图像之间的感知距离在预训练的深度特征的空间中测量，在评估图像相似性方面优于先前的低级，基于像素的指标。虽然众所周知，较旧模型（例如Alexnet和VGG）捕获感知相似性的功能却较少，但研究了现代和更准确的模型。在本文中，我们提出了一项大规模的经验研究，以评估成像网分类器在感知相似性方面的表现。首先，我们观察到成像网的精度与现代网络（例如重置，有效网络和视觉变压器）的感知得分之间的反相关性：更好的分类器达到了较差的感知得分。然后，我们在不同的深度，宽度，训练步骤，重量衰减，标签平滑和辍学时检查了成像网的精度/感知分数关系。更高的精度将感知得分提高到一定点，但是我们在中高精度方面发现了精度和感知得分之间的帕累托前沿。我们使用许多合理的假设，例如失真不变性，空间频率灵敏度和替代感知函数，进一步探索这种关系。有趣的是，我们发现仅在Imagenet上接受少于5个时代训练的浅重新收集和重新注册，其新兴的感知得分与直接受到监督的人类感知判断直接训练的先前最佳网络相匹配。

在过去的几年中，人类视力与卷积神经网络（CNN）之间越来越多的相似之处。然而，香草CNN通常在推广到对抗性或分布（OOD）示例的概括方面表现出卓越的性能。对抗训练是一种领先的学习算法，用于提高CNN在对抗和OOD数据上的鲁棒性；但是，对这些属性，特别是形状偏差和内部特征知之甚少，在对抗性CNN中学到的内部特征。在本文中，我们进行了一项彻底的系统研究，以了解形状偏差和一些内部机制，以使Alexnet，Googlenet和Resnet-50模型的普遍性通过对抗训练进行了训练。我们发现，尽管标准成像网分类器具有较强的纹理偏见，但它们的R对应物很大程度上依赖形状。值得注意的是，对抗性训练在“鲁棒性” CNN的过程中诱导了隐藏的神经元的三个简单偏见。也就是说，R网络中的每个卷积神经元经常会更改以检测（1）像素的平滑模式，即一种机制，该机制可以阻止高频噪声通过网络； （2）更多较低级别的功能，即纹理和颜色（而不是对象）；（3）输入类型较少。我们的发现揭示了有趣的机制，这些机制使网络更具对抗性，并解释了一些最新发现，例如，为什么R网络从更大的容量中受益（Xie等，2020），并且可以在图像合成中充当强大的图像（Santurkar et eT） Al。2019）。

State-of-the-art computer vision systems are trained to predict a fixed set of predetermined object categories. This restricted form of supervision limits their generality and usability since additional labeled data is needed to specify any other visual concept. Learning directly from raw text about images is a promising alternative which leverages a much broader source of supervision. We demonstrate that the simple pre-training task of predicting which caption goes with which image is an efficient and scalable way to learn SOTA image representations from scratch on a dataset of 400 million (image, text) pairs collected from the internet. After pre-training, natural language is used to reference learned visual concepts (or describe new ones) enabling zero-shot transfer of the model to downstream tasks. We study the performance of this approach by benchmarking on over 30 different existing computer vision datasets, spanning tasks such as OCR, action recognition in videos, geo-localization, and many types of fine-grained object classification. The model transfers non-trivially to most tasks and is often competitive with a fully supervised baseline without the need for any dataset specific training. For instance, we match the accuracy of the original ResNet-50 on ImageNet zero-shot without needing to use any of the 1.28 million training examples it was trained on. We release our code and pre-trained model weights at https://github.com/OpenAI/CLIP.

Transfer learning is a cornerstone of computer vision, yet little work has been done to evaluate the relationship between architecture and transfer. An implicit hypothesis in modern computer vision research is that models that perform better on ImageNet necessarily perform better on other vision tasks. However, this hypothesis has never been systematically tested. Here, we compare the performance of 16 classification networks on 12 image classification datasets. We find that, when networks are used as fixed feature extractors or fine-tuned, there is a strong correlation between ImageNet accuracy and transfer accuracy (r = 0.99 and 0.96, respectively). In the former setting, we find that this relationship is very sensitive to the way in which networks are trained on ImageNet; many common forms of regularization slightly improve ImageNet accuracy but yield penultimate layer features that are much worse for transfer learning. Additionally, we find that, on two small fine-grained image classification datasets, pretraining on ImageNet provides minimal benefits, indicating the learned features from Ima-geNet do not transfer well to fine-grained tasks. Together, our results show that ImageNet architectures generalize well across datasets, but ImageNet features are less general than previously suggested.

