Adversarial Robustness Through Artifact Design

Adversarial examples arose as a challenge for machine learning. To hinder them, most defenses alter how models are trained (e.g., adversarial training) or inference is made (e.g., randomized smoothing). Still, while these approaches markedly improve models' adversarial robustness, models remain highly susceptible to adversarial examples. Identifying that, in certain domains such as traffic-sign recognition, objects are implemented per standards specifying how artifacts (e.g., signs) should be designed, we propose a novel approach for improving adversarial robustness. Specifically, we offer a method to redefine standards, making minor changes to existing ones, to defend against adversarial examples. We formulate the problem of artifact design as a robust optimization problem, and propose gradient-based and greedy search methods to solve it. We evaluated our approach in the domain of traffic-sign recognition, allowing it to alter traffic-sign pictograms (i.e., symbols within the signs) and their colors. We found that, combined with adversarial training, our approach led to up to 25.18\% higher robust accuracy compared to state-of-the-art methods against two adversary types, while further increasing accuracy on benign inputs.

The Ultimate Combo: Boosting Adversarial Example Transferability by Composing Data Augmentations

Transferring adversarial examples (AEs) from surrogate machine-learning (ML) models to target models is commonly used in black-box adversarial robustness evaluation. Attacks leveraging certain data augmentation, such as random resizing, have been found to help AEs generalize from surrogates to targets. Yet, prior work has explored limited augmentations and their composition. To fill the gap, we systematically studied how data augmentation affects transferability. Particularly, we explored 46 augmentation techniques of seven categories originally proposed to help ML models generalize to unseen benign samples, and assessed how they impact transferability, when applied individually or composed. Performing exhaustive search on a small subset of augmentation techniques and genetic search on all techniques, we identified augmentation combinations that can help promote transferability. Extensive experiments with the ImageNet and CIFAR-10 datasets and 18 models showed that simple color-space augmentations (e.g., color to greyscale) outperform the state of the art when combined with standard augmentations, such as translation and scaling. Additionally, we discovered that composing augmentations impacts transferability mostly monotonically (i.e., more methods composed $\rightarrow$ $\ge$ transferability). We also found that the best composition significantly outperformed the state of the art (e.g., 93.7% vs. $\le$ 82.7% average transferability on ImageNet from normally trained surrogates to adversarially trained targets). Lastly, our theoretical analysis, backed up by empirical evidence, intuitively explain why certain augmentations help improve transferability.

Group-based Robustness: A General Framework for Customized Robustness in the Real World

Machine-learning models are known to be vulnerable to evasion attacks that perturb model inputs to induce misclassifications. In this work, we identify real-world scenarios where the true threat cannot be assessed accurately by existing attacks. Specifically, we find that conventional metrics measuring targeted and untargeted robustness do not appropriately reflect a model's ability to withstand attacks from one set of source classes to another set of target classes. To address the shortcomings of existing methods, we formally define a new metric, termed group-based robustness, that complements existing metrics and is better-suited for evaluating model performance in certain attack scenarios. We show empirically that group-based robustness allows us to distinguish between models' vulnerability against specific threat models in situations where traditional robustness metrics do not apply. Moreover, to measure group-based robustness efficiently and accurately, we 1) propose two loss functions and 2) identify three new attack strategies. We show empirically that with comparable success rates, finding evasive samples using our new loss functions saves computation by a factor as large as the number of targeted classes, and finding evasive samples using our new attack strategies saves time by up to 99\% compared to brute-force search methods. Finally, we propose a defense method that increases group-based robustness by up to 3.52$\times$.

Scalable Verification of GNN-based Job Schedulers

Recently, Graph Neural Networks (GNNs) have been applied for scheduling jobs over clusters achieving better performance than hand-crafted heuristics. Despite their impressive performance, concerns remain over their trustworthiness when deployed in a real-world environment due to their black-box nature. To address these limitations, we consider formal verification of their expected properties such as strategy proofness and locality in this work. We address several domain-specific challenges such as deeper networks and richer specifications not encountered by existing verifiers for image and NLP classifiers. We develop GNN-Verify, the first general framework for verifying both single-step and multi-step properties of these schedulers based on carefully designed algorithms that combine abstractions, refinements, solvers, and proof transfer. Our experimental results on challenging benchmarks show that our approach can provide precise and scalable formal guarantees on the trustworthiness of state-of-the-art GNN-based scheduler.

