時間：114年2月27日(星期四) 15:10-17:00

地點：大仁樓301室

主持人：彭彥璁老師

演講者：台北科技大學電子工程系黃士嘉教授

演講題目：Machine Learning in Smart Devices: Shifting from Rule Based to Learning Based Approaches

演講摘要：

In this talk, we will explore the transformative shift in smart devices from rule-based to learning-based approaches, focusing on applications such as visibility restoration and object detection in challenging conditions like rain, fog, and low light. We will begin by reflecting on my research journey, starting with early work on smart panels, TVs, and smartphones, and progressing to more recent advancements in smart cars. We’ll discuss how early rule-based systems, while effective in their time, were limited by their reliance on predefined rules and static responses to environmental challenges. As technology evolved, the shift toward machine learning and deep learning allowed for more flexible, adaptive solutions that improve over time. Through real-world examples, we’ll explore how learning-based models are revolutionizing visibility restoration and object detection, making these devices smarter, more responsive, and capable of handling dynamic environments. Finally, we will reflect on the decades of experience that have shaped this evolution, the lessons learned, and the challenges faced along the way. We’ll also discuss the future of smart devices, where machine learning and AI continue to push the boundaries of what these devices can achieve, ultimately enhancing user experience and safety.