以前的工作提出了许多新的损失函数和常规程序，可提高图像分类任务的测试准确性。但是，目前尚不清楚这些损失函数是否了解下游任务的更好表示。本文研究了培训目标的选择如何影响卷积神经网络隐藏表示的可转移性，训练在想象中。我们展示了许多目标在Vanilla Softmax交叉熵上导致想象的精度有统计学意义的改进，但由此产生的固定特征提取器转移到下游任务基本较差，并且当网络完全微调时，损失的选择几乎没有效果新任务。使用居中内核对齐来测量网络隐藏表示之间的相似性，我们发现损失函数之间的差异仅在网络的最后几层中都很明显。我们深入了解倒数第二层的陈述，发现不同的目标和近奇计的组合导致大幅不同的类别分离。具有较高类别分离的表示可以在原始任务上获得更高的准确性，但它们的功能对于下游任务不太有用。我们的结果表明，用于原始任务的学习不变功能与传输任务相关的功能之间存在权衡。

Unsupervised visual representation learning remains a largely unsolved problem in computer vision research. Among a big body of recently proposed approaches for unsupervised learning of visual representations, a class of self-supervised techniques achieves superior performance on many challenging benchmarks. A large number of the pretext tasks for self-supervised learning have been studied, but other important aspects, such as the choice of convolutional neural networks (CNN), has not received equal attention. Therefore, we revisit numerous previously proposed self-supervised models, conduct a thorough large scale study and, as a result, uncover multiple crucial insights. We challenge a number of common practices in selfsupervised visual representation learning and observe that standard recipes for CNN design do not always translate to self-supervised representation learning. As part of our study, we drastically boost the performance of previously proposed techniques and outperform previously published state-of-the-art results by a large margin.

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.

Quantifying the perceptual similarity of two images is a long-standing problem in low-level computer vision. The natural image domain commonly relies on supervised learning, e.g., a pre-trained VGG, to obtain a latent representation. However, due to domain shift, pre-trained models from the natural image domain might not apply to other image domains, such as medical imaging. Notably, in medical imaging, evaluating the perceptual similarity is exclusively performed by specialists trained extensively in diverse medical fields. Thus, medical imaging remains devoid of task-specific, objective perceptual measures. This work answers the question: Is it necessary to rely on supervised learning to obtain an effective representation that could measure perceptual similarity, or is self-supervision sufficient? To understand whether recent contrastive self-supervised representation (CSR) may come to the rescue, we start with natural images and systematically evaluate CSR as a metric across numerous contemporary architectures and tasks and compare them with existing methods. We find that in the natural image domain, CSR behaves on par with the supervised one on several perceptual tests as a metric, and in the medical domain, CSR better quantifies perceptual similarity concerning the experts' ratings. We also demonstrate that CSR can significantly improve image quality in two image synthesis tasks. Finally, our extensive results suggest that perceptuality is an emergent property of CSR, which can be adapted to many image domains without requiring annotations.

本文探讨了贝尔视觉变压器预训练的更好的码本。最近的工作成功地转移了从NLP到视野领域的BERT预训练。它直接采用一个简单的离散VAE作为视觉销售器，但尚未考虑由此产生的视觉令牌的语义水平。相比之下，NLP字段中的离散令牌是自然的高度语义。这种差异激励我们学习一个感知码本。我们惊奇地找到了一个简单而有效的想法：在DVAE训练期间强制执行感知相似性。我们证明，所提出的感知码本生成的视觉令牌确实表现出更好的语义含义，随后有助于预训练在各种下游任务中实现卓越的转移性能。例如，我们在Imagenet-1K上实现了84.5前1个精度，vit-B骨干，优于竞争方法Beit +1.3，具有相同的训练纪元。它还可以通过+1.3框AP和+1.0掩模AP，在ADE20K上的语义细分，在ADE20K上提高对象检测和分割任务的性能，+1.0 miou，代码和型号将在\ url {https：// github.com/microsoft/peco}。

While the Transformer architecture has become the de-facto standard for natural language processing tasks, its applications to computer vision remain limited. In vision, attention is either applied in conjunction with convolutional networks, or used to replace certain components of convolutional networks while keeping their overall structure in place. We show that this reliance on CNNs is not necessary and a pure transformer applied directly to sequences of image patches can perform very well on image classification tasks. When pre-trained on large amounts of data and transferred to multiple mid-sized or small image recognition benchmarks (ImageNet, CIFAR-100, VTAB, etc.), Vision Transformer (ViT) attains excellent results compared to state-of-the-art convolutional networks while requiring substantially fewer computational resources to train. 1

