们提出了自我训练（DST）。首先，伪标签的生成和利用是由两个独立于参数的分类器头解耦，以避免直接误差积累。其次，我们估计自我训练偏差的最坏情况，其中伪标记函数在标记的样品上是准确的，但在未标记的样本上却尽可能多地犯错。然后，我们通过避免最坏的情况来优化表示形式，以提高伪标签的质量。广泛的实验证明，DST在标准的半监督学习基准数据集上的最先进方法中，平均提高了6.3％，而在13个不同任务上，FIXMATCH的平均水平为18.9％。此外，DST可以无缝地适应其他自我训练方法，并有助于稳定他们在从头开始的培训和预先训练模型的训练的情况下，在培训的情况下进行培训和平衡表现。

Pre-training is a dominant paradigm in computer vision. For example, supervised ImageNet pre-training is commonly used to initialize the backbones of object detection and segmentation models. He et al. [1], for example, show a contrasting result that ImageNet pre-training has limited impact on COCO object detection. Here we investigate self-training as another method to utilize additional data on the same setup and contrast it against ImageNet pre-training. Our study reveals the generality and flexibility of self-training with three additional insights: 1) stronger data augmentation and more labeled data further diminish the value of pre-training, 2) unlike pre-training, self-training is always helpful when using stronger data augmentation, in both low-data and high-data regimes, and 3) in the case that pre-training is helpful, self-training improves upon pre-training. For example, on the COCO object detection dataset, pre-training benefits when we use one fifth of the labeled data, and hurts accuracy when we use all labeled data. Self-training, on the other hand, shows positive improvements from +1.3 to +3.4AP across all dataset sizes. In other words, self-training works well exactly on the same setup that pre-training does not work (using ImageNet to help COCO). On the PASCAL segmentation dataset, which is a much smaller dataset than COCO, though pre-training does help significantly, self-training improves upon the pre-trained model. On COCO object detection, we achieve 54.3AP, an improvement of +1.5AP over the strongest SpineNet model. On PASCAL segmentation, we achieve 90.5 mIOU, an improvement of +1.5% mIOU over the previous state-of-the-art result by DeepLabv3+. 1 ⇤ Authors contributed equally. 1 Code and checkpoints for our models are available at https://github.com/tensorflow/tpu/tree/ master/models/official/detection/projects/self_training 34th Conference on Neural Information Processing Systems (NeurIPS 2020), Vancouver, Canada.

We demonstrate that self-learning techniques like entropy minimization and pseudo-labeling are simple and effective at improving performance of a deployed computer vision model under systematic domain shifts. We conduct a wide range of large-scale experiments and show consistent improvements irrespective of the model architecture, the pre-training technique or the type of distribution shift. At the same time, self-learning is simple to use in practice because it does not require knowledge or access to the original training data or scheme, is robust to hyperparameter choices, is straight-forward to implement and requires only a few adaptation epochs. This makes self-learning techniques highly attractive for any practitioner who applies machine learning algorithms in the real world. We present state-of-the-art adaptation results on CIFAR10-C (8.5% error), ImageNet-C (22.0% mCE), ImageNet-R (17.4% error) and ImageNet-A (14.8% error), theoretically study the dynamics of self-supervised adaptation methods and propose a new classification dataset (ImageNet-D) which is challenging even with adaptation.

This work tackles the problem of semi-supervised learning of image classifiers. Our main insight is that the field of semi-supervised learning can benefit from the quickly advancing field of self-supervised visual representation learning. Unifying these two approaches, we propose the framework of self-supervised semi-supervised learning (S 4 L) and use it to derive two novel semi-supervised image classification methods. We demonstrate the effectiveness of these methods in comparison to both carefully tuned baselines, and existing semi-supervised learning methods. We then show that S 4 L and existing semi-supervised methods can be jointly trained, yielding a new state-of-the-art result on semi-supervised ILSVRC-2012 with 10% of labels.

