Neeraj Gupta

The potential for sabotaging Deep Convolutions Neural Networks classifiers by Universal Perturbation Attack (UPA) has proved itself as an effective threat to fool deep learning models in sensitive applications such as autonomous vehicles, clinical diagnosis, face recognition, etc. The prospective application of UPA is for adversarial training of deep convolutional networks against the attacks. Although evolutionary algorithms have already shown their tremendous ability in solving non-convex complex problems, the literature has limited exploration of evolutionary techniques and strategies for UPA, thus, it needs to be explored on evolutionary algorithms to minimize the magnitude and number of perturbation pixels while maximizing the misclassification of maximum data samples. In this research. This work focuses on utilizing an Integer Coded Genetic Algorithm within an evolutionary framework to evolve the UPA. The evolutionary UPA has been structured, analyzed, and compared for two evolutionary optimization structures: (i) constrained single-objective evolutionary UPA and (ii) Pareto double-objective evolutionary UPA. The efficiency of the methodology is analyzed on GoogleNet convolution Neural Network for its effectiveness on the Imagenet data set. The results show that under the same experimental conditions, the constrained single objective technique outperforms the Pareto double objective one, and manages a successful breach on a deep network wherein the average detection score falls to 0.446429. It is observed that besides the minimization of the detection rate score, the constraint of invisibility of noise is much more effective rather than having a conflicting objective of noise power minimization.

The expected fifth industrial revolution or Industry 5.0 (I-5.0) is human-centered and concerns societal values, and sustainability. I-5.0 focuses on human and machine coworking by augmenting human-collaborative intelligent robots. The current developments in information communications and the increasing market need for high agility and innovative ways to tailor products urge the world for an I-5.0 transformation. Artificial intelligence (AI) plays a key role in I-5.0 in offering intelligent components, smart digital twins, smart cyber-physical systems, and cobotics for efficient, precise and human-centered decision-making. In this context, AI boosts human-robot collaboration in production for timely and precise responses to the market and thereof realizing the concept of mass personalization. In addition, since I-5.0 is based on the idea of the widespread connection of devices, and identities in and out of the production process, there is a high level of vulnerability in front of cyberattacks where AI assists in smart cyber defense techniques. This article is a review of I-5.0 human-center concerns, involvement of social values, and I-5.0 technological elements in connection with human-collaborative artificial intelligence. The role of AI in mass personalization and defense in front of cybersecurity threats is analyzed, and the horizon of I-5.0 in AI context is depicted. Finally, the realization challenges are reviewed as I-5.0 aims to make the future instead of waiting to face the future, and there is a variety of challenges on this path.

Electrical energy is critical to a country’s socioeconomic progress. Distribution system expansion planning addresses the services that must be installed for the distribution networks to meet the expected load need, while also meeting different operational and technical limitations. The incorporation of distributed generation sources (DGs) alters the operating characteristics of modern power systems, resulting in major economic and technical benefits, such as simplified distribution network expansion planning, lower power losses, and improved voltage profile. Thus, in this study, an analytical method is used to design the expansion planning of the Addis North distribution network considering the integration of optimal sizes of distributed generations for the projected demand growths. To evaluate the capability of the existing Addis North distribution network and its capability to supply reliable power considering future expansion, the load demand forecast for the years 2020–2030 is done using the least square method. The performance evaluation of the existing and the upgraded network considering the existing and forecasted load demand for the years 2030 is done using ETAP software. Accordingly, the results revealed that the existing networks cannot meet the existing load demand of the town, with major problems of increased power loss and a reduced voltage profile. To mitigate this problem, the Addis North feeder-1 distribution network is upgraded and for each study case, the balanced and positive sequence load flow analysis was executed and the maximum total real and reactive power losses were found at bus 29. The result shows that the upgraded network of bus 29 was the optimal location of DG and its size was 9.93 MW. After the optimal size of DG was placed at this bus, the real and reactive power losses of the upgraded networks were 0.2939 MW and 0.219 MVAr, respectively. At bus 29 the maximum power losses reduction and voltage profile improvements were found. The active and reactive power losses were minimized by 21.285% and 19.633% respectively and the voltage profiles were improved by 8.78%. Thus, in the predicted year 2030, DG power sources could cover 61.12% of the feeder-1 power requirements.

Compressive sensing has the ability of reconstruction of signal/image from the compressive measurements which are sensed with a much lower number of samples than a minimum requirement by Nyquist sampling theorem. The random acquisition is widely suggested and used for compressive sensing. In the random acquisition, the randomness of the sparsity structure has been deployed for compressive sampling of the signal/image. The article goes through all the literature up to date and collects the main methods, and simply described the way each of them randomly applies the compressive sensing. This article is a comprehensive review of random acquisition techniques in compressive sensing. Theses techniques have reviews under the main categories of (1) random demodulator, (2) random convolution, (3) modulated wideband converter model, (4) compressive multiplexer diagram, (5) random equivalent sampling, (6) random modulation pre-integration, (7) quadrature analog-to-information converter, (8) randomly triggered modulated-wideband compressive sensing (RT-MWCS). Request access from your librarian to read this article's full text.

Optimization plays a fundamental role in understanding stability characteristics of optical systems, for example, lenses, mirrors, and their constrained counterparts. For a dynamical object, the authors address the issue of stability of an image formed under fluctuations of optimization variables. As per this analysis, for a given single lens, mirror, or optical instrument, the positivity of pure correlation components of the fluctuation matrix characterizes the local stability about a chosen fixed point, whilst the positivity of the determinant of the fluctuation matrix offers global stability of the image under variations of the object distance and lateral magnification. In this paper, they demonstrate that there exists a clear cut distinction between the positive and negative lenses and mirrors about the line of unit lateral magnification. Moreover, they propose the extension of the above model to a class of constraint optical system optimization. They also propose an extended formulation for an optimized designing of nonlinearly constrained optical systems with finitely many components. Request access from your librarian to read this article's full text.