Deep networks for computer vision are not reliable when they encounter adversarial examples. In this paper, we introduce a framework that uses the dense intrinsic constraints in natural images to robustify inference. By introducing constraints at inference time, we can shift the burden of robustness from training to the inference algorithm, thereby allowing the model to adjust dynamically to each individual image's unique and potentially novel characteristics at inference time. Among different constraints, we find that equivariance-based constraints are most effective, because they allow dense constraints in the feature space without overly constraining the representation at a fine-grained level. Our theoretical results validate the importance of having such dense constraints at inference time. Our empirical experiments show that restoring feature equivariance at inference time defends against worst-case adversarial perturbations. The method obtains improved adversarial robustness on four datasets (ImageNet, Cityscapes, PASCAL VOC, and MS-COCO) on image recognition, semantic segmentation, and instance segmentation tasks. Project page is available at equi4robust.cs.columbia.edu.

Spatial-temporal (ST) graph modeling, such as traffic speed forecasting and taxi demand prediction, is an important task in deep learning area. However, for the nodes in graph, their ST patterns can vary greatly in difficulties for modeling, owning to the heterogeneous nature of ST data. We argue that unveiling the nodes to the model in a meaningful order, from easy to complex, can provide performance improvements over traditional training procedure. The idea has its root in Curriculum Learning which suggests in the early stage of training models can be sensitive to noise and difficult samples. In this paper, we propose ST-Curriculum Dropout, a novel and easy-to-implement strategy for spatial-temporal graph modeling. Specifically, we evaluate the learning difficulty of each node in high-level feature space and drop those difficult ones out to ensure the model only needs to handle fundamental ST relations at the beginning, before gradually moving to hard ones. Our strategy can be applied to any canonical deep learning architecture without extra trainable parameters, and extensive experiments on a wide range of datasets are conducted to illustrate that, by controlling the difficulty level of ST relations as the training progresses, the model is able to capture better representation of the data and thus yields better generalization.

很少有学习模型学习人类注释有限，而这种学习范式在各种任务中证明了实用性数据使该模型无法充分探索语义信息。为了解决这个问题，我们将知识蒸馏引入了几个弹出的对象检测学习范式。我们进一步进行了激励实验，该实验表明，在知识蒸馏的过程中，教师模型的经验误差将少数拍物对象检测模型的预测性能（作为学生）退化。为了了解这种现象背后的原因，我们从因果理论的角度重新审视了几个对象检测任务上知识蒸馏的学习范式，并因此发展了一个结构性因果模型。遵循理论指导，我们建议使用基于后门调整的知识蒸馏方法，用于少数拍物检测任务，即Disentangle和Remerge（D＆R），以对相应的结构性因果模型进行有条件的因果干预。从理论上讲，我们为后门标准提供了扩展的定义，即一般后门路径，可以在特定情况下扩展后门标准的理论应用边界。从经验上讲，多个基准数据集上的实验表明，D＆R可以在几个射击对象检测中产生显着的性能提升。

将异常检测外包给第三方可以允许数据所有者克服资源限制（例如，在轻量级的IoT设备中），促进协作分析（例如，分布式或多方场景下的分布式或多方场景），并受益于较低的成本和专业知识（例如托管安全服务提供商）。尽管有这样的好处，但数据所有者可能会不愿外包异常检测而没有足够的隐私保护。为此，大多数现有的隐私解决方案将面临新的挑战，即保留隐私通常需要消除或减少数据条目之间的差异，而异常检测严重取决于该差异。最近，在本地分析设置下，通过将差异隐私（DP）保证的重点从“全部”到“良性”条目移动，这一冲突是在本地分析设置下解决的。在本文中，我们观察到这种方法不直接适用于外包设置，因为数据所有者在外包之前不知道哪些条目是“良性”的，因此无法选择地将DP应用于数据条目。因此，我们提出了一种新型的迭代解决方案，使数据所有者逐渐“脱离”良性条目的异常条目，以便第三方分析师可以通过足够的DP保证产生准确的异常结果。我们设计并实施了我们对异常检测（DPOAD）框架的差异私人外包，并通过从不同应用域中的真实数据进行实验，证明了其比基线拉普拉斯和无止痛机制的好处。

多变量时间序列预测是一个具有挑战性的任务，因为数据涉及长期和短期模式的混合，具有变量之间的动态时空依赖性。现有图形神经网络（GNN）通常与预定义的空间图或学习的固定邻接图模拟多变量关系。它限制了GNN的应用，并且无法处理上述挑战。在本文中，我们提出了一种新颖的框架，即静态和动态图形学习 - 神经网络（SDGL）。该模型分别从数据获取静态和动态图形矩阵分别为模型长期和短期模式。开发静态Matric以通过节点嵌入捕获固定的长期关联模式，并利用图规律性来控制学习静态图的质量。为了捕获变量之间的动态依赖性，我们提出了基于改变节点特征和静态节点Embeddings生成时变矩阵的动态图。在该方法中，我们将学习的静态图信息作为感应偏置集成为诱导动态图和局部时空模式更好。广泛的实验是在两个交通数据集中进行，具有额外的结构信息和四个时间序列数据集，这表明我们的方法在几乎所有数据集上实现了最先进的性能。如果纸张被接受，我将在GitHub上打开源代码。

