LatIA (Journal)
Not a member yet
134 research outputs found
Sort by
Systematic Review on the Application of Nanotechnology and Artificial Intelligence in Agricultural Economics
Get PDFThe convergence of nanotechnology and artificial intelligence (AI) represents a transformative force in agricultural economics, offering innovative solutions to longstanding challenges such as productivity inefficiencies, environmental degradation, and unsustainable resource use. This study presents a systematic literature review (SLR) aimed at synthesising theoretical frameworks, applications, and economic implications associated with these technologies in agriculture. A structured search strategy was developed using Boolean operators to combine key terms related to nanotechnology, AI, and machine learning. Comprehensive searches were conducted across six academic databases—Springer, IEEE Xplore, ACM, Science Direct, Wiley, and Google Scholar—complemented by manual and snowballing techniques. From an initial pool of 840 records, 55 studies met the inclusion criteria after rigorous screening and eligibility assessment. Findings indicate that nanotechnology enhances nutrient delivery, pest control, and crop monitoring through nanosensors and nano-fertilisers, while AI facilitates data-driven decision-making, yield prediction, and resource optimisation in precision farming. Despite promising results, challenges such as high initial investment, technological complexity, and limited access for smallholder farmers remain significant. The review concludes that the integration of nanotechnology and AI can improve agricultural efficiency, economic viability, and environmental sustainability. However, targeted investments, capacity-building, and interdisciplinary collaboration are essential to bridge the gap between innovation and implementation in developing economies
Ethical AI for Personalized Banking: Addressing Bias and Fairness Challenges
Get PDFIntroduction: The integration of Artificial Intelligence (AI) into personalized banking has enhanced service delivery in areas such as loan processing, credit assessment, and fraud detection. Despite these advancements, ethical concerns, especially algorithmic bias and lack of fairness, pose significant challenges. This study addresses the need for equitable AI systems that promote transparency, fairness, and regulatory compliance in the banking sector.
Objective: This study aims to develop and implement a comprehensive framework for integrating ethical principles into AI-driven banking systems, with a focus on mitigating algorithmic bias, enhancing fairness, and improving transparency in personalized banking services.
Methods: A comprehensive methodology is proposed that integrates bias-aware data collection, fairness-constrained machine-learning models, and explainable AI (XAI) techniques. Tools such as Shapley Additive Explanations (SHAPs) and Local Interpretable Model-Agnostic Explanations (LIMEs) are applied to interpret model outputs. Adversarial debiasing and fairness-aware learning algorithms were employed to identify and mitigate systemic biases in financial data. Alternative data sources, including utility and rental payment histories, were incorporated to enhance inclusivity.
Results: The implementation of the proposed framework demonstrates improved fairness in decision-making without significantly compromising model accuracy. Bias metrics show measurable reductions in disparate impacts across the demographic groups. Explainability tools enhance transparency, enabling a more transparent communication of AI decisions to both users and regulators.
Conclusions: Embedding ethical principles into AI-driven banking systems is critical to ensuring fairness, regulatory alignment, and public trust. The structured framework presented in this study supports the development of responsible AI systems to mitigate bias, enhance explainability, and foster financial inclusion. This approach serves as the foundation for building equitable and accountable AI applications in modern banking
Research Trends and Collaboration Patterns in Quantum Internet: A Bibliometric Mapping Using Biblioshiny, VOSviewer, and CiteSpace
Get PDFIntroductionQuantum Internet has emerged as a transformative technological frontier, attracting increasing global research attention. Its development spans foundational physics, cryptography, network engineering, and applied technologies, demanding systematic evaluation of scholarly progress.
ObjectiveThis study aims to present a comprehensive bibliometric analysis of Quantum Internet research, identifying publication trends, influential contributors, collaborative networks, and evolving thematic directions.
MethodData were retrieved from the Scopus database and analyzed using Biblioshiny, VOSviewer, and CiteSpace. The study examined publication growth, author and journal influence, country-level collaborations, bibliographic coupling, keyword co-occurrence, and thematic evolution.
