Automated Teller Machines (ATMs) are widely used for cash withdrawals and other banking transactions. However, their touchscreens pose a significant health risk due to potential contamination with various pathogens. Studies have shown that ATMs can harbor bacteria and viruses linked to diseases like influenza (25%), E. coli (18%), and Staphylococcus aureus (14%). This research proposes Gesture Vault, a touchless ATM system that leverages facial and hand gesture recognition technologies. Gesture Vault addresses hygiene concerns by eliminating the need for physical contact with the ATM screen. Facial recognition ensures secure user identification, while hand gestures provide an intuitive interface for account selection and transaction initiation. This type of authentication also strengthens security compared to traditional PIN-based systems. Gesture Vault promotes a hygienic and secure banking experience, potentially reducing the spread of infectious diseases and improving user confidence in using ATMs.

According to several studies, it is feasible to significantly raise the detection engine’s effectiveness and accuracy by choosing the right features for a threat detection system. New advances like Distributed Computing and Enormous Information have expanded network traffic, and the danger identification framework must proactively gain and dissect the information delivered by the approaching traffic. Nonetheless, not all elements in an enormous dataset help to portray the traffic, therefore restricting and picking few reasonable highlights might accelerate and improve the danger discovery framework’s exactness. Deep neural networks enhance the detection rates of intrusion detection models, making machine learning-based intrusion detection systems (IDS’s) useful recently. Consumers, however, find it more and more challenging to comprehend the reasoning behind their selections as models become more complex accuracy. Using relevant features from the NSL-KDD dataset, we apply appropriate feature selection mechanisms to implement an intrusion detection system to implement a faster system with increased accuracy. We use Explainable Model (SHAP) to interpret the results of IDS. The interpretation of findings utilizing the Explainable Model (SHAP) for machine learning (ML) and deep learning (DL) models heavily depends on its efficiency. While DL models require more resources, ML models are computationally efficient. Both models, however, gain from SHAP interpretations, which offer perceptions into the significance of features and contributions to predictions. While DL models excel in accuracy, ML models offer efficiency. The decision is based on the particular needs and resources that are available, with SHAP offering greater knowledge of model behavior and feature impact.

Network-based architectures provide accessibility to various hardware elements and information. The Vehicle to Vehicle (V2V) protocol is a Wi-Fi-based communication that potentially exposes the system to a variety of vulnerabilities. Automotive electronics is a rapidly evolving field that combines numerous machine learning algorithms, dependable communication protocols, and precise image processing techniques to create a user-friendly and self-driving autonomous car. A generic network communication vulnerability assessment method is used to undertake a navigational attack against connected vehicles using GPS falsification. This study emphasizes the security risks of using Wi-Fi for V2V (vehicle 2 vehicle) communication without suitable access control measures in place. The tool used for establishing V2V Communication is based on Kali Linux. The breakdown of vulnerabilities within each layer, such as eavesdropping, spoofing, and Malevolent intermediary strike, highlights the diverse range of potential threats vehicles might face. The focus on the communication layer, both within the vehicle and with external infrastructure (V2X communication), aligns with the current trend of connectivity in automotive technology. Providing a state-of-the-art review of attacks and suggesting countermeasures demonstrates a proactive approach to mitigating potential risks. It's essential not only to identify vulnerabilities but also to propose effective solutions to safeguard vehicular communications and, ultimately, the control layer responsible for critical vehicle functions.

The process of software development is complex, and one of its most critical aspects is estimating the amount of work required for various projects. However, accurately defining the exact amount of work needed in the early stages of production can be challenging. Researchers have been working on creating different machine and deep learning models to address this issue. These models, including single-approach models and multi-model ensembles, utilize optimization strategies to provide precise predictions. We propose a hybrid particle swarm optimization (PSO) based artificial neural networks (ANNs) model for software effort estimation (SEE), which has shown to outperform existing models. This model was tested on various datasets such as Albrecht, China Desharnais, Kemerer Kitchenham Maxwell, and Cocomo81. The hybrid PSO-optimized ANNs model has exhibited exceptional accuracy, as evidenced by consistently high R-squared (R2) values across multiple datasets. Additionally, the model has displayed low root mean square error (RMSE) and mean absolute error (MAE) values, indicating precise predictions. These outcomes affirm the model's precision and effectiveness. The model's small MAE further confirms its accuracy in predicting the required work during software development. With these remarkable results, the hybrid PSO-optimized ANNs model will undoubtedly play a crucial role in software development processes, providing accurate and precise predictions of the required work.

The growth of Internet of Things (IoT) devices has created opportunities in multiple sectors, but has also created the threat of growing attacks. Malicious actors are taking advantages of the lightweight nature of IoT devices and targeting them to perform attacks. This paper presents a novel approach to detect malicious traffic at an early stage to block attacks at the entry point of network so devices could be protected from getting infected. An optimized Machine Learning(ML) based algorithm is proposed specifically designed for light weighted IoT devices. The IoT- 23 dataset is used in this study to create a model and test the performance. The proposed model detects diverse categories of IoT malwares, creating security threats and challenges to IoT systems. MQTT (Message Queuing Telemetry Transport) is utilized to simulate behavior of IoT devices. Attack and normal traffic is passed to the IoT network to record latency and throughput in both scenarios. Performance of five different classifiers is compared with the proposed algorithm and had outstanding results with an accuracy 99.98%. The detection model takes around 52.89 seconds and processes 3965.27 samples per second. Utilization of CPU during the entire process is observed around 3.5%. Future research directions and suggestions have also been given to enhance the security of IoT environment.

Large language models are capable of generating source code in a zero-shot manner to develop programs that meet user functional requirements. However, when faced with scenarios involving complex business requirements, the generated source code may fail to satisfy user needs. Addressing the challenge of understanding software requirements, we propose a two-stage code generation approach. Initially, the large language model generates pseudocode based on the user’s functional requirements, refined through an iterative process with user feedback. Subsequently, the model generates source code based on the finalized pseudocode. We conducted empirical studies on an open code generation dataset, and experimental results with models such as GPT-4, Claude Sonnect 3, and Geminipro 1.5 demonstrate that our method outperforms zero-shot prompt learning in scenarios with complex user requirements, with improvements in PASS@K reaching up to 15%.