在这项工作中，我们解决了由三个腔室驱动的软仿生执行器的运动计划的逆动力学问题。由于其内在的柔软性，软仿生致动器已应用于许多应用中。尽管可以得出数学模型来描述该执行器的逆动力学，但捕获材料和系统的非线性和不确定性仍然不准确。此外，如此复杂的模型是耗时的，因此在实时控制单元中应用不容易。因此，在该领域开发一种无模型方法可能是一个新想法。为了克服这些内在问题，我们提出了一个后传播（BP）神经网络，学习在三维空间中移动的软仿生执行器的逆动力学。在使用样品数据进行训练之后，BP神经网络模型可以代表操纵器尖端位置与施加到腔室的压力之间的关系。所提出的算法比分析模型更精确。结果表明，相对于总执行器长度，相对平均误差为2.46％，可以实现所需的末端位置。

Video action segmentation aims to slice the video into several action segments. Recently, timestamp supervision has received much attention due to lower annotation costs. We find the frames near the boundaries of action segments are in the transition region between two consecutive actions and have unclear semantics, which we call ambiguous intervals. Most existing methods iteratively generate pseudo-labels for all frames in each video to train the segmentation model. However, ambiguous intervals are more likely to be assigned with noisy and incorrect pseudo-labels, which leads to performance degradation. We propose a novel framework to train the model under timestamp supervision including the following two parts. First, pseudo-label ensembling generates pseudo-label sequences with ambiguous intervals, where the frames have no pseudo-labels. Second, iterative clustering iteratively propagates the pseudo-labels to the ambiguous intervals by clustering, and thus updates the pseudo-label sequences to train the model. We further introduce a clustering loss, which encourages the features of frames within the same action segment more compact. Extensive experiments show the effectiveness of our method.

Deep learning-based full-reference image quality assessment (FR-IQA) models typically rely on the feature distance between the reference and distorted images. However, the underlying assumption of these models that the distance in the deep feature domain could quantify the quality degradation does not scientifically align with the invariant texture perception, especially when the images are generated artificially by neural networks. In this paper, we bring a radical shift in inferring the quality with learned features and propose the Deep Image Dependency (DID) based FR-IQA model. The feature dependency facilitates the comparisons of deep learning features in a high-order manner with Brownian distance covariance, which is characterized by the joint distribution of the features from reference and test images, as well as their marginal distributions. This enables the quantification of the feature dependency against nonlinear transformation, which is far beyond the computation of the numerical errors in the feature space. Experiments on image quality prediction, texture image similarity, and geometric invariance validate the superior performance of our proposed measure.

凝视估计对于许多科学领域和日常应用至关重要，范围从认知心理学的基本研究到注意力吸引人的移动系统。尽管深度学习的最新进展在建立高度准确的凝视估计系统方面取得了巨大的成功，但相关的高计算成本以及对大规模标记的凝视数据的依赖，以实现对现有解决方案实际使用的监督学习地点挑战。为了超越这些局限性，我们提出了FreeGaze，这是一种用于无监督的注视表示学习的资源有效框架。 FreeGaze在其设计中结合了频域目光的估计和对比度注视表示。前者大大减轻了系统校准和凝视估计中的计算负担，并大大减少了系统延迟。尽管后者克服了现有基于学习的同行的数据标记障碍，并确保在没有凝视标签的情况下确保有效的凝视表示学习。我们对两个凝视估计数据集的评估表明，通过现有基于监督的学习方法，FreeGaze可以在系统校准和注视估计中分别实现高达6.81和1.67倍的速度，以实现可比较的凝视估计精度。

自然图像的统计规律（称为自然场景统计数据）在不引用图像质量评估中起重要作用。但是，人们普遍认为，通常是计算机生成的屏幕内容图像（SCI）不持有此类统计信息。在这里，我们首次尝试学习SCI的统计数据，基于可以有效确定SCI的质量。所提出的方法的基本机制是基于一个狂野的假设，即没有物理上获得的SCI仍然遵守某些可以以学习方式理解的统计数据。我们从经验上表明，在质量评估中可以有效利用统计偏差，并且在不同的环境中进行评估时，提出的方法优越。广泛的实验结果表明，与现有的NR-IQA模型相比，基于深度统计的SCI质量评估（DFSS-IQA）模型可提供有希望的性能，并在跨数据库设置中显示出很高的概括能力。我们的方法的实现可在https://github.com/baoliang93/dfss-iqa上公开获得。