SlideShare a Scribd company logo
Trista Chen, Chief Scientist of Machine Learning of Inventec
Wei-Chao Chen, Co-founder of Skywatch & Head of Inventec AI Center
{chen.trista, chen.wei-chao} @ inventec.com
S9617
About Inventec
• Public company (Since 1975; Taipei, Taiwan)
• Tier 1 electronics manufacturer
• Annual revenue USD $16B+ (2018)
• Factories in Taiwan, Shanghai and Chongqing
GTC 2019
Personal Computers
Laptops
Enterprise Computers
Servers
Solar Energy Smart Devices
Medical Devices
AI for Smart Manufacturing
• Process Automation
– Automatic optical inspection
– Production scheduling
– Automatic testing
• Predictive Analysis
– Order forecast
– Production / yield prediction
– Predictive maintenance
GTC 2019
Smart Manufacturing
Automatic optical inspection
Order forecast
Production scheduling
Test automation
…
Future Product
AIOT / Smart City applications
Medical devices
Robotics
Computation Platform
….
Inventec AI Center
Why Edge AI?
• Privacy
• Reliability
• Network Efficiency
GTC 2019
JETSON TX2 MODULE
•NVIDIA Pascal™ Architecture GPU
•2 Denver 64-bit CPUs + Quad-Core A57
Complex
•8GB L128 bit DDR4 Memory
•32GB eMMC 5.1 Flash Storage
•Connectivity to 802.11ac Wi-Fi and Bluetooth-
Enabled Devices
•10/100/1000BASE-T Ethernet
GTC 2019
Laptop Testing Robot
A0I
Re-inspection
Laptop Surface AOI
Re-inspectionAOI for SMT lines and PCBs
GTC 2019
Laptop Testing Robot
A0I
Re-inspection
Laptop Surface AOI
Re-inspectionAOI for SMT lines and PCBs
1. Laptop Testing Robot
Background
• Millions of laptops manufactured every month
– Similar basic design, many different configurations
– Production tests are mostly automatic
GTC 2019
1. Laptop Testing Robot
Background
• Product approval is labor intensive
– Takes weeks to certify for mass production
– ~1M test assets
– Hundreds of people running the tests
GTC 2019
1. Laptop Testing Robot
Motivation (least -> most important)
• Labor cost and management
– Skilled and trustworthy testers are difficult to find
• Time-to-market
– Regression tests take weeks / months to complete
• Increase confidence
– The management confidence level on manual testing is only
around 70%
GTC 2019
0. 3
3
0 0 . / 0
0
0
0 3
GTC 2019
By: Jimmy Ou, Wei-Chao Chen, et al. of Skywatch
Collaboration with Joseph Shi, Jack Hung & Inventec QA Engineers
1. Laptop Testing Robot
GTC 2019
1. Laptop Testing Robot
Pneumatic
actuators
Mechanical / electrical keyboard and mouse
GTC 2019
1. Laptop Testing Robot
Camera with computer vision capability
Asideoflaptop
Industrial Camera
1. Laptop Testing Robot
• Computer Vision (OpenCV)
– Perspective correction
– Edge detection & denoising
– Template matching
• Tesseract OCR
– LSTM-based NN OCR engine
GTC 2019
1. Laptop Testing Robot
GTC 2019
Test subject Controller
4X real-time demo
1. Laptop Testing Robot
Future Opportunities
• Automatic testing script generation
– from human-readable to machine-readable test cases
• Scheduling optimization
– Reshuffle test levels and orders to speed up process
• Smart random testing
– Aka monkey testing
GTC 2019
GTC 2019
Laptop Testing Robot
A0I
Re-inspection
Laptop Surface AOI
Re-inspectionAOI for SMT lines and PCBs
2: AOI (Automated Optical Inspection) Re-inspection
GTC 2019
Common SMT defects
2: AOI Re-inspection
AOI Machine
Fail
Pass
GTC 2019
High false positive
Low first-pass yield
Re-inspection
by human
2: AOI Re-inspection
AOI Machine
Pass
Fail AOI
Re-inspection
Pass
Fail
GTC 2019
By: , , of NTU & advised by Wei-Chao Chen of Skywatch & NTU & Trista Chen of Inventec
Data & domain knowledge provided by Jessie Huang, Peter Chu @ TAO, among others
High false positive
Low first-pass yield
2: AOI Re-inspection
GTC 2019
Traditional
Computer Vision
defects w/ well-defined rules
Deep Learning
defects w/ intuitive descriptions
2: AOI Re-inspection
GTC 2019
0. Excess Solder (true)
1. Excess Solder
(false positive à pass )
2. Insuff. Solder (true)
2. Insuff. Solder
(false positiveàpass)
GTC 2019
Laptop Testing Robot
A0I
Re-inspection
Laptop Surface AOI
Re-inspectionAOI for SMT lines and PCBs
3. Laptop Surface AOI
GTC 2019
Brand
Logo
Customer
Info
Brand
Logo
GTC 2019
Customer
Info
Brand
Logo
GTC 2019
Customer
Info
Type #2
Dot
Type #1
Scratch
Type #3:
Area
Type #4:
Abrasion
Type #5:
Don’t care
3. Laptop Surface AOI
GTC 2019
Defect type Class A spec Class B spec
Scratch Length: 12mm
Acceptable: 2 lines
Length: 20mm, Acceptable: 2
lines
Dent 0.5 mm2 < size < 0.7mm2
Acceptable: 2 points
0.5 mm2 < size < 1mm2
Acceptable: 3 points
…
Fail
Pass
Human readable rules
Labor intensive
3. Laptop Surface AOI
GTC 2019
Defect type Class A spec Class B spec
Scratch Length: 12mm
Acceptable: 2 lines
Length: 20mm, Acceptable: 2
lines
Dent 0.5 mm2 < size < 0.7mm2
Acceptable: 2 points
0.5 mm2 < size < 1mm2
Acceptable: 3 points
…
Fail
Pass
Labor intensive
Human readable rules
AI
3. Laptop Surface AOI
GTC 2019
Fail
Pass
AI
Training data
3. Laptop Surface AOI
Pass
Fail
AIMobile M1 (Nvidia TX2)
GTC 2019
By: Benson Lin of Skywatch, Trista Chen and Irene Chen of Inventec, Wei-Chao Chen of Skywatch & Inventec
In collaboration with Steven Wang, Sing-Wang Chen, Alfa Shih & Tim Zhang et al @ ICC
Machine manufactured by Jerry Tseng @ Leh-Yeh; Edge machine provided by Mark Lu @ AIMobile
Surface defect classification
3. Laptop Surface AOI
Pass/Fail Classification
DL
detection
Pass
Fail
AIMobile M1 (Nvidia TX2)
GTC 2019
3. Laptop Surface AOI
Pass/Fail Classification
DL
detection Pass
Fail
GTC 2019
Stage 1:
Explainable
Stage 2:
Adjustable
3. Laptop Surface AOI
Pass/Fail Classification
DL
detection Pass
Fail
GTC 2019
Stage 1:
Explainable
Stage 2:
Adjustable
3. Laptop Surface AOI – Explainable
GTC 2019
Why did it pass/fail?
• Defect type
• Defect count
• Defect size
Defect type Class A spec Class B spec
Scratch Length: 12mm
Acceptable: 2 lines
Length: 20mm, Acceptable: 2
lines
Dent 0.5 mm2 < size < 0.7mm2
Acceptable: 2 points
0.5 mm2 < size < 1mm2
Acceptable: 3 points
…
Defect class
cluster analysis
Defect types:
s1 s2 s3 s4 s5
5 classes
a b c
a b c
a b c
a b c
a b c
15 classes
GTC 2019
Stage-1 multi-class detector
5 classes
?
?
GTC 2019
a b c
a b c
a b c
a b c
a b c
Defect types:
s1 s2 s3 s4 s5
15 classes
Laptop categories:
a, b, c
Defect types:
s1 s2 s3 s4 s5
Output
Input
s5 s4 s1 s3
s2
5 classes
Defect class
cluster analysis
3. Laptop Surface AOI
Pass/Fail Classification
DL
detection Pass
Fail
GTC 2019
Stage 1:
Explainable
Stage 2:
Adjustable
3. Laptop Surface AOI – Adjustable
GTC 2019
Class A
2 lines 2 dots
3. Laptop Surface AOI – Adjustable
GTC 2019
Class B
3. Laptop Surface AOI – Adjustable
GTC 2019
Class C
3. Laptop Surface AOI – Adjustable
GTC 2019
Class A
3. Laptop Surface AOI – Adjustable
GTC 2019
Class A+
3. Laptop Surface AOI – Adjustable
GTC 2019
Class A++
3. Laptop Surface AOI
Pass/Fail Classification
DL
detection Pass
Fail
GTC 2019
Stage 1:
Explainable
Stage 2:
Adjustable
3. Laptop Surface AOI – Factory Prototype
GTC 2019
CCD
NB
Benefits of AI for Manufacturing
GTC 2019
• The Big Scope
– Consistency
– Faithful digital
record
– Industry 4.0
OEE
SECS
MES
SECS
1.
2.
3.
4.
MES
/ /
/ /…
/ / &
Idle
1.
2.
3.
4.
1.
2. MES/
3.
4. /
Repair
Source: Taiwan’s Institute for Information Industry (III) 2017
• The Obvious
– Accuracy
– Labor Saving ROI
Benefits of AI for Manufacturing – I4.0
GTC 2019
95.00
96.00
97.00
98.00
99.00
100.00
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Production Efficiency
Peak Productivity Identified Problem
Image source: Hitachi “Factories of the Future”, NEXT 2019
time
Process chain
GTC 2019
Laptop Testing Robot
A0I
Re-inspection
Laptop Surface AOI
Re-inspectionAOI for SMT lines and PCBs
Unexpected Bonus:
to do well while building your AI engine
We’re on a mission to connect 60,000 residential disabled in Taiwan to join
AI work by providing high-quality data fuels to your AI engines.
GTC 2019
Source: Christopher Porter’s Roomba and Cat from Flickr
GTC 2019
Trista Chen, Chief Scientist of Machine Learning of Inventec
Wei-Chao Chen, Co-founder of Skywatch & Head of Inventec AI Center
{chen.trista, chen.wei-chao} @ inventec.com
S9617

