Pakistan agricultural sector forms the backbone of its economy, with wheat being the staple crop, processed predominantly through small-scale flour grinders (Aata Chakkis). Alongside wheat, other grains such as maize (corn), barley, millet, rice, and chickpeas are also ground in these mills for daily consumption. However, these mills consume significant electricity, increasing operational costs and pressuring Pakistans fragile power grid. With rising power costs and frequent outages, mill owners are adopting solar-powered solutions. This study explores how Variable Frequency Drives (VFDs) can optimize the efficiency of solar-powered flour grinders by providing grain-specific motor speed control, reducing electricity consumption, and maintaining [...]

Web application security reports are usually ordered by technical severity but low-resourced organizations want to know which vulnerability will become the most expensive security debt to repay? This paper presents SECURE-DEBT a model used to rank the 2025 OWASP web application risks to prioritize the remediation of web applications in resource-constrained healthcare-style organizations and small to medium enterprises (SMEs)? SECURE-DEBT considers the unresolved vulnerabilities as security debt thus incurring future risk based on evidence, exposure, sensitive information, exploitability, operational impact and time considerations for a fix? The study follows an empirically-driven, framework-based conceptual research article? The scoring considers mapping to [...]

Upper-limb amputations are a significant source of disability, especially in resource-limited environments where expensive commercial-grade prosthetics (often costing over USD 70,000) are unaffordable. This paper describes the design, optimisation and pilot testing of a 7-degree of freedom (DOF) transhumeral motorised prosthesis controlled by TinyML-based myoelectric decoding. The gesture-recognition accuracy (F1-score) of 95.5% across five able-bodied participants (gesture-level accuracy: approximately 92.3% for the most difficult gesture) with an end-to-end response time of 100ms is achieved using an affordable ESP32 microcontroller and a Cost-Complexity Pruned (CCP) Random Forest (RF) classifier. This paper outlines the hardware-software co-design, mathematical formulation of [...]

This paper presents a multilingual spam email classification framework capable of processing both German and English texts using classical machine learning models combined with a translation-enhanced pipeline. The system integrates TF-IDF vectorization, Logistic Regression, Multinomial Naive Bayes, and a probability-calibrated Linear Support Vector Classifier (SVC). A key contribution is the incorporation of an automatic German-to-English translation layer using MarianMT, allowing cross-lingual evaluation and robustness analysis. Multiple models are trained on a manually curated dataset of 3,790 emails, achieving up to 98.55% accuracy and 0.9860 AUC-ROC on the evaluated dataset. A Streamlit-based application is implemented for real-time inference, [...]

Transformer architectures have become a central foundation of modern artificial intelligence because of their scalability, parallel computation, and representation learning across language, vision, code, and multimodal domains. Since the original Transformer, the architecture has evolved into encoder-only, decoder-only, encoder-decoder, efficient attention, vision, multimodal, sparse Mixture-of-Experts, open-weight, state-space, recurrent, and hybrid model families. This review presents a structured and benchmark-informed comparison of Transformer architectures using a defined literature selection protocol, inclusion and exclusion criteria, model-family coding dimensions, and source-verification rules. The paper compares major architectures with respect to training objective, attention or sequence-modeling mechanism, computational complexity, scalability, context length, modality support, [...]