Method of Controlled Random Tests Generation

Objectives. The problem of constructing controlled random tests is solved by two-dimensional scaling of initial templates using Hadamard matrices. The limitations of classical approaches to generating test patterns based on enumeration of candidates for test patterns are shown. With an increase in the threshold values of the difference measures of binary test patterns, the computational complexity of constructing such tests increases. The main goal of this article is to develop methods for constructing tests based on initial templates and their expansion to the required bit size based on the application of formal rules.Methods. For two-dimensional scaling of initial templates with specified Hamming distance thresholds, Hadamard matrices and the Sylvester recursive procedure for their construction are applied. The experimental research employed the method of statistical trials.Results. It is demonstrated that methods for constructing controlled random tests based on templates can be viewed as a procedure for scaling controlled random tests to the required bit size. Both templates characterized by the minimum bit size of patterns and any controlled random tests are used to construct the desired tests. The procedure itself is characterized as one-dimensional scaling, which increases the bit size of patterns while maintaining their quantity. To simultaneously increase the bit size and quantity of test sets, a method based on two-dimensional scaling of templates using Hadamard matrices is proposed. This allows for the construction of controlled random tests without the labor-intensive process of enumerating candidate test patterns and computing their difference measure values. It is shown that the unique orthogonality property of Hadamard matrices, as their order increases, enables achieving ratios of the average Hamming distance between test patterns to their bit size close to 1/2. It is noted that the characteristics of the initial templates do not significantly affect the characteristics of the resulting tests constructed using Hadamard matrices obtained through the Sylvester recursive procedure. The feasibility and efficiency of the proposed approach to constructing controlled random tests are evaluated for the case of testing memory devices. It is demonstrated that controlled random tests constructed using Hadamard matrices have significantly higher coverage capability compared to random tests.Conclusion. An approach for generating test patterns in the formation of controlled random tests using Hadamard matrices is considered. The proposed approach is based on two-dimensional scaling of initial templates using these matrices. It is shown that the use of various templates and their two-dimensional scaling allows for the construction of controlled random tests with the required bit size of test patterns and a larger number of them

Random numbers are essential for communications security, as they are widely employed as secret keys and other critical parameters of cryptographic algorithms. The Linux random number generator (LRNG) is the most popular open-source software-based random number generator (RNG). The security of LRNG is influenced by the overall design, especially the quality of entropy sources. Therefore, it is necessary to assess and quantify the quality of the entropy sources which contribute the main randomness to RNGs. In this paper, we perform an empirical study on the quality of entropy sources in LRNG with Linux kernel 5.6, and provide the following two findings. We first analyze two important entropy sources: jiffies and cycles, and propose a method to predict jiffies by cycles with high accuracy. The results indicate that, the jiffies can be correctly predicted thus contain almost no entropy in the condition of knowing cycles. The other important finding is the failure of interrupt cycles during system boot. The lower bits of cycles caused by interrupts contain little entropy, which is contrary to our traditional cognition that lower bits have more entropy. We believe these findings are of great significance to improve the efficiency and security of the RNG design on software platforms.

Facial recognition systems are increasingly deployed in privacy-sensitive and resource-constrained environments such as smart cards. However, traditional face verification relies on high-dimensional floating-point embeddings, which are unsuitable for compact and efficient matching on such platforms. To address this challenge, this work investigates the generation of binary face templates that retain identity information while reducing storage and computational cost. The objective of this study is to benchmark binary biometric representations derived from mobile-class convolutional neural networks (CNNs), aiming to support reproducible, lightweight face verification pipelines. We evaluate four lightweight CNNs—EfficientNet-B0, MobileNetV2, ShuffleNetV2, and SqueezeNet1_1—trained on the MORPH dataset. Binary templates are generated via Principal Component Analysis followed by Iterative Quantization (PCA–ITQ) at 32, 64, and 128 bits. Models are tested cross-dataset on the Georgia Tech Face Database (GT Face) to assess generalization. At 128 bits, EfficientNet-B0 and MobileNetV2 achieve strong verification performance, with area under the curve (AUC) ≈ 0.895–0.899 and equal error rate (EER) ≈ 0.182–0.185. A Hamming-distance analysis confirms clear separation between genuine and impostor pairs, and the bit-flip rate (~17%) indicates intra-subject consistency. Bit-length scaling further reveals monotonic improvements in AUC from 32 to 128 bits, highlighting a trade-off between accuracy and compactness. These results demonstrate that binary templates from lightweight CNNs can deliver efficient, privacy-preserving authentication with limited performance degradation. The proposed pipeline supports reproducibility and aligns with FAIR data principles, making it suitable for secure biometric deployments on constrained hardware.