More Related Content

What's hot

State of AI Report 2022 - ONLINE.pdf
State of AI Report 2022 - ONLINE.pdfState of AI Report 2022 - ONLINE.pdf
State of AI Report 2022 - ONLINE.pdf
vizologi
 
AI and future Jobs
AI and future JobsAI and future Jobs
AI and future Jobs
David Asirvatham
 
Explainable AI in Industry (WWW 2020 Tutorial)
Explainable AI in Industry (WWW 2020 Tutorial)Explainable AI in Industry (WWW 2020 Tutorial)
Explainable AI in Industry (WWW 2020 Tutorial)
Krishnaram Kenthapadi
 
Introduction to Knowledge Graphs and Semantic AI
Introduction to Knowledge Graphs and Semantic AIIntroduction to Knowledge Graphs and Semantic AI
Introduction to Knowledge Graphs and Semantic AI
Semantic Web Company
 
AI in manufacturing
AI in manufacturingAI in manufacturing
AI in manufacturing
Stephanie Locke
 
Artificial Intelligence - A Brief Scan
Artificial Intelligence - A Brief ScanArtificial Intelligence - A Brief Scan
Artificial Intelligence - A Brief Scan
Hai Siang Chia
 
Artificial intelligence: Driving future growth in Singapore- Accenture
Artificial intelligence: Driving future growth in Singapore- AccentureArtificial intelligence: Driving future growth in Singapore- Accenture
Artificial intelligence: Driving future growth in Singapore- Accenture
Accenture ASEAN
 
Robotic Process Automation (RPA)
Robotic Process Automation (RPA)Robotic Process Automation (RPA)
Robotic Process Automation (RPA)
Flevy.com Best Practices
 
AI for safety critical systems.pdf
AI for safety critical systems.pdfAI for safety critical systems.pdf
AI for safety critical systems.pdf
020PavithraSK
 
ALL ABOUT ARTIFICIAL INTELLIGENCE
ALL  ABOUT ARTIFICIAL INTELLIGENCEALL  ABOUT ARTIFICIAL INTELLIGENCE
ALL ABOUT ARTIFICIAL INTELLIGENCE
Kiran Mandrawadkar
 
人工智慧下的AOI變革浪潮:影像辨識技術的突破與新契機
人工智慧下的AOI變革浪潮:影像辨識技術的突破與新契機人工智慧下的AOI變革浪潮:影像辨識技術的突破與新契機
人工智慧下的AOI變革浪潮:影像辨識技術的突破與新契機
Albert Y. C. Chen
 
AI Readiness: Five Areas Business Must Prepare for Success in Artificial Inte...
AI Readiness: Five Areas Business Must Prepare for Success in Artificial Inte...AI Readiness: Five Areas Business Must Prepare for Success in Artificial Inte...
AI Readiness: Five Areas Business Must Prepare for Success in Artificial Inte...
Kaleido Insights
 
