Jing Yang’s research while affiliated with University of Kuala Lumpur and other places

Prompt engineering significantly influences the reliability and clinical utility of Large Language Models (LLMs) in medical applications. Current optimization approaches inadequately address domain-specific medical knowledge and safety requirements. This paper introduces EMPOWER, a novel evolutionary framework that enhances medical prompt quality through specialized representation learning, multi-dimensional evaluation, and structure-preserving algorithms. Our methodology incorporates: (1) a medical terminology attention mechanism, (2) a comprehensive assessment architecture evaluating clarity, specificity, clinical relevance, and factual accuracy, (3) a component-level evolutionary algorithm preserving clinical reasoning integrity, and (4) a semantic verification module ensuring adherence to medical knowledge. Evaluation across diagnostic, therapeutic, and educational tasks demonstrates significant improvements: 24.7% reduction in factually incorrect content, 19.6% enhancement in domain specificity, and 15.3% higher clinician preference in blinded evaluations. The framework addresses critical challenges in developing clinically appropriate prompts, facilitating more responsible integration of LLMs into healthcare settings.

Lung cancer is considered as the king of the human cancer globally, which is a major concern affecting public health. In the current work, a lab-built high-speed capillary electrophoresis (HSCE) system coupled with an ultra-short identification microchannel (25 mm) and a facile-to-use space domain internal standard quantification method (SDIS) was employed for the high-efficient analysis of the typical lung-cancer tumor (p53 gene). The application performances including reliability, accuracy, and precision of the provided HSCE system were investigated in detail. Remarkably, codon 249 (118 bp) and codon 248 (200 bp) of the p53 gene could be well identified and quantified in an ultra-short time of 60 s simultaneously. The detected concentration of the two target samples (codon 249 and codon 248 fragments) featured a well linearity determined by the SDIS method, and the theoretical detection limit could achieve 0.016 62 and 0.054 67 ng/μl, respectively. This work may provide a promising detection strategy for lung cancer diagnosis in the clinical field.

Skin cancer, one of the most prevalent and lethal cancer types, poses significant challenges for early diagnosis due to the diversity in lesion size, shape, color, and surface reflections. The Internet of Things (IoT) has revolutionized healthcare by enabling real-time data exchange and supporting advancements in automated diagnosis through deep learning (DL) techniques such as convolutional neural networks (CNNs). However, CNNs often require large, labeled datasets, which are costly and time-consuming to compile. To address these challenges, we propose an innovative active learning (AL) framework driven by deep reinforcement learning (DRL) and a novel scope loss function. This framework optimizes classification while reducing reliance on extensive labeled data. Unlike traditional active learning techniques that rely on static selection methods, our model dynamically incorporates deep reinforcement learning (DRL) for strategic sample selection during training. The scope loss function balances the exploitation of labeled data with the exploration of new, unlabeled data, enabling efficient training. Additionally, an enhanced artificial bee colony (ABC) algorithm with a mutual learning strategy optimizes hyperparameter tuning, boosting model performance. Evaluated on the International Skin Imaging Collaboration (ISIC) and human against machines 10000 images (HAM10000) datasets, the proposed framework achieved high accuracy, with F-measures of 92.791% and 91.984%, respectively. This novel approach demonstrates significant potential to advance early skin cancer detection, offering a reliable and efficient tool for healthcare professionals.

Nucleic acid aptamers, selected through the Systematic Evolution of Ligands by Exponential Enrichment (SELEX), are short nucleic acid sequences that exhibit high affinity and specificity towards diverse targets. Over the past three decades, substantial advancements have been made in both the technology and applications of nucleic acid aptamers. This review provides an in-depth analysis of the historical development and defining characteristics of aptamers, highlighting recent technological innovations in SELEX, including Capillary Electrophoresis SELEX, Microfluidic SELEX, Cell-SELEX, and others. We explore the applications of aptamers in therapeutic and targeted drug delivery, emphasizing their advantages over traditional antibodies such as cost-effectiveness, ease of synthesis, and lower immunogenicity. Key challenges such as stability, specificity, and efficient delivery are discussed, with proposed strategies for improvement including advanced chemical modifications and integration with nanotechnology. By integrating advanced technologies, aptamers hold significant promise for enhancing precision medicine and personalized therapeutic interventions, offering new avenues for the treatment of complex diseases.