Positive-Unlabeled (PU) learning aims to learn a model with rare positive samples and abundant unlabeled samples. Compared with classical binary classification, the task of PU learning is much more challenging due to the existence of many incompletely-annotated data instances. Since only part of the most confident positive samples are available and evidence is not enough to categorize the rest samples, many of these unlabeled data may also be the positive samples. Research on this topic is particularly useful and essential to many real-world tasks which demand very expensive labelling cost. For example, the recognition tasks in disease diagnosis, recommendation system and satellite image recognition may only have few positive samples that can be annotated by the experts. These methods mainly omit the intrinsic hardness of some unlabeled data, which can result in sub-optimal performance as a consequence of fitting the easy noisy data and not sufficiently utilizing the hard data. In this paper, we focus on improving the commonly-used nnPU with a novel training pipeline. We highlight the intrinsic difference of hardness of samples in the dataset and the proper learning strategies for easy and hard data. By considering this fact, we propose first splitting the unlabeled dataset with an early-stop strategy. The samples that have inconsistent predictions between the temporary and base model are considered as hard samples. Then the model utilizes a noise-tolerant Jensen-Shannon divergence loss for easy data; and a dual-source consistency regularization for hard data which includes a cross-consistency between student and base model for low-level features and self-consistency for high-level features and predictions, respectively.

糖尿病性视网膜病变（DR）是发达国家工人衰老人群中失明的主要原因之一，这是由于糖尿病的副作用降低了视网膜的血液供应。深度神经网络已被广泛用于自动化系统中，以在眼底图像上进行DR分类。但是，这些模型需要大量带注释的图像。在医疗领域，专家的注释昂贵，乏味且耗时。结果，提供了有限数量的注释图像。本文提出了一种半监督的方法，该方法利用未标记的图像和标记的图像来训练一种检测糖尿病性视网膜病的模型。提出的方法通过自我监督的学习使用无监督的预告片，然后使用一小部分标记的图像和知识蒸馏来监督微调，以提高分类任务的性能。在Eyepacs测试和Messidor-2数据集中评估了此方法，仅使用2％的Eyepacs列车标记图像，分别使用0.94和0.89 AUC。

在本文中，我们在半监督对象检测（SSOD）中深入研究了两种关键技术，即伪标记和一致性训练。我们观察到，目前，这两种技术忽略了对象检测的一些重要特性，从而阻碍了对未标记数据的有效学习。具体而言，对于伪标记，现有作品仅关注分类得分，但不能保证伪框的本地化精度；为了保持一致性训练，广泛采用的随机训练只考虑了标签级的一致性，但错过了功能级别的训练，这在确保尺度不变性方面也起着重要作用。为了解决嘈杂的伪箱所产生的问题，我们设计了包括预测引导的标签分配（PLA）和正面验证一致性投票（PCV）的嘈杂伪盒学习（NPL）。 PLA依赖于模型预测来分配标签，并使甚至粗糙的伪框都具有鲁棒性。 PCV利用积极建议的回归一致性来反映伪盒的本地化质量。此外，在一致性训练中，我们提出了包括标签和特征水平一致性的机制的多视图尺度不变学习（MSL），其中通过将两个图像之间的移动特征金字塔对准具有相同内容但变化量表的变化来实现特征一致性。在可可基准测试上，我们的方法称为伪标签和一致性训练（PSECO），分别以2.0、1.8、2.0分的1％，5％和10％的标签比优于SOTA（软教师）。它还显着提高了SSOD的学习效率，例如，PSECO将SOTA方法的训练时间减半，但实现了更好的性能。代码可从https://github.com/ligang-cs/pseco获得。

这项工作研究了伪标签的偏见问题，一种广泛发生的自然现象，但经常通过先前的研究忽视。当在源数据上培训的分类器被传送到未标记的目标数据时，会生成伪标签。当半监督的学习模型Fixmatch预测未标记的数据时，我们观察到沉重的长尾伪标签即使未标记的数据被策划到平衡。没有干预，培训模型继承了伪标签的偏置，最终是次优。为了消除模型偏置，我们提出了一种简单而有效的方法DebiSmatch，包括自适应脱叠模块和自适应边际损失。通过使用在线更新的队列，可以自动调整脱叠的强度和边距的大小。在ImageNet-1K上基准测试，DebiasMatch分别在半监督学习（0.2％注释数据）和零拍摄学习任务中显着超过26％和8.7％的最先进。