Speech

8/26/2023 Speaker, Winstron ITS (緯創軟體)

感謝您的指導與協助，期待能透過您的專業意見，讓這次的講座更加精彩與具有實質價值

*** 講座主題：智慧製造、AI 於工業4.0應用案例、搭配實作練習

*** 講座大綱：Case Study: 5 x 8 = 40 minutes, 實作練習: 10 minutes, Q&A: 10 minutes

1. 影像辨識中的機器學習應用：探索產品與原料樣式辨識模型的建立，以及在產品瑕疵檢測中的影像辨識應用。(8 minutes)

Case Study: 塑膠射出成形產品瑕疵檢測

2. 數據分析中的機器學習應用：如何透過機器學習提升產品良率，解決物料表與配方預測等問題，並深入探討異常關鍵因素分析等議題。(8 minutes)

Case Study: 物料表與配方預測(紡織業、化工業、食品業。。。)

3. 虛擬量測技術在產業中的應用：以半導體產業為例，探究品質異常關鍵因素分析以及加工件公差的虛擬量測方法。(8 minutes)

Case Study: 加工件公差的虛擬量測

4. 新興應用領域中的影像辨識：從印刷電路板瑕疵偵測、商品影像辨識，到虛擬量測在品質分析與配方預測中的應用。(8 minutes)

Case Study: 印刷電路板瑕疵偵測

5. 融合雲端與地端的解決方案：探討如何在雲端(Cloud)進行訓練與驗證，並將推論與預測部署在地端(Edge)以提供更完整的解決方案。(8 minutes)

Case Study: 印刷電路板瑕疵偵測 (在雲端(Cloud)進行訓練與驗證、推論與預測部署在地端(Edge))

實作練習: 設備預測維修(PHM)與剩餘壽命預測 (10 minutes)

Q&A (10 minutes)

Speech of 2021 (Nvidia GTC 21)

Forecast the Remaining Useful Life of a Lithium-Ion Battery with Features Extraction Using Time Backtracking Method [S31320] [link]

We'll talk about applying NTUT autoML (automatic machine learning) to forecast the remaining useful life of a lithium-ion battery with features extraction using the time backtracking method, a research framework revealing the battery's entire charging and discharging behavior and its history of usage environment, based on its voltage, current, and temperature record. Features extraction using the time backtracking method includes charging (CC and CV), idle, discharging, DOD (depth of discharge), lowest voltage, highest temperature, operation window, and more. An automatic machine learning tool developed by the National Taipei University of Technology will be applied on these data (extracted features) to train the model using NVIDIA’s GPU Tesla v100 to predict a lithium-ion battery's remaining useful life. The research framework can be used in autonomous vehicles to predict their remaining useful battery life, as well as whether the battery will be broken in the next defined cycle time.

About the speaker: Steve Chen, NTUT Industry 4.0 Consulting Group

Steve is an NVIDIA Certified Instructor who's currently leading a team delivering an AI Train & Coach Program, helping companies apply machine learning tools to solve their industries' problems, ranging from AOI+AI defects inspection to finding critical factors causing low yield rate and recipe prediction. Steve's the project manager and co-founder of the NTUT Industry 4.0 Consulting Group. Steve currently serves as an adjunct lecturer in the Department of Vehicle Engineering, National Taipei University of Technology, Taiwan.

Speech of 2018

Topic: From Image Sentiment Classification to Nucleus Segmentation, how do we apply AI to the industry? (從臉部表情辨識、肺部image疾病辨識，到腦瘤癌細胞與細胞核位置的定位，產業實務上我們如何應用機器學習視覺辨識技術?)

Abstract: For the past two years, our group has focused on Big Data practical projects in Electronics, Textile, LED and Semiconductor industry. With the momentum kicking off Artificial Intelligence in 2018, we have shifted the gear to image recognition, from cases studies in Image Sentiment Classification, Chest X-ray Classification to Automatic segmentation of brain tumor. The latest on-going one is 'A Cross-Disciplinary Case Study: Expedite Cancer Treatments and New Drug Development through Nucleus Segmentation using Machine Learning & Edge Computing'. The crucial question is how we are going to apply these matured AI skills to the industry?

In this speech, we will show you our journey transforming to machine learning & vision recognition using the cases study above, and our ways referencing NVIDIA AI Methodology of deploying these matured skills to manufacturing area. 我們正在建立一套PC Based機器學習視覺辨識系統，來處理目前客戶在生產線上所面臨的問題：瑕疵品檢測、標籤列印歪斜及瑕疵、電路板檢測、焊錫檢測、機器手臂定位校正...；以彌補傳統AOI誤判率過高的缺點。Other AI applications, including cardiovascular risk prediction via retina deep learning, lip reading for hearing impaired, Chest X-ray Classification, and emerging innovative fields, will be also addressed.

0-public-From Image Senti...y AI to the industry-v1.0

5/2018 Speaker, 資訊工程系，國立台灣科技大學 (closed)

4/2018 Speaker, EMBA資管個案專題分享，國立台灣科技大學 (closed)

4/2018 Speaker, 資訊管理系，國立政治大學 (cancelled)

3/2018 Speaker, 機器學習概論與應用，財務金融系，國立台北商業大學 (closed)

3/2018 Speaker, 時間數列分析，財務金融系，國立台北商業大學 (closed)

3/14/2018 Speaker, 類神經網路與應用，明志科技大學 (closed)

2/22/2018 Knowledge Sharer in study group, 台灣人工智慧學校@南港中研院 (closed)

3/19/2018 應台灣電路板協會之邀授課，課程名稱：「應用大數據工具提升良率與稼動率」助力PCB產業 (closed)

5/2018 應財團法人紡織產業綜合研究所之邀授課，課程類別：紡織智慧工廠人才培訓，課程名稱：大數據分析與物聯網整合應用 (closed)

7/2018 扶輪社例會中專題演講，題目：產業實務上我們如何應用人工智慧機器學習視覺辨識技術? (engage)

7/12/2017 Speaker, Breakout Sessions II, annual SAP Academic Conference 2017 Asia Pacific Japan (APJ), Seoul, South Korea

Date: July 11-14, 2017

Conference Venue: SAP AppHaus Korea

Topic: Vision: from teaching SAP curriculum to engaging in SAP Consulting Business.

Learn how to shape your vision using SAP curriculum in SAP Consulting Business with National Taipei University of Technology (NTUT) Big Data practical project experience in Electronics, Textile, LED and Semiconductor industry, utilizing Big Data tools to improve yield rate and overall equipment effectiveness.