Deep Neural Networks (DNNs) have been shown to be vulnerable against adversarial examples, which are data points cleverly constructed to fool the classifier. Such attacks can be devastating in practice, especially as DNNs are being applied to ever increasing critical tasks like image recognition in autonomous driving. In this paper, we introduce a new perspective on the problem. We do so by first defining robustness of a classifier to adversarial exploitation. Next, we show that the problem of adversarial example generation can be posed as learning problem. We also categorize attacks in literature into high and low perturbation attacks; well-known attacks like fast-gradient sign method (FGSM) and our attack produce higher perturbation adversarial examples while the more potent but computationally inefficient Carlini-Wagner (CW) attack is low perturbation. Next, we show that the dual approach of the attack learning problem can be used as a defensive technique that is effective against high perturbation attacks. Finally, we show that a classifier masking method achieved by adding noise to the a neural network's logit output protects against low distortion attacks such as the CW attack. We also show that both our learning and masking defense can work simultaneously to protect against multiple attacks. We demonstrate the efficacy of our techniques by experimenting with the MNIST and CIFAR-10 datasets.
The most exciting recent development in nonlinear dynamics is realization that chaos can be useful. One application involves "Secure Communication". Two piecewise linear systems with switching nonlinearities have been taken as chaos generators. In the present work the phenomenon of secure communication with chaos masking has been investigated experimentally. In this investigation chaos which is generated from two chaos generators is masked with the massage signal to be transmitted, thus makes communication is more secure.
Recent work on stealing machine learning (ML) models from inference engines with physical side-channel attacks warrant an urgent need for effective side-channel defenses. This work proposes the first $\textit{fully-masked}$ neural network inference engine design. Masking uses secure multi-party computation to split the secrets into random shares and to decorrelate the statistical relation of secret-dependent computations to side-channels (e.g., the power draw). In this work, we construct secure hardware primitives to mask $\textit{all}$ the linear and non-linear operations in a neural network.…
Oregon State University（俄勒冈州立大学）；Samsung Research（三星研究院）；SRI international（斯坦福国际研究院）
We study the problem of data release with privacy, where data is made available with privacy guarantees while keeping the usability of the data as high as possible --- this is important in health-care and other domains with sensitive data. In particular, we propose a method of masking the private data with privacy guarantee while ensuring that a classifier trained on the masked data is similar to the classifier trained on the original data, to maintain usability. We analyze the theoretical risks of the proposed method and the traditional input perturbation method. Results show that the propose…
We propose a secure and lightweight key based challenge obfuscation for strong PUFs. Our architecture is designed to be resilient against learning attacks. Our obfuscation mechanism uses non-linear feedback shift registers (NLFSRs). Responses are directly provided to the user, without error correction or extra post-processing steps. We also discuss the cost of protecting our architecture against power analysis attacks with clock randomization, and Boolean masking. Security against learning attacks is assessed using avalanche criterion, and deep-neural network attacks. We designed a testchip in…
Duke University（杜克大学）；University of Illinois Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）
Classifiers in supervised learning have various security and privacy issues, e.g., 1) data poisoning attacks, backdoor attacks, and adversarial examples on the security side as well as 2) inference attacks and the right to be forgotten for the training data on the privacy side. Various secure and privacy-preserving supervised learning algorithms with formal guarantees have been proposed to address these issues. However, they suffer from various limitations such as accuracy loss, small certified security guarantees, and/or inefficiency. Self-supervised learning is an emerging technique to pre-t…
University of Campinas（坎皮纳斯大学）；University of Washington Tacoma（华盛顿大学塔科马分校）；Monash University（莫纳什大学）；Aarhus University（奥胡斯大学）
We propose a privacy-preserving Naive Bayes classifier and apply it to the problem of private text classification. In this setting, a party (Alice) holds a text message, while another party (Bob) holds a classifier. At the end of the protocol, Alice will only learn the result of the classifier applied to her text input and Bob learns nothing. Our solution is based on Secure Multiparty Computation (SMC). Our Rust implementation provides a fast and secure solution for the classification of unstructured text. Applying our solution to the case of spam detection (the solution is generic, and can be…
