Biblio

Previous work has shown that a neural network with the rectified linear unit (ReLU) activation function leads to a convex polyhedral decomposition of the input space. These decompositions can be represented by a dual graph with vertices corresponding to polyhedra and edges corresponding to polyhedra sharing a facet, which is a subgraph of a Hamming graph. This paper illustrates how one can utilize the dual graph to detect and analyze adversarial attacks in the context of digital images. When an image passes through a network containing ReLU nodes, the firing or non-firing at a node can be encoded as a bit (1 for ReLU activation, 0 for ReLU non-activation). The sequence of all bit activations identifies the image with a bit vector, which identifies it with a polyhedron in the decomposition and, in turn, identifies it with a vertex in the dual graph. We identify ReLU bits that are discriminators between non-adversarial and adversarial images and examine how well collections of these discriminators can ensemble vote to build an adversarial image detector. Specifically, we examine the similarities and differences of ReLU bit vectors for adversarial images, and their non-adversarial counterparts, using a pre-trained ResNet-50 architecture. While this paper focuses on adversarial digital images, ResNet-50 architecture, and the ReLU activation function, our methods extend to other network architectures, activation functions, and types of datasets.

The Controller area network (CAN) is the most extensively used in-vehicle network. It is set to enable communication between a number of electronic control units (ECU) that are widely found in most modern vehicles. CAN is the de facto in-vehicle network standard due to its error avoidance techniques and similar features, but it is vulnerable to various attacks. In this research, we propose a CAN bus intrusion detection system (IDS) based on convolutional neural networks (CNN). U-CAN is a segmentation model that is trained by monitoring CAN traffic data that are preprocessed using hamming distance and saliency detection algorithm. The model is trained and tested using publicly available datasets of raw and reverse-engineered CAN frames. With an F\_1 Score of 0.997, U-CAN can detect DoS, Fuzzy, spoofing gear, and spoofing RPM attacks of the publicly available raw CAN frames. The model trained on reverse-engineered CAN signals that contain plateau attacks also results in a true positive rate and false-positive rate of 0.971 and 0.998, respectively.

In the process of big data nearest neighbor similar data retrieval, affected by the way of data feature extraction, the retrieval accuracy is low. Therefore, this paper proposes the design of big data nearest neighbor similar data retrieval algorithm based on improved random forest. Through the improvement of random forest model and the construction of random decision tree, the characteristics of current nearest neighbor big data are clarified. Based on the improved random forest, the hash code is generated. Finally, combined with the Hamming distance calculation method, the nearest neighbor similar data retrieval of big data is realized. The experimental results show that: in the multi label environment, the retrieval accuracy is improved by 9% and 10%. In the single label environment, the similar data retrieval accuracy of the algorithm is improved by 12% and 28% respectively.

Nearest neighbor (NN) search is an essential operation in many applications, such as one/few-shot learning and image classification. As such, fast and low-energy hardware support for accurate NN search is highly desirable. Ternary content-addressable memories (TCAMs) have been proposed to accelerate NN search for few-shot learning tasks by implementing \$L\$∞ and Hamming distance metrics, but they cannot achieve software-comparable accuracies. This paper proposes a novel distance function that can be natively evaluated with multi-bit content-addressable memories (MCAMs) based on ferroelectric FETs (Fe-FETs) to perform a single-step, in-memory NN search. Moreover, this approach achieves accuracies comparable to floating-point precision implementations in software for NN classification and one/few-shot learning tasks. As an example, the proposed method achieves a 98.34% accuracy for a 5-way, 5-shot classification task for the Omniglot dataset (only 0.8% lower than software-based implementations) with a 3-bit MCAM. This represents a 13% accuracy improvement over state-of-the-art TCAM-based implementations at iso-energy and iso-delay. The presented distance function is resilient to the effects of FeFET device-to-device variations. Furthermore, this work experimentally demonstrates a 2-bit implementation of FeFET MCAM using AND arrays from GLOBALFOUNDRIES to further validate proof of concept.

A Field Programmable Gate Array (FPGA) is a digital Integrated Circuit made up of interconnected functional blocks, which can be programmed by the end-user to perform required logic functions. As FPGAs are re-programmable, partially re-configurable and have lowertime to market, FPGA has become a vital component in the field of electronics. FPGAs are undergoing many security issues as the adversaries are trying to make profits by replicating the original design, without any investment. The major security issues are cloning, counterfeiting, reverse engineering, Physical tampering, and insertion of malicious components, etc. So, there is a need for security of FPGAs. A Secret key must be embedded in an IC, to provide identification and authentication to it. Physical Unclonable Functions (PUFs) can provide these secret keys, by using the physical properties of the chip. These physical properties are not reproducible even by the manufacturer. Hence the responses produced by the PUF are unique for every individual chip. The method of generating unique binary signatures helps in cryptographic key generation, digital rights management, Intellectual Property (IP) protection, IC counterfeit prevention, and device authentication. The PUFs are very promising in signature generation in the field of hardware security. In this paper, the secret binary responses is generated with the help of a delay based Ring Oscillator PUF, which does not use a clock circuit in its architecture.

