From the 1980s to 2024, modern ceramics in China has undergone significant transformations shaped by cultural policy reforms, global artistic exchange, and shifts in artistic concepts and production practices. Existing studies have largely focused on individual artists, stylistic evolution, or specific art movements, while fewer have examined how broader social, institutional, and policy-related forces have structured the historical development of modern ceramics over time. This study adopts a timeline analysis to map the development of modern ceramics in China from the 1980s to 2024, identifying key milestone events and developmental stages. Drawing on qualitative document analysis and representative case studies, the research examines how cultural policy, international exchange, and evolving artistic paradigms influenced the transformation of ceramics from a primarily craft-based practice to a recognized contemporary art medium. The analysis organizes this historical process into four major stages—early exploration (1980s), institutional formation (1990s), diversification and independent development (2000s), and globalization and technological expansion (2010s–2024). The study employs a qualitative timeline-based document analysis and representative case studies to examine how cultural policy, institutional development, and global artistic exchange shaped modern ceramics in China. In this framework, ‘institutional formation’ refers to the development of ceramic education systems, exhibition platforms, and academic structures during the 1990s, while ‘globalization and technological expansion’ refers to the increasing influence of international exchange, cross-media practices, digital technologies, and global exhibition networks from the 2010s onward. By situating modern ceramics within the contexts of cultural policy and global exchange, this study provides a structured understanding of how modern ceramics in China has been shaped by broader socio-cultural conditions and contributes to socialscience discussions on cultural production and the globalization of contemporary art in China.
This study investigates the structural behavior and fatigue performance of a titanium bicycle frame subjected to a loading condition specified by the ISO 4210 standard. A finite element (FE) model of the frame was developed and validated through experimental static tensile testing. The FE model was then used to evaluate the fatigue strength of the frame using the Sines criterion. For each FE node, the hydrostatic stress associated with the mean load in the cycle and the von Mises stress corresponding to the load amplitude were combined. The analysis revealed high stress levels in the top tube, down tube, and head tube regions. A reinforcement strategy based on increasing the tube thickness in these areas was assessed using FE analysis, demonstrating improved compliance with the fatigue requirements.
This paper presents the theoretical modeling of the microfluidic transport of immiscible, non-Newtonian fluids (specifically, Jeffrey and Casson fluids) through a vertical porous medium by considering magnetohydrodynamic (MHD), Hall current, electroosmotic, buoyancy, thermal radiation, and internal heat generation effects all simultaneously. Current literature in this area has primarily focused on Newtonian or single non-Newtonian fluid flows and has not addressed two distinct non-Newtonian fluid flow models in multiphase porous systems with Hall effects and electroosmosis. To accomplish this goal, governing continuity, momentum, and energy equations were formulated for each of the two fluids; the equations were reduced to non-dimensional form through appropriate scaling and analytical solutions for velocity and temperature distributions were obtained. The key findings of this study include that the velocity profile of both fluids is enhanced at high permeability and at high electroosmotic flow parameters and is reduced by the Lorentz force associated with a high magnetic field strength. Theoretical modeling at this level can provide important information on the dynamics of immiscible fluids that can be applied to the design and operation of microfluidic systems and technological applications associated with energy transport and thermal management for cooling systems.
The increasing prevalence of chronic illness, psychological stress, and environment degradation in the workplace has necessitated a comprehensive digital strategy for employee health. This research introduces a modular Augmented Reality (AR) framework, SEHAT 360°, TENANG 360°, and HIJAU 360°, designed to enhance physical, mental, and environmental well-being among employees. The project employs the ADDIE framework during its analysis and design phases and conducts surveys with 647 employees across five industries to ascertain health priorities and digital engagement preferences. Each module comprises interactive AR content, behavioural monitoring, gamification, and tailored feedback to encourage healthy behaviours and sustained engagement. The system is systematically designed using database schemas and entity-relationship diagrams to ensure scalability. It features a scalable architecture, aligned with ESG indicators and SDGs. It offers a novel, unified approach by combining physical, mental, and environmental health education through immersive AR, an integration not yet explored in existing workplace wellness applications.