AI in science.pptx
AI in science.pptxAI in science.pptx
AI in science.pptx
Dr Ashok Tiwari
 
Artificial Intelligence Machine Learning Deep Learning Ppt Powerpoint Present...
Artificial Intelligence Machine Learning Deep Learning Ppt Powerpoint Present...Artificial Intelligence Machine Learning Deep Learning Ppt Powerpoint Present...
Artificial Intelligence Machine Learning Deep Learning Ppt Powerpoint Present...
SlideTeam
 
BOSTON DYNAMICS by YASH BIHANI
BOSTON DYNAMICS by YASH BIHANIBOSTON DYNAMICS by YASH BIHANI
BOSTON DYNAMICS by YASH BIHANI
Yashbihani4
 
Semantic AI
Semantic AISemantic AI
Introduction to AI & ML
Introduction to AI & MLIntroduction to AI & ML
Introduction to AI & ML
Mandy Sidana
 
AI.pptx
AI.pptxAI.pptx
AI.pptx
tania589588
 
Lesson 1 intro to ai
Lesson 1   intro to aiLesson 1   intro to ai
Lesson 1 intro to ai
ankit_ppt
 
Graphs in Automotive and Manufacturing - Unlock New Value from Your Data
Graphs in Automotive and Manufacturing - Unlock New Value from Your DataGraphs in Automotive and Manufacturing - Unlock New Value from Your Data
Graphs in Automotive and Manufacturing - Unlock New Value from Your Data
Neo4j
 

What's hot (20)

State of AI Report 2022 - ONLINE.pdf
State of AI Report 2022 - ONLINE.pdfState of AI Report 2022 - ONLINE.pdf
State of AI Report 2022 - ONLINE.pdf
 
AI and future Jobs
AI and future JobsAI and future Jobs
AI and future Jobs
 
Explainable AI in Industry (WWW 2020 Tutorial)
Explainable AI in Industry (WWW 2020 Tutorial)Explainable AI in Industry (WWW 2020 Tutorial)
Explainable AI in Industry (WWW 2020 Tutorial)
 
Introduction to Knowledge Graphs and Semantic AI
Introduction to Knowledge Graphs and Semantic AIIntroduction to Knowledge Graphs and Semantic AI
Introduction to Knowledge Graphs and Semantic AI
 
AI in manufacturing
AI in manufacturingAI in manufacturing
AI in manufacturing
 
Artificial Intelligence - A Brief Scan
Artificial Intelligence - A Brief ScanArtificial Intelligence - A Brief Scan
Artificial Intelligence - A Brief Scan
 
Artificial intelligence: Driving future growth in Singapore- Accenture
Artificial intelligence: Driving future growth in Singapore- AccentureArtificial intelligence: Driving future growth in Singapore- Accenture
Artificial intelligence: Driving future growth in Singapore- Accenture
 
Robotic Process Automation (RPA)
Robotic Process Automation (RPA)Robotic Process Automation (RPA)
Robotic Process Automation (RPA)
 
AI for safety critical systems.pdf
AI for safety critical systems.pdfAI for safety critical systems.pdf
AI for safety critical systems.pdf
 
ALL ABOUT ARTIFICIAL INTELLIGENCE
ALL  ABOUT ARTIFICIAL INTELLIGENCEALL  ABOUT ARTIFICIAL INTELLIGENCE
ALL ABOUT ARTIFICIAL INTELLIGENCE
 
人工智慧下的AOI變革浪潮:影像辨識技術的突破與新契機
人工智慧下的AOI變革浪潮:影像辨識技術的突破與新契機人工智慧下的AOI變革浪潮:影像辨識技術的突破與新契機
人工智慧下的AOI變革浪潮:影像辨識技術的突破與新契機
 
AI Readiness: Five Areas Business Must Prepare for Success in Artificial Inte...
AI Readiness: Five Areas Business Must Prepare for Success in Artificial Inte...AI Readiness: Five Areas Business Must Prepare for Success in Artificial Inte...
AI Readiness: Five Areas Business Must Prepare for Success in Artificial Inte...
 
AI in science.pptx
AI in science.pptxAI in science.pptx
AI in science.pptx
 
Artificial Intelligence Machine Learning Deep Learning Ppt Powerpoint Present...
Artificial Intelligence Machine Learning Deep Learning Ppt Powerpoint Present...Artificial Intelligence Machine Learning Deep Learning Ppt Powerpoint Present...
Artificial Intelligence Machine Learning Deep Learning Ppt Powerpoint Present...
 
BOSTON DYNAMICS by YASH BIHANI
BOSTON DYNAMICS by YASH BIHANIBOSTON DYNAMICS by YASH BIHANI
BOSTON DYNAMICS by YASH BIHANI
 
Semantic AI
Semantic AISemantic AI
Semantic AI
 
Introduction to AI & ML
Introduction to AI & MLIntroduction to AI & ML
Introduction to AI & ML
 
AI.pptx
AI.pptxAI.pptx
AI.pptx
 
Lesson 1 intro to ai
Lesson 1   intro to aiLesson 1   intro to ai
Lesson 1 intro to ai
 
Graphs in Automotive and Manufacturing - Unlock New Value from Your Data
Graphs in Automotive and Manufacturing - Unlock New Value from Your DataGraphs in Automotive and Manufacturing - Unlock New Value from Your Data
Graphs in Automotive and Manufacturing - Unlock New Value from Your Data
 

Similar to Edge AI Smart Manufacturing - Defect Detection and Beyond (GTC 2019)

Kojin mould case study
Kojin mould case studyKojin mould case study
Kojin mould case study
Liu PeiLing
 
IRJET- New Generation Multilevel based Atm Security System
IRJET- New Generation Multilevel based Atm Security SystemIRJET- New Generation Multilevel based Atm Security System
IRJET- New Generation Multilevel based Atm Security System
IRJET Journal
 
Accelerating the AIoT @ the EDGE
Accelerating the AIoT @ the EDGE Accelerating the AIoT @ the EDGE
Accelerating the AIoT @ the EDGE
Amazon Web Services
 
IRJET- Design of Mini CNC using Arduino Uno
IRJET-  	  Design of Mini CNC using Arduino UnoIRJET-  	  Design of Mini CNC using Arduino Uno
IRJET- Design of Mini CNC using Arduino Uno
IRJET Journal
 