Security services and share information is provided by the computer network. The computer network is by default there is not security. The Attackers can use this provision to hack and steal private information. Confidentiality, creation, changes, and truthful of data is will be big problems in the network. Many types of research have given many methods regarding this, from these methods Generating Initial Chromosome Key called Generating Dynamic Chromosome Key (GDCK), which is a novel approach. With the help of the RSA (Rivest Shamir Adleman) algorithm, GICK and GDCK have created an initial key. The proposed method has produced new techniques using genetic fitness function for the sender and receiver. The outcome of GICK and GDCK has been verified by NIST (National Institute of Standards Technology) tools and analyzes randomness of auto-generated keys with various methods. The proposed system has involved three examines; it has been yield better P-Values 6.44, 7.05, and 8.05 while comparing existing methods.

Nowadays, physical tampering and counterfeiting of electronic devices are still an important security problem and have a great impact on large-scale and distributed applications, such as Internet-of-Things. Physical Unclonable Functions (PUFs) have the potential to be a fundamental means to guarantee intrinsic hardware security, since they promise immunity against most of known attack models. However, inner nature of PUF circuits hinders a wider adoption since responses turn out to be noisy and not stable during time. To overcome this issue, most of PUF implementations require a fuzzy extraction scheme, able to recover responses stability by exploiting error correction codes (ECCs). In this paper, we propose a Reed-Muller (RM) ECC design, meant to be embedded into a fuzzy extractor, that can be efficiently configured in terms of area/delay constraints in order to get reliable responses from PUFs. We provide implementation details and experimental evidences of area/delay efficiency through syntheses on medium-range FPGA device.

Logical locking is the most popular countermeasure against the hardware attacks like intellectual property (IP) piracy, Trojan insertion and illegal integrated circuit (IC) overproduction. The functionality of the design is locked by the added logics into the design. Thus, the design is accessible only to the authorized users by applying the valid keys. However, extracting the secret key of the logically locked design have become an extensive effort and it is commonly known as key guessing attacks. Thus, the main objective of the proposed technique is to build a secured hardware against attacks like Brute force attack, Hill climbing attack and path sensitization attacks. Furthermore, the gates with low observability are chosen for encryption, this is to obtain an optimal output corruption of 50% Hamming distance with minimal design overhead and implementation complexity. The experimental results are validated on ISCAS'85 benchmark circuits, with a highly secured locking mechanism.

When employing biometric recognition systems, we have to take into account that biometric data are considered sensitive data. This has raised some privacy issues, and therefore secure systems providing template protection are required. Using homomorphic encryption, permanent protection can be ensured, since templates are stored and compared in the encrypted domain. In addition, the unprotected system's accuracy is preserved. To solve the problem of the computational overload linked to the encryption scheme, we present an early decision making strategy for iris-codes. In order to improve the recognition accuracy, the most consistent bits of the iris-code are moved to the beginning of the template. This allows an accurate block-wise comparison, thereby reducing the execution time. Hence, the resulting system grants template protection in a computationally efficient way. More specifically, in the experimental evaluation in identification mode, the block-wise comparison achieves a 92% speed-up on the IITD database with 300 enrolled templates.

Two design techniques improve the robustness of Whitenoise encryption algorithm implementation to side-channel attacks based on dynamic and/or static power consumption. The first technique conceals the power consumption and has linear cost. The second technique randomizes the power consumption and has quadratic cost. These techniques are not mutually exclusive; their synergy provides a good robustness to power analysis attacks. Other circuit-level protection can be applied on top of the proposed techniques, opening the avenue for generating very robust implementations.

Cross-modal hashing, which searches nearest neighbors across different modalities in the Hamming space, has become a popular technique to overcome the storage and computation barrier in multimedia retrieval recently. Although dozens of cross-modal hashing algorithms are proposed to yield compact binary code representation, applying exhaustive search in a large-scale dataset is impractical for the real-time purpose, and the Hamming distance computation suffers inaccurate results. In this paper, we propose a novel index scheme over binary hash codes in cross-modal retrieval. The proposed indexing scheme exploits a few binary bits of the hash code as the index code. Based on the index code representation, we construct an inverted index structure to accelerate the retrieval efficiency and train a neural network to improve the indexing accuracy. Experiments are performed on two benchmark datasets for retrieval across image and text modalities, where hash codes are generated by three cross-modal hashing methods. Results show the proposed method effectively boosts the performance over the benchmark datasets and hash methods.

Searchable encryption will become more important as medical services intensify their use of big data and artificial intelligence. To use searchable encryption safely, the resistance of terminals with embedded searchable encryption to illegal attacks (tamper resistance) is extremely important. This study proposes a searchable encryption system embedded in terminals and evaluate the tamper resistance of the proposed system. This study also proposes attack scenarios and quantitatively evaluates the tamper resistance of the proposed system by performing experiments following the proposed attack scenarios.