ResultsFindings indicate steady growth in publications since 2016, peaking in 2024 with 130 documents. Leading contributors include Angela Sara Cacciapuoti, Marcello Caleffi, Laszlo Gyongyosi, and Stephanie Wehner, while prominent publication outlets are Physical Review Letters, Physical Review A, Scientific Reports, and New Journal of Physics. Country-level analysis highlights the United States, China, the United Kingdom, Italy, and Japan as primary research hubs, with strong international collaborations. Network analyses reveal 15 author and 10 journal clusters, underscoring interdisciplinary connections. Bibliographic coupling and keyword co-occurrence identify critical themes such as quantum key distribution, entanglement distribution, quantum memory, blockchain, and cybersecurity, alongside emerging areas like quantum IoT and quantum machine learning. Thematic maps show a shift from theoretical constructs to application-driven studies, integrating artificial intelligence and advanced sensing. Trend analyses confirm growing attention to scalability, security, and interdisciplinary applications, though gaps remain in resource optimization, experimental validation, and integration with classical infrastructures.
ConclusionsThis analysis provides a knowledge framework and practical insights into the intellectual and conceptual structure of Quantum Internet research. By mapping influential contributors, core themes, and research gaps, the study supports academics, policymakers, and industry in making strategic investments and advancing future research in this rapidly evolving field
Challenges in Sub-Saharan Africa’s Food Systems and the Potential Role of AI
Get PDFSub-Saharan Africa (SSA) faces persistent food insecurity due to low agricultural productivity, limited access to modern technologies, and growing climate variability. This study explores the transformative potential of Artificial Intelligence (AI) to enhance food systems across SSA. The objective is to assess how AI applications—such as machine learning, remote sensing, and big data analytics—can address systemic inefficiencies in cereal crop production, with a focus on barley, millet, and sorghum. Using a systematic review approach aligned with PRISMA guidelines, literature from 2015–2025 was analyzed across multiple databases to identify empirical studies and models related to AI in SSA agriculture. Results reveal that AI can significantly improve crop monitoring, yield forecasting, and resource optimization. However, adoption barriers such as inadequate infrastructure, financial constraints, and the digital divide persist. The study concludes that while AI holds significant promise, its success in SSA depends on inclusive policies, capacity building, and localized data governance. It recommends interdisciplinary research, investment in rural digital infrastructure, and participatory innovation frameworks to empower smallholder farmers and ensure equitable AI deployment. This review provides a roadmap for integrating AI into SSA food systems to enhance resilience, productivity, and food security
Integrating AI-Based Natural Language Processing in Vocational Education: Usability, Learning Gains, and Student Engagement in Indonesia
Get PDFIntroduction: The advancement of Artificial Intelligence (AI) has brought substantial changes to education, particularly through AI-based digital assistants.Objective: This study developed and evaluated an AI-powered digital assistant equipped with Natural Language Processing (NLP) capabilities, specifically designed for Indonesian vocational schools.Methods: Adopting the 4D development model (Define, Design, Develop, Disseminate), the system was created using machine learning algorithms and NLP to enhance interactivity and personalization. The assistant enables natural language interaction, provides real-time feedback, and adapts learning material difficulty to students’ comprehension levels. The system was tested with 100 vocational school students, with usability assessed using the System Usability Scale (SUS) and learning gains measured through pre- and post-tests.Results: Results showed a SUS score of 71.05, indicating good usability, and a significant improvement in post-test scores compared to pre-test scores (p < 0.001), reflecting enhanced conceptual understanding, engagement, and motivation.Conclusions: These findings demonstrate the potential of AI-powered NLP assistants to enrich vocational education and prepare students for technology-driven industrial demands.
Optimizing Resource Discovery in Grid Computing: A Hierarchical and Weighted Approach with Behavioral Modeling
Get PDFParallel programs that require sizeable computational electricity increasingly depend on grid computing structures. Efficient, helpful resource discovery algorithms are critical for optimizing resource allocation and minimizing execution time in these environments. This look presents a unique hierarchical and weighted resource discovery algorithm designed to decorate resource distribution while decreasing communique costs among grid nodes. A behavioural modelling technique systematically establishes the weighted resource discovery algorithm\u27s accuracy and effectiveness. The behavioural model is carried out using StarUML. At the same time, the NuSMV version checker is hired to verify essential residences along with reachability, equity, and impasse-free operation of the resource discovery procedure. Critical overall performance metrics, including the quantity of inspected nodes consistent with request and the frequency of re-discovery operations, are used to evaluate the rules\u27 efficiency and flexibility.
