AI4X 2025 Papers

As semiconductor devices continue to shrink to the nanoscale, ensuring their structural accuracy becomes increasingly critical. Optical imaging techniques play a key role in defect detection in semiconductor manufacturing. However, these optical techniques are fundamentally limited by the diffraction limit, which restricts the resolution of features smaller than half the illumination wavelength. While deep learning has shown promise in pushing beyond this limit, conventional models often rely on pixel-wise loss functions that fail to capture the global structure of semiconductor patterns. This can result in broken lines, missing features, and inaccurate defect detection. In this work, we introduce a topology-preserving deep learning framework tailored for high-resolution optical imaging in semiconductor characterization. By embedding topological constraints through persistent homology in a differentiable loss function, our model enforces structural consistency across the entire image. Experimental results demonstrate an improvement of nearly 30% in the quantifier metric against ground truth compared to traditional methods on the nano-particle dataset, accurately resolving features as small as 0.16λ (100 nm) under 640 nm illumination. The model further reduces false disconnectivities by 5% in nano line localization. This topologically-aware approach provides a robust and non-destructive solution for subwavelength imaging in semiconductor metrology, enabling more reliable inspection in precision manufacturing.

Financial markets, characterized by their volatility, uncertainty, complexity, and ambiguity (VUCA), pose significant challenges for accurate predictions. Investment has become increasingly intertwined with technological advancements, as machine learning models revolutionise the field of stock market trend predictions, offering potential solutions by processing large datasets, identifying trends, and minimizing human bias. While machine learning is increasingly applied in financial forecasting, understanding the relative strengths and weaknesses of different algorithms across varying time frames remains underexplored. This is especially relevant given the rise of algorithmic trading and new stock markets such as cryptocurrencies, underscoring the need for precise, data-driven predictions. The aim of this study is to evaluate the performance of machine learning algorithms in predicting stock prices within Singapore’s banking sector. The study explores how each algorithm performs when trained on different amounts of data, comparing its effectiveness for short-term, mid-term and long-term stock price predictions. To do this, historical stock prices were collected using the Yahoo Finance API, focusing on closing prices as the target variable. Using the data collected from major Singaporean banks, namely DBS, OCBC and UOB, this study evaluated the performance of various machine learning algorithms: Random Forest (RF), Support Vector Regression (SVR), K-Nearest Neighbors (KNN), Artificial Neural Networks (ANN), and Long Short-Term Memory (LSTM). The various models were all trained on different datasets, and its predictions for the closing price on a specific date was recorded. Each model was evaluated using rigorous performance metrics, including percentage error, R² values, mean absolute error, and mean squared error, to determine their efficacy in capturing trends and minimising predictive inaccuracies. Different algorithms have distinct methods of learning patterns and handling data variability, and thus will perform differently under the same conditions. Hence, we hoped to gain greater insight into each model’s performance and assess their adaptability to the various time frames. This study contributes to the growing body of research on AI-driven financial forecasting by providing a comparative analysis of machine learning algorithms in Singapore’s banking sector. It highlights the need for flexibility in one’s approach to algorithmic trading to enhance prediction accuracy across diverse scenarios. The insights gained can aid financial analysts, traders, and decision-makers in developing data-driven strategies for stock market investments, ultimately promoting more informed decision-making and risk management in a volatile financial landscape.

The rapid expansion of scientific literature necessitates developing efficient data extraction and analysis methods. This study presents an innovative approach to automating the extraction of cryoprotectant information from scientific publications using a generative pretrained transformer (GPT) model integrated with a Telegram bot interface. Our system processes and analyzes scientific articles to identify and extract relevant data on cryoprotectants and bacteria, significantly reducing the time required for researchers to gather essential information. Our method optimizes the workflow for researchers in cryopreservation and related fields by utilizing modern artificial intelligence technologies, specifically large language models. The Telegram bot, designed to be user-friendly, provides a comfortable and easy platform for quick data access, enhancing scientific research efficiency. The study's methodology involves data preparation, algorithm development, and system validation using a substantial data set of scientific articles. Results demonstrate the model's capability to accurately recognize and extract critical information, although some limitations in term specificity were noted. Our findings suggest that further refinement and training of the model can enhance its accuracy and reliability for specialized scientific applications.