IRJET - New Generation Multilevel based Atm Security System
IRJET - New Generation Multilevel based Atm Security SystemIRJET - New Generation Multilevel based Atm Security System
IRJET - New Generation Multilevel based Atm Security System
IRJET Journal
 
Clobbi CEO Dmitry Shapovalov Keynote @CRU 2019 Brussels "Practical case-studi...
Clobbi CEO Dmitry Shapovalov Keynote @CRU 2019 Brussels "Practical case-studi...Clobbi CEO Dmitry Shapovalov Keynote @CRU 2019 Brussels "Practical case-studi...
Clobbi CEO Dmitry Shapovalov Keynote @CRU 2019 Brussels "Practical case-studi...
Olexiy Lyzun
 
Clobbi CEO Dmitry Shapovalov Keynote @CRU 2019 Brussels "Practical case-studi...
Clobbi CEO Dmitry Shapovalov Keynote @CRU 2019 Brussels "Practical case-studi...Clobbi CEO Dmitry Shapovalov Keynote @CRU 2019 Brussels "Practical case-studi...
Clobbi CEO Dmitry Shapovalov Keynote @CRU 2019 Brussels "Practical case-studi...
Clobbi
 
Clobbi CEO Dmitry Shapovalov Keynote @CRU 2019 Brussels "Practical case-studi...
Clobbi CEO Dmitry Shapovalov Keynote @CRU 2019 Brussels "Practical case-studi...Clobbi CEO Dmitry Shapovalov Keynote @CRU 2019 Brussels "Practical case-studi...
Clobbi CEO Dmitry Shapovalov Keynote @CRU 2019 Brussels "Practical case-studi...
Clobbi
 
KEYNOTE: Edge optimized architecture for fabric defect detection in real-time
KEYNOTE: Edge optimized architecture for fabric defect detection in real-timeKEYNOTE: Edge optimized architecture for fabric defect detection in real-time
KEYNOTE: Edge optimized architecture for fabric defect detection in real-time
Shuquan Huang
 
IRJET- Metal Crack Detection using Image Processing
IRJET- Metal Crack Detection using Image ProcessingIRJET- Metal Crack Detection using Image Processing
IRJET- Metal Crack Detection using Image Processing
IRJET Journal
 
IRJET - Implementation of SDC: Self-Driving Car based on Raspberry Pi
IRJET - Implementation of SDC: Self-Driving Car based on Raspberry PiIRJET - Implementation of SDC: Self-Driving Car based on Raspberry Pi
IRJET - Implementation of SDC: Self-Driving Car based on Raspberry Pi
IRJET Journal
 
Making Use of a Knowledgeable Design by Design for Manufacturing
Making Use of a Knowledgeable Design by Design for ManufacturingMaking Use of a Knowledgeable Design by Design for Manufacturing
Making Use of a Knowledgeable Design by Design for Manufacturing
VayoInfo
 
Procedure of Proactively Designing Products by Design for Manufacturability (...
Procedure of Proactively Designing Products by Design for Manufacturability (...Procedure of Proactively Designing Products by Design for Manufacturability (...
Procedure of Proactively Designing Products by Design for Manufacturability (...
VayoInfo
 
DevOps Fest 2020. Pavlo Repalo. Edge Computing: Appliance and Challanges
DevOps Fest 2020. Pavlo Repalo. Edge Computing: Appliance and ChallangesDevOps Fest 2020. Pavlo Repalo. Edge Computing: Appliance and Challanges
DevOps Fest 2020. Pavlo Repalo. Edge Computing: Appliance and Challanges
DevOps_Fest
 
B Kindilien Finding Efficiency In Mach 120408
B Kindilien Finding Efficiency In Mach 120408B Kindilien Finding Efficiency In Mach 120408
B Kindilien Finding Efficiency In Mach 120408
jgIpotiwon
 
Bulletproof Development
Bulletproof DevelopmentBulletproof Development
Bulletproof Development
Rob Betcher
 
MIPI DevCon Seoul 2018: Mobile Technologies for a Smart World
MIPI DevCon Seoul 2018: Mobile Technologies for a Smart World MIPI DevCon Seoul 2018: Mobile Technologies for a Smart World
MIPI DevCon Seoul 2018: Mobile Technologies for a Smart World
MIPI Alliance
 
Bluetooth Controlled Garbage Collection Robot Arm
Bluetooth Controlled Garbage Collection Robot ArmBluetooth Controlled Garbage Collection Robot Arm
Bluetooth Controlled Garbage Collection Robot Arm
IRJET Journal
 

Similar to Edge AI Smart Manufacturing - Defect Detection and Beyond (GTC 2019) (20)

Kojin mould case study
Kojin mould case studyKojin mould case study
Kojin mould case study
 
IRJET- New Generation Multilevel based Atm Security System
IRJET- New Generation Multilevel based Atm Security SystemIRJET- New Generation Multilevel based Atm Security System
IRJET- New Generation Multilevel based Atm Security System
 
Accelerating the AIoT @ the EDGE
Accelerating the AIoT @ the EDGE Accelerating the AIoT @ the EDGE
Accelerating the AIoT @ the EDGE
 
IRJET- Design of Mini CNC using Arduino Uno
IRJET-  	  Design of Mini CNC using Arduino UnoIRJET-  	  Design of Mini CNC using Arduino Uno
IRJET- Design of Mini CNC using Arduino Uno
 
IRJET - New Generation Multilevel based Atm Security System
IRJET - New Generation Multilevel based Atm Security SystemIRJET - New Generation Multilevel based Atm Security System
IRJET - New Generation Multilevel based Atm Security System
 
Clobbi CEO Dmitry Shapovalov Keynote @CRU 2019 Brussels "Practical case-studi...
Clobbi CEO Dmitry Shapovalov Keynote @CRU 2019 Brussels "Practical case-studi...Clobbi CEO Dmitry Shapovalov Keynote @CRU 2019 Brussels "Practical case-studi...
Clobbi CEO Dmitry Shapovalov Keynote @CRU 2019 Brussels "Practical case-studi...
 
Clobbi CEO Dmitry Shapovalov Keynote @CRU 2019 Brussels "Practical case-studi...
Clobbi CEO Dmitry Shapovalov Keynote @CRU 2019 Brussels "Practical case-studi...Clobbi CEO Dmitry Shapovalov Keynote @CRU 2019 Brussels "Practical case-studi...
Clobbi CEO Dmitry Shapovalov Keynote @CRU 2019 Brussels "Practical case-studi...
 
