Pacific States University

Tea cultivation holds significant economic and agricultural value across many regions, particularly in developing countries, where it supports millions of livelihoods. However, the productivity and quality of tea leaves are severely threatened by a wide range of foliar diseases, which, if undetected or misdiagnosed, can lead to substantial crop losses. Existing manual and traditional image-based detection methods are time-consuming and often lack the precision required for early-stage identification across multiple disease types. To address these limitations, this study proposes TeaLeafNet-GWO, an intelligent deep learning framework that integrates a hybrid CNN–Transformer architecture with Gray Wolf Optimization (GWO) for automated tea leaf disease classification. The model combines the spatial feature extraction power of CNNs with the global contextual awareness of Transformer blocks (MobileViT), enhanced by channel-wise recalibration using a squeeze-and-excitation network (SENet). GWO is employed to fine-tune critical hyperparameters, ensuring optimal performance and generalization. The dataset comprises 999 high-resolution images representing eight distinct tea leaf conditions, including both healthy and diseased classes. Preprocessing techniques such as contrast-limited adaptive histogram equalization (CLAHE) and median filtering were applied to enhance image quality and suppress noise. The model was trained and evaluated using a stratified split (70% training, 30% testing). TeaLeafNet-GWO achieved an average accuracy of 97.72%, F1-score of 86.66%, and Matthews correlation coefficient (MCC) of 84.44% in the training phase, with similarly high performance in testing, outperforming several state-of-the-art hybrid models in comparative analysis. These results demonstrate TeaLeafNet-GWO’s effectiveness as a lightweight, scalable, and accurate solution for real-time disease monitoring, offering tangible benefits to precision agriculture and smart farming systems.

Abstract This chapter explores the design and implementation of AI course assistants as a strategy to increase equity, engagement, and academic success in online higher education. At Los Angeles Pacific University, a fully online institution serving adult and historically underserved learners, an AI assistant named Spark was deployed across all courses within four weeks following a successful pilot study. The implementation process included a multi-phase rollout, beginning with faculty-led dialog and experimentation, followed by the development of a minimum viable product and continuous improvement based on student and faculty feedback. Spark was designed to support student learning through Socratic questioning, course-specific alignment, and inclusive interaction models that accommodated diverse learners, including those with language or neurodiversity needs. Outcomes from a randomized controlled trial and follow-up studies indicated improvements in GPA, motivation, and self-efficacy, with no demographic disparities in benefit. The chapter also examines how Spark was used to enable active learning through tools such as AI-facilitated think–pair–share discussions and role-playing simulations. Ethical and inclusive design practices were prioritized throughout the development process, and faith-informed considerations were integrated to reflect the institution’s Christian values. Finally, the chapter connects the case study to global priorities in education by aligning outcomes with the goals of inclusive, high-quality learning outlined in Sustainable Development Goal 4. Practical recommendations are provided to help institutions of all types responsibly scale AI tools that support learner success while maintaining human-centered values.

Automated and accurate classification of kidney diseases from medical imaging is crucial for early diagnosis and treatment planning. This study introduces a novel Swin Transformer V2-based hybrid deep learning model designed for classifying kidney diseases (Cyst, Normal, Stone, Tumor) using CT imaging. Our approach incorporates Multi-scale Feature Fusion (MFF) blocks for robust feature extraction, Parallel Convolutional Attention (PCA) modules to enhance spatial hierarchy representation, and Capsule Networks for refined attention-based classification. Evaluated on a comprehensive dataset of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{1 2, 4 4 6}$</tex> expert-verified CT images, the proposed hybrid model achieved state-of-the-art performance with an accuracy of 100%, an AUC score of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{1. 0 0}$</tex> across all classes, and a Dice Similarity Coefficient (DSC) of 0.99. These results substantially surpass traditional CNN-based methods, highlighting the effectiveness and potential clinical applicability of our advanced hybrid architecture.

The integration of artificial intelligence (AI) in education has demonstrated significant potential in enhancing learning experiences, yet many students choose to not engage with AI tools available within their courses. This mixed-methods exploratory study investigates the barriers to the adoption of Spark, the AI course assistant implemented at Los Angeles Pacific University (LAPU). Analyzing 602 student survey responses collected through End-of-Course (EOC) evaluations, this study investigates the reasons behind this non-utilization, utilizing combined sentiment analysis with thematic analysis. The findings reveal that perceived necessity, lack of familiarity, and minimal awareness of Spark are primary factors contributing to non-use, alongside external issues such as time constraints and preference for traditional study methods. Notably, the absence of negative sentiment suggests that resistance is not rooted in dissatisfaction with the tool itself, but in limited perceived value or awareness. The study highlights the critical role of communication strategies, faculty engagement, and integration of AI tools into coursework to improve adoption. Addressing these barriers through improved communication, faculty-led AI integration, and targeted training initiatives can significantly enhance student engagement and learning outcomes. Without intervention, non-utilization may limit students’ opportunity to build foundational AI literacy—an increasingly important competency in academic and professional environments shaped by AI technologies. These insights offer actionable guidance for educators and institutions seeking to maximize the benefits of AI powered learning tools, and foster AI literacy among students in the evolving technological landscape.