Cloud computing environments are increasingly vulnerable to sophisticated cyber threats due to their distributed and dynamic nature. Traditional centralized intrusion detection systems suffer from scalability and privacy limitations, making them less effective in modern cloud infrastructures. To address these challenges, a Federated Learning-based Long Short-Term Memory (Fed-LSTM) framework is proposed for distributed and privacy-preserving intrusion detection. The framework enables collaborative model training across multiple nodes without sharing raw data, thereby ensuring data confidentiality. The model is evaluated using benchmark datasets, including CICIDS2017, NSL-KDD, and UNSW-NB15, which are distributed across federated nodes to capture heterogeneous network traffic patterns. Data preprocessing is performed using Min-Max normalization and entropy-based feature selection to improve efficiency and reduce dimensionality. Privacy is ensured through Differential Privacy and Secure Aggregation, while model aggregation is performed using Federated Averaging. Experimental results show that the proposed Fed-LSTM model achieves 95.1% detection accuracy, outperforming traditional machine learning and deep learning baseline methods in terms of accuracy and stability. The proposed approach enhances intrusion detection performance while preserving data privacy, making it suitable for deployment in distributed and dynamic cloud environments.
Arrhythmia is characterized by irregular heartbeats, often caused by abnormalities in heart rate or rhythm. ECG signals are essential for accurately detecting and classifying arrhythmias, aiding in identifying specific types for appropriate treatment. ECG signals contain sensitive health information, but none of the existing works have concentrated on securing the ECG signal processing in VLSI design. Therefore, this paper proposes Advanced Encryption Shamir's Secret Sharing (AE-3S) and Dynamic Convolutional Structured Sparsity Neural Network (DC2SNN) for both secure transmission and Arrhythmia detection. In the suggested process, Empirical Inter-Cluster Fusion Mode Decomposition (EICFMD) significantly boosts the signal quality through dynamic artifact suppression, whereas Northern Goshawk Quadratic Sine Optimization (NGQSO) fine-tunes the choice of features with the highest discrimination power, thus making AE-3S encryption and DC2SNN-based irregular heartbeat classification more efficient together. Initially, signal is converted from analog to digital and then undergoes preprocessing through various steps. Next, Pan-Tompkins Algorithm (PTA) is utilized for peak detection and a time series is constructed on the preprocessed signal. Subtle characteristics of the signal are also preserved. Features are then extracted from the detected peaks, constructed time series, and preserved characteristics. From these features, NGQSO and AE-3S are utilized for feature selection and information security. Finally, cardiac abnormality classification is performed using DC2SNN. Thus, the proposed model efficiently secures ECG signals and detects arrhythmias with a high accuracy of 99.15%. Additionally, the proposed framework integrates AI-driven natural language processing to enable secure multimodal interaction and real-time health communication in wearable IoT environments.
The combination of Black Swan Optimization (BSO) and Residual Networks (ResNet) forms a robust framework for detecting malicious packets within network traffic. ResNet's deep learning architecture learns intricate patterns through residual connections that mitigate the vanishing gradient problem, enabling effective training of deep models. BSO enhances this framework by optimizing feature selection, reducing redundancy, and improving detection accuracy through lower false positive (FP) and false negative (FN) rates. The resulting system supports two complementary deployment modes: An accuracy-oriented configuration, which leverages a richer optimized feature set and deeper residual learning for security-critical environments, and a latency-oriented configuration, which uses a lightweight feature subset to achieve faster execution and reduced computational overhead for real-time congestion-attack detection in resource-constrained 6G scenarios. This dual capability makes the BSO-ResNet integration suitable for both high-assurance cloud security and ultra-low-latency edge environments.
