Kwan-Hee Yoo

This study examines changes from baseline to midline in ICT competency indicators during an ICT teacher training intervention in Cambodian lower secondary education. The intervention aligned with SDG 4.4.1 and included pre-service and in-service teachers from four Regional Teacher Training Centers and 20 pilot schools, along with Grade 8 students. Teachers completed a self-assessment and an objective ICT knowledge test covering 11 SDG 4.4.1 domains. Students completed objective assessments. Teacher competency pass counts increased from baseline to midline. The largest changes appeared in information verification and programming, while foundational domains changed less. Pre-service teachers had higher objective knowledge scores at midline. A nonparametric check showed the same direction. No significant differences were detected across the four provinces. The findings provide evidence from program monitoring that an ODA intervention combining curriculum reform, textbook development, ICT laboratory deployment, and sustained teacher training was associated with changes in teacher ICT competency.

The integration of artificial intelligence (AI) into industrial processes is a promising method for enhancing operational efficiency and quality control. In particular, contact lens manufacturing requires specialized artificial intelligence technologies owing to stringent safety requirements. This study introduces a novel approach that employs polar coordinate transformation and a customized deep learning model, the Polar Sequence-guided Attention Network (PSgANet), to improve the accuracy of defect detection in the rim-connected zone (RCZ) of contact lenses. PSgANet is specifically designed to process polar coordinate-transformed image data by integrating sequence learning and attention mechanisms to maximise the capability for detecting and classifying defective patterns. This model converts irregularities along the edges of contact lenses into linear arrays via polar coordinate transformation, enabling a clearer and more consistent identification of defective regions. To achieve this, we applied sequence learning architectures such as GRU, LSTM, and Transformer within PSgANet and compared their performances with those of conventional models, including GoogleNetv4, EfficientNet, and Vision Transformer. The experimental results demonstrated that the PSgANet models outperformed the existing CNN-based models. In particular, the LSTM-based PSgANet achieved the highest accuracy and balanced precision and recall metrics, showing up to a 7.75% improvement in accuracy compared with the traditional GoogleNetv4 model. These results suggest that the proposed method is an effective tool for detecting and classifying defects within the RCZ during contact lens manufacturing processes.

In this study, we present a Retrieval-Augmented Generation (RAG)-based pipeline designed to extract key values from scientific literature on Silicon Carbide (SiC) crystal growth using the Physical Vapor Transport (PVT) method. To improve the relevance and completeness of the retrieved context, we implemented a hybrid retrieval strategy that combines dense retrieval via FAISS with sparse retrieval using BM25. We employed two distinct prompting approaches for key value extraction. The first approach addresses interactive user queries by utilizing the retrieved context to generate informed responses. The second approach, intended for bulk extraction, follows a two-step process: a binary classification prompt first checks for the presence of relevant information related to a query. If relevant information is confirmed, a subsequent prompt extracts the value under strict constraints—requiring exact phrasing without guessing or explanation. This binary pre-check significantly enhances the identification of true negative cases, thereby reducing irrelevant or missing data. For the generative component of our pipeline, we evaluated three large language models (LLMs): Llama 8B, Gemma 7B, and Mistral 7B, all operating on a local multi-GPU environment using FP16 precision. The results reveal differences in the efficiency of these models within our customized RAG system, particularly in their performance in extracting over 156 targeted technical key-value pairs from 13 benchmark papers.

Detecting fine, weak-textured defects with discontinuous boundaries on complex industrial surfaces is challenging due to interference from background textures and characters, as well as the scarcity of labeled data. To address this issue, we propose YOLO-SR, an engineering modification of YOLO11 tailored to defect segmentation on smart-card surfaces. Rather than introducing a new detection architecture, YOLO-SR reuses the backbone–neck–head design of YOLO11 and only adjusts a few modules to better capture elongated, low-contrast defects. The approach comprises two key components: first, embedding Strip Pooling (SP) within the C3K2 module to form C3K2_SP; second, a Rectangular Self-Calibration Module (RCM) is interposed after the top-level semantic layer. RCM generates rectangular gates to spatially recalibrate local responses, suppressing interference from complex textures and characters. To mitigate data scarcity and distributional bias, a texture-adaptive procedural defect synthesis strategy was developed. This strategy generates defect samples that conform to the background texture statistics of high-quality backgrounds. Experiments on the integrated circuit chip (ICChip) and signature plate (SignPlate) datasets show that YOLO-SR outperforms the YOLO11 baseline. Results indicate that SP and RCM complement each other by integrating directional priors from mid-to-high layers with top-level shape self-calibration. This enhances the visibility and localization stability of elongated defects while maintaining efficient inference.