Abstract Oral cancer poses a critical global health challenge, with early detection significantly improving patient survival rates and treatment outcomes. This study proposes an advanced deep learning-based diagnostic model, LightSE-MobileViT, specifically designed to classify oral cancer using medical imaging. The Oral Cancer Classification dataset used in this study comprises clinically validated lip and tongue images collected from various ENT hospitals in Ahmedabad. The original dataset consisted of 131 images (87 cancerous and 44 non-cancerous). To address class imbalance and enhance model generalizability, data augmentation techniques were employed, expanding the dataset to 981 images with equal distribution across both classes. Our proposed model, LightSE-MobileViT, integrates a lightweight convolutional neural network (CNN) backbone consisting of sequential convolutional layers enhanced with batch normalization and rectified linear unit activations. To further enrich feature representation and spatial attention, a Squeeze-and-Excitation block is embedded after the third convolutional layer. Subsequently, a MobileViT transformer encoder is employed, effectively capturing global contextual information through efficient multi-headed self-attention mechanisms. Experimental evaluations revealed that LightSE-MobileViT achieved superior diagnostic performance, attaining an accuracy of 98.39%, precision and recall values approaching 1.00 for both cancerous and non-cancerous categories, a macro F1-score of 0.98, and an ROC-AUC of 1.00. Comparative analysis demonstrated notable improvements over benchmark models, including CST-CNN (98% accuracy), MobileNetV2 (97% accuracy), DenseNet121 (97% accuracy), and InceptionV3 (90% accuracy). The exceptional performance of LightSE-MobileViT underscores its robust capability and clinical applicability, suggesting significant potential for deployment in automated oral cancer screening, thus facilitating early detection and timely intervention.

Being formed in accordance with the peculiarities of the geographical, climatic, historical and economic conditions of their origin and territorial growth, such cities have a special morphological organization, which allows us to consider the urban planning specifics of the Far Eastern urban “twins”: Fuyuan (Khabarovsk), Zhaohe (Pokrovka–Bikin), Heihe (Blagoveshchensk), identifying the stages of changing their layout, corresponding to the level of economic development and cultural exchange in coastal settlements located along the demarcation line of China and Russ Based on field studies carried out with the help of a structural analysis of a system object, which is a Chinese border city, the morphological organization of its space is considered and the features of the formation process of a cross-border Russian-Chinese agglomeration are revealed.The use of cartographic and statistical analysis of planning solutions made it possible to analyze the details of the trade infrastructure that determines the formation of border paired cities. Consideration of the trade activity of border towns as a factor conceptually significant for identifying their compositional patterns made it possible to clarify the periodization of the historical transformation of the urban structure of settlements organizing cross-border linear residential construction facilities.

The methodical approach to the integrated assessment of the ecological state of urban environment (on the example of Khabarovsk), oriented on engineering-ecological surveys and consisting in the application of the method of normalization and weighting with the determination of weighting coefficients by means of expert evaluations, was proposed. 16 indicators characterizing the ecological state of the natural environment components and physical impacts, united into groups, were identified. Weighting coefficients for individual factors and their groups were established. The use of this approach in assessing the ecological situation of the territory of the Khabarovsk administrative districts was shown and its suitability for various urbanized areas was noted.

The article presents a new method for molding products from epoxy resin. The technical problem to be solved by the invention developed by the authors of the article is to reduce the duration of the product formation cycle. In this work, the problem is solved by nanopulse treatment of epoxy resin during the curing process.

The rapid advancement of technology in the 21st century has brought intellectual property [IP] to the forefront as a critical asset in the tech industry.Licensing agreements, essential for fostering innovation and collaboration, face significant challenges in efficiency, transparency, and enforcement.Traditional licensing processes are often plagued by inefficiencies, including lengthy negotiations, manual documentation, and disputes arising from ambiguous terms.This paper explores the transformative potential of artificial intelligence [AI] and blockchain-based smart contracts in automating and enforcing IP licensing agreements.Smart contracts, programmable agreements executed on blockchain platforms, offer unprecedented opportunities for ensuring transparency, reducing administrative burdens, and automating royalty payments.When integrated with AI, these contracts can analyse vast datasets, predict licensing trends, and provide tailored recommendations, enabling more informed decision-making for licensors and licensees.Furthermore, the immutable nature of blockchain enhances trust by recording transactions securely and transparently, reducing the likelihood of disputes and fraud.This study also addresses challenges associated with adopting these technologies, including interoperability, scalability, legal implications, and ethical considerations in data use.It presents case studies from the tech industry, illustrating successful implementations of AI-powered smart contracts for IP licensing.By combining AI's predictive capabilities with blockchain's transparency and automation, companies can create robust frameworks for managing complex licensing agreements.The paper concludes that integrating AI and blockchain in IP licensing has the potential to revolutionize the tech industry by enhancing efficiency, fostering trust, and driving innovation.However, strategic implementation and stakeholder collaboration are crucial for realizing these benefits.