One of the methods of generating physical random numbers is the use of electrical signals. Previously, when random numbers have been generated from electrical signals, a threshold value has been imposed on the signal level and “0” and “1” have been assigned depending on whether this threshold value was exceeded. In the present paper, the lowest bit of an A-D converter is used so as to demonstrate that high-speed generation of random numbers, which is not achievable by the conventional method, is now possible. Further, it is demonstrated that random numbers can be generated by fluctuations of an oscillator instead by a special noise generator. © 2006 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 89(6): 13–21, 2006; Published online in Wiley InterScience (www.interscience.wiley. com). DOI 10.1002/ecjc.20215

The approach presented in this paper represents voice recordings by a novel acoustic key composed only of binary values. Except for the process being used to extract such keys, there is no need for acoustic modeling and processing in the approach proposed, as all the other elements in the system are based on the binary vectors. We show that this binary key is able to effectively model a speaker’s voice and to distinguish it from other speakers. Its main properties are its small size compared to current speaker modeling techniques and its low computational cost when comparing different speakers as it is limited to obtaining a similarity metric between two binary vectors. Furthermore, the binary key vector extraction process does not need any threshold and offers the opportunity to set the decision steps in a well defined binary domain where scores and decisions are easy to interpret and implement. Index Terms: binary key, speaker modeling, biometrics

Production test data from more than 500,000 chips is analyzed to understand the correlation between the number of defective chips detected by a set of test patterns and the coverage values of these test patterns with respect to various test metrics. Experimental results show that the gate exhaustive metric has the highest correlation when compared to the stuck-at and the bridge coverage estimate metrics, especially for high coverage test patterns. More than 69% of all test patterns can be removed from the test set without reducing the number of detected chips - more than 99% of these patterns are required to obtain high stuck-at coverage. None of the test metrics are very effective in predicting which subset of a given set of test patterns can be removed from the test set without compromising test quality before the patterns are actually applied to manufactured ICs

Absolutely secure communication should be implemented only through the ‘one-time pad' proposed by Shannon, requires that physical random numbers with rates matched with the associated communication systems be used as secret keys. With the wide application of the WDM technology in optical communication, the single channel rate of the current digital communication system has exceeded 10 Gb/s and developed towards 100 Gb/s. To ensure the absolute security of such a large capacity communication, a large number of real-time, and secure random numbers are needed.#br#Secure random numbers are commonly produced through utilizing physical random phenomena, called physical random number generators. However, conventional physical random number generators are limited by the low bandwidth of the applied entropy sources such as thermal noise, photon-counting and chaotic electrical circuits, and thus have typical low bit rates of the order of Mb/s.#br#In recent years, chaotic lasers attracted wide attention due to their generation of secure, reliable and high-speed random number sequences, and so due to their coherent merits such as high bandwidth, large amplitude fluctuation and ease of integration. There have been lots of schemes based on laser chaos for high-speed random number generation, but most of them execute the random number extractions from the associated laser chaos in the electrical domain and thus their generation rates are faced with the well-known ‘electrical bottleneck'. On the other hand, all-optical random number generation (AO-RNG) methods are all signal processes in the optical domain, so they can efficiently overcome this rate limitation and have a great potential in generating ultrafast random numbers of several dozens or hundreds of Gb/s. However, there is no experimental report on its realization of AO-RNG. One of the obstacles in the way for the AO-RNG achievement is to implement the fast and real-time all-optical sampling of the entropy signals (i.e., laser chaos).#br#In this paper, we present a principal experimental demonstration of the feasibility in the all-optical sampling of the chaotic light signal through constructing a TOAD-based all-optical sampler with a polarization-independent semiconductor optical amplifier (SOA). Specifically, we experimentally generate chaotic laser signals using an optical feedback semiconductor laser and finally complete a 5 GSa/s real-time and high-fidelity all-optical sampling of the chaotic laser with a bandwidth of 6.4 GHz. Further experimental results show that whether the optical sampling period is proportional to the external cavity feedback time or not has a great effect on the weak periodic suppression of the chaotic signal: only when both of them are out of proportion, can the weak periodicity of the original chaotic signal be effectively eliminated; and this is favorable for the generation of high-quality physical random numbers. To the best of our knowledge, it is the first time to realize all-optical sampling of chaotic signal in experiments.

