Yung-Chia Chang’s research while affiliated with National Yang Ming Chiao Tung University and other places

The out-of-control action plan (OCAP) is crucial in the wafer probing process of semiconductor manufacturing as it systematically addresses and corrects deviations, ensuring the high quality and reliability of semiconductor devices. However, the traditional OCAP involves many redundant and complicated processes after failures occur on production lines, which can delay production and escalate costs. To overcome the traditional OCAP’s limitations, this paper proposes a novel OCAP aimed at enhancing the wafer probing process in semiconductor manufacturing. The proposed OCAP integrates proactive measures such as preventive maintenance and advanced monitoring technologies, which are tested and verified through a comprehensive experimental setup. Implementing the novel OCAP in a case company’s production line reduced machine downtime by over 24 h per week and increased wafer production by about 23 wafers per week. Additionally, probe test yield improved by an average of 1.1%, demonstrating the effectiveness of the proposed method. This paper not only explores the implementation of the novel OCAP but also compares it with the traditional OCAP, highlighting significant improvements in efficiency and production output. The results underscore the potential of advanced OCAP to enhance manufacturing processes by reducing dependency on human judgment, thus lowering the likelihood of errors and improving overall equipment effectiveness (OEE).

Artificial intelligence algorithms and big data analysis methods are commonly employed in network intrusion detection systems. However, challenges such as unbalanced data and unknown network intrusion modes can influence the effectiveness of these methods. Moreover, the information personnel of most enterprises lack specialized knowledge of information security. Thus, a simple and effective model for detecting abnormal behaviors may be more practical for information personnel than attempting to identify network intrusion modes. This study develops a network intrusion detection model by integrating weighted principal component analysis into an exponentially weighted moving average control chart. The proposed method assists information personnel in easily determining whether a network intrusion event has occurred. The effectiveness of the proposed method was validated using simulated examples.

This special issue is dedicated to exploring the latest innovations and developments in the fields of Natural Language Processing (NLP) and Fuzzy Logic. These cutting-edge domains are pivotal in the advancement of artificial intelligence (AI), enabling machines to understand and interact with human languages and make decisions in uncertain environments. The significance of NLP is highlighted by its applications in various technologies such as chatbots, translation services, sentiment analysis, and more, which enhance user experience and accessibility. Fuzzy Logic, on the other hand, plays a crucial role in handling uncertainties and modeling complex systems, finding applications in control systems, pattern recognition, and consumer electronics. Despite their advancements, the fields of NLP and Fuzzy Logic face several challenges. In NLP, major issues include the need for more sophisticated language models that can understand context better, the handling of multilingual data, and addressing biases in AI systems. Fuzzy Logic must overcome challenges in creating more accurate and interpretable models, integrating with other AI techniques, and improving computational efficiency. This special issue aims to showcase the current trends, breakthroughs, and innovative applications of NLP and Fuzzy Logic. We invite researchers and practitioners to contribute their original research articles and review studies, shedding light on both theoretical and practical aspects of these technologies.

Productivity in the wafer probing process is heavily influenced by two crucial factors: yield and equipment downtime. Low yield in the wafer probing process can be attributed to either the product itself or the wafer probing equipment. To identify the causes of low yield, it is typically necessary to conduct regular equipment check-ups. However, these check-ups also increase equipment downtime, which negatively impacts productivity. In this study, our objective was to enhance the effectiveness of the current contingency plan known as the Out of Control Action Plan (OCAP), which is implemented when a low-yield situation is detected. The OCAP serves as a comprehensive documentation outlining the necessary actions to identify and rectify the underlying causes, ultimately restoring stability to the wafer probing process. According to the experimental results, it is evident that the probe test yield, machine downtime, and pack out under optimal OCAP process flow conditions are better than the existing OCAP process flow mass-production conditions.

In vehicle leasing industry which presents a great business opportunity, information completed by applicants was assessed and judged by leasing associates manually in most cases; therefore, assessment results would be affected by their personal experience of leasing associates and decisions would be further affected accordingly. There are few researches on applicant credit risk assessment due to not easy to obtain of vehicle leasing data. Further, the difficulty in vehicle leasing risk assessment is increased due to class imbalance problems in vehicle leasing data. In order to address such issue, a research on credit risk assessment in vehicle leasing industry was conducted in this study. The great disparity in the ratio of high risk and low risk data was addressed by applying synthetic minority over-sampling technique (SMOTE). Then, classification effect of risk assessment model was improved by applying logistic regression in a two-phase manner. In the section of empirical analysis, the feasibility and effectiveness of the approach proposed in this study was validated by using data of actual vehicle leasing application cases provided by a financial institution in Taiwan. It is found that the proposed approach provided a simple yet effective way to build a credit risk assessment model for companies that provide vehicle leasing.

An unrelated parallel machine scheduling problem motivated by the scheduling of a printed circuit board assembly (PCBA) under surface mount technology (SMT) is discussed in this paper. This problem involved machine eligibility restrictions, sequence-dependent setup times, precedence constraints, unequal job release times, and constraints of shared resources with the objectives of minimizing the makespan and the total job tardiness. Since this scheduling problem is NP-hard, a mathematical model was first built to describe the problem, and a heuristic approach using a non-dominated sorting genetic algorithm (NSGA-II) was then designed to solve this bi-objective problem. Multiple near-optimal solutions were provided using the Pareto front solution and crowding distance concepts. To demonstrate the efficiency and effectiveness of the proposed approach, this study first tested the proposed approach by solving test problems on a smaller scale. It was found that the proposed approach could obtain optimal solutions for small test problems. A real set of work orders and production data was provided by a famous hardware manufacturer in Taiwan. The solutions suggested by the proposed approach were provided using Gantt charts to visually assist production planners to make decisions. It was found that the proposed approach could not only successfully improve the planning time but also provide several feasible schedules with equivalent performance for production planners to choose from.

Defects on thin film transistor liquid crystal display (TFT-LCD) panel could be divided into either macro- or microdefects, depending on if they are easy to be detected by the naked eye or not. There have been abundant studies discussing the identification of macrodefects but very few on microones. This study proposed a multicategory classification model using a convolutional neural network model to work with automatic optical inspection (AOI) for identifying defective pixels on the TFT-LCD panel. Since the number of nondefective pixels outnumbered the defective ones, there exists a very serious class-imbalanced problem. To deal with that, this study designed a special training strategy that worked with data augmentation to increase the effectiveness of the proposed model. Actual panel images provided by a mobile manufacturer in Taiwan are used to demonstrate the efficiency and effectiveness of the proposed approach. After validation, the model constructed by this study had 98.9% total prediction accuracy and excellent specificity and sensitivity. The model could finish the detection and classification process automatically to replace the human inspection.