International Journal of AI and Machine Learning Innovations in Electronics and Communication Technology

Generative Artificial Intelligence for Visual Applications: Architectures, Applications, and Challenges

2026-01-15T05:02:47+00:00

Generative artificial intelligence (generative AI) represents one of the most transformative advances in modern computing, especially in the domain of visual applications. Its ability to generate, reconstruct, and enhance visual content has redefined the boundaries of creativity, automation, and perception. This study presents a systematic review of existing studies on generative AI in visual domains, focusing on three dominant architectures—Generative adversarial networks (GANs), variational autoencoders (VAEs), and diffusion models. The methodology involves a structured literature search across major databases, screening studies using defined inclusion criteria, and synthesizing results into thematic insights. The review identifies core technical principles, major datasets, evaluation metrics, and application areas such as image synthesis, video generation, and 3D content creation. Moreover, it discusses ongoing challenges related to reproducibility, computational cost, interpretability, and ethical considerations, including bias and misinformation. The paper concludes with emerging trends such as controllable generation, multimodal fusion, and sustainability-oriented generative modeling, aiming to guide future research toward responsible and transparent visual AI.

Explainable AI in Legal Text Analysis: A Comprehensive Review of Legal Threat Assessment Approaches

2026-04-22T12:15:02+00:00

The volume of legal documents continues to expand rapidly, and their language is becoming increasingly intricate, making manual review both time-intensive and vulnerable to individual bias and inconsistency. As a result, recent research has increasingly focused on the use of machine learning techniques to predict threat levels and evaluate legal risks directly from textual data. Although such models have demonstrated promising predictive performance, a major limitation remains the lack of transparency in their decision-making processes. In legal practice, where accountability and justification are essential, predictions without clear explanations are difficult to accept and apply. This review examines studies published between 2018 and 2025 addressing automated legal document analysis, with particular emphasis on methods that combine accurate threat prediction with explainability. The reviewed works cover a broad range of approaches, including natural language processing, deep learning, transformer-based architectures, graph-based models, large language models, and contract analysis techniques. Explainable artificial intelligence methods such as LIME, SHAP, and attention-based visualization are examined for their role in making model predictions transparent and legally meaningful. Recurring design patterns observed across the reviewed systems are synthesized into a conceptual framework of six sequential stages covering input, text processing, feature encoding, threat classification, explainability, and output. The analysis identifies dataset scarcity, inconsistent evaluation practices, and limited generalization across legal domains as the primary barriers to practical deployment. The review concludes that explainability is a foundational requirement rather than an optional feature for any AI system intended to support legal decision-making.

Artificial Intelligence-driven Optimization of Electronic Device Performance and Energy Efficiency

2026-04-27T12:06:17+00:00

Artificial intelligence (AI) is transforming the energy efficiency landscape of modern electronic systems by enabling intelligent optimization across algorithms, hardware architectures, and mixed‑signal interfaces. Through predictive workload management, adaptive task scheduling, and reinforcement‑learning‑based control policies, AI reduces redundant computation and data movement, which are the two dominant contributors to power consumption in mobile, IoT, and embedded platforms. Complementary algorithmic advances, including model compression and lightweight neural architectures, further minimize resource usage while preserving performance. At the hardware level, innovations such as edge AI accelerators, memory‑centric and compute‑in‑memory architectures, and adaptive memristor‑based sensing pipelines improve performance per watt by tightly integrating computation with storage and signal acquisition. System‑level intelligence additionally enables coordinated energy governance in distributed electronics, supporting efficient load balancing, renewable‑energy interaction, and improved responsiveness in smart‑grid and cyber‑physical environments. Despite these advances, critical challenges persist, including reproducible energy‑measurement standards, privacy‑preserving telemetry for device analytics, robust verification of mixed‑signal AI circuits, and the absence of unified cross‑layer co‑design methodologies. Addressing these gaps is essential for the development of next‑generation electronic systems capable of delivering high‑performance intelligence while ensuring sustainable, energy‑responsible operation.

Predictive Shielding: A Machine Learning Framework for EMI Mitigation in High-speed PCB Design

2026-05-02T10:41:24+00:00

Radiated Electromagnetic Interference (EMI) remains a critical bottleneck in high-speed PCB design, driving late-stage respins, prolonged time-to-market, and high EMC testing costs due to the limits of rule-based heuristics and compute-intensive full-wave simulations. This work develops and evaluates an AI-augmented Electronic Design Automation (EDA) framework that elevates EMI compliance from reactive verification to proactive, real-time optimization across the PCB lifecycle. A hybrid dataset was assembled comprising 71.4% HFSS and SIwave simulations, 26.8% physics-informed GAN and PINN synthetic samples, and 1.8% EMC chamber measurements. Multimodal representations, including tabular features, 2D pseudo-image maps, and connectivity graphs, were modeled using an ensemble of gradient boosting methods, convolutional neural networks, and graph neural networks. The framework integrates task-aware model selection, an end-to-end pipeline, layered ecosystem architecture, and continuous learning to enable predictive EMI assessment, hotspot detection, coupling-path analysis, and automated mitigation guidance within commercial EDA tools, including Cadence, Altium, Mentor, and ANSYS. The study introduces a production-oriented AI-EDA ecosystem that unifies physics-constrained data augmentation, multimodal fusion, interpretable ensemble modeling, real-time tool integration, and closed-loop learning for EMI-aware optimization. Gradient boosting ensembles achieved the highest suitability scores between 0.70 and 0.79 with strong interpretability and efficiency, while lightweight ensembles improved topology- and geometry-sensitive tasks by 4 to 7%. The AI-driven workflow reduced lifecycle cost by about 69%, shortened time-to-market from 24 to 9 weeks, increased first-pass success beyond 80%, and projected a five-year ROI above 1100% with payback under five months. The framework demonstrates machine learning as a deployable, high-ROI enabler of right-first-time high-speed PCB compliance and motivates industry adoption toward system-level, physics-informed, continuously learning design flows.

AI-driven Adaptive Human-Computer Interaction Models for Smart Control Systems

2026-05-07T05:03:22+00:00

The rapid evolution of intelligent systems and digital environments has significantly increased the complexity of Human-Computer Interaction (HCI), particularly in smart control systems where adaptability, real-time responsiveness, and decision support are critical. Traditional static HCI models remain inadequate for addressing dynamic contexts, user variability, and data-intensive environments. This study proposes an advanced AI-driven adaptive HCI model designed to enhance interaction efficiency, usability, and system intelligence. The proposed framework integrates artificial intelligence, machine learning, the Internet of Things, and big data analytics to dynamically adapt system interfaces and control mechanisms based on user behavior, environmental conditions, and system performance. A hybrid methodology combining conceptual modeling, simulation-based experimentation, and statistical analysis is employed to evaluate the effectiveness of the proposed approach. The results demonstrate significant improvements across key performance indicators, including a 64% reduction in response time, a 23% increase in user satisfaction, and notable gains in decision accuracy and system efficiency. In addition, the adaptive system reduces cognitive load and improves user engagement, confirming its effectiveness in enhancing both technical performance and user experience. This study contributes to the advancement of intelligent adaptive systems by providing a scalable and integrative HCI framework capable of supporting real-time, context-aware interaction. The proposed model has practical applicability in smart cities, industrial automation, healthcare systems, and cyber-physical environments, thereby offering a robust foundation for next-generation intelligent control systems.