The automotive industry depends on high-quality paint coatings for both aesthetic appeal and functional performance. However, surface imperfections such as scratches, paint runs, and orange peel can arise from process and environmental variations. This study employs machine learning (ML) and exploratory data analysis (EDA) to identify key factors that influence surface defect formation in automotive painting. Using historical production data from Lucky Motor Corporation Limited, models based on linear regression, support vector machines (SVM), and random forests were developed and validated under various process conditions. The best-performing model achieved an R² of 0.94, with a mean absolute error (MAE) of 0.125 and mean squared error (MSE) of 0.033, demonstrating high predictive accuracy. The proposed ML framework offers a data-driven approach for quality control and process optimization, with the potential to enhance production efficiency, reduce waste, and improve overall paint surface quality.
Scholars in the social sciences and spatial humanities have developed a robust body of scholarship touting the benefits for the public and researchers alike in engaging with mapping projects, especially using public participatory geographical information systems (PPGIS) approaches. Developing robust and sustainable engagement from community stakeholders has been a continual challenge, however. Scholars have promoted many different public outreach activities and programs that work to engage the public in mapping projects. Little work has been done, though, to develop a model that guides researchers on how to create and sustain public engagement in these projects. This article aims to fill this void by developing a sustainable public engagement model for digital mapping projects that integrates selected best practices from a wide range of fields including citizen science, social sciences and spatial humanities, public relations and communications, and public history and interpretation.
Large Language Models (LLMs) demonstrate impressive capabilities across many applications but remain vulnerable to jailbreak attacks, which elicit harmful or unintended content. While model fine-tuning is an option for safety alignment, it is costly and prone to catastrophic forgetting. Prompt optimization has emerged as a promising alternative, yet existing prompt-based defenses typically rely on static modifications (e.g., fixed prefixes or suffixes) that cannot adapt to diverse and evolving attacks. We propose Dynamic Deep Prompt Optimization (DDPO), the first jailbreak defense based on deep prompt optimization. DDPO uses the target LLM's own intermediate layers as feature extractors to dynamically generate defensive embeddings via a lightweight multilayer perceptron. These tailored embeddings are then injected into a subsequent intermediate layer, enabling an input-dependent defense without modifying the LLM's weights. This design ensures high adaptability with minimal computational overhead. Experiments on a diverse set of models and attacks demonstrate that DDPO significantly outperforms static prompt optimization methods, particularly on weakly aligned models and when handling semantically ambiguous benign prompts, successfully distinguishing them from genuinely harmful requests.
This paper reviews asphalt-rubber interactions from macroscopic, microscopic, and molecular perspectives, and discusses solutions for high viscosity and poor storage stability in crumb rubber-modified asphalt (CRMA). Crumb rubber (CR) interacts with asphalt through swelling and degradation, where swelling is a physical process that enhances high- and low-temperature performance, while degradation is a chemical process that generally reduces high-temperature resistance but improves low-temperature properties. The dominant interaction influencing CRMA performance is determined by CR characteristics, asphalt composition, and CRMA production conditions. The basket drainage method and molecular simulations provide insights into the affinity between CR and asphalt components. Strategies to reduce viscosity or improve storage stability include production control, warm mix additives, oil pre-soaking, CR devulcanization, grafting, and polymer additives. These methods either accelerate swelling or promote degradation, affecting viscosity, stability, and thermomechanical performance. Developing kinetic models for CR swelling and degradation, linking them to performance, and identifying consistent solutions for viscosity reduction and stability improvement are crucial for advancing CRMA application from recycled tires for sustainable infrastructure.
Imbalanced classification is a significant challenge in machine learning, especially in critical applications like medical diagnosis, fraud detection, and cybersecurity. Traditional oversampling techniques, such as SMOTE, often fail to handle label noise and complex data distributions, leading to reduced classification accuracy. In this paper, we propose GK-SMOTE, a hyperparameter-free, noise-resilient extension of SMOTE, built on Gaussian Kernel Density Estimation (KDE). GK-SMOTE enhances class separability by generating synthetic samples in high-density minority regions, while effectively avoiding noisy or ambiguous areas. This self-adaptive approach uses Gaussian KDE to differentiate between safe and noisy regions, ensuring more accurate sample generation without requiring extensive parameter tuning. Our extensive experiments on diverse binary classification datasets demonstrate that GK-SMOTE outperforms existing state-of-the-art oversampling techniques across key evaluation metrics, including MCC, Balanced Accuracy, and AUPRC. The proposed method offers a robust, efficient solution for imbalanced classification tasks, especially in noisy data environments, making it an attractive choice for real-world applications.
