6. 陳昇瑋 / 從大數據走向人工智慧 9
2016 Global manufacturing
competitiveness index rankings
7. 陳昇瑋 / 從大數據走向人工智慧
1/3 of the GDP
Manufacturing GDP of $178B, almost 1/3 of total
GDP
30% of the employment are in the manufacturing
sector
Cheap labor cost of $9.42/hr with average labor
productivity of almost $60k in GDP/person
17% corporate tax rate
10
8. 陳昇瑋 / 從大數據走向人工智慧
McKinsey’s Four Dimensions in
AI Value Chain
11
Smart R&D and
forecasting
Project
Optimized
production with
lower cost and
higher efficiency
Produce
Products and
services at the
right price, time,
and targets
Promote
Enriched and
tailored user
experience
Provide
9. 陳昇瑋 / 從大數據走向人工智慧
The Four-P Dimensions in Manufacturing
Improve product design
Automate supplier assessment and price negotiation
Anticipate parts requirements
Improve manufacturing processes
Automate assembly lines
limit product rework
Optimize pricing
Predict sales of maintenance services
Refine sales-leads prioritization
Optimize flight/fleet planning and route
Enhance maintenance engineering
Enhance pilot training
12
Provide
Project
Promote
Produce
11. Professor
• Prior to joining Harvard in 1992, Dr. Kung taught
at Carnegie Mellon University for 19 years.
• In 1999 he started a joint Ph.D. program with
colleagues at the Harvard Business School on
information, technology, and management, and
co-chaired this Harvard program from 1999 to
2006.
• Member of National Academy of Engineering
• Guggenheim Fellowship
• IEEE Computer Society Charles Babbage Award
HT Kung
• Academician, Academia
• William H. Gates Professor, Harvard John A.
Paulson School of Engineering and Applied
Sciences
Current
Past Experiences
13. February – November in 2017
17
台塑石化
長春石化
奇美實業
英業達
欣興電子
敬鵬工業
可成科技
致茂電子
永進機械
研華科技
農科院
紡織所
聯發科技
台積電
宏遠紡織
台元紡織
佳和紡織
強盛染整
臺灣塑膠
龍鼎蘭花
經緯航太科技
14. Unmet “Soft” Needs for Nurturing
Next-Generation Industries in the AI Era
Human resource development
Machine learning experts with hands-on experiences
Problem/opportunity identification
Problem identification is the biggest challenge for newcomers
Business transformation
Problem identification and solving strategies
Spin-offs/R&D initiatives
Shared technology infrastructure
Knowledge base, datasets and baseline practices
18
18. 人工智慧發展策略建議
A sad story that AI-assisted AOI can help avoid
14 suicide events in 2010 at Foxconn China factories
Only 2 of the suicides survived
Industry-wide Problem #1:
Human Operators for Optical Inspection
https://theinitium.com/article/20170802-mainland-Foxconn-factorygirl/
19. 人工智慧發展策略建議
Human Operators for Optical Inspection
The factories recruit only workers under 29 years old
Their work involve checking scratches on consumer
products (likely Apple iPhone) for 2,880 times a day
This means 4 times per minute assuming 12 working
hours per day
https://theinitium.com/article/20170802-mainland-Foxconn-factorygirl/
25. Deep Neural Networks
Deep Convolution Neural Networks
Transfer learning
Pre-trained using 14-million image dataset
Resnet with > 8-million parameters
Input images Model training / inference
OK
OK
26. Case study –
Human vs. Neural Inspection
33
4 human inspectors for 23 product lines
Throughput: 300K patches per human per day =
1.2M patches per day
Leakage rate between 5% to 10% while False alarm
rate > 10%
Human
AI
Equipment: A PC with NVIDIA GeForce 1080Ti
(4,000 USD)
Throughput: 167 patches per second = 10 K patches
per minute = 14M patches per day
Leakage rate < 0.01% while False alarm rate < 5%
28. 人工智慧發展策略建議
Case Study: A Chemical Process
12 parameters
Hydrogen (H)
Catalyst
Ethylene (C2H4), Ethane (C2H6), Butene (C4H8)
Pressure, temperature, fluid level, and so on
Output
A quality index of a certain chemical product
36
30. Residual networks
Very similar to Residual network in
Image classification
main stream + residuals
38
Residual network reference
Cardiologist-Level Arrhythmia Detection with Convolutional
Neural Networks, Pranav et.al., 2017
32. 人工智慧發展策略建議
Industry-wide Problem #3:
Predictive maintenance
Especially important for equipment with high failure cost (such as
motors in machine tools)
Also important for expensive consumables (such as blades used in
precision cutting machines)
40
40. Challenges for AI
Development in Taiwan
Wide gap between academia and
industry
Lack of experienced talents
Used to adopt rather than develop
technology
42. http://aiacademy.tw/
Address the “lack of AI talents”
problem
Offer short, intensive and scalable
training courses
Aim to train >= 1500 talents each
year
44. 72
Corporate Partner Program
Corporates provide real-life problems (and
datasets)
Students tackle these problems as term projects
Corporates may recruit students after they finish
the training courses
45. Current class design
Elite Engineer Class (技術領袖培訓班)
12 weeks
9am to 6pm on Monday to Friday
Lectures + hands-on sessions + term projects
Mid-term and final exams
Manager Class (經理人周末研修班)
12 weeks
9am to 9pm each Saturday
Lectures only
73
46. Elite Engineer Class
74
Applications due on Dec 4, 2017. Nearly
500 applicants registered while we can
only accept 208 students.
Two-step filtering:
1. Document review
2. Entrance exam: calculus, linear
algebra, probability, statistics,
programming