Clobbi CEO Dmitry Shapovalov Keynote @CRU 2019 Brussels "Practical case-studi...
Clobbi CEO Dmitry Shapovalov Keynote @CRU 2019 Brussels "Practical case-studi...Clobbi CEO Dmitry Shapovalov Keynote @CRU 2019 Brussels "Practical case-studi...
Clobbi CEO Dmitry Shapovalov Keynote @CRU 2019 Brussels "Practical case-studi...
 
KEYNOTE: Edge optimized architecture for fabric defect detection in real-time
KEYNOTE: Edge optimized architecture for fabric defect detection in real-timeKEYNOTE: Edge optimized architecture for fabric defect detection in real-time
KEYNOTE: Edge optimized architecture for fabric defect detection in real-time
 
IRJET- Metal Crack Detection using Image Processing
IRJET- Metal Crack Detection using Image ProcessingIRJET- Metal Crack Detection using Image Processing
IRJET- Metal Crack Detection using Image Processing
 
IRJET - Implementation of SDC: Self-Driving Car based on Raspberry Pi
IRJET - Implementation of SDC: Self-Driving Car based on Raspberry PiIRJET - Implementation of SDC: Self-Driving Car based on Raspberry Pi
IRJET - Implementation of SDC: Self-Driving Car based on Raspberry Pi
 
Making Use of a Knowledgeable Design by Design for Manufacturing
Making Use of a Knowledgeable Design by Design for ManufacturingMaking Use of a Knowledgeable Design by Design for Manufacturing
Making Use of a Knowledgeable Design by Design for Manufacturing
 
Procedure of Proactively Designing Products by Design for Manufacturability (...
Procedure of Proactively Designing Products by Design for Manufacturability (...Procedure of Proactively Designing Products by Design for Manufacturability (...
Procedure of Proactively Designing Products by Design for Manufacturability (...
 
DevOps Fest 2020. Pavlo Repalo. Edge Computing: Appliance and Challanges
DevOps Fest 2020. Pavlo Repalo. Edge Computing: Appliance and ChallangesDevOps Fest 2020. Pavlo Repalo. Edge Computing: Appliance and Challanges
DevOps Fest 2020. Pavlo Repalo. Edge Computing: Appliance and Challanges
 
B Kindilien Finding Efficiency In Mach 120408
B Kindilien Finding Efficiency In Mach 120408B Kindilien Finding Efficiency In Mach 120408
B Kindilien Finding Efficiency In Mach 120408
 
STS Platform
STS PlatformSTS Platform
STS Platform
 
Bulletproof Development
Bulletproof DevelopmentBulletproof Development
Bulletproof Development
 
MIPI DevCon Seoul 2018: Mobile Technologies for a Smart World
MIPI DevCon Seoul 2018: Mobile Technologies for a Smart World MIPI DevCon Seoul 2018: Mobile Technologies for a Smart World
MIPI DevCon Seoul 2018: Mobile Technologies for a Smart World
 
RESUME.
RESUME.RESUME.
RESUME.
 
Bluetooth Controlled Garbage Collection Robot Arm
Bluetooth Controlled Garbage Collection Robot ArmBluetooth Controlled Garbage Collection Robot Arm
Bluetooth Controlled Garbage Collection Robot Arm
 

Recently uploaded

Generating a custom Ruby SDK for your web service or Rails API using Smithy
Generating a custom Ruby SDK for your web service or Rails API using SmithyGenerating a custom Ruby SDK for your web service or Rails API using Smithy
Generating a custom Ruby SDK for your web service or Rails API using Smithy
g2nightmarescribd
 
FIDO Alliance Osaka Seminar: Overview.pdf
FIDO Alliance Osaka Seminar: Overview.pdfFIDO Alliance Osaka Seminar: Overview.pdf
FIDO Alliance Osaka Seminar: Overview.pdf
FIDO Alliance
 
Software Delivery At the Speed of AI: Inflectra Invests In AI-Powered Quality
Software Delivery At the Speed of AI: Inflectra Invests In AI-Powered QualitySoftware Delivery At the Speed of AI: Inflectra Invests In AI-Powered Quality
Software Delivery At the Speed of AI: Inflectra Invests In AI-Powered Quality
Inflectra
 
Kubernetes & AI - Beauty and the Beast !?! @KCD Istanbul 2024
Kubernetes & AI - Beauty and the Beast !?! @KCD Istanbul 2024Kubernetes & AI - Beauty and the Beast !?! @KCD Istanbul 2024
Kubernetes & AI - Beauty and the Beast !?! @KCD Istanbul 2024
Tobias Schneck
 
Unsubscribed: Combat Subscription Fatigue With a Membership Mentality by Head...
Unsubscribed: Combat Subscription Fatigue With a Membership Mentality by Head...Unsubscribed: Combat Subscription Fatigue With a Membership Mentality by Head...
Unsubscribed: Combat Subscription Fatigue With a Membership Mentality by Head...
Product School
 
FIDO Alliance Osaka Seminar: Passkeys at Amazon.pdf
FIDO Alliance Osaka Seminar: Passkeys at Amazon.pdfFIDO Alliance Osaka Seminar: Passkeys at Amazon.pdf
FIDO Alliance Osaka Seminar: Passkeys at Amazon.pdf
FIDO Alliance
 
From Siloed Products to Connected Ecosystem: Building a Sustainable and Scala...
From Siloed Products to Connected Ecosystem: Building a Sustainable and Scala...From Siloed Products to Connected Ecosystem: Building a Sustainable and Scala...
From Siloed Products to Connected Ecosystem: Building a Sustainable and Scala...
Product School
 
Builder.ai Founder Sachin Dev Duggal's Strategic Approach to Create an Innova...
Builder.ai Founder Sachin Dev Duggal's Strategic Approach to Create an Innova...Builder.ai Founder Sachin Dev Duggal's Strategic Approach to Create an Innova...
Builder.ai Founder Sachin Dev Duggal's Strategic Approach to Create an Innova...
Ramesh Iyer
 
Elevating Tactical DDD Patterns Through Object Calisthenics
Elevating Tactical DDD Patterns Through Object CalisthenicsElevating Tactical DDD Patterns Through Object Calisthenics
Elevating Tactical DDD Patterns Through Object Calisthenics
Dorra BARTAGUIZ
 
GDG Cloud Southlake #33: Boule & Rebala: Effective AppSec in SDLC using Deplo...
GDG Cloud Southlake #33: Boule & Rebala: Effective AppSec in SDLC using Deplo...GDG Cloud Southlake #33: Boule & Rebala: Effective AppSec in SDLC using Deplo...
GDG Cloud Southlake #33: Boule & Rebala: Effective AppSec in SDLC using Deplo...
James Anderson
 