Reversible data hiding in encrypted images (RDH-EIs) faces challenges in embedding capacity and security. This paper proposes an enhanced RDH-EI framework that integrates secret sharing and convolutional neural networks (CNNs). The process begins by encrypting the image, followed by secret sharing, which splits the encrypted image into multiple spatially correlated shares. Data embedding is performed using a CNN, improving capacity while reducing distortion. The method ensures that the original image can be recovered without loss, even if some shares are missing or corrupted, provided enough uncorrupted shares are received. This approach is particularly useful in fields like medical imaging and secure cloud storage, where both privacy and data integrity are crucial. Experimental results show that the proposed method outperforms existing RDH-EI techniques in terms of security, data capacity, and reversibility, offering a robust solution for secure communication and storage.
In the proposed framework, Corollary De-swinging K-Anonymity (CDS-KA) ensures the secure registration and privacy preservation of Internet of Things (IoT) device details, while Bidirectional LeCun Aranda Long Short-Term Memory (BiLeCun-ALSTM) works in tandem within the Intrusion Detection System (IDS) to classify and predict potential attacks. The collaboration between these components enhances both privacy and security, ensuring efficient detection and protection of sensitive data in blockchain-enabled IoT devices. The integration of blockchain technology with IoT devices brings numerous benefits, such as transparency and data integrity. However, it also raises significant privacy concerns. Yet, none of the existing works concentrates on energy-efficient authorized block mining. Hence, this paper proposes an energy-efficient-aware, Authorized Block-Mining-based IDS (ABM-IDS) in blockchain-enabled IoT devices using CDS-KA and BiLeCun-ALSTM. Primarily, the IoT devices are registered using the device details, and then the details are preserved using CDS-KA. At the time of registration, keys and smart contracts are created. The solidity code is used to generate the smart contract. Then, the Merkle tree (MT) is created from the smart contract using GXNOR-BLAKE 512. Also, the solidity functions are split, followed by hash code generation. Then, the generated hash code is updated in the MT. Similarly, the optimal blocks are recognized from the hash code and also verified in the MT. Conversely, the user logs into the network, and then data sensing is done. Thereafter, the data are encrypted and then balanced via Edward Modulo Curve Cryptography (EMCC) and SCC-AZOA, respectively. Now, the balanced data is input to the IDS. In an IDS, the steps such as data collection, pre-processing, feature extraction, feature selection and classification are done. The proposed BiLeCun-ALSTM significantly predicts whether the data is attacked or not. Afterward, the non-attacked data is sent to the destination in a secure manner by verifying the blockchain. Collectively, the proposed framework obtained better security with an accuracy of 98.65%.
The integration of advanced technologies in healthcare has opened new pathways for improving disease detection and diagnosis. Among the others, Breast Cancer (BC) indeed is one of the biggest concerns worldwide, and its early and precise diagnosis is indispensable. The present paper gives an overview of the introduction of modern machinery, such as ML, DL, IoT, blockchain, cloud computing, and data mining, to aid the detection of BC. By systematically reviewing the contemporary literature, we point out the advances accomplished in the application of AI-assisted techniques in the monitoring and diagnosing of BC, whilst at the same time discussing the ongoing issues. Several ML algorithms are examined in detail, whereas the spotlight is on deep learning approaches such as CNNs and RNNs, which are known to be very effective in interpreting medical images. We examine the performance of models such as CNNs (both self-trained and those employing transfer learning), SPWO-based Deep Maxout networks, ShCNN (utilizing FACS features), and GRU-based RNNs across different publicly available datasets. Our findings aim to provide valuable insights for researchers and healthcare professionals by outlining current trends and evaluating the effectiveness of AI-based approaches. The review emphasizes the growing role of intelligent systems in supporting early diagnosis and improving treatment planning for breast cancer patients.
Recent advances in microfluidics have spurred significant innovations in thermal management technologies, particularly in the implementation of nanofluid-enhanced microchannel heat transfer. This study analyzes the thermodynamic irreversibilities in magnetohydrodynamic flow of couple stress nanofluids within an oblique microchannel integrated into a permeable substrate, incorporating thermal radiation effects. The governing equations, which model a microfluidic system featuring a nanofluid exhibiting microstructural effects and subjected to porous media, magnetic field, and radiative heat transfer, are rendered dimensionless and solved using the Hermite wavelet operational matrix method. Graphical analysis elucidates the parametric sensitivity of velocity, temperature, entropy generation and the Bejan number, providing insights into the system's thermodynamic performance and its optimization potential for microchannel heat transfer applications. The measures of irreversibility within the system, quantified by entropy generation and Bejan number, are diminished by an increase in the magnetic field intensity and medium's porosity. In contrast, an augmentation in the couple stress parameter and the Brinkman number enhances both the entropy production and the Bejan number. This study advances the understanding of entropy optimized thermal management in nanofluid-based magnetohydrodynamic microfluidic systems and also provides design insights for enhancing thermal performance and minimizing irreversibility.
