اشترك بالحزمة الذهبية واحصل على وصول غير محدود شمرا أكاديميا
تسجيل مستخدم جديدRepresentation of White- and Black-Box Adversarial Examples in Deep Neural Networks and Humans: A Functional Magnetic Resonance Imaging Study
86
0
0.0
(
0
)
اسأل ChatGPT حول البحث
ﻻ يوجد ملخص باللغة العربية
The recent success of brain-inspired deep neural networks (DNNs) in solving complex, high-level visual tasks has led to rising expectations for their potential to match the human visual system. However, DNNs exhibit idiosyncrasies that suggest their visual representation and processing might be substantially different from human vision. One limitation of DNNs is that they are vulnerable to adversarial examples, input images on which subtle, carefully designed noises are added to fool a machine classifier. The robustness of the human visual system against adversarial examples is potentially of great importance as it could uncover a key mechanistic feature that machine vision is yet to incorporate. In this study, we compare the visual representations of white- and black-box adversarial examples in DNNs and humans by leveraging functional magnetic resonance imaging (fMRI). We find a small but significant difference in representation patterns for different (i.e. white- versus black- box) types of adversarial examples for both humans and DNNs. However, human performance on categorical judgment is not degraded by noise regardless of the type unlike DNN. These results suggest that adversarial examples may be differentially represented in the human visual system, but unable to affect the perceptual experience.
قيم البحث
اقرأ أيضاً
Powerful adversarial attack methods are vital for understanding how to construct robust deep neural networks (DNNs) and for thoroughly testing defense techniques. In this paper, we propose a black-box adversarial attack algorithm that can defeat both
vanilla DNNs and those generated by various defense techniques developed recently. Instead of searching for an optimal adversarial example for a benign input to a targeted DNN, our algorithm finds a probability density distribution over a small region centered around the input, such that a sample drawn from this distribution is likely an adversarial example, without the need of accessing the DNNs internal layers or weights. Our approach is universal as it can successfully attack different neural networks by a single algorithm. It is also strong; according to the testing against 2 vanilla DNNs and 13 defended ones, it outperforms state-of-the-art black-box or white-box attack methods for most test cases. Additionally, our results reveal that adversarial training remains one of the best defense techniques, and the adversarial examples are not as transferable across defended DNNs as them across vanilla DNNs.
Deep neural networks (DNNs) transform stimuli across multiple processing stages to produce representations that can be used to solve complex tasks, such as object recognition in images. However, a full understanding of how they achieve this remains e
lusive. The complexity of biological neural networks substantially exceeds the complexity of DNNs, making it even more challenging to understand the representations that they learn. Thus, both machine learning and computational neuroscience are faced with a shared challenge: how can we analyze their representations in order to understand how they solve complex tasks? We review how data-analysis concepts and techniques developed by computational neuroscientists can be useful for analyzing representations in DNNs, and in turn, how recently developed techniques for analysis of DNNs can be useful for understanding representations in biological neural networks. We explore opportunities for synergy between the two fields, such as the use of DNNs as in-silico model systems for neuroscience, and how this synergy can lead to new hypotheses about the operating principles of biological neural networks.
The vulnerability of deep neural networks (DNNs) to adversarial examples is well documented. Under the strong white-box threat model, where attackers have full access to DNN internals, recent work has produced continual advancements in defenses, ofte
n followed by more powerful attacks that break them. Meanwhile, research on the more realistic black-box threat model has focused almost entirely on reducing the query-cost of attacks, making them increasingly practical for ML models already deployed today. This paper proposes and evaluates Blacklight, a new defense against black-box adversarial attacks. Blacklight targets a key property of black-box attacks: to compute adversarial examples, they produce sequences of highly similar images while trying to minimize the distance from some initial benign input. To detect an attack, Blacklight computes for each query image a compact set of one-way hash values that form a probabilistic fingerprint. Variants of an image produce nearly identical fingerprints, and fingerprint generation is robust against manipulation. We evaluate Blacklight on 5 state-of-the-art black-box attacks, across a variety of models and classification tasks. While the most efficient attacks take thousands or tens of thousands of queries to complete, Blacklight identifies them all, often after only a handful of queries. Blacklight is also robust against several powerful countermeasures, including an optimal black-box attack that approximates white-box attacks in efficiency. Finally, Blacklight significantly outperforms the only known alternative in both detection coverage of attack queries and resistance against persistent attackers.
Most humans have the good fortune to live their lives embedded in richly structured social groups. Yet, it remains unclear how humans acquire knowledge about these social structures to successfully navigate social relationships. Here we address this
knowledge gap with an interdisciplinary neuroimaging study drawing on recent advances in network science and statistical learning. Specifically, we collected BOLD MRI data while participants learned the community structure of both social and non-social networks, in order to examine whether the learning of these two types of networks was differentially associated with functional brain network topology. From the behavioral data in both tasks, we found that learners were sensitive to the community structure of the networks, as evidenced by a slower reaction time on trials transitioning between clusters than on trials transitioning within a cluster. From the neuroimaging data collected during the social network learning task, we observed that the functional connectivity of the hippocampus and temporoparietal junction was significantly greater when transitioning between clusters than when transitioning within a cluster. Furthermore, temporoparietal regions of the default mode were more strongly connected to hippocampus, somatomotor, and visual regions during the social task than during the non-social task. Collectively, our results identify neurophysiological underpinnings of social versus non-social network learning, extending our knowledge about the impact of social context on learning processes. More broadly, this work offers an empirical approach to study the learning of social network structures, which could be fruitfully extended to other participant populations, various graph architectures, and a diversity of social contexts in future studies.
Neural networks are known to be vulnerable to carefully crafted adversarial examples, and these malicious samples often transfer, i.e., they remain adversarial even against other models. Although great efforts have been delved into the transferabilit
y across models, surprisingly, less attention has been paid to the cross-task transferability, which represents the real-world cybercriminals situation, where an ensemble of different defense/detection mechanisms need to be evaded all at once. In this paper, we investigate the transferability of adversarial examples across a wide range of real-world computer vision tasks, including image classification, object detection, semantic segmentation, explicit content detection, and text detection. Our proposed attack minimizes the ``dispersion of the internal feature map, which overcomes existing attacks limitation of requiring task-specific loss functions and/or probing a target model. We conduct evaluation on open source detection and segmentation models as well as four different computer vision tasks provided by Google Cloud Vision (GCV) APIs, to show how our approach outperforms existing attacks by degrading performance of multiple CV tasks by a large margin with only modest perturbations linf=16.
الأسئلة المقترحة
سجل دخول لتتمكن من نشر تعليقات
التعليقات
جاري جلب التعليقات
سجل دخول لتتمكن من متابعة معايير البحث التي قمت باختيارها