From Daily Decisions to Bottom Line: Connecting Product Work to Revenue by VP...
From Daily Decisions to Bottom Line: Connecting Product Work to Revenue by VP...From Daily Decisions to Bottom Line: Connecting Product Work to Revenue by VP...
From Daily Decisions to Bottom Line: Connecting Product Work to Revenue by VP...
Product School
 
UiPath Test Automation using UiPath Test Suite series, part 4
UiPath Test Automation using UiPath Test Suite series, part 4UiPath Test Automation using UiPath Test Suite series, part 4
UiPath Test Automation using UiPath Test Suite series, part 4
DianaGray10
 
Bits & Pixels using AI for Good.........
Bits & Pixels using AI for Good.........Bits & Pixels using AI for Good.........
Bits & Pixels using AI for Good.........
Alison B. Lowndes
 
Smart TV Buyer Insights Survey 2024 by 91mobiles.pdf
Smart TV Buyer Insights Survey 2024 by 91mobiles.pdfSmart TV Buyer Insights Survey 2024 by 91mobiles.pdf
Smart TV Buyer Insights Survey 2024 by 91mobiles.pdf
91mobiles
 
Encryption in Microsoft 365 - ExpertsLive Netherlands 2024
Encryption in Microsoft 365 - ExpertsLive Netherlands 2024Encryption in Microsoft 365 - ExpertsLive Netherlands 2024
Encryption in Microsoft 365 - ExpertsLive Netherlands 2024
Albert Hoitingh
 
GenAISummit 2024 May 28 Sri Ambati Keynote: AGI Belongs to The Community in O...
GenAISummit 2024 May 28 Sri Ambati Keynote: AGI Belongs to The Community in O...GenAISummit 2024 May 28 Sri Ambati Keynote: AGI Belongs to The Community in O...
GenAISummit 2024 May 28 Sri Ambati Keynote: AGI Belongs to The Community in O...
Sri Ambati
 
Mission to Decommission: Importance of Decommissioning Products to Increase E...
Mission to Decommission: Importance of Decommissioning Products to Increase E...Mission to Decommission: Importance of Decommissioning Products to Increase E...
Mission to Decommission: Importance of Decommissioning Products to Increase E...
Product School
 
Monitoring Java Application Security with JDK Tools and JFR Events
Monitoring Java Application Security with JDK Tools and JFR EventsMonitoring Java Application Security with JDK Tools and JFR Events
Monitoring Java Application Security with JDK Tools and JFR Events
Ana-Maria Mihalceanu
 
FIDO Alliance Osaka Seminar: FIDO Security Aspects.pdf
FIDO Alliance Osaka Seminar: FIDO Security Aspects.pdfFIDO Alliance Osaka Seminar: FIDO Security Aspects.pdf
FIDO Alliance Osaka Seminar: FIDO Security Aspects.pdf
FIDO Alliance
 
Leading Change strategies and insights for effective change management pdf 1.pdf
Leading Change strategies and insights for effective change management pdf 1.pdfLeading Change strategies and insights for effective change management pdf 1.pdf
Leading Change strategies and insights for effective change management pdf 1.pdf
OnBoard
 

Recently uploaded (20)

Generating a custom Ruby SDK for your web service or Rails API using Smithy
Generating a custom Ruby SDK for your web service or Rails API using SmithyGenerating a custom Ruby SDK for your web service or Rails API using Smithy
Generating a custom Ruby SDK for your web service or Rails API using Smithy
 
FIDO Alliance Osaka Seminar: Overview.pdf
FIDO Alliance Osaka Seminar: Overview.pdfFIDO Alliance Osaka Seminar: Overview.pdf
FIDO Alliance Osaka Seminar: Overview.pdf
 
Software Delivery At the Speed of AI: Inflectra Invests In AI-Powered Quality
Software Delivery At the Speed of AI: Inflectra Invests In AI-Powered QualitySoftware Delivery At the Speed of AI: Inflectra Invests In AI-Powered Quality
Software Delivery At the Speed of AI: Inflectra Invests In AI-Powered Quality
 
Kubernetes & AI - Beauty and the Beast !?! @KCD Istanbul 2024
Kubernetes & AI - Beauty and the Beast !?! @KCD Istanbul 2024Kubernetes & AI - Beauty and the Beast !?! @KCD Istanbul 2024
Kubernetes & AI - Beauty and the Beast !?! @KCD Istanbul 2024
 
Unsubscribed: Combat Subscription Fatigue With a Membership Mentality by Head...
Unsubscribed: Combat Subscription Fatigue With a Membership Mentality by Head...Unsubscribed: Combat Subscription Fatigue With a Membership Mentality by Head...
Unsubscribed: Combat Subscription Fatigue With a Membership Mentality by Head...
 
FIDO Alliance Osaka Seminar: Passkeys at Amazon.pdf
FIDO Alliance Osaka Seminar: Passkeys at Amazon.pdfFIDO Alliance Osaka Seminar: Passkeys at Amazon.pdf
FIDO Alliance Osaka Seminar: Passkeys at Amazon.pdf
 
From Siloed Products to Connected Ecosystem: Building a Sustainable and Scala...
From Siloed Products to Connected Ecosystem: Building a Sustainable and Scala...From Siloed Products to Connected Ecosystem: Building a Sustainable and Scala...
From Siloed Products to Connected Ecosystem: Building a Sustainable and Scala...
 
Builder.ai Founder Sachin Dev Duggal's Strategic Approach to Create an Innova...
Builder.ai Founder Sachin Dev Duggal's Strategic Approach to Create an Innova...Builder.ai Founder Sachin Dev Duggal's Strategic Approach to Create an Innova...
Builder.ai Founder Sachin Dev Duggal's Strategic Approach to Create an Innova...
 
Elevating Tactical DDD Patterns Through Object Calisthenics
Elevating Tactical DDD Patterns Through Object CalisthenicsElevating Tactical DDD Patterns Through Object Calisthenics
Elevating Tactical DDD Patterns Through Object Calisthenics
 
GDG Cloud Southlake #33: Boule & Rebala: Effective AppSec in SDLC using Deplo...
GDG Cloud Southlake #33: Boule & Rebala: Effective AppSec in SDLC using Deplo...GDG Cloud Southlake #33: Boule & Rebala: Effective AppSec in SDLC using Deplo...
GDG Cloud Southlake #33: Boule & Rebala: Effective AppSec in SDLC using Deplo...
 
