Date
Tuesday September 23, 2025 from 3:30 PM to 4:30 PMLocation
Neuron 0.262Co-organizer
Eindhoven Artificial Intelligence Systems InstitutePrice
freeBuilding
Neuron
Topic
Perspective taking, partner models and user behaviour in conversational interface interaction
Abstract
This lecture explores the challenges posed by class imbalance in machine learning and its applications in manufacturing and logistics. Class imbalance occurs when the distribution of instances across classes is disproportionate, negatively impacting classification performance and learning efficiency. Practical examples include defect detection in semiconductor manufacturing, fraud detection, and medical diagnosis.
The lecture reviews three main approaches to address class imbalance: data-level techniques (e.g., oversampling, under-sampling, hybrid resampling), algorithm-level strategies (e.g., cost-sensitive learning), and ensemble methods. Additionally, it highlights aggravating factors such as noisy data, class overlap, outliers, and dataset shift that further complicate learning tasks.
Real-world applications in facility location and distribution network design are discussed, alongside solution strategies to mitigate data sparsity and imbalance issues. The lecture emphasizes the importance of tailoring AI techniques to specific challenges in manufacturing and logistics for improved efficiency and reliability.
About the speaker
Professor You-Jin Park is a distinguished academic at the Department of Industrial Engineering and Management, National Taipei University of Technology, Taiwan. He holds a PhD in Industrial Engineering from Arizona State University, specializing in optimization using genetic algorithms.
His research focuses on statistics, operations research, and data mining, with applications in semiconductor manufacturing and process optimization. Professor Park has worked both in academia and industry, including roles at Samsung Electronics and as a consultant at Samsung SDS.
He has held positions as Associate Professor at Chung-Ang University and is currently a Full Professor at Taipei Tech. His projects have been funded by Intel Corp, highlighting his strong connection between academia and industry.
Professor Park has received recognition for his work, including the "Best Researcher Award" in data science. His publications include innovative methods like hybrid resampling techniques for improving classification problems with imbalanced datasets.
He contributes to the UN Sustainable Development Goals through research on energy efficiency and productivity enhancement. His work continues to inspire future generations of engineers and push the boundaries of industrial engineering.
Your host
Vinh Dang, Assistant Professor at the Operations, Planning, Accounting, and Control (OPAC) group in the Department of Industrial Engineering and Innovation Sciences (IE&IS) at the 色中色, will host Professor You-Jin Park, Full Professor at Taipei Tech.
Registration is required but free of charge.