The weighted resource discovery algorithm also evaluates the efficiency of finding loose resources with high-bandwidth connections, optimizing overall grid resource allocation. In addition to enhancing resource localization, the observation introduces resource facts tables, which store information crucial for powerful, proper resource allocation. This study aims to develop grid computing competencies by addressing resource discovery challenges. The hierarchical shape and weighted valid resource selection decorate valid resource inspection, adaptability, and high-bandwidth utilization. Behavioural modelling and formal verification verify the algorithm\u27s accuracy and applicability in grid environments. By using weighted resource discovery and resource information tables, this study drastically improves resource discovery\u27s performance and effectiveness in grid computing, optimizing overall performance and proper resource allocation
Evaluating AI and Machine Learning Models in Breast Cancer Detection: A Review of Convolutional Neural Networks (CNN) and Global Research Trends
Get PDFNumerous studies have highlighted the significance of artificial intelligence (AI) in breast cancer diagnosis. However, systematic reviews of AI applications in this field often lack cohesion, with each study adopting a unique approach. The aim of this study is to provide a detailed examination of AI\u27s role in breast cancer diagnosis through citation analysis, helping to categorize the key areas that attract academic attention. It also includes a thematic analysis to identify the specific research topics within each category. A total of 30,200 studies related to breast cancer and AI, published between 2015 and 2024, were sourced from databases such as IEEE, Scopus, PubMed, Springer, and Google Scholar. After applying inclusion and exclusion criteria, 32 relevant studies were identified. Most of these studies utilized classification models for breast cancer prediction, with high accuracy being the most commonly reported performance metric. Convolutional Neural Networks (CNN) emerged as the preferred model in many studies. The findings indicate that both the quantity and quality of AI-based algorithms in breast cancer diagnosis are increases in the given years. AI is increasingly seen as a complement to healthcare sector and clinical expertise, with the target of enhancing the accessibility and affordability of quality healthcare worldwide
Artificial Intelligence in Perovskite-Based Materials for Energy Applications
Get PDFIntroduction; Perovskite-based materials have gained significant attention in energy applications due to their remarkable optoelectronic properties and versatile composition. These materials, characterized by their ABX₃ crystal structure, have demonstrated high efficiencies in solar cells, light-emitting diodes (LEDs), and potential in energy storage systems. Objective; Perovskite solar cells (PSCs) have achieved efficiencies comparable to silicon-based cells, with advantages in cost and fabrication flexibility. Method; A literature review was conducted, including original articles, reviews, and bibliometric studies. The research focused on AI in Perovskite-Based Materials for Energy Applications.Result; AI is driving significant advancements in the field of perovskite-based materials for energy applications.Conclusion; Perovskite LEDs offer high color purity and tunable emission, making them ideal for display technologies. Despite challenges like stability and scalability, ongoing research aims to enhance their performance, positioning perovskites as key materials in sustainable energy technologies. By accelerating material discovery, optimizing manufacturing processes, enhancing stability and performance, and promoting sustainabilit
Analysis of Cyberbullying Behaviors Using Machine Learning:A Study on Text Classification
Get PDFIntroduction:Cyberbullying is a significant concern in today\u27s digital age, affecting individuals across various demographics. Objective: This study aims to analyze and classify instances of cyberbullying using a dataset sourced from Kaggle, containing text data labeled for different types of bullying behaviors. Method: Our approach to tackling these challenges involves several key steps, starting with data preprocessing and feature extraction to identify patterns and improve detection methods, enhancing our understanding of how cyberbullying manifests in online communications.Result: The dataset provides a valuable resource for developing and evaluating machine learning models aimed at detecting sexist and racist content in tweets.Conclusion: This study advances the current understanding of the complexities involved in detecting cyberbullying and paves the way for future breakthroughs in this domain. The binary classification enabled by the \u27oh_label\u27 column streamlines the analysis process, making it particularly compatible with binary classification model
A Machine Learning Model for Diagnosis and Differentiation of Schizophrenia, Bipolar Disorder and Borderline Personality Disorder
Get PDFSchizophrenia, bipolar disorder, and borderline personality disorder present overlapping symptoms, complicating accurate diagnosis. Misdiagnosis leads to inappropriate treatment, increased patient distress, and higher healthcare burdens. This study develops a machine learning model integrating clinical, neuroimaging, and behavioral data to improve diagnostic accuracy. The model utilizes Convolutional Neural Networks (CNNs) for neuroimaging, Gradient Boosting Machines (GBMs) for structured clinical and behavioral data, and Recurrent Neural Networks (RNNs) for speech analysis. The combined model demonstrated superior accuracy (94.1%) compared to individual models. SHAP analysis identified key diagnostic features, including specific brain regions, cognitive measures, and speech patterns. External validation confirmed robustness, highlighting the model’s potential as a clinical decision-support tool. Future research should focus on enhancing model interpretability and real-time diagnostic support