In machine learning, classifiers are used to predict a class of a given query based on an existing (classified) database. Given a database S of n d-dimensional points and a d-dimensional query q, the k-nearest neighbors (kNN) classifier assigns q with the majority class of its k nearest neighbors in S. In the secure version of kNN, S and q are owned by two different parties that do not want to share their data. Unfortunately, all known solutions for secure kNN either require a large communication complexity between the parties, or are very inefficient to run. In this work we present a classifi…
Advanced Encryption Standard (AES) is a symmetric key encryption algorithm which is extensively used in secure electronic data transmission. When introduced, although it was tested and declared as secure, in 2005, a researcher named Bernstein claimed that it is vulnerable to side channel attacks. The cache-based timing attack is the type of side channel attack demonstrated by Bernstein, which uses the timing variation in cache hits and misses. This kind of attacks can be prevented by masking the actual timing information from the attacker. Such masking can be performed by altering the original…
Princeton University（普林斯顿大学）；University of Chicago（芝加哥大学）
Localized adversarial patches aim to induce misclassification in machine learning models by arbitrarily modifying pixels within a restricted region of an image. Such attacks can be realized in the physical world by attaching the adversarial patch to the object to be misclassified, and defending against such attacks is an unsolved/open problem. In this paper, we propose a general defense framework called PatchGuard that can achieve high provable robustness while maintaining high clean accuracy against localized adversarial patches. The cornerstone of PatchGuard involves the use of CNNs with sma…
In this paper, a secure Convolutional Neural Network classifier is proposed using Fully Homomorphic Encryption (FHE). The secure classifier provides a user with the ability to out-source the computations to a powerful cloud server and/or setup a server to classify inputs without providing the model or revealing source data. To this end, a real number framework is developed over FHE by using a fixed point format with binary digits. This allows for real number computations for basic operators like addition, subtraction, and multiplication but also to include secure comparisons and max functions.…
Enhanced Outsourced and Secure Inference for Tall Sparse Decision Trees
Dept. of Computer Science and Engineering（计算机科学与工程系）；New York University（纽约大学）；Dept. of Applied Informatics（应用信息学系）；University of Macedonia（马其顿大学）；NOVA LINCS；NOVA School of Science and Technology（NOVA科学与技术学校）；Advanced Wireless and Security Lab（先进无线与安全实验室）；Florida International University（佛罗里达国际大学）
Relay attacks pose an important threat in wireless ranging and authentication systems. Distance bounding protocols have been proposed as an effective countermeasure against these attacks and allow a verifier and a prover to establish an upper bound on the distance between them. However, secure distance bounding protocols are hard to realize in practice due to stringent implementation requirements. In this paper, we look into a yet unexplored research area and show how the security strength of Ultra Wide Band (UWB) distance bounding protocols can be significantly increased by imposing several a…
The adversarial patch attack against image classification models aims to inject adversarially crafted pixels within a restricted image region (i.e., a patch) for inducing model misclassification. This attack can be realized in the physical world by printing and attaching the patch to the victim object; thus, it imposes a real-world threat to computer vision systems. To counter this threat, we design PatchCleanser as a certifiably robust defense against adversarial patches. In PatchCleanser, we perform two rounds of pixel masking on the input image to neutralize the effect of the adversarial pa…
While modern day web applications aim to create impact at the civilization level, they have become vulnerable to adversarial activity, where the next cyber-attack can take any shape and can originate from anywhere. The increasing scale and sophistication of attacks, has prompted the need for a data driven solution, with machine learning forming the core of many cybersecurity systems. Machine learning was not designed with security in mind, and the essential assumption of stationarity, requiring that the training and testing data follow similar distributions, is violated in an adversarial domai…
Despite the wide use of machine learning in adversarial settings including computer security, recent studies have demonstrated vulnerabilities to evasion attacks---carefully crafted adversarial samples that closely resemble legitimate instances, but cause misclassification. In this paper, we examine the adequacy of the leading approach to generating adversarial samples---the gradient descent approach. In particular (1) we perform extensive experiments on three datasets, MNIST, USPS and Spambase, in order to analyse the effectiveness of the gradient-descent method against non-linear support vec…