Securing sensitive data in the era of digital transformation presents significant challenges, particularly in multimedia content. While watermarking has emerged as a solution, existing techniques often compromise on imperceptibility, robustness, or reversibility. This study presents a new framework for secure, non-destructive, and reversible watermarking that integrates sophisticated cryptographic techniques with auxiliary information to guarantee flawless reconstruction of the original media while ensuring robust security. The framework employs a two-layer approach: encryption for watermark concealment and auxiliary data to support reversibility. Performance metrics, such as peak signal-to-noise ratio (PSNR), structural similarity index (SSIM), bit error rate (BER), mean absolute error (MAE), and root mean square error (RMSE), were used to evaluate the effectiveness of the proposed system. Experimental results demonstrate the system's effectiveness, achieving a mean PSNR of 50 dB, SSIM of 0.98, and minimal perceptual distortion. The system is robust to attacks such as JPEG compression (50%), Gaussian noise (variance 0.01), and 10% cropping, with a low BER (mostly under 0.02). The framework operates efficiently, with embedding times (ETs) of just 0.25s for 512 & times;512 images, making it suitable for real-time applications. This study presents a dependable and secure watermarking approach, showcasing improvements in security, reversibility, durability, and computational efficiency. Future efforts seek to increase the framework's flexibility for intricate media, including audio and 3D content, while strengthening its ability to withstand new threats such as adversarial attacks and deepfakes.
Presently, an increase in the advent of digital age social networking activities leads to contentious issues, and social media is a strategic way of conducting activities in smart cities. This research aims to create an Integrative Political Structural Framework (IPSF) to analyze how people are engaged in contentious issues in a networked society of smart cities. The study focuses particular attention on the Contextual Concept of Problem-Solving (CCoPS), aggressive communication experience, emotional injustice, and the effectiveness of social networking to investigate the future of social media activism and offline activism in smart cities. Hence, an integrative political model is used to address the three areas, such as weapons possession, expatriation, and the usage of defense power in smart cities. The integrative activism framework gives a valuable perspective for understanding engaged communities and provides a framework for more theoretical conversations on activities and dispute resolution in smart cities. The validation of the suggested IPSF model through empirical means was done with the help of Structural Equation Modeling (SEM), which illustrated the combined effect of the communication, emotions, and digital platform factors on the activism behaviors.
Phishing attacks remain a significant challenge to cybersecurity, a hacker's attempt to deceive users into disclosing vital personal information by pretending to be on an authentic platform. This paper provides a novel phishing web site detection system wherein a convolutional neural network-recurrent neural network (CNN-RNN) is utilized in both detecting phishing web sites and detecting phishing email, respectively, with Sparrow Search Optimization Algorithm (SSOA) being used in optimizing the feature selection as well as the model parameters, respectively. The proposed system gathers phishing data. from the Phish Tank database and then does preprocessing steps such as data cleansing, tokenization, and normalization. The extraction features are used according to their CNN-RNN framework, which is sufficient in terms of capturing the spatial features of phishing web pages and the sequential features of phishing emails. In the optimization step, SSOA is used to tune the model more accurately and efficiently. Experimental results indicate that the CNN-RNN-based model attains 99.27% accuracy, 98.63% precision, 98.89% recall, and 99.12% F1-score, which are higher than other conventional methods. Compared with current schemes, including Region-based Convolutional Neural Network (RCNN), Loopy Belief Propagation (LBP), Random Forest (RF), FastText and CNN, and Support Vector Machine (SVM), the given approach always outperforms those approaches that gie better.