From Daily Decisions to Bottom Line: Connecting Product Work to Revenue by VP...
From Daily Decisions to Bottom Line: Connecting Product Work to Revenue by VP...From Daily Decisions to Bottom Line: Connecting Product Work to Revenue by VP...
From Daily Decisions to Bottom Line: Connecting Product Work to Revenue by VP...
 
UiPath Test Automation using UiPath Test Suite series, part 4
UiPath Test Automation using UiPath Test Suite series, part 4UiPath Test Automation using UiPath Test Suite series, part 4
UiPath Test Automation using UiPath Test Suite series, part 4
 
Bits & Pixels using AI for Good.........
Bits & Pixels using AI for Good.........Bits & Pixels using AI for Good.........
Bits & Pixels using AI for Good.........
 
Smart TV Buyer Insights Survey 2024 by 91mobiles.pdf
Smart TV Buyer Insights Survey 2024 by 91mobiles.pdfSmart TV Buyer Insights Survey 2024 by 91mobiles.pdf
Smart TV Buyer Insights Survey 2024 by 91mobiles.pdf
 
Encryption in Microsoft 365 - ExpertsLive Netherlands 2024
Encryption in Microsoft 365 - ExpertsLive Netherlands 2024Encryption in Microsoft 365 - ExpertsLive Netherlands 2024
Encryption in Microsoft 365 - ExpertsLive Netherlands 2024
 
GenAISummit 2024 May 28 Sri Ambati Keynote: AGI Belongs to The Community in O...
GenAISummit 2024 May 28 Sri Ambati Keynote: AGI Belongs to The Community in O...GenAISummit 2024 May 28 Sri Ambati Keynote: AGI Belongs to The Community in O...
GenAISummit 2024 May 28 Sri Ambati Keynote: AGI Belongs to The Community in O...
 
Mission to Decommission: Importance of Decommissioning Products to Increase E...
Mission to Decommission: Importance of Decommissioning Products to Increase E...Mission to Decommission: Importance of Decommissioning Products to Increase E...
Mission to Decommission: Importance of Decommissioning Products to Increase E...
 
Monitoring Java Application Security with JDK Tools and JFR Events
Monitoring Java Application Security with JDK Tools and JFR EventsMonitoring Java Application Security with JDK Tools and JFR Events
Monitoring Java Application Security with JDK Tools and JFR Events
 
FIDO Alliance Osaka Seminar: FIDO Security Aspects.pdf
FIDO Alliance Osaka Seminar: FIDO Security Aspects.pdfFIDO Alliance Osaka Seminar: FIDO Security Aspects.pdf
FIDO Alliance Osaka Seminar: FIDO Security Aspects.pdf
 
Leading Change strategies and insights for effective change management pdf 1.pdf
Leading Change strategies and insights for effective change management pdf 1.pdfLeading Change strategies and insights for effective change management pdf 1.pdf
Leading Change strategies and insights for effective change management pdf 1.pdf
 

Edge AI Smart Manufacturing - Defect Detection and Beyond (GTC 2019)