Image classification with small datasets has been an active research area in the recent past. However, as research in this scope is still in its infancy, two key ingredients are missing for ensuring reliable and truthful progress: a systematic and extensive overview of the state of the art, and a common benchmark to allow for objective comparisons between published methods. This article addresses both issues. First, we systematically organize and connect past studies to consolidate a community that is currently fragmented and scattered. Second, we propose a common benchmark that allows for an objective comparison of approaches. It consists of five datasets spanning various domains (e.g., natural images, medical imagery, satellite data) and data types (RGB, grayscale, multispectral). We use this benchmark to re-evaluate the standard cross-entropy baseline and ten existing methods published between 2017 and 2021 at renowned venues. Surprisingly, we find that thorough hyper-parameter tuning on held-out validation data results in a highly competitive baseline and highlights a stunted growth of performance over the years. Indeed, only a single specialized method dating back to 2019 clearly wins our benchmark and outperforms the baseline classifier.

对共同腐败的稳健性的文献表明对逆势培训是否可以提高这种环境的性能，没有达成共识。 First, we show that, when used with an appropriately selected perturbation radius, $\ell_p$ adversarial training can serve as a strong baseline against common corruptions improving both accuracy and calibration.然后，我们解释了为什么对抗性训练比具有简单高斯噪声的数据增强更好地表现，这被观察到是对共同腐败的有意义的基线。与此相关，我们确定了高斯增强过度适用于用于培训的特定标准偏差的$ \ sigma $ -oviting现象，这对培训具有显着不利影响的普通腐败精度。我们讨论如何缓解这一问题，然后如何通过学习的感知图像贴片相似度引入对抗性训练的有效放松来进一步增强$ \ ell_p $普发的培训。通过对CiFar-10和Imagenet-100的实验，我们表明我们的方法不仅改善了$ \ ell_p $普发的培训基线，而且还有累积的收益与Augmix，Deepaulment，Ant和Sin等数据增强方法，导致普通腐败的最先进的表现。我们的实验代码在HTTPS://github.com/tml-epfl/adv-training - 窗子上公开使用。

We study how robust current ImageNet models are to distribution shifts arising from natural variations in datasets. Most research on robustness focuses on synthetic image perturbations (noise, simulated weather artifacts, adversarial examples, etc.), which leaves open how robustness on synthetic distribution shift relates to distribution shift arising in real data. Informed by an evaluation of 204 ImageNet models in 213 different test conditions, we find that there is often little to no transfer of robustness from current synthetic to natural distribution shift. Moreover, most current techniques provide no robustness to the natural distribution shifts in our testbed. The main exception is training on larger and more diverse datasets, which in multiple cases increases robustness, but is still far from closing the performance gaps. Our results indicate that distribution shifts arising in real data are currently an open research problem. We provide our testbed and data as a resource for future work at https://modestyachts.github.io/imagenet-testbed/.

Vision transformer (ViT) models exhibit substandard optimizability. In particular, they are sensitive to the choice of optimizer (AdamW vs. SGD), optimizer hyperparameters, and training schedule length. In comparison, modern convolutional neural networks are easier to optimize. Why is this the case? In this work, we conjecture that the issue lies with the patchify stem of ViT models, which is implemented by a stride-p p×p convolution (p = 16 by default) applied to the input image. This large-kernel plus large-stride convolution runs counter to typical design choices of convolutional layers in neural networks. To test whether this atypical design choice causes an issue, we analyze the optimization behavior of ViT models with their original patchify stem versus a simple counterpart where we replace the ViT stem by a small number of stacked stride-two 3×3 convolutions. While the vast majority of computation in the two ViT designs is identical, we find that this small change in early visual processing results in markedly different training behavior in terms of the sensitivity to optimization settings as well as the final model accuracy. Using a convolutional stem in ViT dramatically increases optimization stability and also improves peak performance (by ∼1-2% top-1 accuracy on ImageNet-1k), while maintaining flops and runtime. The improvement can be observed across the wide spectrum of model complexities (from 1G to 36G flops) and dataset scales (from ImageNet-1k to ImageNet-21k). These findings lead us to recommend using a standard, lightweight convolutional stem for ViT models in this regime as a more robust architectural choice compared to the original ViT model design.