Artificial intelligence technology such as neural network (NN) is widely used in intelligence module for Internet of Things (IoT). On the other hand, the risk of illegal attacks for IoT devices is pointed out; therefore, security countermeasures such as an authentication are very important. In the field of hardware security, the physical unclonable functions (PUFs) have been attracted attention as authentication techniques to prevent the semiconductor counterfeits. However, implementation of the dedicated hardware for both of NN and PUF increases circuit area. Therefore, this study proposes a new area constraint aware PUF for intelligence module. The proposed PUF utilizes the propagation delay time from input layer to output layer of NN. To share component for operation, the proposed PUF reduces the circuit area. Experiments using a field programmable gate array evaluate circuit area and PUF performance. In the result of circuit area, the proposed PUF was smaller than the conventional PUFs was showed. Then, in the PUF performance evaluation, for steadiness, diffuseness, and uniqueness, favorable results were obtained.

In recent years, deep convolution neural networks (DCNNs) have won many contests in machine learning, object detection, and pattern recognition. Furthermore, deep learning techniques achieved exceptional performance in image classification, reaching accuracy levels beyond human capability. Malware variants from similar categories often contain similarities due to code reuse. Converting malware samples into images can cause these patterns to manifest as image features, which can be exploited for DCNN classification. Techniques for converting malware binaries into images for visualization and classification have been reported in the literature, and while these methods do reach a high level of classification accuracy on training datasets, they tend to be vulnerable to overfitting and perform poorly on previously unseen samples. In this paper, we explore and document a variety of techniques for representing malware binaries as images with the goal of discovering a format best suited for deep learning. We implement a database for malware binaries from several families, stored in hexadecimal format. These malware samples are converted into images using various approaches and are used to train a neural network to recognize visual patterns in the input and classify malware based on the feature vectors. Each image type is assessed using a variety of learning models, such as transfer learning with existing DCNN architectures and feature extraction for support vector machine classifier training. Each technique is evaluated in terms of classification accuracy, result consistency, and time per trial. Our preliminary results indicate that improved image representation has the potential to enable more effective classification of new malware.

In this paper an analog cellular neural network is proposed with application in physical unclonable function design. Dynamical behavior of the circuit and its high sensitivity to the process variation can be exploited in a challenge-response security system. The proposed circuit can be used as unclonable core module in the secure systems for applications such as device identification/authentication and secret key generation. The proposed circuit is designed and simulated in 45-nm bulk CMOS technology. Monte Carlo simulation for this circuit, results in unpolarized Gaussian-shaped distribution for Hamming Distance between 4005 100-bit PUF instances.

In recent years, binary coding techniques are becoming increasingly popular because of their high efficiency in handling large-scale computer vision applications. It has been demonstrated that supervised binary coding techniques that leverage supervised information can significantly enhance the coding quality, and hence greatly benefit visual search tasks. Typically, a modern binary coding method seeks to learn a group of coding functions which compress data samples into binary codes. However, few methods pursued the coding functions such that the precision at the top of a ranking list according to Hamming distances of the generated binary codes is optimized. In this paper, we propose a novel supervised binary coding approach, namely Top Rank Supervised Binary Coding (Top-RSBC), which explicitly focuses on optimizing the precision of top positions in a Hamming-distance ranking list towards preserving the supervision information. The core idea is to train the disciplined coding functions, by which the mistakes at the top of a Hamming-distance ranking list are penalized more than those at the bottom. To solve such coding functions, we relax the original discrete optimization objective with a continuous surrogate, and derive a stochastic gradient descent to optimize the surrogate objective. To further reduce the training time cost, we also design an online learning algorithm to optimize the surrogate objective more efficiently. Empirical studies based upon three benchmark image datasets demonstrate that the proposed binary coding approach achieves superior image search accuracy over the state-of-the-arts.

Robust image hashing seeks to transform a given input image into a shorter hashed version using a key-dependent non-invertible transform. These hashes find extensive applications in content authentication, image indexing for database search and watermarking. Modern robust hashing algorithms consist of feature extraction, a randomization stage to introduce non-invertibility, followed by quantization and binary encoding to produce a binary hash. This paper describes a novel algorithm for generating an image hash based on Log-Polar transform features. The Log-Polar transform is a part of the Fourier-Mellin transformation, often used in image recognition and registration techniques due to its invariant properties to geometric operations. First, we show that the proposed perceptual hash is resistant to content-preserving operations like compression, noise addition, moderate geometric and filtering. Second, we illustrate the discriminative capability of our hash in order to rapidly distinguish between two perceptually different images. Third, we study the security of our method for image authentication purposes. Finally, we show that the proposed hashing method can provide both excellent security and robustness.