Cesium lead halide perovskites (CsPbX3) have become superior candidates for prospective optoelectronic applications. However, the currently reported CsPbX3 quantum dots entail complex production processes and high environmental requirements. In this work, all-inorganic perovskite CsPbBr3 films were prepared by a facile printing strategy, and corresponding photoelectric detection devices were designed and their optical response characteristics investigated. The results showed that the pure CsPbBr3 film is relatively smooth and can maintain high stability under an ambient environment. Furthermore, the thin film prepared by the printing strategy has the advantages of convenience, uniformity, and high photoluminescence, with good application prospects in the field of CsPbBr3 quantum dots.

Objective Early and accurate detection of COVID-19 and pneumonia through medical imaging is critical for effective patient management. This study aims to develop a robust framework that integrates synthetic image augmentation with advanced deep learning (DL) models to address dataset imbalance, improve diagnostic accuracy, and enhance trust in artificial intelligence (AI)-driven diagnoses through Explainable AI (XAI) techniques. Methods The proposed framework benchmarks state-of-the-art models (InceptionV3, DenseNet, ResNet) for initial performance evaluation. Synthetic images are generated using Feature Interpolation through Linear Mapping and principal component analysis to enrich dataset diversity and balance class distribution. YOLOv8 and InceptionV3 models, fine-tuned via transfer learning, are trained on the augmented dataset. Grad-CAM is used for model explainability, while large language models (LLMs) support visualization analysis to enhance interpretability. Results YOLOv8 achieved superior performance with 97% accuracy, precision, recall, and F1-score, outperforming benchmark models. Synthetic data generation effectively reduced class imbalance and improved recall for underrepresented classes. Comparative analysis demonstrated significant advancements over existing methodologies. XAI visualizations (Grad-CAM heatmaps) highlighted anatomically plausible focus areas aligned with clinical markers of COVID-19 and pneumonia, thereby validating the model's decision-making process. Conclusion The integration of synthetic data generation, advanced DL, and XAI significantly enhances the detection of COVID-19 and pneumonia while fostering trust in AI systems. YOLOv8's high accuracy, coupled with interpretable Grad-CAM visualizations and LLM-driven analysis, promotes transparency crucial for clinical adoption. Future research will focus on developing a clinically viable, human-in-the-loop diagnostic workflow, further optimizing performance through the integration of transformer-based language models to improve interpretability and decision-making.

In this study, Mg-doped Cu2ZnSnS4 (CZTS) thin films were prepared by the ionic solution spraying method, which is simple, low cost, and low temperature. The samples were annealed at the optimal doping level. In this study, the relationship between the magnesium doping level, the annealing process conditions (temperature and duration), and various aspects of the sample properties was systematically investigated and discussed. These properties include film crystallinity, optical absorption properties, atomic ratios, and elemental composition. The results show that the Cu2Mg0.2Zn0.8SnS4 thin sample annealed at 500 °C for 10 min exhibits the best thin-film properties and has a relatively ideal atomic ratio. The replacement of Zn²⁺ ions with Mg²⁺ ions in the Cu2ZnSnS4 lattice leads to improved surface uniformity and densification, significant grain coarsening, and a reduction in the material’s optical bandgap to 1.37 eV. These materials demonstrate promising scalability for solar energy conversion applications, with cost-effective processing protocols and enhanced performance characteristics suggesting viable pathways for industrial implementation.

The Industrial Internet of Things (IIoT) seeks to improve smart factory productivity by leveraging automation and scalability. For automation in industry, optimization, collaboration and protection, and scalability, the IoTs paradigm, technology for communication and information, and intelligent systems are integrated as a single organism. This article presents a blockchain-assisted safe data-sharing mechanism that provides security requirements in the industry using IoT. End-to-end authentication is developed based on the blockchain’s reputation, and the smart contract is used to validate nodes’ security measures. Through integrity verification and categorization in node terminals and industry, the blockchain paradigm manages data collection and dissemination. An efficient proof of authentication (PoAh) consensus mechanism is created using the blockchain network to build a collaborative network to preserve logs and verification data in the industrial IoT. It accomplishes trustworthy authentication and endpoint activity tracing, and edge computing is used in blockchain nodes to offer device authentication processes that utilize smart contracts and PoAh. According to an experimental study, the proposed architecture lowers the authentication time and obtains a high response rate. The proposed system’s service time shows the efficiency of PoAh-based blockchain architecture for industrial IoT compared with existing works. Finally, we evaluated various block sizes to ensure an efficient transaction rate. The outcomes demonstrate the practicality of the suggested and put-into-practice architecture, distinguished by enhanced data audibility, device data ownership, security, and privacy while utilizing decentralized storage.