The chronic evolving neurodegenerative disorder Alzheimer's Disease (AD) presents with memory deficits, cognitive impairment, and loss of abilities. AD prevalence is increasing as the world ages, necessitating more precise and easily obtainable diagnostic and treatment approaches. Artificial Intelligence (AI) technologies, and more specifically, machine learning and deep learning, have become game-changers in Alzheimer's disease patient care, including optimizing care, enabling early diagnosis, supporting differential diagnosis, and predicting disease progression over the last few years. To detect amyloid plaques and hippocampal atrophy, this study investigates how AI-driven imaging analysis, specifically convolutional neural networks applied to MRI (Magnetic Resonance Imaging) and PET (Positron Emission Tomography) scans, provides higher sensitivity and specificity. Artificial intelligence models are also being used to analyze clinical and genomic data to identify biomarkers and support risk stratification. AI-assisted cognitive tests provide scalable, noninvasive, and real-time screening. Telemedicine platforms and AI-based Clinical Decision Support Systems (CDSS) are also improving patient management, particularly in remote or underserved areas. Heterogeneity of data, model explainability, ethics, and regulatory guideline requirements remain issues, despite these latest developments. Beyond recent developments such as federated learning and digital twins, the study comprehensively reviews AI's contributions to AD diagnosis and therapy. It also establishes a guide for future research directions for the ethical and equitable integration of AI in clinical practice.
Teacher's professional development is in a fundamental and critical position in the progress of higher education in China. The experience of successive industrial revolutions and educational changes tells us that the key to the success of talent cultivation in colleges and universities in the era of intelligent education lies with teachers. This paper applies data mining techniques and the Apriori algorithm to examine teacher's career development paths and education management using 2023 data from the academic affairs management system of a higher vocational college as the sample dataset. The Apriori algorithm is a powerful tool for discovering hidden patterns and correlations among teacher characteristics, such as education level, workload, and teaching effectiveness, thereby supporting data-driven decision-making in staff development. The results of the study show that encouraging young teachers to pursue further studies, encouraging experienced teachers to lead and help young teachers to study professional technology, enabling teachers to complete the transformation of 'dual-teacher' teachers, and creating a teacher team suitable for higher vocational colleges and universities are the main points of the work of higher vocational colleges and universities in the construction of the teacher team.
Recommender systems are essential in enhancing user experience by making precise preference prediction and retrieval of suitable items. Classical collaborative filtering-based approaches typically miss capturing the semantic richness of the textual reviews and break down in ranking quality, generating less customized recommendations. Furthermore, current deep learning-based models like Bert4Rec or composite methods like J-NCFc continue to have high prediction errors and poor ranking accuracy. To fill these voids, this paper introduces a new CF+BERT model that incorporates collaborative filtering with contextualized review representations obtained from BERT. The originality of this method stems from the integration of user-item interaction patterns and deep semantic representations of reviews to improve both prediction resilience and ranking performance. Experimental outcomes indicate that the model performs excellent rating prediction accuracy with RMSE = 0.4594 and MAE = 0.4588 while guaranteeing 92.3% of the predictions within +/- 0.5 tolerance and 100% within +/- 1.0. When considering ranking tests, the model provides Precision@10 = 0.423, Recall@10 = 0.292, and NDCG@10 = 0.347, further proving its effectiveness in retrieving items that are relevant. Most significantly, the CF+BERT model achieves a state-of-the-art NDCG = 0.9534 over baselines including SENT-ROBERTA (0.6403), J-NCFc (0.4065), and Bert4Rec (0.135). These results show that the introduced approach significantly improves recommendation quality, providing both enhanced predictive accuracy and better ranking performance, thus establishing a new benchmark for future recommender system design. The framework has brought forth the possibility of combining context-aware embedding models with collaborative filtering techniques and toward providing intelligent next-generation recommenders. The future extension would be in the direction of multimodal fusion and large-scale deployment to increase adaptability.
