Chien Hsin University of Science and Technology

Powder metallurgy (PM) technology is extensively used in high-value industries for its energy efficiency, precision, and cost-effectiveness. However, detecting mini-defects in PM production remains challenging, particularly in highly customized and small-batch productions. This study evaluated defect detection in PM parts, PMPDv1 and PMPDv2 datasets comprising 457 and 1521 images, respectively. Automated optical inspection (AOI) and image augmentation techniques were applied to enhance image quality and model learning. The YOLO series models were employed for automated defect detection.Results demonstrated that YOLOv4 achieved a mean average precision (mAP) of 93.94% at a resolution of 1600 but required 31 GB of GPU memory and 881,443 GFLOPs. YOLOv5s, under the same conditions, achieved an mAP of 92.7% with just 12.1 GB of GPU memory and 15.8 GFLOPs, making it suitable for resource-constrained environments. This study confirms the efficacy of YOLO models for PM defect detection and suggests further exploration of transfer learning and generative AI techniques to enhance detection efficiency and accuracy in other products.

Robots are increasingly used in education, but challenges limit their effectiveness: limited numbers, interactions restricted to class time, uniform designs that prevent ownership, and difficulty maintaining engagement over a semester. To address these issues, we propose a novel AI-powered personalized robot framework. Each robot is paired with a digital companion, accessible via students' mobile devices and enhanced by large language models (LLMs) for dynamic, personalized interactions. Students nurture their digital companions by studying, earning virtual currency, and customizing the robot's appearance and behaviors. The nurtured companion's “spirit” is later embodied in a shared physical robot during class, integrating digital and physical learning experiences while using only one robot per classroom. A quasi-experimental study in Taiwan involved 90 Hospitality Management students training for Japanese restaurant service. Three groups were compared: AI-Powered Personalized Robots (with LLM-enhanced companions), Standardized Robots (non-personalized but accessible inside and outside class), and a Conventional Classroom (no robots). Results showed that students in the AI-Powered Personalized Robots group achieved higher learning outcomes, demonstrated greater ownership toward their studies, and engaged more frequently than other groups. This framework lays the groundwork for scalable, AI-integrated education by reimagining how learners interact with intelligent toys, fostering engagement across an entire semester.

Graphene's incorporation into polymers has enabled the development of advanced polymer/graphene nanocomposites with superior properties. This study focuses on the use of a microcellular foamed polystyrene (PS)/graphene (GP) nanocomposite (3 wt%) for nickel (II) ion removal from aqueous solutions. Adsorption behavior was evaluated through FTIR, TEM, SEM, TGA, and XRD analyses. Key factors, including initial ion concentration, pH, temperature, and sorbent dosage, were examined. Results showed optimal nickel removal at specific pH levels with removal efficiency decreasing from 91 to 80% as Ni (II) concentrations increased from 10 to 100 mg/L. The adsorption capacity improved from 11 to 130 mg/g. Equilibrium data aligned with Langmuir and Freundlich isotherm models, while adsorption kinetics followed a second-order kinetic model. These findings highlight the potential of PS/GP nanocomposites for nickel ion removal, offering a promising solution for small-scale industrial applications.

Purpose The social dimension of sustainable development has garnered increasing attention. As universities embrace their social responsibility and consider the interests of various stakeholders, the potential issue of social washing has emerged as a critical topic. This study aims to investigate the presence of social washing in university sustainability reports. Design/methodology/approach The study examines three key stakeholder elements: full-time faculty salaries, weekly teaching hours for full-time faculty and hourly wages for part-time faculty. A content analysis was conducted on the 2022 sustainability reports published by all private universities registered in Taiwan. Findings Results indicate that only 30% of private universities published independent sustainability reports for 2022, and of those, only 62.5% adhered to global reporting initiative guidelines. The study raises concerns about selective disclosure and the concealment of negative information, suggesting the possibility of social washing. This investigation offers an overview of social washing in the sustainability reports of higher education institutions, thereby contributing to the academic discourse on comprehensive and transparent communication with stakeholders. Social implications Universities should take into account the interests of stakeholders and embrace greater social responsibility in their sustainability initiatives. This study analyzes the content of university sustainability reports and encourages higher education institutions to foster balanced communication with their stakeholders. Originality/value Social washing is difficult to detect. This study uses objective indicators to assist higher education institutions in identifying potential social washing behaviors and provides guidance for universities to avoid misleading communication in their sustainability reports.