Constrained Gradient Descent: A Powerful and Principled Evasion Attack Against Neural Networks

Minimal adversarial perturbations added to inputs have been shown to be effective at fooling deep neural networks. In this paper, we introduce several innovations that make white-box targeted attacks follow the intuition of the attacker's goal: to trick the model to assign a higher probability to the target class than to any other, while staying within a specified distance from the original input. First, we propose a new loss function that explicitly captures the goal of targeted attacks, in particular, by using the logits of all classes instead of just a subset, as is common. We show that Auto-PGD with this loss function finds more adversarial examples than it does with other commonly used loss functions. Second, we propose a new attack method that uses a further developed version of our loss function capturing both the misclassification objective and the $L_{\infty}$ distance limit $\epsilon$. This new attack method is relatively 1.5--4.2% more successful on the CIFAR10 dataset and relatively 8.2--14.9% more successful on the ImageNet dataset, than the next best state-of-the-art attack. We confirm using statistical tests that our attack outperforms state-of-the-art attacks on different datasets and values of $\epsilon$ and against different defenses.

Optimization-Guided Binary Diversification to Mislead Neural Networks for Malware Detection

Motivated by the transformative impact of deep neural networks (DNNs) on different areas (e.g., image and speech recognition), researchers and anti-virus vendors are proposing end-to-end DNNs for malware detection from raw bytes that do not require manual feature engineering. Given the security sensitivity of the task that these DNNs aim to solve, it is important to assess their susceptibility to evasion. In this work, we propose an attack that guides binary-diversification tools via optimization to mislead DNNs for malware detection while preserving the functionality of binaries. Unlike previous attacks on such DNNs, ours manipulates instructions that are a functional part of the binary, which makes it particularly challenging to defend against. We evaluated our attack against three DNNs in white-box and black-box settings, and found that it can often achieve success rates near 100%. Moreover, we found that our attack can fool some commercial anti-viruses, in certain cases with a success rate of 85%. We explored several defenses, both new and old, and identified some that can successfully prevent over 80% of our evasion attempts. However, these defenses may still be susceptible to evasion by adaptive attackers, and so we advocate for augmenting malware-detection systems with methods that do not rely on machine learning.

$n$-ML: Mitigating Adversarial Examples via Ensembles of Topologically Manipulated Classifiers

This paper proposes a new defense called $n$-ML against adversarial examples, i.e., inputs crafted by perturbing benign inputs by small amounts to induce misclassifications by classifiers. Inspired by $n$-version programming, $n$-ML trains an ensemble of $n$ classifiers, and inputs are classified by a vote of the classifiers in the ensemble. Unlike prior such approaches, however, the classifiers in the ensemble are trained specifically to classify adversarial examples differently, rendering it very difficult for an adversarial example to obtain enough votes to be misclassified. We show that $n$-ML roughly retains the benign classification accuracies of state-of-the-art models on the MNIST, CIFAR10, and GTSRB datasets, while simultaneously defending against adversarial examples with better resilience than the best defenses known to date and, in most cases, with lower classification-time overhead.

On the Suitability of $L_p$-norms for Creating and Preventing Adversarial Examples

Much research effort has been devoted to better understanding adversarial examples, which are specially crafted inputs to machine-learning models that are perceptually similar to benign inputs, but are classified differently (i.e., misclassified). Both algorithms that create adversarial examples and strategies for defending against them typically use $L_p$-norms to measure the perceptual similarity between an adversarial input and its benign original. Prior work has already shown, however, that two images need not be close to each other as measured by an $L_p$-norm to be perceptually similar. In this work, we show that nearness according to an $L_p$-norm is not just unnecessary for perceptual similarity, but is also insufficient. Specifically, focusing on datasets (CIFAR10 and MNIST), $L_p$-norms, and thresholds used in prior work, we show through online user studies that "adversarial examples" that are closer to their benign counterparts than required by commonly used $L_p$-norm thresholds can nevertheless be perceptually different to humans from the corresponding benign examples. Namely, the perceptual distance between two images that are "near" each other according to an $L_p$-norm can be high enough that participants frequently classify the two images as representing different objects or digits. Combined with prior work, we thus demonstrate that nearness of inputs as measured by $L_p$-norms is neither necessary nor sufficient for perceptual similarity, which has implications for both creating and defending against adversarial examples. We propose and discuss alternative similarity metrics to stimulate future research in the area.

Adversarial Generative Nets: Neural Network Attacks on State-of-the-Art Face Recognition

In this paper we show that misclassification attacks against face-recognition systems based on deep neural networks (DNNs) are more dangerous than previously demonstrated, even in contexts where the adversary can manipulate only her physical appearance (versus directly manipulating the image input to the DNN). Specifically, we show how to create eyeglasses that, when worn, can succeed in targeted (impersonation) or untargeted (dodging) attacks while improving on previous work in one or more of three facets: (i) inconspicuousness to onlooking observers, which we test through a user study; (ii) robustness of the attack against proposed defenses; and (iii) scalability in the sense of decoupling eyeglass creation from the subject who will wear them, i.e., by creating "universal" sets of eyeglasses that facilitate misclassification. Central to these improvements are adversarial generative nets, a method we propose to generate physically realizable attack artifacts (here, eyeglasses) automatically.