Machine learning based malware detection techniques rely on grayscale images of malware and tends to classify malware based on the distribution of textures in graycale images. Albeit the advancement and promising results shown by machine learning techniques, attackers can exploit the vulnerabilities by generating adversarial samples. Adversarial samples are generated by intelligently crafting and adding perturbations to the input samples. There exists majority of the software based adversarial attacks and defenses. To defend against the adversaries, the existing malware detection based on mach…
The evaluation of robustness against adversarial manipulation of neural networks-based classifiers is mainly tested with empirical attacks as methods for the exact computation, even when available, do not scale to large networks. We propose in this paper a new white-box adversarial attack wrt the $l_p$-norms for $p \in \{1,2,\infty\}$ aiming at finding the minimal perturbation necessary to change the class of a given input. It has an intuitive geometric meaning, yields quickly high quality results, minimizes the size of the perturbation (so that it returns the robust accuracy at every threshol…
Biometric verification has been widely deployed in current authentication solutions as it proves the physical presence of individuals. To protect the sensitive biometric data in such systems, several solutions have been developed that provide security against honest-but-curious (semi-honest) attackers. However, in practice attackers typically do not act honestly and multiple studies have shown drastic biometric information leakage in such honest-but-curious solutions when considering dishonest, malicious attackers. In this paper, we propose a provably secure biometric verification protocol to …
Side-channel attacks impose a serious threat to cryptographic algorithms, including widely employed ones, such as AES and RSA. These attacks take advantage of the algorithm implementation in hardware or software to extract secret information via side channels. Software masking is a mitigation approach against power side-channel attacks aiming at hiding the secret-revealing dependencies from the power footprint of a vulnerable implementation. However, this type of software mitigation often depends on general-purpose compilers, which do not preserve non-functional properties. Moreover, microarch…
Virginia Tech（弗吉尼亚理工大学）；University of Maryland（马里兰大学）；The Ohio State University（俄亥俄州立大学）
Semantic communications seeks to transfer information from a source while conveying a desired meaning to its destination. We model the transmitter-receiver functionalities as an autoencoder followed by a task classifier that evaluates the meaning of the information conveyed to the receiver. The autoencoder consists of an encoder at the transmitter to jointly model source coding, channel coding, and modulation, and a decoder at the receiver to jointly model demodulation, channel decoding and source decoding. By augmenting the reconstruction loss with a semantic loss, the two deep neural network…
University of Central Florida（中佛罗里达大学）；University of California, Los Angeles（加利福尼亚大学洛杉矶分校）
The vulnerability of deep networks to adversarial attacks is a central problem for deep learning from the perspective of both cognition and security. The current most successful defense method is to train a classifier using adversarial images created during learning. Another defense approach involves transformation or purification of the original input to remove adversarial signals before the image is classified. We focus on defending naturally-trained classifiers using Markov Chain Monte Carlo (MCMC) sampling with an Energy-Based Model (EBM) for adversarial purification. In contrast to advers…
Massachusetts Institute of Technology（麻省理工学院）；University of California, Berkeley（加州大学伯克利分校）
We identify obfuscated gradients, a kind of gradient masking, as a phenomenon that leads to a false sense of security in defenses against adversarial examples. While defenses that cause obfuscated gradients appear to defeat iterative optimization-based attacks, we find defenses relying on this effect can be circumvented. We describe characteristic behaviors of defenses exhibiting the effect, and for each of the three types of obfuscated gradients we discover, we develop attack techniques to overcome it. In a case study, examining non-certified white-box-secure defenses at ICLR 2018, we find ob…
Intrusion Prevention Systems (IPS), have long been the first layer of defense against malicious attacks. Most sensitive systems employ instances of them (e.g. Firewalls) to secure the network perimeter and filter out attacks or unwanted traffic. A firewall, similar to classifiers, has a boundary to decide which traffic sample is normal and which one is not. This boundary is defined by configuration and is managed by a set of rules which occasionally might also filter normal traffic by mistake. However, for some applications, any interruption of the normal operation is not tolerable e.g. in pow…
Domain Generation Algorithms (DGAs) are used by adversaries to establish Command and Control (C\&C) server communications during cyber attacks. Blacklists of known/identified C\&C domains are often used as one of the defense mechanisms. However, since blacklists are static and generated by signature-based approaches, they can neither keep up nor detect never-seen-before malicious domain names. Due to this shortcoming of blacklist domain checking, machine learning algorithms have been used to address the problem to some extent. However, when training is performed with limited datasets, the algo…