Good robot performance often relies upon the selection of design parameters that lead to a well conditioned Jacobian or impedance "design" matrix. In this paper, a new design matrix normalization technique is presented to handle the problem of nonhomogeneous physical units and to provide a means of specifying a performance based design goal. The technique pre- and post-multiplies a design matrix by scaling matrices corresponding to a range of joint and task space variables. The task-space scale factors are used to set the relative required strength or speed along any axes of end point motion while the joint-space scale factors are treated as free design parameters to improve isotropy through nonhomogeneous actuation. The effect of scaling on actual designs is illustrated by a number of design examples using a global search method previously developed by the authors.

There are numerous areas relied on random numbers. As one knows, in Cryptography, randomness plays a vital role from key generation to encrypting the systems. If randomness is not created effectively, the whole system is vulnerable to security threats where an outsider can easily predict the algorithm used to generate the random numbers in the system. Another main application where one would not touch is the role of random numbers in different devices mainly storage-related like Solid State Drives, Universal Serial Bus (USB), Secure Digital (SD) cards random performance testing. This paper focuses on various novel algorithms to generate random numbers for efficient performance evaluation of different drives. The main metrics for performance testing is random read and write performance. Here, the biggest challenge to test the random performance of the drive is not only the extent to which randomness is created but also the testing should cover the entire device (say complete NAND, NOR, etc.). So, the random number generator should generate in such a way that the random numbers should not be able to be predicted and must generate the numbers covering the entire range. This paper proposes different methods for such generators and towards the end, discusses the implementation in Field Programmable Gate Array (FPGA).

Both partition testing and random testing methods are commonly followed practice towards selection of test cases. For partition testing, the programpsilas input domain is divided into subsets, called subdomains, and one or more representatives from each subdomain are selected to test the program. In random testing test cases are selected from the entire programpsilas input domain randomly. The main aim of the paper is to compare the fault-detecting ability of partition testing and random testing methods. The results of comparing the effectiveness of partition testing and random testing may be surprising to many people. Even when partition testing is better than random testing at finding faults, the difference in effectiveness is marginal. Using some effectiveness metrics for testing and some partitioning schemes this paper investigates formal conditions for partition testing to be better than random testing and vice versa.

3D face recognition technique has gained much more attention recently, and it is widely used in security system, identification system, and access control system, etc. The core technique in 3D face recognition is to find out the corresponding points in different 3D face images. The classic partial Iterative Closest Point (ICP) method is iteratively align the two point sets based on repetitively calculate the closest points as the corresponding points in each iteration. After several iterations, the corresponding points can be obtained accurately. However, if two 3D face images with different scale are from the same person, the classic partial ICP does not work. In this paper we propose a modified partial Iterative Closest Point (ICP) method in which the scaling effect is considered to achieve 3D face recognition. We design a 3x3 diagonal matrix as the scale matrix in each iteration of the classic partial ICP. The probing face image which is multiplied by the scale matrix will keep the similar scale with the reference face image. Therefore, we can accurately determine the corresponding points even the scales of probing image and reference image are different. 3D face images in our experiments are acquired by a 3D data acquisition system based on Digital Fringe Projection Profilometry (DFPP). A 3D database consists of 30 group images, three images with the same scale, which are from the same person with different views, are included in each group. And in different groups, the scale of the 3 images may be different from other groups. The experiment results show that our proposed method can achieve 3D face recognition, especially in the case that the scales of probing image and referent image are different.

The statistical nature of numerous problems in mathematics, physics and engineering have led to the development of methods for generating random data for a given distribution. Ancient methods include, dice, coin flipping and shuffling of cards. Today, various pseudo, quasi and true random generators (RNGs) are being proposed for their improved properties. In this work, test metrics for goodness-of-fit and randomness are reviewed. The method of uniform sampling (MUS) is modified for improving the randomness without harming the goodness-of-fit qualities. The test results illustrate that very high goodness-of-fit can be obtained even when the number of observed samples is as small as 10.

Fuzzy logic based Truly Random number generator for high-speed BIST applications

A least-squares penalty method algorithm for inverse problems of steady-state aquifer models

In industry, software testing and coverage-based metrics are the predominant techniques to check correctness of software. This paper addresses automatic unit test generation for programs written in C/C++. The main idea is to improve the coverage obtained by feedback-directed random test generation methods, by utilizing concolic execution on the generated test drivers. Furthermore, for programs with numeric computations, we employ non-linear solvers in a lazy manner to generate new test inputs. These techniques significantly improve the coverage provided by a feedback-directed random unit testing framework, while retaining the benefits of full automation. We have implemented these techniques in a prototype platform, and describe promising experimental results on a number of C/C++ open source benchmarks.