PurposeThe mental health issues faced by young people can significantly hinder a nation’s development. The purpose of this study is to integrate the Self-Determination Theory and the Theory of Planned Behavior to examine the factors influencing the utilization of mental health days among young individuals.Design/methodology/approachAn online survey was conducted to collect data from young people.FindingsThe results indicated that more positive attitudes, favorable subjective norms and a stronger sense of perceived behavioral control regarding mental health days are all associated with a higher intention to use them. Young individuals who have used counseling resources are more likely to endorse the use of mental health days. Satisfaction of the three basic psychological needs of autonomy, competence and relatedness has a diminishing effect on both motivation and intention to request mental health days. However, there is a positive relationship between motivation and intention to apply for mental health days. When competence needs are less well satisfied, the motivation and, in turn, the intention to request mental health days are enhanced.Research limitations/implicationsShifting the leave framework from a disease-centered approach to a psychological leave model focused on mood adjustment can maximize the positive impact of mental health leave.Originality/valueThis study contributes to the theoretical understanding of the relationships between needs and motivation within the framework of self-determination theory. Furthermore, it integrates components from the theory of planned behavior to examine the influence of social pressures and self-efficacy on the intention to use mental health days. Mental health days can be seen as a cry for help from young adults. The utilization of mental health days serves as an effective means of managing one's mental state.

ABSTRACT Pile foundations are widely used for tall buildings, but determining their capacity and behavior under load often involves expensive and lengthy field tests or imprecise numerical models. This research employed a centrifuge to perform vertical loading tests on single and grouped piles in dry sand, subjected to 100 times gravitational acceleration, to investigate pile capacity and load transfer mechanisms. The results from centrifuge modeling, numerical simulations, and field tests were consistent. The total pile capacity for a single pile was 5800 kN/m², while short and long pile groups had capacities of 6800 kN/m² and 9200 kN/m², respectively. End-bearing accounted for 40% of the capacity in single and short pile groups, and 13% in long pile groups. Surface friction was 95 kN/m² for single piles and 140 kN/m² for grouped piles. A method for estimating total pile capacity based on soil properties showed good agreement with empirical results.

Three-dimensional printing is a non-conventional additive manufacturing process. It is different from the conventional subtractive manufacturing process. It offers exceptional rapid prototyping capabilities and results that conventional subtractive manufacturing methods cannot attain, especially in applications involving curved or intricately shaped components. Despite its advantages, metal 3D printing will face porosity, warpage, and surface roughness issues. These issues will affect the future practical application of the parts indirectly, for example, by affecting the structural strength and the parts' assembly capability. Therefore, this study compares the qualities of the warpage, weight, and surface roughness after milling and grinding processes for the same material (316L stainless steel) between rolled steel and 3D-printed steel. The experimental results show that 3D-printed parts are approximately 13% to 14% lighter than rolled steel. The surface roughness performance of 3D-printed steel is better than that of rolled steel for the same material after milling or grinding processing. The hardness of the 3D-printed steel is better than that of the rolled steel. This research verifies that 3D additive manufacturing can use surface processing to optimize surface performance and achieve the functions of lightness and hardness.

This study examined the effects of students’ argumentation standpoints on their argumentation learning in the context of socio-scientific issues (SSIs). To that end, four kinds of argumentation standpoints were defined: affirmative standpoints, oppositional standpoints, multiple standpoints, and non-standpoints. These four kinds of standpoints allow for six possible combinations of any two of the different kinds of standpoints, which enabled us to conduct six kinds of 2-team format debates. The resulting differences of students’ four types of arguments (i.e., claims, warrants, rebuttals, and qualifiers) generated in six types of debates were examined. This study invited 208 10th-grade students to participate in a quasi-experimental research design. The results showed the affirmative group students demonstrated superior performance in terms of claims and warrants, and the oppositional group students had the largest number of rebuttals. The students in the debate with combinations of affirmative and oppositional groups exhibited the best performance regarding the generation of claims, warrants, and rebuttals. Based on the results, the present study concluded the standpoint had significant effects on the students’ argumentation learning, which suggests that teachers can investigate students’ standpoint on the learning topic of SSI and their prior knowledge about the standpoint before teaching. Keywords: argumentation learning, science education, socio-scientific issue, student standpoint.