  • 1. Trista Chen, Chief Scientist of Machine Learning of Inventec Wei-Chao Chen, Co-founder of Skywatch & Head of Inventec AI Center {chen.trista, chen.wei-chao} @ inventec.com S9617
  • 2. About Inventec • Public company (Since 1975; Taipei, Taiwan) • Tier 1 electronics manufacturer • Annual revenue USD $16B+ (2018) • Factories in Taiwan, Shanghai and Chongqing GTC 2019 Personal Computers Laptops Enterprise Computers Servers Solar Energy Smart Devices Medical Devices
  • 3. AI for Smart Manufacturing • Process Automation – Automatic optical inspection – Production scheduling – Automatic testing • Predictive Analysis – Order forecast – Production / yield prediction – Predictive maintenance GTC 2019
  • 4. Smart Manufacturing Automatic optical inspection Order forecast Production scheduling Test automation … Future Product AIOT / Smart City applications Medical devices Robotics Computation Platform …. Inventec AI Center
  • 5. Why Edge AI? • Privacy • Reliability • Network Efficiency GTC 2019 JETSON TX2 MODULE •NVIDIA Pascal™ Architecture GPU •2 Denver 64-bit CPUs + Quad-Core A57 Complex •8GB L128 bit DDR4 Memory •32GB eMMC 5.1 Flash Storage •Connectivity to 802.11ac Wi-Fi and Bluetooth- Enabled Devices •10/100/1000BASE-T Ethernet
  • 6. GTC 2019 Laptop Testing Robot A0I Re-inspection Laptop Surface AOI Re-inspectionAOI for SMT lines and PCBs
  • 7. GTC 2019 Laptop Testing Robot A0I Re-inspection Laptop Surface AOI Re-inspectionAOI for SMT lines and PCBs
  • 8. 1. Laptop Testing Robot Background • Millions of laptops manufactured every month – Similar basic design, many different configurations – Production tests are mostly automatic GTC 2019
  • 9. 1. Laptop Testing Robot Background • Product approval is labor intensive – Takes weeks to certify for mass production – ~1M test assets – Hundreds of people running the tests GTC 2019
  • 10. 1. Laptop Testing Robot Motivation (least -> most important) • Labor cost and management – Skilled and trustworthy testers are difficult to find • Time-to-market – Regression tests take weeks / months to complete • Increase confidence – The management confidence level on manual testing is only around 70% GTC 2019
  • 11. 0. 3 3 0 0 . / 0 0 0 0 3 GTC 2019 By: Jimmy Ou, Wei-Chao Chen, et al. of Skywatch Collaboration with Joseph Shi, Jack Hung & Inventec QA Engineers 1. Laptop Testing Robot
  • 12. GTC 2019 1. Laptop Testing Robot Pneumatic actuators Mechanical / electrical keyboard and mouse
  • 13. GTC 2019 1. Laptop Testing Robot Camera with computer vision capability Asideoflaptop Industrial Camera
  • 14. 1. Laptop Testing Robot • Computer Vision (OpenCV) – Perspective correction – Edge detection & denoising – Template matching • Tesseract OCR – LSTM-based NN OCR engine GTC 2019
  • 15. 1. Laptop Testing Robot GTC 2019 Test subject Controller 4X real-time demo
  • 16. 1. Laptop Testing Robot Future Opportunities • Automatic testing script generation – from human-readable to machine-readable test cases • Scheduling optimization – Reshuffle test levels and orders to speed up process • Smart random testing – Aka monkey testing GTC 2019
  • 17. GTC 2019 Laptop Testing Robot A0I Re-inspection Laptop Surface AOI Re-inspectionAOI for SMT lines and PCBs
  • 18. 2: AOI (Automated Optical Inspection) Re-inspection GTC 2019 Common SMT defects
  • 19. 2: AOI Re-inspection AOI Machine Fail Pass GTC 2019 High false positive Low first-pass yield Re-inspection by human
  • 20. 2: AOI Re-inspection AOI Machine Pass Fail AOI Re-inspection Pass Fail GTC 2019 By: , , of NTU & advised by Wei-Chao Chen of Skywatch & NTU & Trista Chen of Inventec Data & domain knowledge provided by Jessie Huang, Peter Chu @ TAO, among others High false positive Low first-pass yield
  • 21. 2: AOI Re-inspection GTC 2019 Traditional Computer Vision defects w/ well-defined rules Deep Learning defects w/ intuitive descriptions
  • 22. 2: AOI Re-inspection GTC 2019 0. Excess Solder (true) 1. Excess Solder (false positive à pass ) 2. Insuff. Solder (true) 2. Insuff. Solder (false positiveàpass)
  • 23. GTC 2019 Laptop Testing Robot A0I Re-inspection Laptop Surface AOI Re-inspectionAOI for SMT lines and PCBs
  • 24. 3. Laptop Surface AOI GTC 2019 Brand Logo Customer Info
  • 26. Brand Logo GTC 2019 Customer Info Type #2 Dot Type #1 Scratch Type #3: Area Type #4: Abrasion Type #5: Don’t care
  • 27. 3. Laptop Surface AOI GTC 2019 Defect type Class A spec Class B spec Scratch Length: 12mm Acceptable: 2 lines Length: 20mm, Acceptable: 2 lines Dent 0.5 mm2 < size < 0.7mm2 Acceptable: 2 points 0.5 mm2 < size < 1mm2 Acceptable: 3 points … Fail Pass Human readable rules Labor intensive
  • 28. 3. Laptop Surface AOI GTC 2019 Defect type Class A spec Class B spec Scratch Length: 12mm Acceptable: 2 lines Length: 20mm, Acceptable: 2 lines Dent 0.5 mm2 < size < 0.7mm2 Acceptable: 2 points 0.5 mm2 < size < 1mm2 Acceptable: 3 points … Fail Pass Labor intensive Human readable rules AI
  • 29. 3. Laptop Surface AOI GTC 2019 Fail Pass AI Training data
  • 30. 3. Laptop Surface AOI Pass Fail AIMobile M1 (Nvidia TX2) GTC 2019 By: Benson Lin of Skywatch, Trista Chen and Irene Chen of Inventec, Wei-Chao Chen of Skywatch & Inventec In collaboration with Steven Wang, Sing-Wang Chen, Alfa Shih & Tim Zhang et al @ ICC Machine manufactured by Jerry Tseng @ Leh-Yeh; Edge machine provided by Mark Lu @ AIMobile Surface defect classification
  • 31. 3. Laptop Surface AOI Pass/Fail Classification DL detection Pass Fail AIMobile M1 (Nvidia TX2) GTC 2019
  • 32. 3. Laptop Surface AOI Pass/Fail Classification DL detection Pass Fail GTC 2019 Stage 1: Explainable Stage 2: Adjustable
  • 33. 3. Laptop Surface AOI Pass/Fail Classification DL detection Pass Fail GTC 2019 Stage 1: Explainable Stage 2: Adjustable
  • 34. 3. Laptop Surface AOI – Explainable GTC 2019 Why did it pass/fail? • Defect type • Defect count • Defect size Defect type Class A spec Class B spec Scratch Length: 12mm Acceptable: 2 lines Length: 20mm, Acceptable: 2 lines Dent 0.5 mm2 < size < 0.7mm2 Acceptable: 2 points 0.5 mm2 < size < 1mm2 Acceptable: 3 points …
  • 35. Defect class cluster analysis Defect types: s1 s2 s3 s4 s5 5 classes a b c a b c a b c a b c a b c 15 classes GTC 2019 Stage-1 multi-class detector 5 classes ? ?
  • 36. GTC 2019 a b c a b c a b c a b c a b c Defect types: s1 s2 s3 s4 s5 15 classes Laptop categories: a, b, c Defect types: s1 s2 s3 s4 s5 Output Input s5 s4 s1 s3 s2 5 classes Defect class cluster analysis
  • 37. 3. Laptop Surface AOI Pass/Fail Classification DL detection Pass Fail GTC 2019 Stage 1: Explainable Stage 2: Adjustable
  • 38. 3. Laptop Surface AOI – Adjustable GTC 2019 Class A 2 lines 2 dots
  • 39. 3. Laptop Surface AOI – Adjustable GTC 2019 Class B
  • 40. 3. Laptop Surface AOI – Adjustable GTC 2019 Class C
  • 41. 3. Laptop Surface AOI – Adjustable GTC 2019 Class A
  • 42. 3. Laptop Surface AOI – Adjustable GTC 2019 Class A+
  • 43. 3. Laptop Surface AOI – Adjustable GTC 2019 Class A++
  • 44. 3. Laptop Surface AOI Pass/Fail Classification DL detection Pass Fail GTC 2019 Stage 1: Explainable Stage 2: Adjustable
  • 45. 3. Laptop Surface AOI – Factory Prototype GTC 2019 CCD NB
  • 46. Benefits of AI for Manufacturing GTC 2019 • The Big Scope – Consistency – Faithful digital record – Industry 4.0 OEE SECS MES SECS 1. 2. 3. 4. MES / / / /… / / & Idle 1. 2. 3. 4. 1. 2. MES/ 3. 4. / Repair Source: Taiwan’s Institute for Information Industry (III) 2017 • The Obvious – Accuracy – Labor Saving ROI
  • 47. Benefits of AI for Manufacturing – I4.0 GTC 2019 95.00 96.00 97.00 98.00 99.00 100.00 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Production Efficiency Peak Productivity Identified Problem Image source: Hitachi “Factories of the Future”, NEXT 2019 time Process chain
  • 48. GTC 2019 Laptop Testing Robot A0I Re-inspection Laptop Surface AOI Re-inspectionAOI for SMT lines and PCBs
  • 49. Unexpected Bonus: to do well while building your AI engine We’re on a mission to connect 60,000 residential disabled in Taiwan to join AI work by providing high-quality data fuels to your AI engines.
  • 50. GTC 2019 Source: Christopher Porter’s Roomba and Cat from Flickr
  • 52. Trista Chen, Chief Scientist of Machine Learning of Inventec Wei-Chao Chen, Co-founder of Skywatch & Head of Inventec AI Center {chen.trista, chen.wei-chao} @ inventec.com S9617