Exploring the Landscape of Spatial Robustness

By Logan Engstrom, Brandon Tran, Dimitris Tsipras, Ludwig Schmidt, Aleksander Madry. 2017

564 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Safety, Robustness & Interpretability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Safety, Robustness & Interpretability

Citations in the corpus

Measuring Robustness to Natural Distribution Shifts in Image Classification (2020, methodology): “Encouragingly, the past few years have seen substantial progress in robustness to these distribution shifts, e.g., see [13, 28, 34, 55, 57, 66, 93, 96, 105, 114, 115] among many others.”
On Buggy Resizing Libraries and Surprising Subtleties in FID Calculation (2021, background): “Despite these advances, generative methods continue to be detectable [10, 74], and discriminative networks continue to be sensitive to small perturbations, such as shifts [3, 22] and JPEG compression [29].”
Toward Causal Representation Learning (2021, background): “So far, there has been no deﬁnitive consensus on how to solve these problems, although progress has been made using data augmentation, pretraining, self-supervision, and architectures with suitable inductive biases with respect to a perturbation of interest [60], [64], [137], [170], [206], [233].”
Fundamental Tradeoffs between Invariance and Sensitivity to Adversarial Perturbations (2020, background): “Interestingly, Engstrom et al. (2019a) show an opposite effect for models’ internal representations.”
Wasserstein Adversarial Examples via Projected Sinkhorn Iterations (2019, background): “For example, simple rotations and translations form a fairly limited set of perturbations that can be quite large in `p norm, but are sometimes sufficient in order to fool classifiers (Engstrom et al., 2017).”
Ensemble Adversarial Training: Attacks and Defenses (2017, background): “Of course, we could augment the training set with such perturbations (see Engstrom et al. (2017)).”
Provable defenses against adversarial examples via the convex outer adversarial polytope (2017, background): “For example, the work in Sharif et al. (2016) was able to break face recognition software by using manufactured glasses, which is clearly not bounded in `∞ norm, and the work in Engstrom et al. (2017) was able to fool convolutional networks with simple rotations and translations.”
Towards Causal Representation Learning (2021, background): “In computer vision [75, 228], changes in the test distribution may, for instance, come from aberrations like camera blur, noise or compression quality [106, 129, 170, 206], or from shifts, rotations, or viewpoints [7, 11, 63, 282].”
Robustness May Be at Odds with Accuracy (2018, background): “ves as a standard benchmark. It is worth noting though that several other notions of adversarial perturbations have been studied. These include rotations and translations (Fawzi & Frossard, 2015; Engstrom et al., 2019), and smooth spatial deformations (Xiao et al., 2018). In general, determining”
Identifying Statistical Bias in Dataset Replication (2020, background): “A step in the journey towards this goal is precisely characterizing the kinds of distribution shifts under which models fail—examples include rotations and translations of natural images Engstrom et al. (2019), or corrupted natural images (Hendrycks & Dietterich, 2019). Our findings imply that the d”