This study presents an AI-based vision inspection system for the automatic detection of various defect types in membrane electrode assemblies (MEAs), a core component of polymer electrolyte membrane fuel cells (PEMFCs). The system targets a wide range of defects, including electrode line defects, junction damage, wrinkles, bubbles, foreign material contamination, and gasket-to-electrode misalignment. High resolution fuel cell images are preprocessed and analyzed using a deep learning pipeline that integrates classification models and a DETR(DEtection TRansformer)-based object detection framework enhanced with slicing techniques to accurately detect both visible and fine-grained defects. The system supports real-time deployment in manufacturing environments, providing defect classification, traceability, and statistical reporting. This approach significantly improves inspection precision, reliability, and scalability in PEMFC production processes.

This study presents a digital twin-based virtual laboratory designed to optimize the manufacturing processes of electrode materials for Solid Oxide Fuel Cells (SOFC). Leveraging artificial intelligence models (R2=0.89, RMSE=0.03), the system accurately predicts critical material properties such as electrode thickness, dispersion stability, and polarization resistance. The virtual simulation of key processes—material selection, combining, coating, and sintering—significantly enhances process accuracy, efficiency, and product quality, effectively reducing experimental time and costs.

Vision-based inspection systems are essential for quality control in manufacturing industries, and advances in artificial intelligence (AI) have significantly enhanced their accuracy. However, the high-precision requirements of products such as contact lenses demand even more robust inspection methods. This paper introduces a novel defect-adaptive hierarchical structure convolution neural network (DHS-CNN) model based on InceptionV4. The proposed model architecture reflects the manufacturing process and defect types, and we developed a custom loss function to suit this multi-output hierarchical design. Experimental results on a dataset of 2800 contact lens images revealed that the proposed model improved accuracy by 2.08% over the baseline model. These findings suggest that the defect-adaptive hierarchical structure and customized loss function offer substantial improvements in the vision-based inspection of contact lenses and may enhance AI-driven quality control processes in other manufacturing sectors.

Holography is an emerging technology where phase-only spatial light modulator (SLM) is better than amplitude-only SLM because of its enhanced diffraction efficiency and rendering quality. However, phase-only interference pattern reconstruction is an ill-posed inverse problem, difficult to solve by standard supervised neural network approaches. Accurately representing the target input during reconstruction of phase-only hologram (POH) can become challenging given the inherent high spatial frequency of phase data and the complexity of its fringe patterns, expressing the inverse nature of the problem. Therefore, the model may be trained improperly, inducing major compromise in the predicted hologram’s quality. As a result, an unsupervised approach can be effective to resolve such a problem. This paper presents a novel combined U2-net-based computer-generated hologram (CUSNet-CGH) architecture, an end-to-end deep learning model comprising two distinctively designed combined convolution networks that accurately generate 2D POH. The proposed framework contains U2-net which is known for performing unsupervised learning tasks without relying on pre-trained backbones from labeled dataset, making it an ideal approach for conducting unsupervised training. The numerical reconstructions are compared with several other state-of-the-art methods such as holoencoder, holonet and CCNN-CGH to verify the enhanced quality. The average PSNR of the proposed end-to-end approach is as high as ~34 dB with an average SSIM of 0.91, featured by DIV2K dataset, outperforming the state-of-the-art methods analyzed in the paper. Additionally, the time required to reconstruct the predicted output of the proposed model is verified as competitive, challenging the latest benchmarks, thanks to its minimized number of trainable parameters, and ensuring that the model is computationally lightweight. Finally, experiments also demonstrate that the proposed model yields high-quality optical reconstruction.

With increasing industrial sophistication and complexity, workplaces are increasingly prone to occupational accidents, causing negative impacts on workers and employers, including economic losses and decreased productivity. South Korea occupational safety and health has implemented new policies addressing potential risks to overcome stagnation in industrial accident reduction and predict site accidents from past cases. Cases are human-classified according to rules, including occurrence type or original causal materials. However, human errors, subjective judgments, synonyms, and terms incorrectly used by classifiers reduce original data quality and impede developments or applications of policies, technologies, and methods preventing accidents based on past accidents. This study proposes three artificial intelligence models to objectively classify the occurrence type of accident cases. Models are developed based on a natural language processing model (KoBERT), which considers Korean language characteristics. Each model is tested by sequentially performing sentence preprocessing, keyword replacement, and morphological analysis. The proposed Model 3 exhibits 93.1% accuracy, which was the highest among tested models. Up to three classification categories for occurrence type are allowed to assist objective classification. The accident case-based occurrence type classification model is effective for industrial accident prevention, aiding in strategy development and reducing social costs.