As Internet of Things (IoT) networks are growing rapidly and being applied in a vast array of fields, they increasingly become targets for cyber-attacks needing reliable IDSs. This perspective paper presents a highly innovative hybrid IDS system based on Convolutional Neural Network (CNN), Bidirectional Gated Recurrent Units (BiGRU), Attention Mechanisms, and Binary Snake Optimizer (BSO) for enhanced IoT environment intrusion detection. The CNNs then automatically take important features out of network traffic, while at the same time, BiGRUs capture past and future temporal dependencies about the traffic patterns. Attention Mechanism focuses on different parts of the input feature, enabling the model to detect subtle and evolving attacks. The BSO optimizes the hyperparameters of our model so that convergence time is reduced, improving efficiency for real-time detection. Accordingly, the experimental results have shown that the proposed system performed better with 99.86% accuracy, 99.69% precision, and 99.57% recall, especially compared to current IDS solutions. The ability of the system to classify emerging threats with great efficiency within dynamic IoT networks, along with its scalable and adaptable features, makes it a strong asset for securing IoT systems.
Low-resource language machine translation remains an ongoing issue because of the limited parallel corpora and restricted linguistic diversity. While existing multilingual models such as mBERT or XLM-RoBERTa attain high performance on high-resource languages, they do not reliably characterize the morphological, syntactic, or code-switching features of low-resource languages. The contribution of the proposed transformer-based multilingual natural language processing (NLP) model is based on several innovative methodologies and adaptation strategies that include a unique combination of back translation, nCr-IGAN-based data augmentation, and multilingual subword tokenization for effective generalization across diverse linguistic structures. Moreover, contextual embeddings and dimensionality reduction techniques (e.g., GC-DOA and PCA) were adapted for the extraction and optimization of linguistic features. The proposed multilingual model knows how to learn complex syntactic and semantic structures in low-resource language pairs and exhibits low Mean Square Error (MSE) values, moderate cosine similarity, and high BERTScore in the context of English-Hindi translation tasks. High BERTScore indicates a strong level of semantic alignment between the source and target languages, representing an important contribution toward developing comprehensive, high-value multilingual translation systems for endangered language regimes in NLP. When compared with current multilingual models such as mBERT and XLM-RoBERTa, the system shows numerical improvements, particularly regarding BERTScore and cosine similarity, on English-Hindi pairs. These are new, promising, and useful types of architecture that can deal with the complexities of capturing cross-lingual semantics in low-resource environments.
Accurate disease prediction in corn and soybean crops is required to enhance agricultural productivity and ensure food security. The changing texture of the leaf and illumination create errors in the segmentation process of the conventional disease identification methods. Even though deep learning methods suggest promising outcomes, they tend to get stuck in local solutions and overfit, thus they generalize poorly across different illnesses. Lack of data and unbalanced classes are problems for deep learning in plant disease diagnosis, which leads to biased models and inaccurate generalization. This paper introduces an original approach of Adaptive Feature Segmentation and Dual-Stage Filtering (AFS-DST) and a Hierarchical Residual Attention Network (HRAN) to address the above-mentioned problems. The AFS-DST technique is better because it guarantees the accuracy of the segmentation in the identification of the disease, and it dynamically optimizes features based on color and texture. The HRAN model has a residual attention block and hierarchical feature fusion to improve the generalization and robustness of the model by focusing on disease-specific patterns. Different with the previous methods, the Adaptive Feature Segmentation (AFS) with Dual-Stage Filtering and HRAN is a superior method because it is dynamically updated to adapt the segmentation as per the lighting, the disease severity, and the textures of the leaves. HRAN enhances the predictive accuracy and resistance of the models to environmental variation by attending to multi-level model-specific patterns and therefore enhancing the generalization as well as the robustness of the models. As results of the experiment on the corn and soybean datasets indicate, the proposed model is far more superior to other classic models, with a precision of 97.6 and a little loss incurred in corn, and 88.1 and better results in soybean. These results indicate that the proposed model reduces overfitting and local minima but also sets a new benchmark for disease detection on these crops while providing a strong basis for future research in agricultural disease prediction.