哥内克人Sentinel Imagery的纯粹卷的可用性为使用深度学习的大尺度创造了新的土地利用陆地覆盖（Lulc）映射的机会。虽然在这种大型数据集上培训是一个非琐碎的任务。在这项工作中，我们试验Lulc Image分类和基准不同最先进模型的Bigearthnet数据集，包括卷积神经网络，多层感知，视觉变压器，高效导通和宽残余网络（WRN）架构。我们的目标是利用分类准确性，培训时间和推理率。我们提出了一种基于用于网络深度，宽度和输入数据分辨率的WRNS复合缩放的高效导通的框架，以有效地训练和测试不同的模型设置。我们设计一种新颖的缩放WRN架构，增强了有效的通道注意力机制。我们提出的轻量级模型具有较小的培训参数，实现所有19个LULC类的平均F分类准确度达到4.5％，并且验证了我们使用的resnet50最先进的模型速度快两倍作为基线。我们提供超过50种培训的型号，以及我们在多个GPU节点上分布式培训的代码。

我们介绍了几个新的数据集即想象的A / O和Imagenet-R以及合成环境和测试套件，我们称为CAOS。 Imagenet-A / O允许研究人员专注于想象成剩余的盲点。由于追踪稳健的表示，以特殊创建了ImageNet-R，因为表示不再简单地自然，而是包括艺术和其他演绎。 Caos Suite由Carla Simulator构建，允许包含异常物体，可以创建可重复的合成环境和用于测试稳健性的场景。所有数据集都是为测试鲁棒性和衡量鲁棒性的衡量进展而创建的。数据集已用于各种其他作品中，以衡量其具有鲁棒性的自身进步，并允许切向进展，这些进展不会完全关注自然准确性。鉴于这些数据集，我们创建了几种旨在推进鲁棒性研究的新方法。我们以最大Logit的形式和典型程度的形式构建简单的基线，并以深度的形式创建新的数据增强方法，从而提高上述基准。最大Logit考虑Logit值而不是SoftMax操作后的值，而微小的变化会产生明显的改进。典型程分将输出分布与类的后部分布进行比较。我们表明，除了分段任务之外，这将提高对基线的性能。猜测可能在像素级别，像素的语义信息比类级信息的语义信息不太有意义。最后，新的Deepaulment的新增强技术利用神经网络在彻底不同于先前使用的传统几何和相机的转换的图像上创建增强。

本文首先揭示令人惊讶的发现：没有任何学习，随机初始化的CNN可以令人惊讶地定位对象。也就是说，CNN具有归纳偏差，以自然地关注物体，在本文中被命名为Tobias（“对象是在视线处的”）。进一步分析并成功地应用于自我监督学习（SSL）的经验感应偏差。鼓励CNN学习专注于前景对象的表示，通过将每个图像转换为具有不同背景的各种版本，其中前景和背景分离被托比亚引导。实验结果表明，建议的托比亚斯显着提高了下游任务，尤其是对象检测。本文还表明，托比亚斯对不同尺寸的训练集具有一致的改进，并且更具弹性变化了图像增强。代码可在https://github.com/cupidjay/tobias获得。

We introduce Bootstrap Your Own Latent (BYOL), a new approach to self-supervised image representation learning. BYOL relies on two neural networks, referred to as online and target networks, that interact and learn from each other. From an augmented view of an image, we train the online network to predict the target network representation of the same image under a different augmented view. At the same time, we update the target network with a slow-moving average of the online network. While state-of-the art methods rely on negative pairs, BYOL achieves a new state of the art without them. BYOL reaches 74.3% top-1 classification accuracy on ImageNet using a linear evaluation with a ResNet-50 architecture and 79.6% with a larger ResNet. We show that BYOL performs on par or better than the current state of the art on both transfer and semi-supervised benchmarks. Our implementation and pretrained models are given on GitHub. 3 * Equal contribution; the order of first authors was randomly selected.

自我监督学习的最新进展证明了多种视觉任务的有希望的结果。高性能自我监督方法中的一个重要成分是通过培训模型使用数据增强，以便在嵌入空间附近的相同图像的不同增强视图。然而，常用的增强管道整体地对待图像，忽略图像的部分的语义相关性-e.g。主题与背景 - 这可能导致学习杂散相关性。我们的工作通过调查一类简单但高度有效的“背景增强”来解决这个问题，这鼓励模型专注于语义相关内容，劝阻它们专注于图像背景。通过系统的调查，我们表明背景增强导致在各种任务中跨越一系列最先进的自我监督方法（MOCO-V2，BYOL，SWAV）的性能大量改进。 $ \ SIM $ + 1-2％的ImageNet收益，使得与监督基准的表现有关。此外，我们发现有限标签设置的改进甚至更大（高达4.2％）。背景技术增强还改善了许多分布换档的鲁棒性，包括天然对抗性实例，想象群-9，对抗性攻击，想象成型。我们还在产生了用于背景增强的显着掩模的过程中完全无监督的显着性检测进展。